The National School Lunch Program: Is It Working?

Published by the Government Accountability Office on 1977-07-26.

Below is a raw (and likely hideous) rendition of the original report. (PDF)

                          DOCUMENT RESUME

03051 - (12093200)

Xhe National School Lunch Program: Is It Working? PAD-77-6;
8-1;1810. Jujy 26, 1977. 137 pp • • 3 appendices (28 pp.).

Report to the Congress; by Elmer B. Staats, Coaptroller Generala
Issue Area: Incoae Security Prograas: Program Effectiveness
    (1302).' Prograll Bvaluation Syst.ems: New Techniques and
    Standards (2602); rood: Do.estle Food' Assistance Progralls
    (1707) •
Contact: Proqraa Analysis Div.
Budget Function: Agriculture: Far. Income Stabilization (35-');
    Incoae Security: Public Assistance and Other Income
    Supple.ents (604); Health: Health Research ann Education
    (552) •               .
Organization concerned: Department of Agriculture; Department of
    Healtb~ Education~ and Welfare.
congressional Relevance: House Co••ittee on Education and Labor;
    Senate Com.ittee on Agriculture. Nutrition. and Forestry;
Authority: National School Lunch Act of 19ij6 (P.L. 79-396).
   Agricultural Act of 19ij9, sec. ij16 (7 U.S.C. 1431). Chila
   Nutrition Act of 1966 (P.L. 89-642). Child Nutrition let of
   1972 (P.L •. 92-4331. P.L. 92-153. P.L. 87-823, sec. 11. P.L.
   74-320. P.L. 91-248. P.L. 93-326. P.L. 94-105. 85 Stat. 419.

         The National school Lunch Prograa is designed to
safeguard schoolchild health by improving and/or maintaining
levels of nutrition and to strengthen the agricultural econoay
by stimulating food de.and. Findings/Conclusions: The scbool
lunch progra. provides adequately for the large-scale feeding of
children. but it could be much ~ore effective and efficient than
it is. Alth~ugh studies shOll that the schoo'! lunch. when paired
with a nutritional supplement or with th'e school breakfast, can
affect the nutritional levels of schoolchildren. their findings
about how the lunch itself affects nutritionally deprived and
nutritionally adequate participants are inconclusive. There are
consistent indications that the program.. has strengthened overall
demand for far~ products. although the possibility of a conflict
bet~een the program's agricultural and nutritional proYisions
was no~ed. Shifting eating habits and needs over the past 30
years suggest that the program's objectives should be
reassessed. Recoamendations: Congress should: provide policy
guidance indicating specifically what the purpose of the program
should be and have the program evaluated accordingly; define the
priority of each purpose and direct how the program is to be
evaluated; require tbe Department of Health. Education, and
Welfare to assist the Department of AgricultQre in determining
the progra m' s contribution to chil,dren' s health; review
AgriCUlture's program evaluation plan to be sure it vill support
the needs of congressional oversight; and require AgricQlture to
report to the Congress the results of its evaluation. (SC)

              OF THE UNITED STATES

              The Nationa I School
              Lunch Program--
              Is It Working?

              Departments of Agriculture and
              Health, Education, and Welfare

              This report identifies shortcomings in both
              the evaluation and performance of the School
              Lunch Program. It recommends specific ac-
              tions for improving thp. effectiveness i)nd ef-
              ficiency of program servicb.
              Areas discussed include
                   --schoolchild health,
                   ··children in need of nutrition,
                   ··operating efficiency, and
                   ·-relationship of the program to the Na-
                    tion's agricultural economy.

              PAD-77·6                                          JULY 26,1977
              COMP'TI'tOLLEI't IoiENEftAL 01"' THE UNITED STATES
                            WM~lNGTON.O.C. _ _


To ~he President of the Senate and the
Speaker of the House of Representatives
     This report describes what is known about the National
School Lunch Program's effectiveness in achieving legislative
objectives. It is being released concurrently with a com-
panion summary (PAD-77-7).
     Officials of the Department of Agriculture and the Depart-
ment of Health, Education, and Welfare have been given the op-
portunity to review and- comment on this report. Their vi.ews
have been incorporated where appropriate.
     Our review was made pursuant to the Budget and Account-
ing Act, 1921 (31 U.S.C. 53), the Accounting and Auditing
Act of 1950 01 U.S.C. 67), and the Legislative Reorganiza-
tion Act of 1970 as amended by title VII of the Congressional
Budget Act of 1974 (31 U.S.C. 1154).
     Copies of the report are being sent to the Director,
Office of Management and BUdget, the Sec~etary of Agricul-
ture; and the secretary of Health, Education, and Welfare.

                                      Comptroller General
                                      of the United States
                                       Departmen~s of Agriculture and
                                       Health, -Education, arid Welfar.e

            DIG EST
            The N:ational School Lunch Program is designed
            --safeguard schoolchild health by improving
              and/or ma1ntaining levels of nutrition and
            --streng~hen  the a9~icultural economy by
               stimulating food demand.
            To these ends, the Secretary of Agriculture
            requires that lunches served under the pro-
            gram meet a specified food pattern (type A)
            providing, on average-, one-third of each
            child's daily diet.
            Federal assistance to Sta~es in serving the
            lunches is based on the number of meals
            served. This, in cash and commodities,
            amounted t9 more than $1.7 billion in fis-
            cal year 1975. Over 4 billion lunches were
            served, 8t;lOUt 1.6 billion being provided
            free or at reduced prices to children from-
            economically needy families.
            The program is avail~ble in most of the Na-
            tion's schools and_ -is the largest of .everal
            federally supported child-feeding programs.
            THE CONGRESS
            The scnool lunch program provides adequately
            for the large-scale feeding of children, but
            it could be much more effective and effi-
            cient than it isw
            GAO thinKS the Congress should:
            --Provide policy guidance indicating speci-
              fically what the purposes of the program

Tear Sheet. Upon removal, the report
cover date should be noted hereon.
                                       i                       PAD-77-6
  should be and have the program evaluated

--Define the priority of ea~h purpose and
  direct how the prog~am is to be evaluated.

--Require the Department of Health, Educa-
  tion, and Welfare (HEW) to assist the
  Department of Agriculture in determining
  the program's contribution to children's

--Review Agriculture's program evaluation
  plan to be sure it will support the needs
  of congressional oversight.

--Require Agriculture to report to the Con-
  gress the results of its evaluation.

Since legislation prohibits the school lunch
program from imposing any requirement rela-
tive to the teaching of nutrition to school-
children, the effectiveness of nutrition ed-
ucation programs is not addressed in this
report. Such programs are, however, cur-
rently b~ing reviewed by GAO on a broader


GAO examined the question of whether the
program is meeting its legislative objec-
tives and analyzed available information
on what is known and not known about the
effectiveness of the program. Areas dis-
cussed include:

--Health impact. DoeS the program support
  the health of schoolchildren?  (See part

--Agricultural impact. Does the program
  increase demand for agricultural commod-
  ities?  (See part II.)

--Participation. Do children in need of
  nutrition eat the lunches? (See part
  III. )

              --Oper~ting  efficiency. Are lunches pro-
                vided as inexpensively as possible? (_See
                par t IV.)
              The, school lunch ,program has been in opera-
              tion ·fot three decades. Though many reports
              over the years_ contain important information
              about the program, this in£ormation generally
              has not been assembled in a way to assist the
              Congress in reviewing the program's b~dget,
              iri considering: program revisions, and in
              overseeing prog~am admi~istration.
              Shortcomings in program evaluation are not
              necess G1 [.ily indi.cative of· actual ·per formance
              or of the prq,g-rarn's potential to support
              schoolchild health. Nutrition--the lack,
              excess,- or quality o-f it--appears to be a
              problem -for millions: of schoolchildren.
              (See chs. 2 and 3.)
              HEALTH IMPACT
              The type A lunch is a critical factor in the
              program's success. -The quantity and type of
              food included in the lunch largely determine
              cost and. the amount of commodities eaten;
              thp. price and presentation ee the lunch de-
              tern:i-ne how well the program reacht!s child-
              ren; and the nutritional ~ualities of the
              lunch determine how well the program safe-
              guards health.
              Although- studies show that the school lunch,
              when pai_red with a nutr i tienal supplement, Or
              with the school br~akfast, can affec,t the
              nutritional levels of schoolchildren, their
              findings about how the lunc-h itself affects
              nutritionally deprived and- nutritionally
              adequate participants are inconclusive.
              (See ch. 4.)
              Although the type A lunch appears to be ef-
              fective in increasing food consumption', GAO
              is not c'onvinced that it is the best choice
              for a nutritional standard. The absence of
              any indication that the prog_ram i~ having a
              net benefit on the health of either needy or
              nonneedy children raises questions about the
              nutritional value of the lunch.

Tei)[ Sheet                             iii
In comparison with other types of lunches,
the type A lunch:

--Appears relatively ineffective in achiev-
  ing the program's nutritional objective.
  The lunch, a standard meal served to all,
  does not appear effective in combating
  children's diverse nutritional problems.
  An alternative standard-·providing more
  flexibility in the pattern and/or portion
  sizes--may improve the program's nutri-
  tional impact.    (See p. 38.)

--May increase the cost of program lunches
  (thereby reducing participation of stu-
  dents who pay). (See ch. 7 and p. 123.)

--Is often presented in a form or content
  which discourages student participation
  and contributes to food waste. (See
  pp. 68 to 76.)
GAO found consistent indications that the
program has strengthened overall demand
fOL farm products. However, possibility
of conflict between the program's agricul-
tural and nutritional provisions was noted.
(See chs. 5 and 11.)

Shifting eating habits and needs over the
past 30 years suggest that the program's
objectives should be reassessed. Present
agricultural conditions are considerably
different. Conditions of oversupply are
less frequent and concerns about the agri-
cultural economy have generally tended
toward the problem of shortages.
Between 1971 and 1975, an expanded free/
reduced-price program substantiully in-
creased the participation of low-income
children; but, because much of the in-
crease was offset by declines in the
participation of regular-price stUdents,
overall participation tended to remain
constant. (See ch. 6.)
             The shift toward low-income children (the
             population group with the greatest preval-
             ence of nutritional problems) probably in-
             creased theprogram',spotential as a nutri-
             tion aid. On the other hand, the program
             became less ,effective in reaching the
             regular-price students'. (See pp. 38 and 88.)
             Although many authorities have expressed
             a desire to improve participation levels,
             the question remains: How? Available stud-
             ies, though beneficial in identifying some
             of the "factors" affecting participation,
             help little in estimating the impacts of
             various policy alternatives. (See ch. 7.)
             For example:
             --Price-participation relationships are an
               extremely weak forecasting tool.
             --The relative importance (rank) of the in-
               dividual factors affecting participation
               has not been clearly defined.
             --Better information is needed to assess the
               effects a change in participation would have
               and to direct the program toward children in
               greatest need.
             While it is true that the school lunch pro-
             gram's operating expenses increased rapidly
             over the 1973-75 period, the main cause was
             inflation. The real cost of producing a
             program lunch actually declined. (See
             ch. 9.)

             A potential exists for Agriculture to reduce
             program food costs by more than $100 million
             per year without sacrificing nutritional im-
             pact.     (See ch. 9.)

             Agriculture1s commodity distribution pro-
             gram helps small school systems save on
             food costs. A flat-rate disbursement of
             cash in lieu of commodities would provide
             a disproportionate benefit for large school
             systems because of economies of scale in
             procurement.     (See p. 115.)

Tear Sheet                            v
The Secretary of Agriculture should':

--Require a formal, systematic evaluation
  of how well the school lunch program
  meets legislative objectives. The evalu-
  ation sh.ould use the expertise and re-
  sources of the Deaprtment of Health, Ed-
  ucation, and Welfare and should report
  its results to the Congress in a timely

--With assistance from HEW, (1) determine the
  nutritional standards needed to best s~fe­
  guard schoolchild health, (2) if found de-
  sirable, revise the program's meal regUla-
  tions to reflect nutriti.onal requirements
  that will give menu planne~s planning
  flexibility,    (3) improve the pcogram's
  cost-effectiveness, (4) encourage higher
  levels of student participation, and (5)
  reduce food was-te.
--Determine the effect of commodity distri-
  bution surges on the school lunch program's
  nutritional objective and, if surges are
  determined to have an important effect, im-
  pleffient corrective procedures so that
  agricultural considerations do not cOmpro-
  mise the program's nutritional effective-
--Improve the accuracy of participation
   forecasts and determine the relative im--
   partance of individual factors (including
   price) which affect participation.
--Determine how a change in program partici-
  pation affects the magnitude and characteris-
  tics of unmet nutritional needs in the non-
  participant population.
--Examine approaches and take actions to im-
  prove the economy- of small and mediUI'I-
  sized school systems' food procureffi ~t.

             HEW concurred with GAO's recommendations and
             said that it would assist Agriculture- in de-
             veloping meal standards and in evaluating the
             program's nutritional impacr. HEW also pr~­
             vided technical comments pertaining to the
             relationship between nutrition and health
             which were used in preparing this report.

             Agriculture generally agreed with the recom-
             mendations that do not involve the prog-ram's
             nutritional impact, and has acted or agreed
             to act to implement them.   (See pp., 79, 105,
             and 124.)  It did not respond- to the recom-
             mendation for determining the relationship
             betwe~n participation levels ahd the unmet
             nutritional needs of the target population.
             (See p. 105.)

             Agriculture said it recognized the need for
             a program evaluation and that a draft of
             the Food and Nutrition Service's research
             program for the next 5 years, now under
             review, includes the development of a
             methodology for assessing the school lunch
             pro-gram's nutritional impact. It:. also
             said that Nit is questionable that such
             a study would be successful in accomplish-
             ing its objectives." Since GAO has not
             reviewed the research plan, it has no
             means of assessing whether or not the pro-
             gram evaluation will be effective. Agri-
             CUlture, however, made no mention of a
             positive commitment to start the evalua-
             tion, nor did it reply to GAO' 5 recommenda-·
             tion that such an evaluation be coordinated
             with congressional oversight needs and sup-
             ported by HEW.    (See p. 51.)

             Agriculture disagreed with the recommenda-
             tion pertaining to nutritional standards.
             It said that such standards would be dif-
             ficult to determine and that, while it
             shared GAO's concerns regarding the type A
             pattern's effect on participation and food
             waste, there were ways of addressing such
             concerns short of abandoning nationally es-
             tablished meal standards.  (See p. 79.)

Tea, Sheel                          vii
There is no doubt that the nutritional as-
pects of the school lunch program are comp-
licated and difficult to evaluate. The
consequence of not doing such evaluations,
however, is to leave some very important
issues to chance. Therefore, GAO believes
that Agriculture should take positive ac-
tion on the recommendations outlined in
this report.

                          c on ten t s

DIGEST                                                      i

   1      INTRODUCTION                                      1
              Legislative history                           1
              Program administration                        5
              Scope of review and report organization       7

                              PART I
                    ABILITY TO SAFEGUARD HEALTH             9

            HEALTH?                                        11
              What is good nutrition?                      12
              Benefits of improved nutrition               14
              What is the health threat?                   15
                  Overconsumption recognized as health
                    threat                                 16
                 Deficiencies in RDA nutrition may also
                    affect health                          17
              Implications for NSLP                        19
              Complexities of measurement                  21
              Ten-State Nutrition Survey                   22
                  Nutrient intakes increased by school
                    lunch program~                         27
              Preschool Nutrition Survey                   30
              Health and Nutrition Examination Survey      32
              Implications for NSLP                        37

              Nutritional impact undetermined              40
                  Existing studies                         41
                  Evaluation is possible                   46
              Need to upgrade nutrition education          48
              Need for further evaluation                  49
              Recommendation to the Secretary of
                Agriculture                                51
              Agency comment. and our evaluation           51
              Recommendations to the Congress              52
CHAPTER                                                      Page
                            PART II

            THE SCHOOL LUNCH PROGRAM                           55
              NSLP's consumption of agricultural products      56
                  Local market purchases                       57
                  An outlet for surplus food                   57
              Impact on the agricultural economy               59
                  Commodity-by-commodity approach              60
                  Sector-by-sector approach                    63
              possible conflict with nutrition objectives      68
                  Differences in type A pattern and the
                    program's nutritional target               68
                  Some food service directors believe mod-
                    ifications in meal pattern will im-
                    prove participation and decrease waste     70
                  USDA recognizes the need for improve-
                    ments in the NSLP lunch                    72
                  Timing of commodity distributions has
                    important impact on nutritional
                    objectives                                 77
              Have agricultural goals been satisfied?          77
              Recommendations to the Secretary of
                Agriculture                                    78
              Agency comments and our evaluation               79
              Recommendation to the Congress                   80
                            PART III
                        PROGRAM COVERAGE                       81
   6      PARTICIPATION TRENDS                                 82
              NSLP availability in schools                     82
              Student participation                            85
                  Regular and reduced-price/free
                    participation                              87

   7      FACTORS AFFECTING PARTICIPATION                      89
              Influence of noneconomic factors                 89
                  Alternative food sources                     89
                  Attitude of school administrators            90
                  Menu choice and preparation                  91
                  Regular-price participation rates not
                    affected by size of free lunch program     92
              Influence of price                               92
                  Parents aware of relative prices             93
              Price-participation relationships provide a
                weak forecasting tool                          95
              Better information needed to assist
                dec isionmci"ker s-                            98
                  Multifac~or  relationships                   99
                  Regular-price participation declines
                     as per capita income advances            102
                  Matching resources with needs               103
              Recommendations to the Secretary of
                Agr.icul ture                                 104
              Agency comments                                 105

                                PART IV
                             PROGRAM COSTS                    106

              Determining the cost savings of USDA food
                distributions                                 108
                  Cost of USDA foods increased by intra-
                     state distr ibution expenses             109
                   USDA foods less expensive than schools'
                     open market purchases                    109
              Di.ffering opinions on cash in lieu of
                commodities                                   113
              Impact of suspending commodity program          115

   9      FACTORS AFFECTING COST GROWTH                       117
              Growth in cost of prodQcing an NSLP lunch       118
                  Component costs                             118
                  Cost variations due to economies of scale   121
              Implications for cost savings                   123
              Recommendation to the Secretary of
                Agr icul ture                                 124
              Agency comments                                 124

            LUNCH PROGRAM                                     125
              Sources of NSLP funding                         125
              Federal funds carry increasing share of
                NSLP costs                                    127
              Observations on the design of school lunch
                subsidies                                     128

                               PART V
                      PROGRAM EVALUATION ISSUES               131

            MATTERS FOR CONSIDERATION                         132
              Major program issues remain unresolved          132
              Issue discussion                                133
CHAPTER                                                     Page

  11                Health considerations                    133
                    Agricultural considerations              134
                    Participation cons-ideratio-ns           136
                    Cost considerations                      137

       I           April 20, 1977, letter from the Acting
                    Administrator, FNS                       138

                   March 14, 1977, letter from the Acting
  III                Deputy Administrator, ERS               154

                   April 14, 1977, letter from the Inspector
                    General. HEW                             163


CPI        Consumer Price Index

DMF        decayed, missing, and filled

ERS        Economic Research Service
FNS        Food and Nutrition Service

GAO        General Accounting Office

HANES      Health and Nutrition Examination Survey

HEW        Department of Health, Education, and Welfare

NSM        Nutrient Standard Menu

NSLP       National School Lunch Program

PNS        Preschool Nutrition Survey

RDA        Recommended Dietary Allowances

SES        socioeconomic status
TSNS       Ten-State Nutrition Survey

USDA       Department of Agriculture
                         LIST OF TABLES
2.1    USDA estimates of annual savings through improved
         nutritionl                                           14
2.2    The 10 leading causes of death, United States,
         1974 (based on a 10-percent sample of deaths)        16
'3.1   RelatiV,e importance of nutri,tional problems--
         Ten.,.State Nutrition Survey (1968-19,70)            25
3,2    Distt(buti6n of nutrient intakes for males (12-14
         yrs. )--Ten-State' Nutrition Survey (1968-1970)      26
3.3    Comparison of, mean nutrient intake of persons 10
         through 16 years of age participating in school
         lunch programs with those not participating,
         Ten-State Nutrition Surve~ (1968-1970)               28
3.4    Estimated mean nutrient intakes as a percent of
         RDA for 6-11 ana l2-l7-year-olds: United States,
         1971-72 (HANES preliminary)                          34
3.5    Low biochemical indices for 6-11 and l2-l7-year-
         olds: United States, 1971-72 (HANES preliminary)     36
3.6    Estimated number of 6-17-year-olds with low bio-
         chemical indices: United States, 1971-72
         (HANtS preliminary)                                  37
4.1    Changes in classification of children from fall
         to spr ing--Cornell study                            44
5.1    NSLP food costs, fiscal year 1971-75                   57
5.2    Iowa school full lunch program support for certain
         commodities                                          61
5.3    Changes in U.S. business receipts and gross na-
         tional product associated with Federal contribu-
         tions to NSLP, fiscal year 1974                      64
5.4    Net changes in job numbers resulting from Federal
         contributions to NSLP, fiscal year 1974              66
5.5    School Lunch Pattern, 1976 (proposed)                  73
5.6    Concerns about the type A pattern and changes to
         help alleviate the concerns in the 1976 pattern      74
6.1    Total enrollment and participating students in
         NSLP schools, fiscal year 1971-75                    85
6.2    Eligible and participating students in the
         regular-price and free or reduced-price lunch
         program, fiscal year 1971-75 (estimated)             88
7.1    Sample sack lunches used in Washington State study
         (1970 prices)                                        94
7.2    Number of States participating in NSLP, by regular
         prices charged and daily participation rates.
         fiscal year 1973                                     95
7.3    Regular-price participation as a function of per
         capita income and price, United States, fiscal
         year 1973                                           100
 8.1   USDA commodity distributions to NSLP, fiscal year
         1971-75                                               107
 8.2   Average prices paid by school systems for foods
         by source of purchase, July 1973-Apri1 1974           110
8.3    USDA food cost per lunch--fisca1 year 1974              112
8.4    Comparison of food costs: USDA and open market          112
9.1    NSLP cost growth--fiscal year 1970-75                   118
9.2    Component costs of NSLP Lunch--fiscal year
         1970-74                                               119
 9.3   Highlighting the factors affecting cost growth of
         NSLP lunch (meal cost for fiscal year 1970 ad-
         justed to fiscal year 1974 prices and compared
         with actual fiscal year 1974 costs)                   120
 9.4   Average enrollment and daily lunches served per
         NSLP school, fiscal year 1970-74                      122
10.1   School lunch federal assistance rates(Jan.-June 1~76)   126
10.2   Sources of NSLP funding, fiscal year 1970-75            127
10.3   NSLP participation categories, fiscal year
         1972-75                                               128
10.4   Funding components 0f the fiscal year 1974
         school lunch (estimated)                              129

                        LIST OF FIGURES
 3.1   Percent obesity in ado1escents--Ten-State
         Nutrition Survey                                       26
 3.2   Mean decayed, missing, and filled (DMF) permanent
         teeth for persons 10 through 16 years of age
         by grams of carbohydrate consumed between
         meals--Ten-State Nutrition Survey                      27
 5.1   Total nutrients served and eaten: elementary
         schools                                                75
 6.1   Growth in NSLP availability, fiscal years 1950-75        83
 6.2   NSLP participation as percent of U.S. enrollment,
         fiscal year 1974                                       86
 7.1   Price-participation relationships for regular-price
         students, United States, fiscal year 1973              97
                          CHAPTER 1
      The National School Lunch Program (NSLP), authorized by
the, National School Lunch Act of 1946 (Public Law 79-396) and
ex~arided, in more recent legislation, is the largest of sever-
al 'fed,,'rally funded child-feeding programs.
      AS stated in the authorizing legislation, NSLP's objec-
ti,ves a're ". • • to safeg,uard the health and well-being of
the Nation's children and to encourage the domestic consump-
tion of nutritiou's agricultural commodities and other food."
To do this, the Federal Government encourages and assists
public and noriprofi t pr ivate schools below college level to
serve well-balanced luncheS to children. This assistance
     --A basic cash and donated food subsidy for all lunches,
       with additional cash reimbursement for meals served
       free or at reduced prices to children who cannot pay
       the full price.
     --Nonfood assistance funds to help needy schools acquire
       food service equipment.
     --State administrative expense funds to partially reim-
       burse States for undertaking the additional adminis-
       trative activities required by the program.
     --Limited funds to undertake program-related nutritional
       education and training projects, studies, and surveys
       of food service requirements, and special development
      From 1947 to 1975 NSLP has increased in Federal expendi-
ture from less than $100 million to more than $1.7 billion
(cash and commodities). In fiscal year 1975, about 88,800
schools (approximately 81 percent of the Nation's total) were
members of N5LP, making program lunches available to almost
88 percent of all schoolchildren. Over 25 million children
(56.7 percent of the NSLP enrollment) participated in the
program; nearly 39 percent of these children received free
or reduced-price lunches.
     Federal assistance in feeding schoolchildren has ex-
isted for roughly four decades. The U.S. Department of
Agriculture (USDA), authorized by section 32 of Public Law
74-320, began food distribution to schools during the mid-
19305. One objective was to provide nutritious, low-cost
meals to children, but the primary design was to create an
outlet for foods acquired under surplus-removal programs.
When wartime demands in the early 1940s drained these sur-
~luses, USDA initiated the Nation's first cash assistance
program by partially subsidizing schools for local food
     After the war, school lunch participation was approxi-
mately 4 million children. Uncertain of year-to-year funding,
schools were reluctant to enter the program or to expand
existinq programs. Recognizing this shortcoming, the Congress
enacted the National School Lunch Act of 1946 (Public Law 79-
396). This act authorized the creation. of NSLP. The act
established three basic operating standards:
     --School lunches should conform to nutritional standards
       established by USDA.
     --Free or reduced-price lunches should be provided to
       children unable to pay the regular price.
     --The program should be operated on a nonprofit basis.
     USDA's food distribution authority was further expanded
by section 416 of the Agricultural Act of 1949 (7 U.S.C.
1431), which authorized donations of food acquired by the
Commodity Credit Corporation under price-support programs.
     In 1962 the criteria for apportioning funds were re-
vised to provide a more equitable distribution and to en-
courage program expansion. At the same time, section 11,
Public Law 87-823, was added (but not funded until 1966)
to provide special financial assistance to schools serving
students from impoverished areas.
     The Child Nutrition Act of 1966 (Public Law 89-642),
recognizing lithe demonstrated relationship between food
and good nutrition and the capacity of children to develop
and learn," further amended and expanded the National
School Lunch Act of 1946.
     In 1970 the NSLP legislation was again amended and
expanded (Public Law 91-248) to include administrative and
procedrral changes. Strengthening the program's provisions
for serving free or reduced-price lunches to economically

neeqy ch ildren, th is act (1) manda ted that free lunches be
served to needy children iind (2) provided specific guide-
lines tO,be used in determining eligibility for free and
reduced-,pr iC,e lunches. (The serving of reduced-pr ice 1 unches
remained a State opt-ion-, however.)

     Public Law 92-153, approved November 5, 1971 (85 Stat.
419), raised the reimbursement rates to 6 cents for each
regular-price lunch and an additional 40 cents for each free
or reduced-price lunch served-.
     Further refinements were added with the passage of the
Child Nutrition Act of 1972 (Public Law 92-433). This legis-
lation guaranteed a m-inimum Federal subsidy on a "performance
funding" basis (i.e., ~r meal served) and increased the reim-
bursement rate from 6 cents to not less than 8 cents per
lunch. It also rescinded USDA's authority to regulate selling
food items in competition with programs iiuthorized under the
Child Nutrition Act and the National School Lunch Act.
Before this legislation, selling of competitive foods had
been prohibited while the school lunch was being served. The
Congress emphasized that this action was not intended to show
disapproval of existing regulations but to more appropriately
vest regulatory authority in State and local agencies.
     Public Law 93-150, enacted in November 1973, extended
the performance funding concept to section 11 special cash
assistance funds. It provided an escalator concept by which
average Federal payment rates are to be adjusted semi-annually
to reflect changes in the Consumer Pr ice Index for food away
from home.   It authorized cash payments to make up shortages
in commodity distributions, and it required the Secretary of
Agriculture to conduct a comprehensive study of child nutri-
tion programs.
     In 1974 NSLP legislation was again amended (Public
Law 93-326). This act prescribed a minimum level of com-
modity assistance at 10 cents per lunch, or cash payments
in lieu thereof, with provisions that the rate be adjusted
on an annual basis to compensate for changes in the Consumer
Price Index for food away from home. The act also raised
the eligibility criteria for reduced-price lunches to 175
percent of the Secretary's income poverty guidelines, though
the States retained the ootion of whether or not to offer
reduced-price lunches.   -
     Public Law 94-105, enacted in October 1975, represents
the most recent school lunch legislation. This act

     --expanded the NSLP coverage to include the Trust
       Territory of the Pacific Islands and, in addition
        to schools, any public or licensed nonprofit private
        residential child care institution, such as orphanages
       and homes for the mentally retarded.
     --revised--effective January 1976--the formula under
       which the Secretary of Agr.iculture determines income
       poverty guidelines.
     --establiLhed a mandatory reduced-price lunch program.
       Children from households with an annual income level
       which falls between the applicable income guidelines
       prescribed by the State for free lunches and 95 per-
       cent above the income poverty guidelines prescribed
       by the Secretary are to be served NSLP lunches at a
       price not to exceed 20 cents.
    --excluded Federal funds received by a State to provide
       free and reduced-price lunches from the general re-
      quirement that States match every dollar of Federal
       funds with three dollars of State and local funds.
    --directed the Secretary to establish, in cooperation
      with State educational agencies, administrative pro-
      cedures to diminish plate waste without endangering
      the nutritional integrity of the NSLP lunch. In this
      regard, the act further specifies that senior high
      school students will not be required to accept foods
      which they do not intend to consume, but that the
      failure to accept offered foods will not affect the
      student's charge or the amount of Federal reim-
     The 1975 act also authorized the Secretary of Agriculture
to carry out a nutrition program staff study to
    "* * * determine how States are utilizing Federal funds
    provided to them for the administration of the child
    nutrition programs * * * and to determine the level
    of funds needed by the States for administrative
    purposes. * * * As part of this study, the Secretary
    shall also examine the degree and cause of plate
    waste in the school lunch program. The Secretary
    shall examine possible relationships between plate
    waste and (1) lack of adE'quate menu development,
    (2) the service of competitive foods, and (3) the
    nature of the type A lunch pattern. The Secretary

     shall review the study design with the appropriate
     congressional committees prior to its implementation,
     and shall report his findings together with any recom-
     mendations he may have with respect to additional legis-
     lation, to the Congress no later than March 1, 1976."1/

     The Food and Nutrition Service (FNS), USDA, is respons-
ible for the national administration of NSLP. The program
is normally administered in cooperation with State depart-
ments of education. When the State agency is prohibited by
law or otherwise unable to disburse Federal funds to private
schools, an FNS regional office acts as the administering
agency. Participation at the local level is voluntary.
     Program responsibilities are divided among the National,
State, and local levels as follows.
     At the national level, FNS headquarters and six regional
     1.   Supervise the States' administration of the program.
     2.   Administer the program for private schools in those
          States where the State educational agencies are pro-
          hibited from disbursing funds to private schools.
     3.   Distribute commodities to the States and private
          s~hools   where applicable.

     4.   Review State and local school ol?erations.
     5.   Apportion funds to the States.
     6.   Provide technical and administrative assistance to
     7.   Fund the Food and Nutrition Information and Educa-
          tional Materials Center at the National Agricultural
     8.   Set standards for nutritious   me~ls.

l/USDA officials said the study will be issued in the latter
.- part of 1977.

     At the State level, educational agencies administer the
program in public schools, and private schools where per-
mitted.   Each age ncy:

     1.   Submits an annual State plan of child nutrition
          operations for FNS a~proval.
     2.   Establishes a system of accounting under which
          school food authorities will report program infor-
     3.   Maintains current records on schools' operations and
          accounts for program funds.
     4.   Oetermines whether the matching requirements of the
          act are being met.
     5.   Provides sUgervisory assistance to   loc~l   schools.
     6.   Provides the schools with monthly information on
          foods determined by USDA to be in plentiful supply.
     7.   Investigates complaints.

     At the local level, schools or school districts operate
the program and determine which stud~nts are eligible for the
free or reduced-price lunches. In order to receive r deral
funds each school:
     1.   Ooerates on a nonprofit basis and observes limita-
          tions on the use of program funds.
     2.   Serves lunches meeting the minimum nutritional re-
          quirements as prescribed by the Secretary of

     3.   Offers lunch to all children attending school.
     4.   Provides free and reduced-price lunches for children
          from families with incomes below the applicable
          guidelines prescribed in legislation.
     5.   Complies with all requirements of the Civil Rights
          Act and related program regulations.
     6.   Purchases, and uses to the extent possible, com-
          modities designated as being in abundance, and foods
          donated by USDA.

     7.    Maintains full, accurate records for supporting
           r.eimbursement cTaims.

      A consid~rabl~ amount of research has been done on
topic·s either directly or indi.rectly affecting NSLP. Some
of the r",search wa.s flawed, but many repor ts contain impor-
tant inforfl:lation 6'ri program   perform~nce.   The information,
however, has generally not been evaluated and assembled in
a synthesized form for uSe in determining program policies.

     The purpose of this study was to scrutinize and orga-
nize available research in a way that would be useful for
committees to consider in their oversight functions and re-
sponsibilities under the Congressional Budget and Impound-
ment Control Act of 1974 (Public Law 93-344). However, in
a program as diverse as NSLP, some study limitations had to
be made to fit within the manageable context of project re-
     In this regard, we focused on what we believed to be
the principal issue of an NSLP evaluation--the program's
effectiveness in meeting its stated legislative objectives
(safeguarding health and increasing food demand). Other
aspects of the program, such as its economic impact on
localities and its relationship to income maintenance pro-
grams, were e.:cluded from the scope of our work.
      In addition to using available evaluation studies and
research reports, we also interviewed a number of persons
knowledgeable in the fields of nutrition and NSLP and dis-
cussed USDA's current and projected research on child nu-
trition with officials of FNS.
     Our report uses the sequence shown on the next page to
present a five-part study of NSLP.  Parts I and II focus on the
legislative goals of the program ana whether or not existing pro-
gram policies and procedures contribute to the attainment of
those goals. Parts III and IV focus on the program's ability
to encourage student participation and to achieve
cost-effective operation. Finally, part V brings these
independent findings together and, in context, presents an
evaluation synthesis of NSLP overall.

               Legislative Objectives of NSLP
Safeguard health and well-             Encourage domestic
being of the Nation's children         consumption of nutritious
                                       agricultural commodities
                                       and other food

            Part I
          I\bilib' to
       Saf~uard Health
                                              Part II
                                     AgrIcultural Objectives

-What constitutes good               -Do children consume more
 nutrition?                           commodities under NSLP
-Does improved nutrition con-         than if it did not exist?
 tribute to good health?             -Does NSLP's consumption
-Can the effectiveness of a           of commodities assist the
 nutrition intervention               Nation's agricultural
 program be evaluated?                economy?
-What are the characteristics        -Do NSLP' s ag r icul tur al
 of nutritionally needy               provisions contribute to
 children?                            nutritional objectives?
-Does NSLP contribute to
 good health?

                             Part III
                         Program Covera~
    -What factors influence participation?
    -What do we know about nonparticipants?
    -Does nonparticipation jeopardize schoolchild health?

                             Part IV
                          Proqram Costs
     -Do USDA commodity distributions provide cost savings?
     -What are the factors affecting program cost growth?
     -Is the present system of Federal assistance effective
      in encouraging student participation?

                          •Part V
                  Program Evaluation Issues

                             PART I
     Although several studies have been done on NSLP, there is
presently still a lack of an adequate evaluation of the pro-
gram's ability to safeguard health. The impact of the program
has not been isolated from external factors. As a result, we
don't know for sure whether the program is having a favorable,
neutral, or adverse impact on children's health. However, be-
fore proceeding, it is important to note that tne terms
"health," "nutrition," and nnutritional status U are not synony-
mous. Their definitions, as used in this report, are:
     --Health, freedom from disease or ailment.
     --Nutrition, the process by which plants and animals take
       in and utilize food materials.
     --Nutritional status, the condition of an individual's
       health as influenced by the intake and utilization of
        For example, the intake and utilization of protein,
vitamins, and other nutrients can be acceptable (implying
good nutritional status), while at the same time dietary de-
ficiencies in nonnutrient fiber (a nutrition problem) may,
at. least in the opinion of some authorities, increase the
risk of bowel cancer (a health problem).
    (Technical note: Present knowledge of nutrition-health
relationships is incomplete. It is not definitely known, for
example, how low the intake of iron can be without affecting
health or growth and development. Similarly, there are a
number of hypothesized relationships now under stUdy (such
as a link between fiber deficiencies and bowel cancer) which
have yet to be confirmed or denied. Nevertheless, authori-
ties do have opinions on many of these issues. Several such
opinions alleging specific nutrition-disease felationships
are cited, as opinions or as what some authorities believe,
in this report. The reader is cautioned that such statements
cannot be considered scientific fact.)
     Because NSLP is intended to safeguard schoolchildren's
health and because the generally accepted associations between
nutrition and health may conceal important technical differ-
ences, we believe the definitions given above should be kept
in mind when considering the following evaluative issues.

     --What is good nutrition and how does it affect health?

     --What are the nutr ition problems to be countered by NSLP?

     --Can NSLP, throuqh one-sixth of a student's annual meals
       and classroom instruction, produce a quantifiable im-
       provement in some index of schoolchild health?

     Chapters 2 thrnugh 4 focus on the above questions. Chap-
ter 2 provides insights into the complex relationshi~s between
nutrition and health. Chapter 3 presents findings from three
of the Nation's major nutrition surveys as a means of defining
the schoolchild's nutrition problem(s) and determining the
diet modifications needed to safeguard health. Chapter 4 de-
scribes previous attempts to evaluate NSLP's health impact and
explain some of the factors that complicate such an evaluation.

                          CHAPTER 2
                    NUTRITION AND HEALTH?
     While it has been generally accepted that many Americans
suffer from health problems which can be ameliorated through
diet modification; it is important to recognize that our nu-
trition problems are not the same as those found in less de-
veloped nations. The modifications required to achieve opti-
mum nutrition may be very different.
     Primary malnutrition and lack of food pose serious
health problems in many countries of the world. Since the
diets of these peoples are insufficient to satisfy physio-
logical needs and combat disease, higher intakes of the es-
sential nutrients (vitamin C, protein, etc.) have often
directly benefited health. In the United States, however,
extreme undernutrition is rare and the classical deficiency
diseases (e.g., scurvy, beriberi) are virtually nonexistent.
Even so, many authorities consider nutrition (the lack, ex-
cess, or quality of it) a major public health problem.
     Because the National School Lunch program's design is
based on the premise that nutritious lunches will safeguard
children's health, and since different opinions exist as to
the specific aspects of nutrition which have the greatest
health impact, this chapter explores the following topics.
     --What is good nutrition?

     --What are the benefits of improved nutrition?
     --What is the health threat?

     Published in 1943, the first edition of Recommended
Dietary Allowances provided standards for gooo-nutrftfon. As
knowledge has ImProved, the standards have been refined. In
the Food and Nutrition Board's latest release, recommended
dietary allowances (RDA) are described as the levels of intake
of essential nutrients considered adequate for meeting the .
known nutritional needs of practically all healthy
persons. 1/, 2/ These standards are used by nutritionists,
physicians, dIetitians, consumers, and NSLP, whose nutri-
tional target for each meal is to approximate one-third of
the RDA. Recognizing the RDA's merits, the Board also
pointed out some limitations and difficulties in establishing
universal nutrition standards. It explained:
     --While a diet made up of ordinary foods meeting the RDA
       standard should maintain health, present knowledge of
       nutritional needs is incomplete. The requirements for
       many nutrients have not been set. Because of unrecog-
       nized needs RDA should be provided from as wide a se-
       lection of foods as practicable.
     --RDA are established for healthy people and do not
        give any consideration to special needs because of
        infections, disease, metabolic disorders, or other
        factors requiring special diets.
     --Nutritional requirements differ with age, sex, body
       size. physiological state, and genetic makeup.
     --RDA's are estimates of acceptable daily nutrient in-
        takes in the sense that although the needs of most
        individuals will be less than the RDA standard, there
       will be some who require more. For example, the
       Board believes that most nutrients can be tolerated
       well in advance of allowances by 2 or 3 t'mes.

!/National Academy of Sciences, Food and Nutrition Board,
  National Research Council, ~~~~~~~nded Dietary
  ~!!~~~~, 8th Ed., Wash., D.C., 1974.

2/Essential RDA nutrients considered are calories; protein;
- vitamins A, D, E, B6, and B12; ascorbic acid (vitamin Cl;
  and folacin, niacin, riboflavin, thiamin, calcium, phos-
  phorus, iodine, iron, magnesium, and zinc.

       However, excessive intake of calories is undesirable,
       for -it leads to obesity and excessive intakes of
       vitamins A and D, and certain trace elements can be
       tox ic-.

     --Allowances are frequently estimated on limited infor-
       mation because experiments on humans are costly and
       often of long duration; certain types of experiments
       are not possible for ethical reasons; and only a small
       number of persons can usually be studied in a single

     --There is not always agreement as to the criteria that
       should be used to establish requirements.
     Although many authorities believe improved nutrition
will greatly reduce medical expenses and will enable the gen-
eral public to enjoy many health benefits, including longer,
more active 1 ives, it should also be noted that
     "In general, our approach to good nutrition has been to
      provide the recommended dietary allowance of nutrients
      for everyone and more recently to restrict excess ca-
      loric intake to reduce the risk for certain diseases.
      Even though this approach through public health mea-
      sures and education has been extremely beneficial, it
      is inadequate in providing optimal nutrition for the
      individual." !!
      In contrast to a diet deficient in RDA nutrients, some
authorities believe the major diet problems are those of
over consuming certain foods.   They say the population is
confronted with a whole new spectrum of diseases in which
nutritional factors either are the prime cause or else are
highly contributory to the development of a disease state.
These diseases include:
     --Heart and allied diseases, together with diabetes
       mellitus, in which high intakes of calories and cho-
       lesterol may be a contributing dietary factor.

l/u.S. Department of Health, Education, and Welfare (HEW),
- Public Health Service, Report of the President's Bio-
  medical Research Panel:--XppendlX A, The Place-of-BIo-
  medfcal science In- Medlclne and the State-or the scIence,
  HEW Pub:-I05) 76=sOI:-Apr:-I976:------------------------

     --Hypertension (high blood pressure),    ~robably   related
       to diets high in salt.
     --Bowel cancer, apparently caused by fiberless diets.
     --Obesity, caused by overeating and the lack of exercise.
     --Liver disease, caused by excessive usage of alcohol.
     --Tooth decay, caused by high intakes of sugar.
     A 1971 USDA report 11 estimated potential savings from
improved diets (e.g., reauctions in absenteeism and medical
and dental expenses) to be in excess or $12.1 billion an-
nually. (See table 2.1.) It also estimated that improved
diets would reduce the incidence of obesity, cancer, diabetes,
and respiratory and infectious diseases by 80. 20, 50, and 20
percent, respectively. Many benefits, including improved
work efficiency and learning ability, were not expressed in
dollar savings.
                            Table 2.1
                            - -----

Nutrition-related                Magnitude      Poten tial sav ings
~~~~~~_~~oble~~__                of loss        !E~~!~£roved ~ie!

Heart and vasculatory                $31.6                $ 6.3
Arthritis                              3.6                   .9
Den tal heal th                        6.5                  3.2
Alcoholism                              2.0                  .7
Digestive                               4.2                 1.0
    Total                           ~47.9                 $12.1
a/The omission of many qualitative benefits causes a sizeable
- understatement of potential savings.

l/Weir, C. E., An Evaluation of Research in the United States
- on Human NutrItion:-Report NO:-2: Beneflts from Human
  NurtIfion~esearcn:    Agrlculture-Research-SerVlce~-USDA,
       ImprOved nutrition may already be providing health bene-
fits. Dr. J,. Stamler, Professor of Cardiology at Norhtwestern
University', recently commented on the decline in death rates
frain' he'art disease of middle-aged men in the United States. II
Fr'om 1968 'to 1972, the coronary death rate dropped 8.7 per- -
cent for white men aged 35 to 64.\ Downward trends were also
noted for black men and for all women in the same age group.
These findings reflect a reversal of trends which had been
increasing since 1940. Although the precise cause of this
reversal is unknown, Dr.\ Stamler believes the major influ-
ences are a [educ~ion in cigarette smoking, less incidence of
high blood pressure, and improved eating habits; for example,
less intake of saturated fats from animal sources and in-
creased intake of polyunsaturated fats from vegetable sources.
     Some authorities believe that nutrition is a contributing
factor for five diseases included in the 10 leading causes of
death in the United States. (See table 2.2 on next page.)
Overconsumption recoqnized

     The health threat from overconsuming certain foods
(e.g., cardiovascular disease, obesity, tooth decay) is found
at all income levels of American society. Many of these prob-
lems are easily recognized and preventable.
     In 1972 the Food and Nutrition Board of the National
A=ademy of Sciences and the Council on Foods and Nutrition of
the Arner ican Medic,al Association issued a joint statement on

!/"The Recent Decline in De,ath Rates from Premature Coronary
  Heart Disease in the United States,'l address before
  American Heart Association's Science Writers Forum, Jan.

                           Table 2.2

                                       Death rate
                                       per 100,000    Percent of
 Rank and cause of death               Eopulation    total deaths
--             ------                                ------
 1.   Diseases of heart (note b)          353.1          38.6
 2.   Malignant neoplasms, including
      neoplasms of lymphatic and
      hematopoietic tissues               169.5          18.5
 3.   Cerebrovascular diseases
      (note b)                             97.2          10.6
 4.   Acc iden ts                          48.9           5.3
 5.   Influenza and pneumonia              25.7           2.8
 6.   Diabetes mellitus (note b)           17.4           1.9
 7.   Cirrhosis of the liver
      (note b)                             16.0           1.8
 8.   Arteriosclerosis (note b)            15.2           1.7
 9.   Certain causes of mortality
      in early infancy                     13.2          1.4
10.   Suicide                              12.5          1.4
      All other causes                    145.7         16.0
                                          ----          ---
           Total                          914.4        100.0
a/Source: HEW, MonthlX Vital Statistics R~ort. Provisional
- Statistics, Annual summary-ror-Ehe-UnIted States, 1974;
  vol. 23, No. 13, May 1975.
b/Some authorities believe that nutritional factors contri-
- bute to the onset or severity of this disease.

diet and coronary heart disease. 11 The statement indicated
that although investigations have-identified a number of
"ri"sk factors," inc"lud.i'ng some which can be modified by diet,
not enough evfdence exists to quantify the benefits that may
come from modifying the individual factors of this set. There
was, however~ enough evidence to recommend that measurement of
plasma lipids (such as cholesterol) be included as a routine
part of physical examina-tions and that persons in a   It[   isk
categorylf receive appropr iate dietary advice 6

     The White House ~0nference on Food, Nutrition, and Health
indicated concern about excessive consumption of calories by
schoolchildren. Their final report stated, "For obese chil-
dren of age 12, the odds against being normal weight adults
are 4 to 1 and if weight reduction does not occur by the end
of adolescence, the odds rise to 28 to 1." 21 To avoid card-
iovascular problems, other reports suggest encouraging good
nutrition and weight control beginning at birth. The first
changes in the vascular system may occur by the age of 3.
although coronary heart diseases may not be diagnosed until
the fortieth year of life.
Deficiencies in RDA nutrition
~~t a~=arrect-hearth

     Many authorities believe RDA deficiencies, some of which
can be combated by vitamin and mineral supplementation, have
an important effect on health, development, and growth. But
perhaps because extreme malnutrition is rare in America, it
has been difficult to demonstrate the health impact of slight
RDA deficiencies. Therefore, while it may be possible to cor-
rect RDA deficiencies and improve one's nutritional status,
the precise health impact of an improvement in nutritional
status has not been completely identified.
     Some school food service directors have noted better
mental performance by participants in the school lunch and
breakfast programs, but the comparative results of these
programs have not been documented in learning and behavior
in school settings. As a result of limited studies, the

l/Food and Nutrition Board, National Academy of Sciences,
- National Research Council, ~~~~_~~~ Coronary Hea!!_P!~~~~~'
  Wash., D.C., July 1972.
2/White House Conference on Food, Nutrition and Health,
- Final Report, Dec. 1969.

National Academy of Sciences found that investigations have
not completely addressed the relative importance of malnu-
trition versus social-environmental factors on intellectual
development. Both have been found to be significant. 1/
Despite serious methodological shortcomings in the studies
that have been made (e.g., difficulties in isolating a
portion of any observed changes to nutritional factors alone),
the Academy noted that the evidence indicated that early and
seVQre malnutrition is an important factor in later intellec-
tual development-and that it was above and beyond the effects
of social-familial influences. The Academy was less certain
about the effects that mild to moderate protein-calorie
malnutrition, or chronic sUbnutrition, had on later intellec-
tual development. -AS a case -ut-point, anemia. in the pre-
school years appears to adversely affect motivation and abil-
ity to concentrate for extended periods of time.
     A 1968 USDA food consumption report noted that the per
capita consumption of vitamin A and ascorbic acid decreased
considerably between 1945 and 1966. In 1971, Dr. Murray of
the Canadian Food and Drug Directorate amplified this finding.
Going beyond the capabilities of food consumption reports and
the biochemical evaluations of nutrition surveys (which
measure serum vitamin A in the blood), Dr. Murray focused on
the autopsy examination of livers. He found that (1) many
people did not have any vitamin A in their livers at ~eath
and (2) an even greater number had very little vitamin A
liver stores. These findings, which were subsequently con-
firmed for the United States population, led Dr. Murray to
sta te :
     "Now, as far as I know, there is nothing decisive about
     liver stores of vitamin A.  It has never been demon-
     strated that there is any direct immediate advantage
     to having a reserve of vitamin A, nor any direct im-
     mediate disadvantage in not having liver stores. It
     would be foolish, however, to be unconcerned at their
     absence in a substantial proportion of the population.
     I t is difficult to believe that a diet which permits
     the dissipation of liver vitamin A will, in every case,
     be sufficient to maintain blood levels." ~I

l/Food and Nutrition Board, National Academy of Sciences,
- National Research Council, The Re1ati~nsh~_2!_~~~~~~~~~
  !~Br~~~_~~~~E~~~t and Beh~~£I' Wash., D.C., June 1973.

3/"Vitamin A Nutriture inNorth America,"   ~!:.~£ee~~2..0f    the
     A gene~al relationship between adult stature and socio-
economic status has- often been reported. A New York study
noted that infants of poor mothers were 15 percent smaller
than other infants. The extent of permanent stunting due to
early malnutr-ition dep',lds on factors such as timing in re-
spect to growth periods and the duration of malnutrition.
Short stature has been associated with increased risks dur ing
childbir th.
     As a nutrition intervention program, NSLP seeks to pre-
vent rather than to cure a disease state. Its health impact
usually does not become apparent for a long time, and then
only in comparison with what otherwise might have been.
     It is difficult to measure the school lunch's effective-
ness in terms of a discrete change in a participant's health.
Many health effects do not become apparent for years. but
short-term studies (e.g., spanning a school year) which com-
pare the nutrition-health relationships between participants
and nonparticipants may provide insights to the health impact
of the program. (See ch. 4.) This applies to studies assess-
ing both
     --health parameters responsive to short-term change
       (e.g., the designers of an NSLP evaluation should con-
       sider the feasibility of detecting the program's in-·
       fluence on features such as: the incidence and dura-
       tion of illness, obesity, tooth decay, periaontal ais-
       ease, etc.) and

     --nutritional status 1/ (which, while not a direct mea-
       surement of current-health, is believed to have long-
       term health consequences).
     The design of the NSLP lunch warrants special attention
throughout this report. Although program regulations require
each lunch to approXdmate one-third of the RDA, it should be
remembered that the lunch is but a supplement to the home

l/Nutritional status is defined as the condition of an
- individual's health as influenced by intake and utilization
  of nutrients, determined from the correlation of informa-
  tion obtained from physical, biochemical and dietary

diet. Its effectiveness should be considered in terms of
how well it fulfills the actual diet needs of children.
     Schoolchildren exhibit a variety of nutritional condi-
tions (e.g.; some are well nourished, some are underfed,
and some fall short in RDA nut,riture and/or overeat the
wrong foods). A meal designed to reinforce the caloric
(energy) intake of underfed children could, in addition to
improving the nutriture of these children, have undesired
side effects on the health of those who overeat or who are
already well nourished (e.g., promote obesity).

                             CHAPTER 3
                        NUTRITION PROBLEM
      To examine how a National School· Lunch Program meal can
best supplement children's diets and to see i f the free and
reduced-price program is targeted to reach those in greatest
need of nutritional assistance, we focused on two questions:

      --What are the specific nutritional problems affecting
        schoolchildren (e.g., the types of problems, and the
        proportion of schoolchildren affected)?
      --Can "nutritionally at risk" children be identified on
        the basis of vlsual or socioeconomic characteristics?

     Although more research is needed to completely answer
either question, many insights can ae gained by comparing the
findings of three surveys that evaluated nutritional status
in large segments of the U.S. populatior,. Pertinent
findings--and their implications for NSLP--are presented in
this chapter.
--------            ----------
     The purpose of a survey of nutritional status is to
assess health as influenced by the intake and utilization
of nutrients.     Accordingly, its design may differ from that
of a health survey (e.g., one which emphasizes a direct
measurement of health by determining the presence of com-
municable diseases, cancer, etc.).       A survey of nutritional
status usually limits direct observations of health to
features such as signs of the classical deficiency diseases,
dental health, and anthropometric considerations. 1/       Its
principal thrust is to obtain information about inairect
health parameters such as iron nutriture.
      Since there is little evidence of the classical de-
ficiency diseases in the United States, the methods used
in assessing nutritional status are based on the assump-
tion that most of the malnutrition encountered will be
early subclinical malnutrition with or without physical

l/The study of human body measurements on a comparative basis
-   (e.g., head circumference, height, weight).

signs. These methods require extensive coordinated sur'veys
to obtain dietary intake data, biochemical specimens, clini-
cal examinations, and anthropometric measurements. Even with
these provisions, the interpretation of nutritional status
remains complicated. For example:
     "Standards of what constitutes good health and adequate
     nutritional status have not been precisely defined by
     medical and other scientjfic research. It is not defin-
     itely known, for example, how Iowa hemoglobin value
     can be without affecting health or growth and develop-
     ment. Similarly, the ideal growth rate for children,
     compatible with the longest useful and healthy life,
     has not been definitely determined." !/
     "* * * to our knowledge there have been no defin-itive
     efforts to develop a system f~r 'rating' or 'scoring'
     the nutritional status of populations or individuals,
     We do not have available any substantive data which
     allow us to group dietary, clinical or biochemical
     data (weighted or unweighted) to signify the degree
     of subclinical malnutrition. Certainly, the presence
     of two or more 'unacceptable' biochemical indices in
     some segment of a population does not necessarily denote
     severity of malnutrition or even of a greater potential
     for malnutr ition to develop at some future time." 't./
     The Ten-State Nutrition Survey (TSNS),3/ previously
called the National Nutrition Survey, was tne first compre-
hensive study ever developed for evaluating the nutritional

l/U.S., General Accounting Office,"Observabions on Evaluation
- of the Special Supplemental Food program"(B-176994), Wash.,
  D.C., Dec. 1974.
2jOwen, G. M., Kram, D. M., Garry, P. J., Lowe, J. E., and
- LUbin, A. H., "A Study of the Nutritional Status of Pre-
  schoOl Children in the United States, 1968-1970,"
  ~~~~~ri~, Supplement to Vol. 53, No.4, Apr. 1974.

3/U. S. Depar tment of Heal th, Education, and WeI fare, .Heal th
- Services and Mental Health Administration, Ten-State
  Nutrition Survey, 1968-1970, HEW PUblications-THSM~72­
  rna to 1FHH;Atlanta-;-Ga., 1972.

status of a large segment of the U.s. population. Supervised
by the Department of Health. Education. and Welfare (HEW).
the study gathered data during the period 1968-70 and was
intended to determine the incidence and location of serious
hunger and malnutrition in the United States. S~Yere time
and cost constraints led to deficiencies in obtaining a
representative sample of low-income households, a necessity
for projecting sample findings into the overall U.s. popu-
lation.!/ This shortcoming led us to conclude that
     "* * * the TSNS data should not be considered as reli-
     able estimates of the prevalence o~ serious hunger and
     ~alnutrition in any of the survey populations, including
     members of low-income families. * * * "~/
     Although statistical shortcomings prevent TSNS from
rendering precise estimates about the prevalence of school-
child malnurition. the survey did provide useful information
on the nutritional status of over 14,000 school-age children.
By considering the dominant characteristics of the TSNS
sample (e.g., persons suspected to be at high risk in the
low-income areas of 10 States), looking at the differences
of the children surveyed, and comparing TSNS's finding with
related research, we found a number of points which we be-
lieve have an important impact on NSLp·s effectiveness.
These points can be summarized as follows.
     1. In regard to identifying nutritionally needy
children, TSNS reported:
         --Clinical examinations did not provide a useful
           means of identification.

         --Characteristics of malnutrition were often unique
           to the local situation and specific sUbsegments of

l/Because of constraints of time and money, the study was
- limited to 10 States: Wash., Calif., Tex .• La., S.C., Ky.,
  W.Va., Mich., Mass., and N.Y.   Within each State, TSNS sur-
  veyed families from the Census enumeration districts which
  had the lowest average income (lowest quartile) according
  to the 1960 Census.
2/U.S General Accounting Office,"Evaluation of Efforts to
  Determine Nutritional Health of the U.S. Population,"
  Report B-16403l(3}. Wash., D.C., Nov. 1973, p. 19.

          the population (as social, cultural, and economic
          differences) .
        --Education attainment (years of school completed)
          of the person buying and preparing the famil¥'s
          food was related to the nutritional, status of
          children under the age of 17.
        --Evidence of malnutrition increased a3 income level
          decreased. Within each ethnic group, nutritional
          deficiencies were often more prevalent in the low-
          income-ratio States. 1/ (It should r8 kept in mind
          that TSNS was primarily a study of low-income
          families. The income-malnutrition relationships
          for higher income famil.ies will· be described later
          in this chapter.)
     2. In regard to identifying the nutrition problems
affecting schoolchildren. TSNS reported:
        --Adolescents between the ages of 10 and 16 showed
          the highest prevalence of unsatisfactory nutri-
          tional status.
        --Iron, vitamin A, and riboflavin nutrition was a
          public health problem of medium-to-high importance
          among some groups. Protein, thiamin, vitamin C,
          and iodine nutriture was a low-to-minimal, public
          heal·th problem. (See table 3.1 on following page.)
        --There was evidence that many persons made poor
          food choices that led to inadequate diets and to
          poor use of the money available for food. Families
          seldom used foods rich in vitamin A. placed heavy
          emphasis on meat rather than less expensive protein
          sources, and generally showed low levels of iron
          intake. As shown in table 3.2 on page 26 onl¥ 4
          percent of the l2-l4-year-old males in
          high-income-ratio States had protein intakes below
          80 percent of the RDA; 87 percent consumed over
          120 percent of the RDA for protein.
l/States classified as "low-income-ratio States had more

  than half of the surveyed families living at a "below
  povertyn level~ "high-income-ratio States" had more than
  half of the families living "above poverty."

   --Weight data indicated an excess of both overweight
     and underweight children in all segments of the
     surveyed population. ~Ihi.te male adolescents showed
     a higher prevalence of obesity than black males.
     (See figure 3.1. on p. 26.) These findings were
     consistent with the wide range of caloric intakes
     shown in dietary data.
   --Increased levels of food consum~tion would have
     brought nutrient intakes closer to the RDA stand-
     ards, but these increases might also have provided
     excessive intakes of calories in many segments of
      the population. Foods with better nutrient-to-
     calorie ratios were needed to increase intakes of
      iron, vitamin A, and riboflavin.
   --Tooth decay was found to be closely associated
     with the intake of refined carbohydrates (foods
     with high quantities of sugar such as pastries,
     candies, soft drinks) and the amount that remains
     on thp teeth. Black children in the low-in
     come-ratio States and all qroups of children
      in the high-income-ratio States showed a positive
     association between tooth decay and the
     between-meal consumption of high carbohydrate
     foods. (See the chart on page 27.1

                                 ~lll:.:!Ulr;li~~Ll!.!<;8-1~!OI IIIOt~ ~l

Ethnic 'J,O<Jl!
1\9" (1...>r51             i=r~H- Io=lfi        ~:,·--H~It·-n=n                             ~:~:'~!!H~::£!'{~~
!!~-                 ____ J!ll_~~.!L_!!>!l£.._ _~!!.. __ !'~!:'~!!-_ E!!:                   ~l..!~   _!'£!~ !J__ .:~.!!
                                                           L.. ~-illco~C-r&IIO St&IC~

I',ot .. il\
                                 " - '"
                                          A        A
                                                                 "C -,-,"      "             A               A
'-'it .. ",I" 1\
Vit.o:in C
                                                   r>           0
                                                                        A     '"
                                                                                            -"A              -
                                     8             8
C.O"Oth , d.y.lllP!Hnt
OI>foS   lty
                                   "" 8
                                                   A           "
                                                                        ,-     ':.
                                                                                            .-""            ""     nta

                                                          Nioll-llI('lI~-ltHIO SUI"':

                                                                 A      A      A             A               A     A
lIit ...;n ,.
Vit~",in C
                                    C     "
Iodin ..
G<o~th I
               dov,lo~ ..   nt     .-""   0        "
                                                                .-"'"   J                   .,.
                                                                                                            n/a    "i"
                                                                                             -ff.     ~IQ"
                                                                                              . . . . . .01'.",-
                                                                                               -.     l.,~

~Th ..     51abols   t"o<"s~n~ r.I~~i~..      d"o ... ,s or laportanc"                  -          • 'uni .... l
   as public h.~ltb p.obl .. ~s as d.. t"roir... d by "valuation                        n'a -         -~.    ~v.,hOI'
   of clini"al. ~;"ta.v. "nthrooo....«l".

                                              Table 3.2

           Distribution of Nutrient Intakes for Males (12-14 yrs.)--
                           TSNS (1968 1970) (note a)

                                            Intake groups
                      Low income ratio States             High income-ratio States
                 Intakes below       Intakes over
                      80\       RDA      120\
Nutrient             of RDA    + 20\    of RDA
                            --(percentage distribution within States' group)
Calories                    61      23:!.6                   37       35                                    28
Protein                     12      17     11                  4       9                                    B7
Calcium                     157     19     14                 41      27
                                                                       __                                   32
                            68      +-32~                    57
                                                                                                       4 3 ~

Vitamin A                   68      10     22                60       17                                     23
Thiamin                     49      27     24                38       31                                    31
Riboflavin                  29      28     43                15       20                                    65
Vitamin C                   50      13     37                 42      13                                    45
~/Based on eighth edition of Recommended Dietary Allowances. Actual levels
  of nutrient intake may be somewhat understated in this table. Although
  TSNS gathered limited data on the use of vitamin/mineral supplementsl
  this data was neither integrated with overall levels of intake nor
  presented in the final report. Dietary intake data is based on 24-hour
  recall. Althouah the data reflect only the day of recall, the number of
  persons consuming intakes below the RDA standard suggests that many diets
  are apt to be inadequate over time. Source: Ten-State Nutrition Survey
  1968-1970, DHEW Pub. (HSH) 72-8133.

                                               FIGURE 3.1

                             PERCENT OBESITY IN ADOLESCENTS _ TSHS            Q

      PERCENT OBESE                                         PERCENT OBESE
      J<l                                                   30       FEMALES

      "                                                     "
      20                                                    20

                -      /"
                                                                        p                     \ __ .=.i'


          oL--,--,--,----,-.-----1                               o L ........-...-.......--.-----,_-l
                \,     \3     14    15   16   11                       \,         \3    14        lS   \.   \7
                              AGE                                                       AGE

           o SOURCE,   TSNS 1968-1970, CHEW PUB. (HSM) 72_B131.

                                                FIGURE 3.2

                     OF C"RIOHYDRATE CONSUMED IETWEEN MEALS _ TSNS          0

          LOW IMCOIllE RATIO STATES                             HIGH INCOME RATIO STATES


 3••                                                    3••
 2••                                                    2.0
 1.0_                                                   I ••

   •                                                      •
                                                                    o                  >149

                                                               GRAMS OF CARBOHYDRATE CONSUMED
TEETH                                                TEETH

a SOURCE: TEN.STA.TE NUTRITION SURVEY 1968_1910,                                •   WHITE
                                                                                o   8LACK

                                                                                fJJ SPANISH AMERiCAN
Nutrient intakes increased
by school lunch programs
     In comparing the nutrient intakes of persons who did and
those who did not eat the school lunch, TSNS indicated:ll
        "School lunch programs were found to be a very important
         part of nourishment for many children. Particularly in
         the low-income-ratio states, school lunches contributeo
         a substantial proportion of the total nutrient intake
         of many school children. The contribution of school
         lunch to overall nutrition was particularly important
         among black children.             11

I/Based on the dietary recalls of 4,106 individuals between
- the ages of 10 and 16. Since many respondents did not know
  whether an existing school lunch program was USDA supported,
  the term Itschool lunch" h~s been used to describe any orga-
  nized food service provided at the nooo period except
  vending machine or concession service.

        We believe that TSNS, supported by similar findings in
   other studies, provides a reasonable basis for assuming that
   NSLP participation increases the nutrient intakes of school-
   children. 11 (See table 3.3 below.) However, since average
   intake values conceal wide variations of nutrient intakes,
   and since TSNS did not compare the health and nutritional
   status of these children, we cannot be sure of the school
   lunch's net impact.

                        ~DmOartSO"     or   Me~n.Nutrl~~! Jnta~~        ot   P~!g~~.J~-!hrOU9h~£2~~_g!~~~
                               Partic~ting_!n      School    ~unC~_~!DQr~~S ~!~~_!~!£_~2!_t2~~~

                                                Protein      Calcimll    Iron     Vlt.I:nin"    ThiJ.~in      Riboflavin   ~i"c:in      Vita~in     C
                                    £!l~!!!     .l2l__       _!!!~L_     1.~2.1   _.!.!':"lL:..L. __ l~'!L_   ___ LiI'2L    !....!!!s.L__ __-i!!!i.L.
Nonschool lunch        1~19
    l'Iltbn daily Int.:l\e           1916       6$.6':1        6l)j     11.11       j.l,'j6       1.21                     13.loj         57."6
School lunch 110'11
    Mean dally Intake                2011       76. hI         ':IOJ    11.24       ')2)2         l.1B                     !l.llS         5'.71
    Mean ,ntake contr!-
      buttd by sebool
        lunc~                         6~~       211. J'i       420       L 1)       19'.lJ                      0.8J        -I.6J         20. $1
    Peecent of daily
      1I1ta~t <:'onu lDut·      j
      by school lunch               H.7$        J7.2,)       -16.;1     JLU        38.0'1        31. J$        H.7(1       JJ. JS

~onscbool       luncn IIJJ9
    110/11/\   dllily Intake         2J11       ,hl.li;                                           1.]7          2.02       16.41

School I uncll (11.;
    ~tan Intake                      26Ul      102.tlll                                           1, 42         2. SI      UI.'J1
    SchoOl lunch ~~an
      contrlbuti"n                    676       28.61          ~18       J.2J       lH8           O.l!          U.72        ~.J'J         17016
    Percent contlibuted
        by 5\:1'1001                25.99       27.dO        32.;2      23.22      22.97        21.83          2ll.6tl     23.21          21.1 !I
a/Records reflect nutrient Inta_e of 1 day prior to Interviaw eKcludinq w£ekends and non$cl'lool session.
- Source, Ten~LI1.!!1!..iU~~!L!!~!:.!E2.Dtt£''' Puo. 11lSM) 72-,IlJJ.

   !/In fiscal year 1970. nearly 74 percent of the schoolchildren
         1n these 10 States were enrolled in NSLP schools.

     To analyze how well the school lunch program satisfied
the nutritional needs of children, we compared the average
nutrient content of a lunch with the nutritional problems
reported in the same study, as shown in table 3.1. We found
the following:
    --Mean intakes for all nutrients, except for thiamin
      in the low-incorne-ratio States, were greatest for
      persons who ate the school lunch.
    --School lunches provided roughly one-half of the RDA
      for protein, calcium, riboflavin, and vitamin C~ one-
      third of the RDA for vitamin A; one-fourth of the RDA
      for calories and thiamin; and one-fifth of the RDA
      for iron. 1/
    --In regard to problem nutrients identified in table
      3.1, school lunch programs provided a diet supplement
      1.    Effectively raised the mean intakes of vitamin A
            and riboflavin to RDA standards (e.g., 4000-5000
            IoU. and 1.3-1.8 mg., respectively).
      2.    Used foods with iron-to-calorie ratios below those
            of the child's home diet and were relatively in-
            effective in raising mean iron intakes to RDA
            standards (18 mg.).
       3.   Increased mean calorie intakes, which probably
            benefited growth and development at the cost of
            increased obesity.
    --The strength of nutritional reinforcement bore little
      resemblance to need.    Protein supplementation, for
      example, a relatively expensive food source and one
      for which mean intakes were well in excess of the RDA
      standard, was greater than that provided for iron
       (raising questions as to how much consideration was
      given to need in the design of the school lunch).

l/since the TSNS sample was not categorized by age and sex,
  a precise statement of RDA requirements is not possible.
  Our estimates are based on a uniform age-sex distribution.

      The Preschool Nutrition Survey (PNS) 1/, 2/ was designed
to provide an overview of descriptive data-on the nutritional
status of a cross-sectional sample of preschool children.
This study was limited to a specific age group; it sampled
from a broader geographic and income base than that used in
     The PNS designers reasoned that while an adequate income
might provide the opportunity to eat well, it did not auto-
matically insure a nutritious diet. For this reason, PNS
used an index of socioeconomic status (SES) 3/ other than
income to compare relationships between nutritional status
and the overall lifestyle of a household.
     In describing eating practices and food preparation in
the household, PNS indicated:
     --Poor families did not spare meat in children's diets.
       Despite differences in the money available for food,
       total consumption of meat and poultry varied little
       by SES.
     --Mothers in higher SES groups indicated they enjoyed
       cooking, frequently tried new foods and used printed
       recipes. At lower SES levels, someone other than the
       mother assumed a greater share of the responsibilities
       for procurement and preparation of the family's food.

!/See footnote 2, p. 22.)
2/PNS was sponsored by the Maternal and Child Health Service,
- HEW.
3/Warner Index Status Characteristics was used, which is
  based on ratings of occupation, source of income, dwelling
  type and dwelling area. While per capita income and SES
  were generally related, inconsistencies occurred predomin-
  antly in the highest and lowest income groups. Where
  these inconsistencies were noted, dietary intakes and
  biochemical indices of children were generally more in
  keeping with SES than with income.

       These families made less frequent use of cookbooks
       and tended to rely on neighbors and relatives for
       information about food.
     --Higher SES groups tended to have more established
       daily ~atin9 patterns, were less permissive in
       catering to childrenJs-food preferences, and showed
       less tendency to use food as a means of reward or
     --Relatively few children had nothing to eat in th~
       morning, although with increasing age more children
       ostensibly prepared their own breakfasts at least some
       of the time.
     Comparing the relationships between SES and children's
diets, the study found that:
     --The percentage of children using vitamin/mineral
       supplements tended to increase as SES improved and to
       decrease with advancing age. Of 3,441 children sur-
       veyed, 1,731 took supplements. The majority of these
       children used mUltivitamin preparations; 486 took
       preparations containing iron; and only 15 used prep-
       arations containing calcium.
     --~lthough children in the lowest SES group consistently
       consumed less food than other children, there was
       little difference between SES groups with respect to
       average nutritive quality of diets (i.e., the nutrient-
       to-calorie relationships were about the same).
     --~sSES increased, fruits contributed progressively
      more energy and nutrients; vegetables generally con-
      tributed less.
    --Cereal grains were a major source of iron and calories.
      ~s SES increased, the amount of energy contributed by
       breads, cereals, pastas, etc., declined; and energy
      derived from.cakes, cookies, sweet rolls, etc., in-
      creased. The proportion of iron provided by cereals
      increased with SES and appeared to reflect the con-
      sumption of heavily iron-fortified breakfast cereals.
     --"People purchase and consume food to meet energy needs
       and when income is limited, there is less likelihood
       of buying foods such as fruIts which are relatively

       expensive sources of energy, but happen to be good
       sources of ascorbic acid."

     The nutritional quality of the diet for most nutrients,
therefore, varied little by socioeconomic group_ However,
because evidence that "nutritional risk" (i.e., lower dietary
intakes, lower biochemical indices, and smaller physical size
for age) was clustered among preschool children of lower SES,
PNS concluded that:
     "Having examined dietary, clinical, anthropometric and
     biochernica~data in some detail, * * * the major
     nutritional problem confronting those children 'nutri-
     ionally at risk' was insufficiency of food.'1

     In 1969, while the earlier studies were being performed,
the Secretary of HEW established a National Nutrition Surveil-
lance System. That system--the Health and Nutrition Examin-
ation Survey (HANES)--was intended to examine, in continuing
3-year cycles, a sample of the civilian, noninstitutional
population between the ages of 1 and 74 who reside in the
contiguous 48 States.
     The HANES sample design, which was developed jointly by
the Bureau of Census and the National Center for Health
Statistics, established explicit quality control requirements
so that sample results would be capable of providing reliable
estimates of nutritional status in the United States. At the
same time, probabilistic design features were included to
permit more detailed analysis of data for certain high-risk
groups--namely, the poor, preschool children, women of child-
bearing ages, and the elderly. The first examination cycle,
which used nutritional assessment methods nearly identical
to those employed in TSNS, examined about 20,000 Americans
from 1970 to 1974. The second examination cycle is scheduled
to begin in the first half of 1976.
     While it has been reported that clinicians in the United
States are occasionally confronted with cases of overt malnu-
trition, the first HANES cycle did not find any evidence of
such deficiencies in the surveyed population. 1/ It therefore

!/Habicht, J. P., ~ Nutrition_~!fect_£!~~-~-~~
  Pr~£~~, PresentatIon for Food and NutrItIon Serv1ce,
  USDA Executive Retreat, Feb. 1975.

appears that severe malnutrition does oc~ur, but is suffic-
iently rare as to avoid detection in HANbS.
     A comprehensive analysis of this data is not
yet available; however, HEW published a preliminary
report in 1974 which was based on a representative subset
of the total sample.l! That subsample was more closely
representative of the civilian noninstitutionalized population
of the United States than any previous survey of nutritional
     The HANES preliminary report compared the mean intakes
of selected nutrients by age, sex, race, and income groups.
Major findings included (see table 3.4 on following page):
     --White persons in the income group above poverty level
       had the highest caloric intakes, while blacks in the
       lower income group had the lowest intakes. Substan-
       tial numbers of individuals had low caloric intakes.
       (A more meaningful analysis of the prevalence of
       under- and overweight children should be presented
       in the survey's final reports.)
     --Mean protein intakes for all population subgroups
       exceeded dietary standards. Although white persons
       had the highest overall intakes, mean protein intakes
       per 1,000 calories showed little or no variation by
       race or income (e.g., protein consumption was closely
       related to caloric intake).
     --In ali age groups and at both income levels, the
       mean calcium values were consistently higher for
       whites than blacks. However, since mean calcium in-
       takes per 1,000 calories were essentially the same in
       all subgroups, the differences in calcium values
       were primarily due to higher caloric intakes.
     --Mean vitamin C intakes were adequate in all population
       subgroups. vitamin C intakes per 1,000 calories were
       higher for blacks than for whites, indicating that
       differences in vitamin C intakes between these groups

l!HEW, National Center for Health Statistics, preliminary
   Findings of the First Health and Nutrition ExaminatIOn
   ~~£vey; ~nited_~~at~~-l~71~1972: Dletary Intake and
   Bi~~mical Find~~, DHEW Pub.      (HRA) 74-1219 1,
   Rockville, Md., Jan. 1974.

  were more related to food choice than to total food

--Mean vitamin A intakes were lowest in adolescents.
  Variations in intake levels between subgroups were
  related to choice of nutrient consumed rather than
  total caloric intake.
--Iron intakes were below standard for adolescents                                and
  for black children aged 6-11 in the lower income                                group.
  Mean iron intakes per 1,000 calories were higher                                for
  blacks in most age groups than for white persons                                re-
  gardless of income.

                                     Ta,ble 3.4

              esti~ated "ean Nutrient Intakes as a Percent of RDA
             for 6-11 and 12-17-Year-Olds: United States, 1911-72
                           (HANES Pcelilllinaryl (note a)

 (note b)         Inco~e   below poverty level       Income above poverty level
                     WhIte           BJ~ck               W61 ttl        Black

                     --·-------{intakes as a percent of RDA)---------
 Calories              as               72                    91          83
 Protein              211             171                    21'         197
 CalciUM              11.              87                    133          95
                      ,.."             "                      "           95
 Vitardo A            130             114                    121         112
 Vitamin C                            162                    19'         182
                                         12 17-y'ar-olds
 Calor les             81                                                 8.
                                      14.                     "
                                                              08          .,
 VitalDin A

 Vitamin C            158             164                    177         165
 a/Intake standards Are based on the eighth edition o( Recommended
 - Dietary Allowances and a uniform age-sex distribution wlthln-rhe
   6-11 and 12-li age groups. Since RDA standards vary by age and
   sex and since these characteristics have not been reported tor
   the HANES sample, values shown 1n this table are SUbject to
   si2eable error. They should be used only as an order of magni-
   tude esti~ate in determining the adequacy of nutrient intakes.
   Source: Preliminary Findings of the First Health and Nutrition
   Exa~ination Survey, DHEW Pu . (HRAl 74-1219-1.

 b/Dietary intake data does not include the etfects ot             vita~in/
 - ~ineral supple~nts.

     HANES found that 22 percent of the subsample took
vitamin/mineral supplements on a regular basis and that an-
other 10 percent took supplements irregularly. But, like
the TSNS, its dietary recall process was limited to askir
whether or not a vitamin or mineral supplement was used.
As a result, the nutrient values added by supplements were
not included in dietary intake values and, for the prelim-
inary analysis, the sample's size precluded evaluation of
how vitamin/mineral supplements affected dietary and bio-
chemical findings.
     The HANES study also analyzed biochemical indices for
6-11 and 12-17-year-olds. Biochemical values classified as
"low" were used to identify groups of people who were more
likely to be "at risk" of developing deficiency diseases.
Data collected by race, age, and income group indicated
the following (see table 3.5 on the following page):
     --There was evidence of iron deficiency with anemia
       as measured by the number of children with low
       hemoglobin, hematocrit, serum iron, and transferrin
       saturation levels~
     --In the 12-l7-year age group, t:le percent of low
       values for hemoglobin and hematocrit was 3 to 6 times
       higher in blacks than in whites and did not appear
       to be associated with Incom r   Similar differences
       were observed for serum iron and transferrin satura-
       tion, though with a lower magnitude.
     --A high proportion of low transferrin saturation values
        (a measure of iron stores) was found in all ~ ?ulation
       subgroups. For the 6-11 age group, low values were
       most prevalent among white children. In adolescents,
       however, the proportion of low values was greatest
       among blacks.
     --No low serum albumin values were observed in children
       aged 6-17.
     --Although white children had a greater percentage of
       low serum protein values than black children, there
       was no clear-cut evidence of nutritional protein
     --Low Serum vitamin A levels were noted mostly among
        white low-income children between the ages of 6 and 11.
       Vitamin A deficiencies, as measured by mean serum

        vitamin A levels, were found to decrease with age
        in all race and income groups.

                                   Table 3.5

           Low Bioche~ical Indices fQr 6-11 and 12-17-Year-Olds:
            United States, 1971-72 (HANES Preliminary) (note a)

                        Income below poverty        Income above poverty.
Biochemical test          level (note b)               level (note b)
    (note c)              whit~    black               white    black          Total
                          ---{percent of low values by population group)---
                                                 6-11-   ·.'~ar-olds

Hemoglobin               0.96        7.06                  1.55         7.58    2.59
Hematocrit               2.21        3.Bl                  2.65         8.08    3.08
Serum iron               3.69        1.95                  2.24         2.73    2.37
Transferc in saturation 17.66        8.89                 11.06         9.99   11.63
Serum protein            1.08        0.00                  4.13         1.57     (dl
Serum albumin            0.00        0.00                  0.00         0.00    0.00
Serum vitamin A          2.94        1.49                  0.25         0.69    0.73
Hemoglobin                 3.67     20.40                  2.51        15.02    4.68
Hematocr i t               6.82     27.79                  6.71        18.56    9.01
Serum iron                 1.94      6.04                  1.61         3.41    1.97
Transfenin saturation      6.78     12.54                  6.33         7.39    6.49
Serum protein              1.56      0.00                  3.23         0.19    2.69
Serum albumin              0.00      0.00                  0.00         0.00    0.00
Serum vitamin A            0.00      0.00                  0.22         0.21    0.17
a/Source: Preliminary Findings of the First Health and Nutrition
- Examination Survey, DHEW Pub. (HRAI 74-1219-1.
!!/Exc!udes personS .... ith unknown incomes.
£/Low biochemical indexes indicate the prevalence of groups who are
  more likely to be at risk of developing deficiency diseases. Hemo-
  globin and hematocrit are used to measure iron deficiency ane~ia.
   Both measurements are general rather than specific indicators of the
  cause of anemia and there is a close relationship between the tvo.
   Serum iron and transferrin saturation measurements give some indication
  of the amount of iron present in the blood. Serum protein and albumin
   are both affected by the level of protein intake in the d'let and t.here-
   fore may be low if there is a protein deficiency. They may, however,
   also be influenced by various diseases not directly related to
  nutrition. Serum vita:'in 1\ is a measure of vitalllin 1\ deficiency.
~/Not   available.

     The study also provided rough estimates of the number of
children experiencing low biochemical values. (See table 3.6
below.) It is interesting to note that although the poor
gene~ally experience a high prevalence of low biochemical
values, the actual number of children with low values is
considerdbly higher in the "above poverty" group.
                            Table 3.6
              Estimated Number-of-6-l7-Year-Olds
               with Low Bfochemical Indices:
     United States, 1~7l-7i (HANES Preliminary) (note a)
              Income-below poverty      Income above poverty
                 l£vel (note b)            level (note bl      Total
Hemoglobin                  582                  1,086         1,801
Hematocrit                  745                  2,129         2,995
Serum iron                  296                    771         1,076
Transferrin saturation    1,100                  3,298         4,494
Serum protein                68                  1,437               (c)
Serum albumin
Serum vitamin A             128                     98            223
Estimated population      8,920                 38,389         49,582
a/Source: Prelimin!!y Findin~of the First Health and
- Nutrition ExemInatlon Survey, DHEW Pub. (HRA) 74-1219 1.
~/Excludes   persons with unknown income.
E/Not available.
     By bringing together what we believe to be the most im-
portant studies into the health and nutritional status of the
Nation's schoolchildren, this chapter provides an important
foundation f(_r considering how NSLP can best achieve its nu-
tritional objectives.    It is important, however, to note that
these studies focused on nutriture, rather than diet.          The
nonnutrient part of diet (e.g., salt, fibre, saturated fat,
etc.) has an important role in safeguarding health, but has
received minimal attention in large-scale nutrition surveys.

     We believe the implications of the three studies--TSNS,
PNS, and HANES--as they affect NSLP, can be summarized as
     1. Income-poverty guidelines, as presently used in the
free lunch program, represent the best known means of selec-
tively targeting NSLP to reach those children in greatest
need of nutritional assistance~ At present, income criteria
provide the best available means for targeting NSLP to reach
the group of schoolchildren having the highest prevalence
of nutritional deficiencies (children from low-income
families). Nevertheless, it should also be noted that there
are probably several times as many nutritionally needy
children among the higher income groups--groups for which
"targetable" characteristics of nutritional need have not
yet been established.
     2. School lunch programs increase the nutrient intakes
of participants. TSNS, supported by related research, pro-
vides evidence that NSLP is effective in increasing both
the nutrient intakes and the quantity of food consumed by
participants. This finding suggests that the progra~ is a
very important part of nourishment for needy children. It
also cautions that the increased levels of caloric intakes
can cause undesired side effects (obesity).
     3. Modifications to NSLP's nutritional standards may
improve program effectiveness. Revisions to NSLP's nutri-
tional standards of one-third RDA would enable program
lunches to better supplement the schoolchild's home diet.
    --Each survey indicated sizeable numbers of both under-
      weight and overweight children. If the program is to
      meet the needs of underfed children without providing
      excessive caloric intakes in other segments of the
      population (e.g., increasing the risk of obesity),
      program pOlicies should permit the selective rein-
      forcement of caloric intakes as appropriate to the
      individual child's needs.
    --Iron deficiency or iron deficiency with anemia was a
      problem in all populations surveyed. In many in-
      stances, the deficits in mean iron intakes were
      greater than one-third RDA.   However, !'.i.nce intake
      levels for most vitamins and minerals can be tolerated
      well in advance of the RDA, across-the-board increases
      in iron supplementation do not pose a threat such as
      that associated with calories. If NSLP is to make up

 the deficit between RDA standards and the home diet,
 reinforcement levels greater than one-third RDA are
 needed for nutrient iron as well as other vitamins!
 minerals for which major deficiencies have been found
 in some segments of the population. Such reinforce-
 ment, however, would have to be carefully planned to
 safeguard against some children receiving excessive
 supplementat ion of those nutr ients wh ich t   in excess,
 are toxic.
--All surveyed populations had mean protein intakes well
 in excess of RDA standards. In this regard, it appears
 that NSLP's nutritional standaEds place undue emphasis
 on protein, usually the most expensive component of
 the NSLP lunch.

                          CHAPTER 4
                 ---------------    -------
                     NUTRITIONAL IMPACT
     Since 1946 National School Lunch Program legislation has
expressed congressional intent to "safeguard the health and
well-being of the Nation's children." Effective implementa-
tion of this policy requires the administering agency--USDA--
to establish a sequ~nce of derivative program objectives and
to define the means for their accomplishment. In developing
these program objectives, USDA has an implicit requirement to
     --understand the schoolchild's nutrition problems,
     --identify target groups with special needs,
     --establish priorities, and
     --develop standards for program evaluation.
Unfortunately, NSLP's effectiveness in satisfying legislative
goals remains unresolved.

     In 1973 the Congress enacted Public Law 93-lS0, seeking,
among other things, information on NSLP's nutritional effec-
tiveness. Section 10 of the law directed the Secretary of
Agriculture to carry out a comprehensive study to determine
if the benefits of NSLP were:
     "* .. * accruing to the maximum extent possible to all of
     the nation's school children, including a study to de-
     termine if those most in need are receiving free
     lunches" .. *."
     USDA's "Comprehensive Study of the Child Nutrition
Programs" II was delivered to the Congress in July 1974. Due
to the short deadline provided in law, USDA's study relied on
existing literature that demonstrated the importance of nutri-
tion on child development as well as on reports relating to
the school feeding programs. The study noted that compara-
tively few carefully designed surveys have been undertaken to
evaluate the effects of these programs on the nutritional

l/USDA, CO~EE~~~~~!y~~~~~~_~!-~~~-Ch!ld-~~~E!~!£D-Pr~SE~~'
   Committee Prlnt of the Commlttee on Agrlculture and
   Forestry, U.S. Senate, Wash., D.C., Sept. 1974.

status of participating children and that
     "* * * it is doubtful that a study can be expected to
     measure quantitatively the impacts of a specific food
     program on the basis of nutritional status of children
     who receive only one-sixth of their annual meals from
     the program."

Nonetheless, USDA judged the child nutrition programs as
being extremely effective.
    We believe the USDA study was a worthwhile undertaking
which synthesized in a meaningful way various isolated, yet
related, facts and information about the school feeding pro-
grams. However, even though personal and anecdotal evidence
supports the nutritional benefits accuring to NSLP partici-
pants, documented studies are inconclusive. The lack of a
substantive program evaluation precludes an objective ap-
praisal of what the program accomplishes, how these accom-
plishments compare with intended objectives, and how effec-
tively program resources are managed.
Existing studies

     Several studies have investigated the nutritional impact
of school feeding programs. While most of these studies have
made a valuable contribution to existing knowledge, their
findings about NSLP's nutritional impact tend to be incon-
clusive. For example:
     --The Ten-State Nutrition Survey's dietary intake eval-
       uations (see ch. 3) generally found that children
       participating in school lunch programs had higher nu-
       trient intakes and consumed greater quantities of food
       than those not participating. These findings lend merit
       to assuming beneficial program effects on nutrition-
       ally deprived children. They also indicate a poten-
       tial for promoting obesity. However, since no attempt
       was made to test for differences in the nutritional
       status of participants and nonparticipants, the study
       did not provide direct evidence of the program's net
       impact on the health or nutritional status of parti-
       cipating children.
     --A 1970 study employed three nutritional indices--
       height, weight, and hematocrit--to evaluate the school
       lunch program in four schools located in Baltimore's

       lowest economic strata. 1/ Children were followed
       throughout the school year to determine whether there
       was any benefit to the participants as opposed to
       similarly matched children who did not participate.
       The study concluded that ". • • nutritionally disadvan-
       taged children participating in an institutional
       school feeding program fared no better than those
       comparably matched children who were not participating
       in the organized school feeding program." The authors
       attributed this lack of success to several factors,
       including high rates of absenteeism. incomplete con-
       sumption of lunch, poor nutritional reinforcement at
       home, and a variety of educational and economic deter-
       minants. They suggested that NSLP's nutritional stand-
       ards (i.e., one-third RDA) ignore these requirements
       and that higher levels of nutritional reinforcement
       will result in a greater physiological impact on nu-
       tritionally deprived children. It is important to
       recognize, however, that the study focused on whether
       or not the NSLP lunch upgraded the status of nutri-
       tionally disadvantaged schoolchildren.    Its conclu-
       sions may not be representative of the program's im-
       pact on children who regularly consume most of the
       NSLP lunch or, for that matter, on the overall schoo~
       child population. Nonetheless, the study's suggestion
       of a need to increase the program's nutritional stan-
       dards merits serious consideration. This suggestion
       is strengthened by the results of a later study 2/
       which found that a nutritionally enriched supplement
       in addition to the NSLP lunch provided major improve-
       ments in the status of nutritionally deprived children
       and may have been associated with a lower rate of
       absenteeism among supplemented youngsters.
     --A study at Cornell University measured the impact of
       the school lunch on the nutritive intake, biochemical
       indices, and physical growth of elementary school

l/Paige, D. M., "The School Feeding Program: An Under-
- achiever," JO~£!!~!_2L~££!_!!~alth, 42: 392-395, 1972.
2/Paige, D. M., Cordano, A. and S. Huang, Nutritional~­
- plementation of Disadvanta~ed ElementaEX-~chool cfiiroren,
  Presentation-for-rne-lOlst Annual Conventron-01 the-----
  American Public Health Association, Nov. 1973.

      children during the 1970-71 school year. 1/ Children
      were divided into three groups--nutritionally needy,
      intermediate, and nutritionally adequate--and exa-
      mined in the fall and the spring. Of the children
      eligible for the fr~e lunch program, twice as many
      were judged nutritionally needy as were considered
      nutritionally adequate. In the group of higher eco-
      no~i~ status, these proportions were about equal.  Few
       biochemical measurements were made, causing the re-
      search team to rely almost exclusively upon dietary
      recall comparisons with the RDA to evaluate nutri-
      tional status. Diets of nutritionally adequate
      children showed little change over the school year,
      except for vitamin A, which was considerably lower
      in the spring. Conversely, nutritionally needy chil-
      dren had larger supplies of all nutrients in the spring
      than in the fall. The bulk of this increase was sup-
      plied by home feeding (15 to 26 percent came from
      school lunches). Students were classified by nutri-
      tional status in the fall and again in the spring, as
      shown in table 4.1. The authors noted that fewer
      children were classified as nutritionally needy in the
      spring. They also reported that many children classi-
      fied as nutritionally adequate in the fall "had be-
      come overweight and, therefore, no longer met all of
      the criteria for nutritional adequacy." 2/ We believe
      the reclassifications toward "intermediaEe" nutrition
       levels introduce a question as to how well the pre-
       sent NSLP lunch complements the home diet. It should
       be noted, however, that dietary recall provides a com-
       paratively weak basis for judging an individual's
       nutritional status. Seasonal variations in dietary

l/Emmons, L., Hayes, M., and Call, D., "A Study of School
- Feeding Programs, II Journal of The American Dietetic
   ~~~£iation, Vol. 6l-,-~ept:-r97~~-pp:~62=275~----

2/Identical procedures were used in the Cornell study to
- assess changes of nutritional status in children that
  participated in both school lunch and school breakfast
  programs. Compared with the lunch alone, the combination
  of programs had a greater impact on nutritionally needy
  children. The proportion of children classified as nu-
  tritionally needy declined from 30.3 percent in the fall
  to 10.4 percent in the spring. The proportion of nutri-
  tionally adequate children also declined (28.2 to 16.4
  percent) .

       intakes and the absence of a control group further
       handicap attempts to assess the significance of this
       report's findings.
                          Table 4.1

Nutritionally needy                    25.9%        22.5%
Intermediate                           49.8         64.5
Nutritionally adequate                 24.3         13.0
                                      ---           ----
      Total                           100.0%       100.0%
    --A Harvard study used 12-to-18-year-olds in a large
      boarding school to evaluate the feasibility of modi-
      fying blood cholesterol levels in adolescent chil-
      dren. 1/ The study demonstrated that an association
      between diet and blood cholesterol exists in males as
      early as the second decade of life and that serum
      cholesterol can be lowered by means of a modified diet.
      The study concentrated on comparatively simple dietary
      changes such as using low-fat milk with extra skim
      milk solids, replacing butter with a highly polyun-
      saturated margarine, and using polyunsaturated oils and
      shortenings in baked goods and for frying. While the
      study achieved a IS-percent reduction in serum choles-
      terol, the author cautions that these features might
      not be available to schools which rely heavily on con-
      venience and commercially prepared foods. Although
      this study was not directly associated with the school
      lunch, we believe it demonstrated that dietary factors
       in addition to nutrients can be used in an evaluation
       of child-feeding programs. Considering the importance
       of the nonnutrient part of diet on individual health,
       we believe that future evaluations of NSLP should in-
       clude the program's impact on the total diet, rather
       than simply those aspects of diet related to nutrients.

liFord, C., et al., "An Institutional Approach to the
- Dietary Regulation of Blood Cholesterol in Adolescent
  Nales," preve'2!!~_t!~di£!!!~, 1:3:426-445, 1972.

     Although USDA's report raised an important question as
to whether or not NSLP's nutritional impact can be evaluated,
the studies cited above appear to indicate that it can. 1/ In
addition, the studies provide some basis for concluding that
    --the NSLP lunch, if properly designed, should provide
      a recognizable improvement in the nutritional status of
      schoolchildren (an implication of the Baltimore study's
      success in combining the NSLP lunch with a low-lactose
      supplement); and
    --it may be possible for NSLP to safeguard health
      through techniques in addition to RDA supplementation
      (an implication of the Harvard study's success in
      modifying serum cholesterol levels in adolescent
      children) .
On the other hand, the Cornell study reported an increased
incidence of obesity among NSLP participants. This indicates
a need for an evaluation not only of the program's ability to
increase nutrient intakes, but also of the extent and nature
of ics possible side effects.

     In commenting on our report (see app. III), HEW stated:
     "The report criticizes the regular Type A school lunch
     because it contributes to obesity in some children and
     has not been able to improve iron nutriture. Since the
     report elsewhere concluded that present studies of NSLP
     are inadequate to evaluate nutritional impact, it is
     premature to implicate the program on these grounds.
     This is particularly true since, as the report points
     out elsewhere, the school lunch provides only one-sixth
     of the meals of the participants and can, therefore,
     only be a supplement to home meals."

l/In commenting on this report (see app. I), USDA stated:
   "The program is designed to provide a maximum of five
  meals per week. Assuming that an average of 1/3 RDA is
  provided through the lunch over the five day period this
  would be only 20% of the child's total nutritional re-
  quirements for that period. * * * Because of the rela-
   tively small proportion of the total nutritional require-
  ments the NSLP is expected to provide and the complexities
  associated with determining nutritional status, it is
  questionable that [an evaluation] would be successful in
  accomplishing its objectives."
     We wish to give special emphasis to the fact that our
report states a need for further evaluation of NSLP's health
impact. It does not provide scientific evidence that NSLP
as a whole fails to improv~ iron nutriture, or that it pro-
motes obesity. On the other hand, it does bring together
evidence that (1) obesity and iron deficiencies constitute a
nutritional problem among schoolchildren, (2) the NSLP lunch
increases food consumption without distinguishing between the
needs of underfed and overweight children, and (3) where
studied, the NSLP lunch has been found to provide less than
one-third of a schoolchild's RDA for iron.
     Most of the studies which have attempted to evaluate
NSLP's health impact focused on its ability to improve iron
nutriture. Aside from being inconclusive, not one of the
studies showed any indication of improving iron nutriture.
On the other hand, such studies have found indications of
an increased prevalence of obesity among NSLP participants.
These findings, while not necessarily representative of NSLP
as a whole, are a cause for concern and jUGtify further evalu-
ations of the program's health impact.


     Dr. George Graham, Professor of International Health at
Johns Hopkins University, believes that evaluating NSLP's
nutritional impact is difficult, but that it is possible.
Dr. Graham emphasizes the need to examine carefully the total
nutrient intake--in the home and at school--together with an-
thropometric and biochemical indices as used for the Ten-
State Nutrition Survey. Since some schools are still joining
the program, he suggests this approach could be tried on
students before entering the program and then repeated one
year after joining. Further, he states:
     lIThe subject of controls is of course extremely impor-
     tant. If one documented the existence of significant
     undernutrition in a school population, if one proved
     that their total nutrient intake improved with parti-
     cipation, and if one demonstrated significant improve-
     ment in nutritional status, then one might be able to
     suggest that the participation did improve the nutri-
     tional status. For absolute proof, however, it would
     be necessary to have a control group of similar nutri-
     tional status who did not participate in the program. II

     Dr. Graham's comments are echoed by other experts. They
agree that such an evalu.ation is feasible and that, giving
adequate consideration to design complexities, approximately
1 year would be required to formulate an appropriate evalua-
tion design. When asked if such an evaluation could assess
the impact of school lunch participaton on longer term health
benefits--useful life expectancy, lifetime earnings, medical
expenses, etc.--expert opinion was divided. Especially note-
worthy, however, is the concern by experts about adverse pro-
gram effects:
     "r doubt very much there is a positive effect of the
     school lunch on any of these parameters, as there is com-
     pelling evidence to suggest that moderate undernutrition
     [low calorie intake relating to weight] will prolong
     useful life expectancy and there is the real danger that
     these programs are continuing to p-romote the overnu-
     trition which is this country's greatest health and nu-
     trition problem." !/
     "* * * evaluation could be designed to capture short and
     long term aspects. Short term benefits could, in some
     situations, be less important relative to their long
     term implications. For example, provision of calories
     which may be important for some may have some long-
     term negative implications insofar as obesity is con-
     cerned. In some situations where participants are only
     thinking of satisfying their hunger and need more energy
     sources the supply of calories may be important, but to
     other participants this could have an adverse
     effect." ~/
     ". • • the class A school lunch includes a slab of butter
     and a cup of whole milk. Neither of these are essential
     ingredients of a balanced diet. * * * there is consider-
     able evidence that the excessive ingestion of saturated
     fats and cholesterol may predispose certain individuals
     to premature cardiovascular disease. In addition, autopsy
     studies have shown that many healthy American males
     already have moderate coronary disease at a very early

l/Letter of Dr. G. G. Graham, M.D., School of Hygiene and
- Publ ic Health, The Johns .Hopkins U"iversi ty, Oct. 1974.
2/Conversation with Dr. G. M. Owen, Department of Pediat-
- rics, University of New Mexico, Feb. 1975.

       age. In the light of this evidence, perhaps the class
       A school lunch should be examined more thoroughly." !/
     Many of the Nation's leading experts are suggesting that
greater emphasis be placed on the subject of nutrition educa-
tion. Dr. Graham, commenting on USDA's report, states:
     lIThe whole section on nutriton education* * * is based
     on the assumption that the typical American diet is
     ideal and desirable, when all the overwhelming evidence
     points to its disastrous effect on our health. Some
     of the present food fad ism of young people in particular
     is assumed to be all wrong. Much of it may be very
     right. Although they made many mistakes, many of these
     young people are turning against the gluttony of their
     parents and many of their own contemporaries. They de-
     serve being listened to and perhaps guided, where they
     are making mistakes. Many of them have read the modern
     scientific nutrition literature much more carefully than
     many of the nutrition ~professionals'l who are trying to
     impose traditional patterns on the younger generation.   II

     Nutrition education receives comparatively low priority.
In view of the fact that presumably l'normal" American dietary
practices may predispose a relatively large percentage of our
population to premature cardiovascular disease and possibly
other acute and chronic debilities (see ch. 2), it may be
desirable to shift the emphasis on nutrition education f'om
conceiving it as a passive, abstract discipline to a vi~_le,
active part of preventive health. We believe nutrition
education needs to deal with current food trends. It needs
to identify food as more than a mere composite of RDA nutri-
ents. Improved nutrition education involves disseminating
appropriate knowledge on extenders, saturated fats, fibers,
preservatives, and other food constituents present in today's
     Associating diet practices with day-to-day health is
felt to be more relevant for schoolchildren, who, made

l/Statement of Dr. S. Sc)ultz, University of Pittsburgh
- School of Medicine, before the Senate Select Committee on
  Nutrition and Human Needs, May 1973.

aware of health problems in their environment, may see
direct application of nutrition instruction in their daily

     The em9hasis given nutrition education varies among
State and local governments.   Comprehensive legislation which
allows for teaching nutrition education in the Nation's
schoo13 is presently lacking; the program is therefore highly
dependent on the attitudes of State and local authorities.
Describing the extent to which the program is misunderstood
at these levelS, Mr. C. F. Olsen, Director of NSLP in Idaho,

     "I've always said that if a school lunch program is doing
     nothing more than a gastronomical filling station, then
     it has no need for existence. And as I made that ob-
     servation in education meetings, some of the superin-
     tendents kind of crossed their eyes a little and said,
     'Really what else do we do, that's what we have it for.
     It's there to feed them.'  It's not there just to feed
     them.  It shouldn't be.   It should be an integral part
     of the school and every part of that program can be
     impleme~lted into some phases of the curriculum* * *." !/


     The problems of health and nutrition differ among school-
children.   Some children are well n0urished, some overeat the
wrong foods, ar.d others are underfed. While it may be that
the NSLP lunch is a valuable source of nutritious food for
needy children, there is an inherent danger that this same
meal promotes overeating in other children. In fact, j t
would be naive not to expect some adverse side effectJ in any
large-scale feeding program which stresses a standardized
menu pattern and portion sizes.     Because of these consider-
ations, the question of whether NSLP safeguards health re-
quires more than a simple "yes" or "no" answer.     It requires
a comparison of both beneficial and adverse health influ-
     At present the Nation lacks an adequate evaluation of
NSLP's impact on schoolchild nealth.      There is little ob-
jective evidence on which to undertake either a strengthen-
ing of the program's health impact or the elimination of

undesired side effects. Similarly, legislative bodies have
little substantive evidence on which to compare the program's
resource requirements against anticipated health benefits--
a disadvantage in allocating scarce budget resources among
competing needs.

     These considerations, and a lack of objective evidence
on NSLP's nutritional impact, indicate a need for further
program evaluations. However, these needs should be balanced
against the cost of the evaluation, which would not be known
until after the preliminary design has been approved. Such
an evaluation may require about 4 years: 1 year for devel-
oping the survey design, 2 for data collection in the field,
and 1 for analyzing findings. We believe a multidisciplinary
team offers the greatest potential for compressing time-cost-
scale factors, for providing assured re~iability of findings,
and for introducing scientific objectivity. The effort could
be authorized in two phases: the first would establish a
survey design, while the second would involve field data
collection and analysis, which would begin only after ap-
proval of a satisfactory design.
     Some considerations which we believe are important to
the design are:
    --An evaluation of NSLP--even though providing useful
      information to gauge the program's performance--can
      only serve as an objective impetus for improvement
      if its information can be used by managers.
    --Since measures of nutritional status include only a
       portion of the diet variables that influence health,
       an evaluation of NSLP's health impact shoUld, when
       feasible, incorporate additional means of appraising
       program performance (e.g., its ability to reduce the
       incidence and duration of illness, hypertension, tooth
       decay, elevated blood cholesterol levels, etc.).
    --The evaluation process should focus on selected diet-
      related health variables which are considered to be
      the most strategic to NSLP goals, either in the sense
      that they have the greatest impact on individual
      health or that they, better than any others, show
      whether NSLP is safeguarding the overall le~el of
      schoolchild health as expected.
    --Even though the evaluation process may be constrained
      to a comparatively small sample of children, there's

       an implicit requirement for evaluation results to be
       expressed in terms of their impact on the overall NSLP
       population. To facilitate this requirement, a sample
       stratified by some index of health and/or nutritional
       status may provide better statistical estimates than
       one which groups children by socioeconomic character-
Authorization to begin the second phase--field data collection
and analysis--should be predicated on the review and approval
of a satisfactory design.
     As a means of resolving existing uncertainties and im-
proving program effectiveness, the Secretary of Agriculture
should require a formal, systematic evaluation of NSLP's
performance in meeting legislative objectives. The evaluation
should be coordinated to utilize the expertise and resources
of HEW in all matters pertaining to the health and nutri-
tional status of schoolchildren; and to provide effective and
timely reporting of information needed for congressional
----------       --------------
     HEW advised us by letter dated April 14, 1977 (see app.
III), that it was willing to assist USDA in carrying out the
intent of our recommendation.

     USDA, in a letter dated April 20, 1977 (see app. I),
told us that it recognized the need for a comprehensive
evaluation of NSLP's effectiveness in meeting legislative ob-
jectives. USDA stated that an evaluation plan projecting
FNS's research plans over the next 5 years has been drafted
and is currently under review. It said that the plan calls
for developing a methodology for assessing NSLP's nutritional
impact but that, since the plan was under review and subject
to change, it was not presently available for our review.
     Since we have not reviewed USDA's plan, we have no means
of assessing whether or not it will provide for an effective
program evaluation. We note, however, that USDA's comments
do not make any reference to the considerations that we be-
lieve are important to the design of such an evaluation or
to the recommended coordination with HEW.

     USDA also stated, "It has never been the philosophy of
the Department that the basis for the NSLP is to serve as a
nutrition intervention program to prevent a state of disease"~
it added:
     "Because of the relatively small proportion of the
     total nutritional requirements the NSLP is expected
     to provide and the complexities associated with de-
     termining nutritional status, it is questionable
     that such a study would be successful in accomplish-
     ing its objectives.II

     We view the apparent contradictions in USDA's position
with some concern. In our opinion, the Congress has pro-
vided substantial funding and a clear mandate for the pro-
gram to safeguard schoolchild health. It is possible that
NSLP is safeguarding health but, based on present informa-
tion, it is equally likely that Federal funds are being
spent on a program that is not meeting its objectives. We
believe that NSLP can and should have a beneficial influence
on schoolchild health. To insure this effect, positive ac-
tions must be taken toward evaluating the program's perform-
ance. Such actions and priorities are not obvious in USDA's

     In view of the emphasis that the Budget Impoundment
Control Act of 1974 places on progam evaluation, and con-
sidering the Congress' overall desire for meaningful over-
sight information, the Congress should:

     --Require HEW, the department primarily responsible for
       research related to schoolchild health, to assist USDA
       in evaluating NSLP's health impact.
     --Review USDA's program evaluation plan before imple-
       mentation to make certain that it will provide ade-
       quate information for program oversight and that it
       uses the resources and expertise of USDA and HEW in
       a manner which benefits the evaluation and is in
       keeping with the respective missions of each agency.
     --Require the Secretary of Agriculture, on completion
       of the NSLP evaluation, to provide a comprehensive
       report of his findings, together with any recommen-
       dations he may have with respect to improving program

     The Congress should also be aware that legislation
prohibits NSLP from imposing any requirement relative to the
teaching of nutrition to schoolchildren. The effectiveness
of nutrition education programs is therefore not addressed in
this report. Such programs are, however, currently being re-
viewed by us on a broader scale.

                          PART II
     The second objective expressed in the National School
Lunch Act is "* * * to encourage the domestic consumption
of nutritious agricultural commodities and other food. * * *"

     Though we found indications that children probably con-
sume a larger quantity and variety of commodities under NSLP
than would otherwise be expected, we noted that comparatively
little has been done to determine the program's impact on the
agricultural economy. We do not know for sure how the pro-
gram affects the farm and market price of food, and we can-
not be certain as to the program's effectiveness as a price
support mechanism.
     Chapter 5 describes the major program provisions which
encourage the consumption of agricultural commodities and
finds that some of these provisions may operate to the detri-
ment of NSLP's nutritional goals. In fact, in view of recent
changes in the Nation's agricultural economy (transition
from a period of oversupply toward a state of general
equilibrium), the program's emphasis on stimulating the de-
mand for farm products may no longer be desired.

                         CHAPTER 5
     Relatively small changes in food supply can have a
dramatic impact on the market price of food. If the supply
of food keeps pace with demand, prices tend to remain stable.
But if the supply-demand balance is upset, large price fluc-
tuations can occur, posing a threF _ to farm incomes and to the
consumerts budget.
     Beginning in the 1930s a number of Federal programs were
initiated to stabilize farm and market prices. Surplus farm
production was absorbed in the form of government-held or
supported reserves and released from these stocks in times of
shortages. To keep the growth of accumulated reserves within
manageable proportions, it became necessary to find an outlet
for surplus foods.   II0 ne of the most obvious outlets was
presented by the need for this food by the children of the
Nation, many of whom were malnourished to the point of physi-
cal and mental deterioration. 1/ As a result, in the mid-

thirties, USDA initiated a practice of donating surplus food
to schools for use in providing free lunches to needy
children. This practice benefited both the nutritional well-
being of needy children and the Nation's agricultural poli-

      In World War II, the Nation's agricultural production
was greatly expanded to assist our European al1ie8_ After
the war (in 1946), the European demand for U.S. f,rm products
slackened and the agricultural economy was threatqned with
oversupply. Federal price stabilization efforts faced an
era of rising program costs and huge crop surpluses. As a
result, the National School Lunch Act of 1946 included this
agricultural objective: "~ * * to encourage ttle domestic
consumption of nutritious agricultural commodities and other
food." This objective intended the National School Lunch
program to be both an outlet for surplUS foods and a vehicle
which, by making sizeable purchases of foods in local markets,

l/U.S. Senate, "Providing Assistance to th~ States in the
- Establishment, Maintenance, Operation, and Expansion of
  school-Lunch Programs," 79th Cong., 1st Sess., Rep. No.
  553, July 1945.

would stimulate the overall domestic demand for food, thereby
helping to stabilize farm and market prices.


     The Senate Committee on Agriculture and Forestry sup-
ported the establishment of NSLP indicating:

     "The school-lunch program becomes an organic part of the
     agricultural program of the United States. The amount
     and dollar value of food which would be consumec in
     * * * [the NSLPJ * * * is in itself not inconsiderable.
     * * * an estimate has been made on the basis of meal
     standard and matching requirements which would indicate
     that in the first year of its operation approximately
     $186,000,000 would be expended for foods, and this would
     increase as the matching requirements [local and State
     contributions required by the act] increased.  There is
     also the indirect result of increasing the use of foods
     through education. An established and regular market
     for the agricultural production of this country is pro-
     vided which would, in great measure, not exist if this
     legislation were not enacted. An organized outlet is
     established for the occasional surplus in production
     which exists in almost every agricultural commodity
     field." !/

     As   foretold by the Committee, NSLP has come to repre-
sent an   important market for the Nation's agricultural pro-
ducts.    The program's food costs in fiscal year 1975 amounted
to more   than $2.2 billion, as shown on the next page.

!/See footnote 1, p. 55.

                             Table 5.1

                         NSLP Food Costs,
                    Fiscal Year 1971-75 (note a)

Fiscal        commodity           Local market
 year       distributions        food purchases             Total

                - - - - - - - - ( m i l l ions ))-- - - - - - - - -

1971           $277.3              $1,132.5               $1,409.8
1972            312.1               1,250.8                1,562.9
1973            260.2               1,408.4                1,668.6
1974            316.1               1,615.2                1,931.3
1975 (est.)     421.3               1,808.4                2,229.7

~/Source:     FNS/Program Reporting Staff.

Local market purchases

     Since 1971 over 80 percent of NSLP's annual food needs
have been purchased by schools in local markets. These pur-
chases, by 1973 estimates, represent about 1.5 percent of the
overall U.S. food market. Their size, coupled with some
evidence that the program increases food intakes, provides
some basis for assuming that NSLP strengthens the demand for
agricultural products.  However, we do not know of any study
which has determined the precise impact of these increases
on the Nation's agriculture economy--either in terms of the
farm and market price of food or as me.ns of fostering tho
productio~ of needed commodities.

An outlet for surplus food

     NSLP has also supported the Natiun's agricultural poli-
cies as an outlet for foods acquired through direct market
support operations 'of the USDA.  In 1945 USDA described the
value of such outlets as follows:

     "Last October, a hurricane blew tremendous quantities
     of apples off the trees in the Northeastern States.
     Growers were faced with the prospect of a substantial
     loss. The War Food Administration purchased about
     400,000 bushels of those apples to support the market
     and protect growers from what appeared to be an almost
     certain loss. Although school-lunch programs were only
     beginning to get under way at the time, they absorbed

     half the apples we purchased--apples that would surely
     otherwise have been wasted before they could have moved
     in the normal channels of trade." .!/
     During the late 1940s and in the 1950s the Nation's farms
produced surpluses which were in part purchased in Federal
price support programs and donated to schools, other institu-
tions, and needy families. Although these donations were put
to good use, they were nevertheless a by-product of policies
designed to protect farm incomes until an acceptable balance
between farm production and consumer demand could be achieved.
     In the late 1960s, u.s. agricUltural policies were
changed. As government loan and storage programs were ~ur­
tailed, th~ general pressure for surplus donations decreased
and most c£ the Federal food distribution programs were
phased out. NSLP then became the Nation's primary outlet for
USDA's food distributions.
     The utility of the NSLP outlet was again demonstrated
in 1974 when a set of unusual conditions caught U.S. beef
producers in a severe cost-price squeeze. A summer drought
and an unexpected short q".i," of feed grains raised cattle
production costs while ca;;_;:<tp prices declined.   To assist
this threatened industry, the administration stepped up
purchases of beef and pork for school lunches as a way of
shoring up meat prices and warding off bankruptcies among
the Nation's livestock feeders.
     In recent years, however, there has been a sharp decline
in the quantities and types of foods qualifying for purchase
under price stabilization and surplus removal programs.
Concerned with maintaining a distribution system for the
smaller volume of commodities, USDA sought an alternative
means of supporting NSLP--by replacing commodity donations
with a cash subsidy. The Congress, on the other hand, en-
couraged States to continue their commodity distributions
by (1) mandating a prescribed level of commodity support for
NSLP and (2) authorizing USDA's purchase of nonsurplus com-
modities when necessary to meet the mandated distribution
requirements. The continuing use of this distribution

l/See footnote, p. 55.

system, in addition to assuring an outlet for foods acquired
in USDA's market support operations, has provided sizeable
savings in NSLP food costs (as described in ch. 8).
It should be noted, however, that USDA's food purchases are
primarily directed toward supporting the Nation's agricultur-
al market rather than satisfying NSLP requirements.
     What would happen to the agricultural economy if NSLP
did not exist?   How would the production of needed commodi-
ties be affected, and what would happen to the farm and market
price of food? The answers to such questions are, in our
opinion, the "bottom line" for assessing NSLP's effectiveness
as an agricultural program.

     Over the years, USDA and other organizations have pub-
lished a substantial amount of research on NSLP food costs,
student dietaries, and food preferences.   Much of this re-
search supports the conclusions that NSLP (1) increases food
demand, (2) modifies consumption patterns (possibly enhancing
the utility of price support and surplus removal programs),
and (3) constitutes an important market for agricultural pro-
ducts; and these conclusions lead to a presumption that NSLP
affects the agricultural economy. The presumption is prob-
ably correct; however, it is important to realize that
neither the presumption itself nor its supporting research
constitutes an evaluation of NSLP's impact on the agricul-
tural economy.

     To date, we know of only two studies that have actually
looked at NSLP's effectiveness as an agricultural program.
Both of the studies were sponsored by USDA, and were brought
to our attention by USDA's Economic Research Service (ERS).
One study used a commodity-by-commodity approach to evaluate
NSLP's impact on the demand for particular commodities (e.g.,
carrots, lettuce, milk). The other study--a sector-by-sector
approach--assessed the program1s influence on the business
receipts of various economic sectors (e.g., agriculture,
meat and poultry manufacturing, wholesale trade).
     While these studies do not answer all of the questions
we posed earlier, we believe that they represent important
approaches for determining NLSP's impact on agriculture.
To illustrate these approaches, we have included a summary
of each study below. We must point out, however, that we
have not made an indepth appraisal as to the validity of
either study's findings.   The summaries are intended solely
to illustrate the approaches used and the types of findings
they provide.

Commoaity-by-commodity approach

     Using the commodity-by-commodity approach, a 1950 study 11
analyzed the Iowa school lunch program's impact on the demand-
for particular commodities during the 1947-1949 scnool years.
The study determined the number of NSLP lunches served during
each of the school years, the purchases made by the schools
for lunches in each year, and the purchases made by homes for
the same number of lunches.   (Homes purchases were estim~ted
from a sample of student dietaries; NSLP purchases included
surplus commodity shipments and were estimated from the pur-
chase records of a sample of Iowa .chools.)

     The total food purchases made by the schools for lunches
were subtracted from the corresponding purchases made by the
homes. This difference, or market support, was computed both
upon a per meal basis and a total basis. Table 5.2 shows
the total support given certain commodities by the Iowa school
lunch program for the school years 1947-1948 and 1948-1949.

I/Nelson, P. E., "Market Support Given Certain Commodities by
- the Iowa School Lunch Program,'1 Journal of Farm Ecor.omics,
  May 1950.                                           -

                                                                    Table 5.2
            Iowa School Full Lunch Program Support For
                       certain Commodltles (note a)

                                                      J!l17-W:S                                                     I!W!-I!IJO

                                     J·l•.~ili\·l·                  :"Jq:llliw"                           l·u.~ilh·c                 Kl"!:':tti\1;
                                                         I\mnoj.;                                                        J'ulllul<

                                                                           C:"lllnl \'1:1:'         11l1.],·.~
            e"l"lI                        ~.~:;;l                                                             V,filj
            (;l'I"'U J:":'l'~          l!lll, I rr:                                                        2i!i ,!II;)
            I'..a-I                     (:~J   ,   -,,::.j                                                   ill, i!li
                                                                    Jo'rl'."h   Fnlil~   :'11111    \'1·t:dHhl'·~
            Apl'k~                                                    iiG,t:!lj                                                       0.">,011
            1:,111:111:\,:                                           lR!J,IIGI                                                       2l;';,71;1
            (::'.hl,.:,,:.,             :,,!. ;;i,                                                           GO,SiI'-.
            (::Ifi"<,ls             I, l'j'G,!Ii"~                                                    I       , •!~':!l
                                                                                                          ,:~ ~
            ('..I,-r;'                .Jtil',lH                                                            S3$,-t!17
            IA.II.it"(·                                              I-?:J.:l.'il                                                    HI,lil:/
            O"i":i'l                    ':'0 ,'!':7                                                          '!:I,·J11
            1','1!t{0(·~               Rli1, U:~                                                           u:n, i!f;~
                                                                                    !\ 1"1'1..
            nrt"r                                                    119,MS                                                          1::1(; ,': II
            }'( •.•j;
            J'uuHQ'                      O:l,!,ii
                                                                     '1 t Y.l,Rn
                                                                                                    103,:;             ,             4:~S.ij!1i

                                                                     ~1 j.~~,·ll~lH'lIl1~   Cr':I:I!l(" lili,-,
            EI',;:.<                                                   S!,5i..:i                                                       !Ir.,I::H
            FI"lIl"                      f\.",~l·l                                                           r,' .1~1
            ~fill.              bh!3,x'"                                                           ~/.i:I.~,!l:l-l
            SI.orh·.· .Ii:      -        C.... ::;-!·                                                    7;'. '.!~.!

a/A number of commodities were excluded from the study, either
- because the available records were not adaptable to study
  purposes, or because they were too fragmentary. Source:
  "Market Support Given Certain Commodities by the Iowa
  School Lunch Program,'l Journal of Farm Economics, May 1950.

     As stated in the study,
     IIAn examination of these results shows that the current
     Iowa IIfu11 lunch" program gives negative support to
     several commodities, perhaps the most important being
     beef and pork. However, thase items are well represented
     in the students' total dietaries, at least for certain
     age groups.    [Although NSLP patrons consumed less protein
     for lunch, the study's dietary component found that the
     overall (24-hour) protein intakes of all students a~­
     proximated the reccmmended allowances.]
        *                                      *                                    *                                        *                       *

     "When only the lunch meal is considered, the school
     program appears to contribute both market support and nu-
     tritional supplementation in the case of milk. However,
     it is interesting to note that • • • the total daily
     calcium intake * * * was equivalent for both school and
     non-school lunch students.  Thus, it is possible that
     the apparent support is not equivalent to actual support,
     as has been the case for all other commodities.

       •           •           •            *                •
     "The support given leafy green and yellow vegetables
     [with the exception of lettuce] presents a clear cut ex-
     ample of support combined with dietary supplementation.
     However, a complete analysis of these Iowa results [pos-
     sibly not representative of the more industralized States]
     suggests that the school lunch prQgram's effectiveness as
     an vverall price support mechanism may have been overem-
     phasized. * * *',

     Although this study is dated and deals with a program
much smaller than NSLP as it is is currently designed, it does
illustrate a number of important points.  For example,

     --NSLP's net impact on demand should be assessed in
       terms of changes in the overall (24-hour) diet, not
       in terms of changes in lunch intakes. l/

     --The school lunch program's "demand for agricultural
       products " is, in fact, a collection of demands for
       specific commodities.   (Raising the question as to
       whether the focus of an evaluation should be on

l/In commenting on our report, FNS stated (see app. I):
   ll* * * the Department is conducting a survey to de-
   termine the kinds and amounts of food used in the
  nation's schools. This national probability sample
  of food use will provide information to further
  evaluate the impacts of the NSLP (and the School
  Breakfast Program) on the demand for agricultural prod-
  ucts.    An outside contract for the conduct of this
  study has been underway for some time.    Data collection
  has been completed and analysis is in progress. '1
  In subsequent contacts with FNS, we found that the survey
  does not include information relating to the at-home diets
  of children. We doubt, therefore, that the work will sup-
  port a reliable appraisal of NSLP's net impact on the de-
  mand for agricultural products.

       NSLP's suvport of particular commodities or on its
       support of the overall agricultural economy.)

     The commodity-by-commodity approach is the only means
that we know of for determining NSLP's effectiveness in sup-
porting a particular commodity.  It appears to be worthy of
further consideration both as a technique for assessing the
program's influence on selected commodities, and as a tool
for evaluating the performance of the commodity procurement
and distribution programs.

     A 1976 study 1/ by ERS used a computer model to analyze
NSLP's impact on the national economy during calendar year
1972 and fiscal year 1974. The study also analyzed three
alternatives to the 1974 NSLP:  a universal free-lunch program,
a free-lunch program for those eligible and a reduced-price
lunch for other Gtudents, and a free-lunch for children from
poverty households only.

     Program comparisons were based on the assumption that
personal inco~e taxes were increased by the amount needed to
pay the Federal Government's share of NSLP costs.   The tax
increase for each program variation (treated as a transfer
of ~expenditure power" from taxpayers to schools and/or the
suppliers of commodities) was used to compute NSLP's net
impact on
     --the business receipts among industry sectors (e.g.,
       agriculture, retail trade),

     --gross national product, and
     --the number of jobs.

     Table 5.3 shows NSLP's impact on business receipts and
gross national prOduct for fiscal year 1974. The effects of
cash contributions and the commodity program are shown
separately.   As explained by ERS:

                                                                                        Table 5.3

                                            Changes in u.s. Business Receipts
                                       and Gross Natlonal Product Associated with
                                 Federal Contributlons to NSLP, Fiscal Year 1974 (note a)
                                                                 ,:or.tr         1:,:,;tlO:l
                                                                                                                                                                               (ont.( lout iQn
                                                                                                                                                                       :.:lsa                    CC;'I.~Ot.llt0

                                                                                                                                                                               (tnous"n ..IS)
"lIrl"lIltllr~.      for~"tr)'
                            " fj"h"rl"s           .
                                                            ",   lll,51j
                                                                                 .,   ~Ol,.. 5(,   I   ColllhlnOld ~louseh"l,t ~'''tor,,:
!'tIn tn/!. ••••••.••••.••••••••••••••••••••••••••
r..... ,lt rllct Ion •• , ••••.•..•.. , .....•...•...•.••    ,      9. :~"                  '"
                                                                                                          InccOfll..: h".. 1'01" J'nhs.           • ....•...••
                                                                                                          rl", "'h·r~.'·"~ til t,,~ .. ~ ~ "'l\II"r.~ ......•.
                                                                                                                                                                     .;>     )1B,h~t1
                                                                                                                                                                                          .":>     :~.16~
                                                                                                                                                                                                   ) .. ,(,(\[1
!i~". f ilctllr I nit:                                                                                    ~ln,,~ t.1~ In.:r(' •• ~r (0 f""d 1',',1"'''1
   food    :unuf~c:turln8_                                                                                 "1;10",'1 I;.... 'rll""t Io·n ••••••••••••••••••••
      ~;u "      poultry procluce,. ... ,_ ..•.••.••.. _    ~    1)1.'->62        ,   tl6,f>1)           !':qur.ls d'i1n,~ In cl"""u<I(>(lnn e.;>            ..
                                                                                                                                                                                                  •.,11. '\}
      O.. lrj' ;>roduc:t.!l •••.•••••••••••••••••••••••          255.161         •     66,1)71
      Gr.. tl\ .111 pnhlua                            +           2(, ,2 JIi     •     67,5811
      ,,,kery ;.roduc:t5......................... •               21,2112                l,n8]
      Cann".t , p."se.vlld '_.1" ••••••••••••....                100.61'')        •    18,MI           School lunch ~",ctOf;
      l)l;h~r roods' t>euer:lll.c!I ..•••••••••••••••
        Tutal •........•..•.•.•...•••••.••••..•
                                                            •,    18.0114
                                                                                 •     41,2)0
                                                                                                         Ch"n~.t' In "d.nol l""ch ( t'"""~lt"r.'
                                                                                                         !'!Inus d",rr",,""'~ In r.~etl" ~:l"="n~" C/
                                                                                                                                                                           I ,(Ull.lltZ
                                                                                                                                                                                7 ,1,~8                  '"   1
   :Ionfuod "'.Ill .. fattllrln~·-                                                                       Pl,,~ ~· ...lu·, In fu"d,,, r.tv"",' sc:''''<>ls.:      .                                ~1'1.1..~.!.!!
      Cluthlnl: ••.•.•...•.•...••••.•••.•.•.• ' 0 '               21,OB                HI.))I)           [ .. u~ I, ch.ln!\'" In consu"'l,t I OIl .. "I', ......•          ~~~            +       )1$. \6(,
      Oth". nonfood ull11fJlc:t"rln(l, •• _, •••••••• ,           29.2711               8,873
        Total                                         .            56,2')~             19,20)
          Tota, .,nufactuf'ln8 ••.••••••..•.••.•
l..lca1 , IiUbUfb~" tf~n"po't~tlO11 •.•••••.•••••
                                                            •    502,179
                                                                                 •    JJO.O!4
                                                                                        I. )'}O        Set ,;hanl\" In cothln"<l ,..·('tor I; .. ns"'";ltf,,"
All Oth'H tnn5poftu111n                                 ..         l1,9~2        •      9,004           "'''l''Indlt.ur'''!I "'1",.1,,:
C4., el~cctrlcc, vater" aJnhary vtlUtlell •••               +
                                                                                        7 .M;)
                                                                                                          Ch.:>n(tI! In l\n' .. ~ nat lu'tt~1 rr",suet •......• +-:i         lQJ.))4      +"        ~'.Ztl

"hole".. l .. tr~<le
~lllil trade
                                                            +     51 ,':ll)to
                                                                 1~5, 111'J
                                                                                 •       2,61~
                                                                                       U .476
Fln.nc        111,ur..oce.' r l "st .. t            .            lJ4.1.13              49 .11~
'e ..son.,1 soJrvlce',                             ..             17.716                6 .8~1I
'hy,l~l"n,'" dtntlat"                               .              Ill.OM               7. I 'IS
Ilospl t .. ls" l.aboratory serviceli ••....... , •••              16.10]               6,>',41
Educ.,tlon (prlv.,te) •.••••.••••••••• , •••.•.••                   'J,Ll2-              ),~Sl
Othtr seCIOU C/.                                    .              ')I), 'i')6           8. J)'i
   Total ChJIl.e -In btlS10e" reccel pt             .            3Tf:T'ir             ~;m

    ~/Sou(ce:        Ccor:OitiC tfLects of i'e.t.ier~tl ":ontrj.:>utiO:l$ to the U.S. Scnool Luncn ~roQra;r.:
      -Calcnoc:r Year 1~tl2 ..no r·'I.""c.JI ·~earl",-,,,, ,;:.iUI\ Puc.. ~\l.:~ ;;o.-J:JjJ.-.:>e?t~-r;~---~--
   bjn.;-;'onp;-n1clpoint, hnusehol";;.ktOI";,;-und' 51,Oii5 million tel funIS the federal Ct:S:1 cont'dbutlon to. 5chool~. The school"
 1tIl~"ndlt\lre gf tbe t'i1~h cgn.trlbutlgn     ,",I'        t-fUted ll5 an loer"ull 1n Unal dll"'Sf\d of thts.48lOunt'. Mellt1n.. this Increut' 1n final
 a~lII.. n<.lTe"ulp~daddl.t\l.'n,d econo"'le Ilct1vltll. This tncr':!II~ed".~tivl.t)' nsulted In'._ contrlhutlon to' !tfOU natl"n;l.1 product.;.f
 S)9LSH t:>o\lund ..
     cr.):!e'n,;.n;>:>rl1c.l,,3nt !'tUUS',hal" 5ettOr.IIY aud si19.211l tll.>u. .nd to f"nd the teder",1 (_dlty cnfltrlh•• :IO('I t'" 5(hlX'I.s. Th",
 r ..- aer.11 putchIl5':;~,.?e co..-..h.1es vas ~U:;':tl!d II' II:' tncre.lse In ftntll des.'·...! of till' ........lIt. ~ett"'r. thIs InCh',"'" In f1n.11 ,s,'...."nd
 h,"ulfed .d.H.t.}<;'n... .;;~":on.,.. lr ·<>oct Ivlty . .'.~rhl, Incnas..d >tr.llvlty' re~ .. lud In a l;ontrll>vll.oo tu ltrMS. n.n I"",.. \. pr",Ju<'t of SVI. 21) lhOUS<lnd.
    ~/;]thel' 1i.t::I."n 1_'..., "'~~r"~3rt' ~g""",';"'d of: DIrect ~nd tt>Hl'~ferred I..mns: bv~lnc~9 tu.vtl's:itu: ,.Hlo;" ",,\>;dlt,~· ~' ..d.:r"l.
 Stilt" , lot31 Il<>vcr::'""nt ~nl"r;"1~e,; .. nd "tlua servlc~J.
    ~/rn thl' .. l·t'';'" It l.s c.·.. i'~"'d prl:t-lrllr lie dtp.r...clnl .... uf Schuol lunch equlp::wnt.
     "In 1974, the cash transfer of about $1.1 billion re-
     sulted in a net increase in business receipts of $573.2
     milli'on'a'nd Tii'":'GNP of $397.5 million. With respect to
     ~~siness .rece£pt~, ~ome sectors gained while others
     would-have g",ined' more i f there had been no program.
     Thus.,- agriculture, food manufacturing, and the whole-
    ,sale trade sectors gained a total of $942.6 million;
    whe-re-astother sectors such as the retail trade sector
     woula;have gained $106.1 million more without the pro-
     gram. Schools buy primarily from wholesalers and food

     "Commodity distr ibution yielded analogous resul ts. In
     qscalye"r 1974 when the USDA purchased $319.2 million
     of food:,pr()ducts which were CJistributed to the schools,
     the businiss receipts for the Nation rose by a net
     $409.2 mil,lion and GNP by $50.2 million. Agriculture,
     food' :mailUt,actur ing, and the wholesale trade sectors
     gained $556.;~ million in business receipts while retail
     trade would"'have gained $41.5 million more in business
     receipts~wiitlout   it."

      Table 5.4 on the following page presents the number of
new jobs associated with the change in business receipts due
to the cash and commodity contributions. Whenever the change
in business receipts is greater with the Federal contribution,
a n+t1 is shown. When the increase in business receipts is less
~ith the eOntribution, the job figure is cited with a 11_."
For all sectors, 26,383 jobs were added by the cash contri-
bution, and 12,052 jobs were added by the commodity program.
      The economic impact of a cash contribution to the school
lunch program is not the same as the impact of a commodity
contribution. Cash contributions, for example, tend to in-
crease employment i~- the "Manufacturing: Bakery products"
sector. Commodity ~ontributions have just the opposite ef-
fect.   (See table 5 ;'4. )
     Notice also that the economic impact reported for cash
in the table~ is based on a funding level about 3.4 times
greater than that used for commodities. (See footnote b,
table 5,.3: $1,085 million divided by $397.5 million.) Thus,
while ~able 5.3 shows the cash contribution to have the
larger overall impact on meat and poultry manufacturing
($137.7 million vs. $136.5 million), the commodity contri-
bution has the greatest impact ·per dollar of Federal funds"
contributed to the school lunch program.

                                   Table 5.4

      ~et    Changes in Job Numbers Resulting from
 Peaeral~ContrIEutlons to     WSLP, fiscal Year 1974

                                                                    :;(.ol chan'lfC ir. nUmtH:r
                       Sector                            .      .       9.L....1£~~::!!L!2.           _
                                                             C uS il                (omIT'Oll i t Y
                                                         contritlution              contrioutian
Agriculture, forestry. & fisheries ...••.•••..•. :             + 26,389              + 16.281
~linlng ••.••••.•••.••.••.••••••••••••••••••••••• :            +     15                        12
Construction ..••••.•.•••••••••••••••••••••••••• :             +    269                        13
  Food manufacturing--
    Heat & poultry products ••••••.••••••••••••• :             +_1.942               + 1.925
    Dairy p .... ducts .•••.•••••••.•••••••••••••••• :         + 4,665               + 1.225
    Grain mill products •••••••••••••••••••••••• :             +    304              +   786
    Bakery products •.•.•••••••.••••• _ ••••••••••   :         +    815                  118
    Canned & preserved food·s •••••••••••••••••••    :         + 2,447               +   943
    Other foods [, beverages ••.•••••••••••••••••    :         +    272              +   636
      Total ......••.. , ......•......•.••.•...••.   :         + 10,445              + 5.397
  Nonfood   ~~nufacturing--
    Clothing ••••••••••••••••••••••••••••••••••• :                     2.744             1.049
    Other nonfood manufacturing ...••.•••...•..• :                       945               286
      Total .•.•.................•..••.•••••.••• :                     3.689             1.335
        Total manufacturing .......•......•••••• :             + 6,756               +   4,062
Local & suburban transportation .•.•..•.••.••.•• :                       276                  111
All other transportati0n ...•.•.•............... :             +         584         +        293
Communications ................•............••.• :                       372                  186
Gas, electric, water, & sanitary utilities ..••• :             +    56                        122
Wholesale trade ......•.......••..••..•....•.• ~.:             + 2.392               +     120
Retail trade ..•.•••.•.......••..•..••.••.••••.• :             - 11,265                  4.403
Finance, insurance, & real estate •.•..••..•••.• :                2.694                       987
Personal services ...•.......•••••.•••••.••.•.•• :                2.648                  1,019
Physiciang & dcntigts ............•..•.••..••... :                  613                       242
Hospitals & laboratory services .....••••.••.•.• :                2.121                       844
Education (private) .•.•....••..••.••••.••...••• :                  262                       102
School lunch ........••....•....•••..•• , •...•..• :           + 11.806                         7
Other sectors (nl,t':! a l                         :              1.633                       656
  Total number of new jobs ...•.....•..•••.•...• :             + 26,383              + 12,052
  al Other sectors is an aggragate composed of direct and transferred imports;
business travel and gifts; office supplies; Federal, State. and local govern-
ment enterprises; and other services.
Source:     Lconomic    ~[f0ctS    of   Fe~erQl   Contributions to the u.s.
            ~~Fio~!LUn~h Pr £9..ra~:.._ caI£~~     aE.       '{('a£_!.,il.~ - a~~_ r'T~cCl!
             '{e~.!.~74,   USDA   Pun. ALR No. 1::ilf";'"Scpt. 10G.

     An important capability of the sector-by-sector approach
is that it can be used to compare the economic (and to some
extent, the income-maintenance) impact of proposed NSLP fiscal
actions. Such impacts, however, are measured within the con-
text and specifications built into the model. 1/ The model
does not describe NSLP's impact on the consumption of a par-
ticular commodity (e.g., lettuce, cabbage), nor does it
describe the program's impact on the farm and market price
of food. ~/
     We believe that ERS' initiative in developing the sector-
by-sector approach is commendable. While the approach may not
address the specific issues that we believe are essential for
evaluating NSLp·s effectiveness as an agricultural program,
it does proVide an important perspective of NSLP's contribu-
tion to the Nation's economic policies. The ERS model may
be of use to decisionmakers in appraising the economic effects
of alternative NSLP fiscal policies.

l/In the case of the ERS model, economic activity is primarily
- measured in the context of business receipts (or sales)
  among industry sectors; computational ?rocesses are guided
  by specifications describing the demand interactions between
  sectors. The demand specifications (describing how each
  sector apportions its "income from sales" toward purchases
  from other sectors) are perhaps the most critical part of
  any sector-by-sector model. The 1976 study, based on a
  retrospective analysis of NSLP's economic impact during the
  1972-74 era, was supported by ERS' evidence that there
  were no fundamental differences between the model's speci-
  fications and actual data for that period. The use of the
  same specifications in computing present or future impacts,
  however, implies the assumption that purchase patterns
  will remain constant and that technological change will
  not materially affect the real cost of goods.
2/The ERS model, like any input-output analysis, implicitly
- assumes that all additional demand is "real." In"an in-
  flationary economy, characterized by "tight resources" and
  inflexibility of supply, the additional demand incorporates
  "price increases ll which do not represent "real additional

     NSLP's effectiveness as a nutrition program depends on
student participation, which to some extent depends on the
form and content of the lunch itself. Since participation
in the program is voluntary. the lunch, in addition to being
a nutritious meal, must be presented in an appetizing manLer.
In appraising the success of this endeavor, authorities have
expressed concern about
     --the low levels of student participation (implying
       difficulties in getting children to eat the NSLP
       lunch) and
     --the excessive amounts of plate waste (implying that
       the lunch is unappetizing or too large and that only
       a portion of its nutritional value is consumed).

     There are two aspects of NSLP in which agricultural
considerations may be retarding the program's nutritional
effectiveness. These are (1) the Secretary's prescribed meal
standard and (2) the method of distributing commodities. In
each instance the potential problem is one of administrative
practices rather than legislative provisions and appears to
adversely affect student participation.
Differences in type A pattern and
the programis nutritIonal target

     Legislation requires that lunches served by schools
participating in NSLP shall meet the minimum nutritional
requirements prescribed by the Secretary. Although the
nutritional target is one-third RDA per meal, the Secretary
also requires the use of specified quantities and groups
of foods.
     The current meal standard, the type A pattern, is
composed of: !/
     --One-half pint of fluid milk.   ~/

l/These requirements are based on the food needs of 10- to
- 12-year-olds. Size of servings should vary in relation to
   the age of the children. Substitutions may be made to meet
  special medical needs.
2/The definition of milk was expanded in 1973 to include
- fluid forms of whole, low-fat, and skim milk, cultured
  buttermilk, and flavored forms of these milks.
     --Two ounces (edible portion) of lean meat, poultry, or
        fish; an equivalent quantity o£ an alternate such as
       chee-se, cooked dry beans or peas or peanut butter, or
       an equivalent combination of any of these.
     --Three-fourths cup serving of two or more vegetables
       or fruits (full-strength fruit or vegetable juices may
       be counted as part of this requirement).
     --One slice of whole grain or enriched bread, or an
       acceptable equivalent.   l/
     --One teaspoon of butter or fortified margarine.  (This
       requirement was deleted from the pattern in June 1976.)
This food-based pattern (developed in 1946) provides a prac-
tical means for insuring that all food service personnel,
regardless of their training, can understand the program's
nutritional requirements. The pattern also reflects the
fact that, until recently, most schools prepared NSLP lunches
primarily from raw food ingredients.
     Although the type A pattern may be well suited for using
donated commodities, its flexibility as a meal-planning tool
is essentially limited to a choice of items within specified
food groups. The pattern limits the form and content of
school lunches. One-third RDA can be provided in alternative
ways. As stated by Dr. Jean Mayer, Professor of Nutrition
at the Harvard School of Public Health:
     "American eating habits have changed drastically in the
     last 20 years and today's typical lunch is not usually
     a full-sized meal. Peanut butter or ham and cheese On
     whole-grain bread, a glass of milk and fruit or a glass
     of orange juice, is a nutritious and well-balanced meal,
     and more in keeping with today's eating habits. Food
     does little good unless it is eaten. And, now, of all
     times, we can i l l afford to waste either money or food.
     Perhaps we should begin to change the school lunch
     program by trying to save food and money. We will be
     better able to feed every child in need."

     In commenting on our report, FNS stated, IIDr. Mayer's
example of a typical lunch supports the Type A Pattern with

l/rn 1974, the definition of bread was expanded to include
- crackers, taco shells, pizza crust, etc.

the exception of one less fruit or vegetable." (See app. I.)
We might add that milk is also an optional item in Dr. Mayer's
lunch. This illustrates our point relating to the differences
between the type A pattern and the program's nutritional tar-
get. In point of fact, and irrespective of its nutritional
value, Dr. Mayer's "typical lunch" does not meet USDA's pattern
requirement. It would not qualify for Federal reimbursement
as an NSLP lunch.
Some food service directors believe
modlficatlons In meal pattern will
improve partlclpation and decrease

      In 1975 the Senate Select Committee on Nutrition and
Human Needs asked State School Food Ser-vice Directors: "What,
if any, modifications in the meal pattern should be made to
help increase participation and decrease waste in the lunch
program?" 1/ Opinion was divided. About half believed that
the type A-pattern was needed to safeguard the program's
nutrition standards and that greater emphasis on nutrition
education would improve participation. The other half sought
definite changes in the type A pattern. Some of the responses
     "There should be more diversity permitted * * *. The
     current pattern of the Type A lunch contributes to
     food waste and discourages paying students from partic-
     ipating in the program.'1

     IISerious considerations should be given to re-structur-
     ing the meal pattern requirements to permit children to
     receive the basic nutrients without having a specified
     component, such as milk, as a daily requirement. It is
     possible for a single component to become prohibitive
     because of price."
     "Permit a dairy alternate for fluid milk just as we
     do with meat alternate.II

     "Remove the butter requirements in the Lunch Program.
     Also change the fruit and/or vegetable requirement that

l/D.S. Senate, Select Committee on Nutrition and Human Needs,
- School Food Program NeedB--1975, D.S. Government Printing
  Offlce, Wash., D.C., Apr. 1975.

     stipulates two sources of such fruit and/or vegeta-
     bles." 1/
     "Offer either j'uice or milk, not both.   11

     t1Accord-ing to the requirement, in order to be counted
     in meeting the meat/meat-alternate requirement the
     high protein foods must be served in a main dish or
     main dish and one other item. We fail to see the
     nutritional advantage of this for it seems immaterial
     where the protein source is used as long as the required
     level is achieved for the total meal."
     By eliminating the pattern requirement and prescribing
meal standards as simply "one-third RDA" (1) the type A lunch
would continue to be an acceptable meal and (2) nutritionists
would be accorded greater flexibility in designing menus.
This action might provide lower cost lunches and, at the same
time, be more effective in achieving the program's nutritional
objectives (e.g., provide higher levels of student partici-
pation and a more complete consumption of program lunches).

     In commenting on this report (see app. I), FNS described
as apparently unfounded our contention that one-third RDA can be
mer in many forms and that the inflexibilities of OSDA's food
pattern contribute to higher costs, food waste, and a meal
design which is not representative of today's eating styles.
Further, in regard to TSNS (see ch. 3), indicating that iron-
to-calories ratios were lower for school lunches than for
children's home diets, FNS stated:
     "A recent nutrient calculation of the Type A Pattern,
     based on foods representative of frequency of service
     to 60 test groups over a four week period shows that
     the Pattern furnishes approximately 8 mg. iron per
     1,000 calories. This amount is well over the 6 mg.
     iron per 1,000 calories which is the amount expected
     from a varied, well-balanced diet as specified by the

     As we pointed out in our description of TSNS and will
mention again in describing figure 5.1 of this chapter,
there are indications that the type A pattern provides too

liThe butter-margaine requirement was deleted from the
- type A pattern in June 1976.

little iron. We examined the research 1/ supporting FNS'
statement and found that the statement ~as not based on the
type A pattern at all, but rather on the School Lunch Pat-
tern, 1976. The School Lunch Pattern, 1976 (hereafter
referred to as the 1976 Pattern), was developed by the
Agricultural Research Service in consultation with FNS and
proposed as a replacement for the type A pattern.

     The 1976 Pattern (see table 5.5), and the concerns
about the type A pattern which it attempts to alleviate (see
table 5.6), support our position about inflexibilities in
the type A lunch; and it should be noted that the 1976 Pat-
tern did two things:  it reduced the fat content of lunches
and it increased the use of iron-enriched products (e.g.,
bread and rice). Thus, it not only improved the iron-to-
calorie ratio of lunches, it also improved the total amount
of iron served per lunch.
     On the other hand, the butter-margarine requirement
has been removed from the type A lunch. This action, while
improving the meal's iron-ta-calorie rutic (by decreasing
calories only), has not increased the total amount of iron
provided by the NSLP lunch.

     USDA also expressed concern about the acceptability of
the school lunch program in its present form. Changes in
NSLP's meal pattern have been the subject of much review.
The principal alternative proposed thus far is the Nutrient
Standard Menu (NSM).   Using the NSM procedure, a meal is
designed to achieve a specified nutrient goal (e.g., one-
third RDA). The emphasis is on the nutritive value of in-
dividual food items rather than on a meal pattern. One
means of assessing the NSM altern. eive is by analysis of
"plate waste," reported as a problem with the type A pattern.
A study by Colorado State University 2/ compared NSM with the
type A pattern. Data was collected {rom students consuming
type A lunches at 58 schools in the fall of 1972. The
following spring, schools changed to the NSM t~chnique and
the survey was repeated.   Figure 5.1 on page 75

l/USDA, Consumer and Food Economics Institute, Agricultural
- Research Service, School L~nch~~~£~~_!~2~~ July 1976.

~/Harper, J. M., and G. Jansen, ~mEari~of_!~~~~_NSM
  Menus in the National School Lunch Pro~m: Phase I I
  ~~E2rt~-Colorado State university, FNS Contract No:-r2-35-
  600-65, Aug. 1973.

                                                                              Table 5.5
                                                          School Lunch Pattern, 1976

                                                                  ----r;::;-t,;;;j                   I                t:1c1:::i\i·~I:O;;I-----r--:if-c7."j;""<l:;r-I---

_ _ _ _ _ _ _ _ _..,.od _...                                                   ch,ld                                         ,htld                   so',,,,,] r l!d
                                       ------                =-i~~i..)!::i'.r....:~·,-;::f-,-;;:.:1-;.-_r   (;·6 ,,7.',::,-;' ~--9-:iT~-.·;'·'--i:'::Y;~ ... r-~;~---·
COt-YI":! .'(';It (It·,,I''I.). pr'ultc\',                                        -------                      ------              -- --        ._-- - - - - - - - - - -
 fhl,: ~/                                                          1 Dlluce        J OIlIlCr.               1-1/2 o"nct~             2 (lunet',;      J ('une ... ,>
    Att"r1"'lc,; .",,)' <:<,pla:,. ,Ill "r
    p.Jrt of the C,".H, l'uulle)',               (or
    [htL: 1 ct:r., 1/2 r"O' c.-.,kcd
    dry h"nIl5 "r (I""". J t.l!ll,,-
    "1'''''1\'; .,,,,,nul !.outler. 1 uunCe
    clwdtl:lr-lY,,~ cl;,·c~l'!.       ("1.1          ,;
    1 o"n-::,' uf c",,),-ll 1""n          ~'   lL.
    poultry, or Chll.

\'Cl;ct.ll.oll' mul/or fruit:                                      1/4 cup          1/2 cup                 1/2      CU!'              3/" cur'
    ~:<lSI    ~n<:l,llolc   ill least t,.. o idndll

lirei'll (vhn1;:o p,ralr! n         (,llrl:~",J)           £/}/2      ~jl('"                                                                             1-)   GJJt·c~
    I,] U"(I,.,l"l" t,,,y r'·:'J;'l~<: nIl       01"
    1';11 t l,r till:' bl"":l.J:.
    Il~k"ry      I'ro'I"C:~J1
                           n,adt· .,( ~·hJl ..-
    r.raln       l'nl"ld,,'d «(",,1 Ill"
    flOlll" b/~ In e'l' pI ('IHlched
    r·lcl:' ul" r~st" t.lllr ro:'l,l,'Icc 1
    ~11e(: (Of ltread.

tl1lk. :.                                                                           3/4   cup    !/J/4                             1/2 plnt              1/2 pint
     Alternate!! tI.1Y rcplMI' all or                                 '"'                                     CUI'

    part of thl! ",:H.:.        J{4 (01,1:11:'''
    cl"',hJ,,r-t)'l'c ch<:,'~.. ej. 3/4
    cup or CN (olf,C d, .. ,·,:;;. J/t,.
    cui' of Ie.. en:;,;. "I' j,"" r.dlk,
    1/2 Cu!' of1/2
    rCI,jdt·CS   ul,n;,..Ol"(,j_'_.,,_i:UI"._'
                   CU!' of cd II:.                           I.
--_.__ .-                                                                                                                    ----------_.-

                     for 'r)·!'c A Scl>ool Lunches," (1',\ 119) Hay 1914 rcv15!on, par.r. 11.

             £/      Serve 111 OunC,' Or Plor., or precool.:,'d dry infant cel"...\1 or otill'r cereal ..:itl> ;tdded JrQ:l In place of ot'

                     1n ilIddltlo>:I to    cr.. ad        wo or elor" dolYS per 5-day           "'""i<;,   if such CCtc.,l 1,; not l:cn.·,'d to the child ~J1

                     part of II    brCil~folst        or inaek served ilIt the center.

          ~/         Or e1tht slices of brr ...I, or e<.jui,,;l!cnt, pel' 5-d.lY                     "'""k.
          ;::./ If ehel'.:;e n::placcs :'Cillt In the lund., it c.anrtol rc,,1.,c:>' 10ilk al~o.

             !/ S... rve      1/2 pint if It is Impr.lctlc"l to ser"e 1/2 Cllf' or )/4 cu!'.

              Rf.H~:;Il::R:     Thel'll a"'::lU"ts of foods tlrc :"'pottJllt tiS the fO'llldatIoll for                      :I.   nutritlous lUllC'"h.   TIlclr u~e Is

             Source:                  School Lunch Pattern, 1976.

                                                       Table 5.6
         Concerns About the T e A Pattern and Chan es
  to Help A leviate the Concerns 1n the     Pattern (note a)

        Exeessive      pl~te V~$;C     cecues     bcc~use
        the   P"ltefl\~-

          Requires     I~     Coods lh3t 3ft not                A-ouncs DC Vtect"blc. and Crult~ ror
          well liked,      not~bly     vcr.cc3ulcs              children of ,o~c "r.c. arc reduced.
          lIud huiu.

          kcquircs     ~orc   rood    th~n so~e                 Pattern i , specific Cor S different
          children ClIn C"l.                                    lI&C level,. '''ttern provides only
                                                                18 to 25 pcrt~nt or tht RDA for
                                                                food'cncrtr (caloricl).
          aC,\uh'e.:;. ;'Illk, "hlch SOlilC children            Hilk   prDdutt~          that u.unlly tan      ue
          c~nl1ot ~olcr:ltc \I~ll Cot" phy5l.o-                 lolcr3lrd. such "r; c:heeso auJ. Y0\l.urt.
          lotteal rc"sons.                                      =7 bc used in "htt ot .Uk.

          kesulu in lunches ~nl1kc lho$c                        flexibility         i~   added:     Rice and     pa~ta
          chiJdren usually cat.                                 E3y replace brC"d. 3rcad rt,\uir~nt
                                                                ':,\7 be ."rcad over 5"d:l)' .,('('1:.
                                                                Alt('lc:ltes for .IJ~ NAy be $erv~4.

        Co~u   of lunehes :Ire hi&h because
        the:: Jlattern--

                                                                /ur'lollnu   tof   lrIeat (or chUdren of         '('lilt
                                                                act's ore reduced.
          MequiN$ tile     ~\uch     other Cood.                A~ol.lnts    of food :Ira   s~cct(lel.l     to    ~cct
                                                                needs of 5         ~tO l~vels,

        1'1lIn"t"c vllrl<:l.l lunches is dHCicl.llt
        beeau~e thc rlttCrn--

          RequitQs atlk 0$       bevc~ar.c       every   d~y.   Altern~tes IrI~Y         be    $~rvcd.

                                                                Altcrn~te:;        tl:lY be served.

        IlutrSt:lf'll:l1 lill,11lty <If the lunch is
        'l.lJelltIOl'l,1ble orr.!us,- the pllll"rn--

          Is o:lsed on 1"Jl,\ rh .. t .He out or date           :::o~t recent tOil 091:') "'cre use,l.

          F'tovl,Jc~   too little      ~ron.                    Well OVN ,          :tor. or   IrOn pcr 1000
                                                                kelilorjts-~,,~a~nt e~Jlc('ted tlO~ "
                                                                v.. rlrd, .... ~ll-h.'hnecd ,Hel-·,iI: ,'ruvid:('J.

          I-ro.'~ote$ I"M'loe:; th.. t   OIlC I>tell        b/Butlcr-a,ltr;,1rlll... rc"uirC:II('rI( $1: dd('tcd
          '" l.u (""tenl.                                   -  rr"", p:otu'rn. r:lt In P.:sUC:CIl I:;
                                                               II~ft~J to rr~vld~ 35-40 pCfctnt of

          I'I'O~l('s lu"ch",s thOlt      .1[('   Meh            LUntl!.c5 c:~nutnlnr, rJc~ :lnd 1.:I:;la Olr~
          tn t,)rbo!oy.:r.'l ....                               not rc~~lrrd to (OntOlln brc~d also.

~/30urce:        School Lunch Pattern,                                   1~76.

b/~he butter-margarine requirement was delete6 from the type
- A pattern in June 1976.

shows represen-tative nutrient values for these meals as
served to abo~t 1,500 fifth-graders. During the same period
Rutgers University examined the nutrient content of 160
sack lunches in New Jersey elementary schools. 11 For com-
parative purposes, we have included the nutrient values
of these sack lunches with those of NSM and type A lunches
shown in figure 5.1.

                                                         'ICURE   ~.1


      a tYPl" A LUNCH
      IE NS/Il LUNCH
                 •     H<»IE-MADE SAC"' LUNCH
  I.S            0     PLATE WASTE


   ''-='"                                                                       VITAJotIK A   VITA-lIlIK C


     In terms of the one-third RDA requirement, both NSM and
the type A pattern evidenced shortfalls in the amount of
nutrients actually consumed by children. While NSM in-
dicated a slight increase in nutrient consumption, this ap-
pe_ted largely a function of portion size. Clearly, both NSM
and the type A lunch came closer to meeting the one-third RDA
requirement than the sack lunch sample. (It should be noted,
however, that sack lunches can be but were not structured
to meet NSLP's nutrition standards).

l/U.S. Department of Agriculture, Food and Nutrition Service,
- Proceedings of the National School Food Service Conference,
  June 27-29, 1972.

     IFNS told us that two additional studies 1,2/ have com-
pared computer assisted nutrient standard menuS-against the
type A menu and have found that students consumed proportion-
ately more of the meals planned by the type A pattern.)
     with respect to waste, figure 5.1 indicates that the NSM
technique did not provide a marked advantage over the type
A pattern. However, while 494 of the 675 NSM menU3 in the
study deviated from the type A pattern, most of the devia-
tions were minor; most of the food components were selected
from the type A school lunch recipe file. 3/ The comparative
amounts of waste indicate that children preferred the foods
and/or pattecns characteristic of sack lunches.
     We believe NSM's emphasis on designing meals to fulfill
a nutritional standard is desirable, particularly if NSLP is
to provide an optimum supplement for the schoolchild's home
diet.  (See ch. 3.) However, the essential objectives of NSM
can probably be achieved by simgly promulgating "nutritional
standards" together with a series of sample menus (or pat-
terns).   Overregulation or unnecessary requirements in the
program's meal standard would probably inhibit nutritionists
in designing lunches which are more representative of today's
eating styles.   i/
l/Food and Nutrition Department, Dade County Public Schools,
   Miami, Fla. Co~ar!~£~of-!le~_~~~~_££~~ter_As~isted
   Nutrient Standara-Menus, USDA Contract No.-r2-!5-600=T16,
2/Division of School Food Service, Memphis City Schools,
- Memphis, Tenn., Comparison Of Type A and Nutrient Standard
  Menus., USDA Contract No. 12"35-600-115, Feb. 1975.
3/Following the Colorado State Study, 29 participating menu
- planners were authorized to use the NSM in their normal
  school feeding programs. Many have since reverted to the
  type A pattern because of its simplicity and because the
  number of recipes provided in NSM planning guides lacked
  the flexibility of the type A pattern. Only two menu
  planners continue using the NSM in 1976. FNS is supporting
  these planners in expanding the variety of foods available
  in NSM planning guides and in appraising the ,ystem's per-
  formance and support requirements in a school.

4/1n compliance with Public Law 94-105, enacted in Oct. 1975,
- USDA is now conducting a study which will, among other
  things, examine possible relationships between plate waste
  and the nature of the type A pattern. The results of this
  study are expected to be available in the latter part of
Timing of commodity distributions has an
important impact on nutritlonal obJectlves

     Although commodity distributions play an important role
in providing low-cost meals, improvements in the timing and
quantity of deliveries may be needed. The following state-
ments, reported by school food service directors from Kansas
(which changed to a program of cash SUbsidy in lieu of com-
modities during fiscal year 1975), provide an illustration
of how commodity distributions may interfere with menu
planning and students' acceptance of the NSLP lunch: !!
     "* * * [we] received 21,120 frankfurters to use in the
     month of May. They could not be held over the summer.
     Students were very unhappy with the lack of var aty in
     the menus. Particip~tion declined."

     "We received 120 cases of orange juice February 15 of
     1974 and 142 cases of orange juice in August. We will
     still be using this orange juice most of 1975-1976."
     "It was such a pleasure to plan menus and not have to
     worry about a surge of commodities."

     Similar problems are noted in other reports. In most
instances, these "commodity surges" appear to be lvoidable.
They are both a misapplication of food and a detriment to
NSLP's effectiveness as a nutrition program.  It would be
appropriate for USDA to investigate these situations and,
where necessary, to implement corrective procedures to see
that NSLP's nutritional objectives are not unnecessarily
compromised by administrative difficulties or support of the
agricultural market.
     We believe NSLP is successful in strengthening the
domestic demand for agricultural products. This belief is
based on (1) indications that school lunch participants
consume a greater quantity and variety of commodities than
would otherwise be expected, (2) substantial purchases of
program foods in local markets, and (3) the program's
demonstrated capability as a commodity outlet. However,
the degree to which this strengthened demand affects the
market price of food, or contributes to NSLP's effectiveness
as a price support mechanism, is unknown. Such a deter-
mination would require further evaluation.

l/George, I., Testimony before the Senate Subcommittee on
- Agricultural Research and Gene~al Legislation, Apr. 1975.

      However, we are not sure that an evaltia:tidn' of NSLP I S
impact on the agriculture market will provide information
appropriate to its cost. In view of recent changes in the
Nation's agricultural economy (while significant market im-
balances still occur, concern has tended to shift to the
problem of shor tages and away from the problem of surplus(!s'),
the program's ability to increase demand (and hence, :the
market price of food) may no longer be an appropriate state-
ment 0f NSLP's purpose. Therefore, we believe' the question
as to whether or not the program's agricultural objective
should remain operative is an issue requiring overall p·o-licy
determination by the Congress.
     In order to determine the nutritional standards needed
for the National School Lunch Program, we recommend that the
Secretary of Agriculture
     --Oetermine, with assistance from HEW, the nutritional
       standards needec for NSLP to best safeguard school-
       children's health; and, if found desirable, revI-se
       the program's meal regulations to reflect nutritional
       requirements that will provide meal planners with
       planning flexibility, improve the program's cost~
       effectiveness, encourage higher level. of student
       participation, and reduce plate waste. I/
     --Determine the effect of commodity distribution surges
       on NSLP's nutritional objective and," if surges are
       determined to have a significant effect, implement
       corrective procedures so that agricultural consider-
       ations do not compromise the prog:ram I s nutr itional

l/In an earlier report entitled "~he Impact of Federal
- Commodity Donations on the School Lunch Program" (CED-
  77-32), we recommended that the Secretary of Agr icul ture
  include a nutrient standard as an option to the typ~'A
  lunch pattern to provide menu planners with greater flex-
  ibility in using commodities. In making that review, we
  did not evaluate the adequacy of the type A pattern in
   improving the nutrition of students. However, on the
  basis of our current review of various studies of
  NSLP it would appear that an evaluation of the nutritional
  standards for NSLP should be done.

     HEW' advised us by letter dated April l~, 1977 (see
app. ItI), that it was willing to assist USDA in carrying
out the intent of our recommendation rega'rding NSLP' s
nutritional standards.
      In a letter dated April 20, 1977 (see ,,['p. I), USDA
expressed'.   c9nc~rns,   Similar to ours reg'arding prog'rarn partici-
pation and plate waste, but noted that "there are ways of
addre'ssin,j' these concerns short of abandoning nationally
establ'ished mea"! standards.     II   USDl~   cited various complexities
associated with determining nutritional stand'ards for NSLP
beyond the present goal of providing one-third or more of
the RDA for children of various ages. It suggested a list
of activities for expanding program puticipation and re-
ducing food wastp. which included working with food service
personnel to improve the appearance and quality of food
served;' revising the type A pattern to allow smaller portion
sizE!$ for elementary school students; and eliminating the
sale of snack foods during lunch.
     We believe that USDA is earnestly attempting to im-
prove NSLP's performanc~. However, there appears to be a
reluctance on the part of FNS to consider administrative
changes in the program's meal standards that might improve
NSLP's effectiveness in meeting legislative objectives, es-
pecially those concerning the type A pattern. We have not
recommended the type A pattern be E!liminatE!d out of hand,
but rather that nutritional standards bE! dE!tE!rminE!d and that,
based o'n such standards, needed revisions be made.

     In regard to our recommendation concerning the effect
of commOdity distribution surges on NSLP's nutritional ob-
jective, USD~ cited their response to our earlier report
(CED-77-32) which statE!d that
      "the Department is required, for the most part, to
      givE! first priority to itE!ms in surplus and in need
      of priCE! support, so that controls OVE!r the timing
      and availability of deliveries are often restricted.
      * * * Greater efforts will continue to be made to
      achieve improvements and we will encourage the States
      to establish similar procedures to the extent possiblE!
      in making deliveries to their local districts."
     We believe USDA's actions are beneficial. However, in
view of the fact that NSLP's E!ffectiveness as a price support

mechanism has not been ascertained, we continue to:have ques~
tions regarding how USDA allocates NSLP's priorities between
agricultural and nutritional objectives.
     A typical problem arising in programs which have mul-
tiple goals is tha t, tlnder certain cond i tions, goa.l confl lrts
may precipitate undesired side effects within and outside
of the program. As indicated earlier, NSLP is a case in
point. In addition, a desire to use the program in support
of emerging Federal pOlicies may have introduced add:H;ional;
unwritten objectives which influence the scope and purpose
of NSLP. For example, though not explicitly given an income
security objective by legislation, the program is currently
classified as an Income Security function within the Presi-
dent's Budget.
     In view of the Congressional Budget Act of 1974's em-
phasis on clearly stated legislative intent, and the changes
in national priorities since enactment of the National School
Lunch Program, the Congress should provide policy guidance
indicating specifically what the goals of the program should
be, prioritize them, and have the program evaluated accord-

                                          PART III

                               PROGRAM COVERAGE

    ',D,ur ing:, ~he 1975 school
                               year, the, NSLP lunch was served in
about,88~800,schools (,about   81 percent of the Nation's total) with
an'e,rir'oilmento-f approXImately 44.8 mill ion students. The number
of'stude'ntsiparticipGitinq in NSLP was 25.4 million and, of this
riumb'er,'9~9:million r'ecelved free or reduced-price lunches.

                           PERCENT OF' U.S. SCHOOLCHILOREN
                            BY PARTICIPATION CATEGORY _
                                     FY 1975 (EST'''''''' TED)

              IN NSLP SCHOOLS
                              12.1                   R fCiU LA R·P ~l CE

                                                             ---- --
                                                     ---- --

      While the number of schools serving the NSLP lunch has in-
creased in recent years, there has not been a proportionate in-
crease in the number of participating students. The participation
of regular-price students has declined, but, since the number of
children eligible for free and reduced-price meals has increased,
overall participation levels have tended to remain constant.
      Because program coverage is an important measure of NSLP ef-
fectiveness and bec~use student participation in the NSLP lunch
has been generally low, we thought a review of participation
studies might provide insights for future program improvements.
Cha~ter 6 introduces this issue by describing the growth of
NSLP availability and recent trends in student participation.
Chapter 7 provides a specific look at the factors which affect
student participation4
                          CHAPTER 6

                    PARTICIPATION TRENDS
     Between 1946 and 1970, there was continuous growth in
the number of U.S. schoolchildren. That trend peaked at 52.1
million students in 1970, and by 1975 school enrollment had
declined by about 1.2 million students. The decline, associ-
ated with a drop in birth rates during the 1960s, has to date
affected only elementary school enrollment.

     Current census projections indicate further declines in
school enrollment.  Compared with 50.9 million students en-
rolled in 1975, the 1980 enrollment in regular day schools is
expected to be between 45 and 47 million.

      The continuinq decline in U.S. enrollment and the
current shift of students from elementary to secondary schools
(where lunch program participation has traditionally been
lower) creates downward pressures on National School Lunch
Program participation levels. Actually, of course, many
other factors, such as expanded program availability, changes
in lunch prices, and improvements in the attractiveness of
program lunches may interact to change participation.


      In 1950 only 54,000 schools participated in NSLF. (See
figure 6.1 on the following page.) By 1975, however, the
program had grown to include nearly 89,OO~ schools with a
combined enrollment of about 44.8 million students--approxi-
mately 81 percent of the Nation's schools and 88 percent of
the schoolchildren.

                                                              'IWlI:E 6.1

                                                 GROWTH IN NSLP AVAILABILITY
                                                   FISCAL YEARS 1950_75 0

       us; SCHOOLS                '69
       (THOUSANDS)                              '50            ,,.

       U.S. STUDE"MTS

                                                               "'.,          ".2

                           '950                         "50            ""                          1975
   o 50u«:e:   Fi~l.  Yeor 1973 SfOtistic.s olld Hist«iCtlI Tobin, FNS/Prog,om Rtpo,ling SllIff.
               1915-dolo bosed lin prelirnillory reports of lhe FNSiP,ogrorn Reperling SloH.
               1950-65 NSLP enrollment not a"ai[cble.

     In 1974 USDA reported that 86 percent of the Nation's
schoolchildren were enrolled in NSLP schools. Of the remain-
ing 14 percent, 4 percent attended schools with other tyoes
of food service and 10 percent attended schools without any
food service at all, except perhaps for a milk program.

USDA characterized the 18,000 schools without food service
as follows:
     --Over half were private, nonprofit schools with a
       combined enrollment equal to 49 percent of all
       children attending such institutions.

     --One-third of the schools had enrollments of less than
       100 students.
     --Their students were more likely to live in urba~ areas
       and/or come from more affluent families. Although 25
       percent of the children in NSLP were eligible for
       free or reduced-price lunches, only 10 percent of the
       children in schools without food servi'ce we,re consid-
       e~ed economically needy.

     To expand the program's availability, USDA placed great-
est priority on assisting schools without a food service
capability, especially schools with a high proportion of
needy children. Although Federal funds for nonfood (equip-
ment) assistance 1/ have provided an important means for over-
coming physical or  financial constraints and bringing new
schools into NSLP, progress has been slow. 2/ The Department,
indicating increased difficulty in overcoming the attitudinal
reaso~s for schools not joining NSLP, described some of these
reasons as follows:

     --School administrators and/or teachers are against
       school lunches.
     --Children walk home for lunch.
     --Some private schools have too many competing demands
        for available resources to be used in a lunch program
        or simply want to operate free of Federal or State

l/Nonfood assistance funds are used to help schools in low-
- income areas establish, maintain, or expand food service
  programs. State and local sources must match at least 25
  percent of the equipment costs (the matching requirement
  may be waived for especially needy schools without food
  service) .

2/The number of NSLP schools increased by 3,048 in the 1972-
- 73 period; 820 in 1973-74; and 1,603 in 1974-75.

In   1~?4,    as a resul.t' of these difficulties, USDA reported:
       IIT~e,'D€,pa[tm~nt now          be'lieves that thece are a number of
       schools, particularly private schools, that will never
       join the nation"l school lunch program. The Department
       and States will continue to make reasonable efforts to
       re:ach. thes'e school's, however, and to document the rea-
       sans for, nonparticipation."                 !/
     Although the student enrollment in NSLP schools increased
over the 1971-75 period, the number of participating students
remained fairly constant. ·In fiscal year 1975, 25.4 million
students (56.7 percent of the NSLP school enrollment) parti-
cipated in the program:
                                        Table 6.1
              To !2~:L!~:£E.~g!!!~!2L~£sU.>~E.!;.!c£.!c12~!;.!c!N~!; ~~'=..!:!!; ~

                        Students                                       Rate of
Year             ~~K£II~~ ~~rtr£IE~~!~~                     12~E.!;.!c£.!cPa!2.!c~£_iE~E.£~~!2l
1971              43.1                 24.6                               57.1
1972              44.0                 24.9                               56.7
1973              43.6                 25.2                               57.4
1974              44.9                 25.0                               55.7
1975              44.6                 25.4                               56.7

~/Source:        FNS/Program Reporting Staff publications.
     In 1975 the "nonparticipants in NSLP schools r• (numbering
about 19.4 million children) accounted for 75.9 percent of
all U.S. schoolchildren who did not eat the NSLP lunch. It
appears, therefore, that the NSLP enrollment itself presents
the greatest opportunity for further increases in program

!/USDA, ~~!!!12E.~~~~~.!c~~_~!2~~l_~!_!2~~Ch.!c!~_~~!2E..!c!2.!c~~_~E.£~E.~!!!~L
   Committee Prlnt of the Commlttee on Agrlculture and
   Forestry, U.S. Senate, Wash., D.C., Sept. 1974.

     It should be noted, however, that USDA's II p,articipation''
values essentially represent the number of lunches which
would be eaten daily if the entire enrollment were in attend-
ance (e.g., no absentees). Since many students takeluhch
less than 5 times a week, these values may substantially
understate the number of children who, at one time or-another,
benefit from program services. !/
     Participation rates of students in NSLP vary greatly
among States (see figure 6.2 below). Most of the States
with participation rates below 55 percent of enrollment tend
to be located in the West and Northeast.


                                                 ~. 75"'.ndeb~

    .      .,.'.-'
                                                 [Mj. 55" 10 74%
   -       ~                                          35" 10 54%

                                                .D    34" .nd und'r

a/Participation values are based on preliminary estimates.
  Source: !~22._~'!£9.~~_Ex£!~~at£El_~£~S'V£~_!II, USDA.

l/USDA is now undertaking a study on the frequency of student
- participation in NSLP. This study is scheduled for comple-
   tion in the second half of 1977 and is expected to provide
   a better estimate of the number of children utilizing pro-
  gram services.

-, ,   ·~·ronl·.i,ts   inception,'in' 1946, NSLP has prov ided for
servirig·-;-tree."or.'~re'duced';;;'pricelunches to children unable to
pay'. the;'r:egu,iar pric'e charged in participating schools. As
of 1968.~·however" these. provisions were not being effectively
carrfed' out. Accordingly, the Congress first authorized
substantial spec.ia) furiding for free lunches in 1968; and,
in' 1·970:, 'PubHc Law 91-248 was enacted which (1) mandated
that free lunches be served to needy children and (2) pro-
vided specific guideHnes to be used in determining eHgi-
bi.lity for free and reduced-J:'rice lunches. II (The serving
of reduced-price lunches remained a State option, however.)
     The impact of Public Law 91-248 has been impressive.
Prior to 1968 only 3 million needy children were receiving
free lunches. But in the 5 years beginning in 1971, the
number of children participating in the free and reduced-
price program increased from 7.1 to 9.9 million, a gain of
2.8 million children. 21 (See table 6.2 on the following
page.)                -

liThe Secretary of Agriculture issues annual poverty-income
- guidelines based on changes in the Consumer"price Index.
  States are authorized to set eligibility levels for free
  lunches up to 125 percent of the Federal guidelines. In
   the 1971-75 period, the reduced-price program was optional
  and little used. Since 1976, however, NSLP schools have
  been required to provide reduced-price lunches, at a price
  not to exceed 20 cents, for children who are not eligible
  for free lunches, but whose family's income is below 195
  percent of the Secretary's guideline.
~/In 1971-75, FNS program reports consolidated free and
  reduced-price students into a single participation
  category. Only 1-2 percent of the total meals served in
  this period were reduced-price lunches.

                      !.='!j~~!,!! _~!!~.!!! !-.!£.!!?~ t i~ ..;?t u_d~~;§_   !!t _~t'! _ !!i'~1" r=?!l£!
                                  !!'I~_f!£~_~!:-E~~!:!!<s~.::.flli~!-~~!L !'!.22!.E!:,;.
                                    ~J ~£~~ .r.£EL 192..!:1~H:!Hir-!!~~J1!:!~5!_:H

                   ___        Re9u~!.!£.e_~!~_1                   __               Prc~    ~nd   rcduccd-~riel!   stu~~nts
                                                        Partie D-              ----------------- Par t iei'!):'
                                                           olt hHI                                                   alian
     i'illc"l                        P,1rtl ..... I"-      tall!                                  I'':HllCip-        rate
         Y-S:!!!   ~!!I2l!-S:~         ~.u!!9           lE-S:!S£!1.H            !2!1!21l.s5!        ~!.!!!1        iE~I££ntl
                             (ll1illionn)                                                 (millions)
                    JS. )
                                        J 1. ')
                                        16. 'J
                                                           48.7                    ,. ,
                                                                                   7. ,             7.1
                    ll. ')
                                                                                 10. J
                                                                                 11. 1
     ol/USO~    does not publish scoaratc pacticio,1tlon ralps COt reaular-price and (rep!
     -     reduc~~-pricc students. OU~ c5ti~t(>s Wl!re co~puted (~o~ FNS!Proqre~ P.eporting
           Stafr puhl kat ions.

     In contrast, there has been a marked reverse trend in
the number of regular-price participants. From 1971 to 197·5,
participation in the regular-price program declined from 17.5
to 15.5 million, a drop of 2 million students.
     The increasinq share of children eliaible for the free
and reduced-price programs, together with· the higher partici-
nation rates of those categories (85.9 percent versus 46.6
percent regular-price participation in fiscal year 1975), has
been an imoortant influence in ~aintaining NSLp1s overall
level of participation.

                           CHAPTER 7


      Nearly 88 percent of the Nation's schoolchildren
at'tended National School Lunch Program schools in fiscal year
1975. Yet, on an average school day, about 43 percent of the
Nsr;p enrollment did not eat program lunches. USDA's "Compre-
hensive Study of the Child Nutrition Programs" has character-
ized non?articipating children as more likely ~o
     --live in urban areas,
     --be economically non needy , and/or
     --attend secondary schools.l/

     Although there is general agreement on the need to im-
prove NSLP participation levels, the questiun remains as to
how this can best be accomplished. One of the most widely
known factors affecting participation is the price charged
for the NSLP lunch. Many school administrators believe that
"price ll is very important and that increased Federal sub-
sidie3 would lower student payments, thereby improving par-
ticipation levels. This may be true. However, price is not
the sale factor influencing participation; nonprice factors
should also be considered.


     Several studies have shown that noneconomic tactors have
an important influence on daily participation levels. Some
of the more important factors are:
     --the availability of alternative food sources;
     --attitudes of schoo! ~dministrators; and
     --menu variety, choice, food preparation, and food
       qual ity.

Alternative food sources
     USDA, noting that nonparticipating children usually ob-
tain lunch from alternative food sources, has identified the
following "major sources of competition" for the NSLP lunch
(based on the analysis of 1972 survey data):l/

l/USDA Comprehensive Study of the Child Nutrition Programs,
   Committee Print of the Committee on Agriculture and
   Forestry, U.S. Senate, Wash. D.C., Sept. 1974.

     A la carte items--In schools with a la carte food avail-
     able (primarily secondary schools) the percentage of
     students participating in NSLP was only slightly more
     than half of the comparable percentage for NSLP schools
     without a la carte (11 percent of enrollment compared to
     61 percent).
     Sack lunches--Eighty-seven percent of NSLP schools had
     students who brought sack lunches from home. About 18
     percent of the students in NSLP schools ate sack lunches
     on the day of the survey. Sack lunches were consider-
     ably more common among elementary students than among
     secondary students.
     Off-campus foods--Almost half of the Nation's schools
     permitted students to leave school_ during lunchtime.
     NSLP participation in schools allowing studen~s to
     leave campus at lunchtime averaged about 10 percent
     lower than those not permitting students to leave.
Attitude of school administrators
     Attitudes of school administrators, teachers, and workers
also have a bearing on participation. The effect of attitude
on the quality of food preparation (and hence, on participa-
tion) is reasonably obvious. There is, however, an entirely
different aspect of NSLP which is seldom addressed, but
greatly affected by ~dmiminstrators' attitudes and deci-
sions--the social climate of the lunchroom.
     A student's choice of where and what to eat involves
both the food itself and the social aspects of dining. Stu-
dents, for example, have indicated a need to socialize during
the lunch period. Of the many factors affecting the lunch-
room's social atmosphere, two appear to have a particularly
important impact on participation. ~hese factors, as re-
ported in a USDA study of high school participation, are:
     --Length of lunch period and fast service. The length
       of the lunch period in SOme schools ranged from 21 to
       40 minutes. In the schools with shorter time periods,
       many students didn't eat or obtained a quick snack
       from the a la carte line. Over 54 percent of the stu-
       dents indicated that the lunch line was too long and
       they did not have time to enjoy lunch.

     --Merchandising th~ type A lunch. The majority of
       schools with high participation made a special effort
       to make certain that program lunches were well dis-
       played, attractively served, and easily accessible
       to students.  In most of the low-participation schools,
       this was not done.

Menu choice an2-E!epara~ion

      Participation is also dependent upon the schoolchild's
appraisal of the NSLP lunch--its acceptability in terms of
food quality, preparation, and presentation. Existing re-
search has indicated the following influences of these fac-

   -Food quality and preparation.     In general, studies indi-
    cate that students perceive the quality of food used in
    NSLP as being "average. 11 l-1any students compla in of
    poorly prepared foods and a dislike of the basic food
     items used in NSLP menuS.

   -Menu variety and choice. A USDA study of 20 high schools
    found that menu variety and choice of items within the
    menu were somewhat limited.  Fifty-six percent of the
    students had no choice in the style of the lunch (e.g.,
    soup, sandwich, salad, plate, etc.), and 49 percent had
    no choic~ in the components Of the lunch.

     The impact of the above findings on program participa-
tion is underscored by the following excerpt from a 1973
GAO report.

     "In a needy secondary school, which had converted its
     lunch program from a 1a carte service to a type A lunch
     during the 1970-71 school year, general participation
     fell from an average 850 st~dents daily during the 1968-
     69 school year to about 630 students daily in December
     1971. The principal of this school told us that he con-
     sidered this drop in participation remarkable because,
     under a 1a carte service, no free or reduced-price
     lunches had been served and that atout 75 percent of the
     students were eligible for free or reduced-price lunches
     und~r the type A lunch program.  He said that, when the

       type A lunches were served, students had no choice of
       what they could eat and lost interest in the lunches."   1/

     Some school administrators believe that the presence of
a large number of students receiving free or reduced-price
lunches within a school tends to inhibit,or reduce partici-
pation of students who pay the full price for their lunches.
A study of NSLP participation in North Carolina schools
tested this possibility and concluded that
       '1*   * * the proportion of students participating who pay
       full price for their lunches is not affected by the num-
       ber of students receiving lunches free of charge or at
       reduced prices." ~/

     As a further test, we compared the participation rates
of regular-price students with those of students eligible for
free or reduced-price lunches. Our analysis, based on fiscal
year 1973 data for the 50 States and -the District of Columbia,
indicates that the participation rates of the two categories
are, for practical purposes, independent of one another.
This finding, because it is derived from statewide totals,
should not be considered conclusive until validated by fur-
ther research. It does, however, suggest that regular-price
students and students eligible for free or reduced-price
lunches may respond differently to a particular set of par-
ticipation stimuli--a feature to be considered in further
studies of NSLP participation.
     Information relating lunch prices to participation
levels has significance from a policy viewpoint since the
regular price is set by local school officials and

l/U.S. General Accounting Office,"Progress and Problems in
- Achieving Objectives of School Lunch Program,·
  8-178564, Wash., D.C., June 1973.
2/Nicholson, R.H., Some Economic Aspects of the National
-   Sc~££!_~~~~~_~rog£~§=I~:~£Ith_£~£~II~~,-EconomIcs
    tion Report No. J2, North CarolIna State University,
    Raleigh, N.C., July 1973.

.dministrators ~possibly estimated as the price required to
make up the annual difference between the cost of preparing
meals and the subsidies of Federal, State, and local govern-
ments). These price decisions affect the cost of NSLP lunches
to parents .and children in individual schools and collec-
tively influence the scope of the program's benefits on a
national basis.

      FNS administrative reports and studies long have docu-
mented that agency's cdncern for increasing and for accurately
forecasting the numbers of program participants. Published
and unpublished reports have identified factors which have
been viewed as being associated with participation and as
being of use for forecasting numbers of participants. In con-
sidering rp=earch directed toward identifying the relative im-
portance of factors which explain participation, the emphasis
has been directed toward those factors over which the school
has substantial degrees of control, e.g., prices and costs of
lunches, and also to those which could be meaningfully quanti-


     We mentioned earlier that nonparticipants normally ob-
tain lunch from alternative food sources. Although such
meals are usually nutritionally inferior to the type A lunch,
they still have a cost. The parents of these children, whether
providing food from home or funds for the purchase of food at
school, are paying for the child's lunch.   In this regard, it
is not the price of the NSLP lunch per se, but rather its
price relative to alternatives that affects student partici-
pation (i.e~, differences in prefe-rence are expressed on the
market in monetary terms).

     In order to compare NSLP lunch prices with a readily
available alternative, the authors of a 1973 study 11 used
as a test instrument a sample of 4 sack lunches which met
the minim~m nutritional requirements specified for NSLP.
(See table 7.1 on the following page.) The food cost in these

samples ranged from 26 to 41 cents, roughly comparable to
NSLP's prevailing price of 2S to 40 cents. !/
                               Table 7.1

                       Price                                 Price
Number 1              i£ent~l.         Number 2              (cen~~l.
2 slices of bread        3.65          2 slices of bread        3.65
2 oz. bologna           13.87          2 oz. salami            16.33
1 tablespoon butter      2.68          1 tablespoon butter      2.68
Carrots, 3/16 lb.        3.28          Cucumbers, 3/16 lb.      5.15
Banana, 3/16 lb.         2.94          Grapes, 3/16 lb.         8.15
1/2 pint of milk         4.94          1/2 l?int of milk       4.94
Total                   31.36          Total                   40.90
Number 3                               Number 4
-------                                -----
2 slices of bread        3.65          2 slices of bread         3.65
2 oz. liver sausage     13.56          2 oz. peanut butter       7.88
1 tablespoon butter      2.68          1 tablespoon butter       2.68
Celery, 3/16 lb.         3.74          Celery, 3/16 lb.          3.74
Apple, 3/16 lb.          4.06          Banana" 3/16 lb.          2.94
1/2 pint of milk         4.94          1/2 pint of milk          4.94
                        -  -
                        32.63          Total
a/Each lunch conforms to NSLP's type A' pattern requirement
- and is designed to provide one-third of an elementary school
  child's RDA. Source: Pricing and Partic!.\2ation Rates in
  the National School Lunch programs-rn washIngton Public--
  FhoorDIstrlcts, washIngton StateUnIversl ty, Oct. I9i3
     Mothers of elementary school children were aware of the
sack lunch cost. In school districts charging 30 to 35 cents
for the NSLP lunch, over 75 percent of the mothers interviewed
believed the price was reasonable; less than a third said
they could provide a sack lunch for less money. In districts

liThe sack lunches, designed as a test instrument, should
- not be considered representative of the cost or content
  of sack lunches eaten by schoolchildren. The samples
  are presented here to illustrate flexibility in the type A
  pattern, to show that a sack lunch style meal can provide
  one-third of the schoolchild's RDA.

where .the NSr;p lunch price was 40 cents, only 40 percent of
thEi. r'espondents' believed the price was reasonable; 70 percent
said they couid provide a sack lunch at less cost. This find-
ing (that p..'rents believed the sack lunch less costly
than NSLP lunChes in the higher priced districts) suggests
that the NSLP meal competes, on a price basis, with the sack

2££y:'!;' e a .w.eak~£!casting ~!
     The relationship between prices and the rates of daily
participation which existed in the United States during fiscal
year 1973 are shown in table 7.2. A diagonal pattern is
formed by indicating the number of States associated with
each regular-price and participation level. In general, this
pattern indicates that the higher a child's payment, the
lower the level of participation.

                                                          T"bl~   7.2
                                          of Statel Participating In NSLP.
                          by Regular Prices Chargfd and Daily Participation             Rat~5.
                                          risc.l Year 197) (note 01
                    Rate of dot!'1 participation. p!rcent of                r~oul.r-price cntoll~.nt

      Aver''ll! child's                                                                                Toul
        p'y•• nt
        {note bl


                           " "
                               lO-    ll-

                                                           ,,- 51- ,,-

                                                         " " " " " " "
                                                                       ,,- 71-          62-


      56-60                                                                                                  3
      46-50                                    2
                                                                               3                             ,
                                                                        1      ,,
                                                                                         1                 12

      Totll 00.
       States             0               ,    3     ,            ,            , , , •                     51

      (note bl

      !!T.bvl.r riqllr~$ i~dic.tc the nu~ber of States Wh05~ avera'lc price is betkcen
        the I_opnts Jlsted in the left-hand colv~n and whose partlci~tion rate
        for req"hr-pr lc. students is bet..een thl! ~rcent.ge$ indicated at till: top.
      g/oet.t~ined by divJdtn9 State-tepofted f19Vles on total lunches served--less
        total tlee and ledaeed-price lunehes served--into lotal ehildten's pay_ents.
        This sU9htly overstates the lunch prices· sinee a $llall but unknown, portion
        of the childten's payeenls ale {or reduced-price lunches.
      ~/Includes   the   Dis~fict    of   Col~bi~.

     USDA's Comprehensive Study of the Child Nutrition Program
informed the Congress that variations in participation depend
significantly upon the relationship of price to participation.
The study went on to describe a mathematical relationship be-
tween price and participation which was used to compare a
series of program alternatives. This relationship has since
been used in congressional debate on school lunch legislation.

     We believe the specific relationship reported by the
study would be technically correct only for NSLP lunches
priced in the 20 to 35 cents range. 1/ Since NSLP prices
were considerably above that range in fiscal year 1974--
and remain so today--the study's values should not be used
for projecting participation levels. More importantly, we
believe that the association between price and participation
levels is an extremely weak forecasting tool.
      To illustrate this point, figure 7.1 on the next page
shows the linear relationship between prices and the partici-
pation rates of regular-price students which existed in fiscal
yea," 1973. The straight line drawn there is the one that best
fits a l'scatter" of points, each point representing the aver-
age price-participation levels of one State (computed from
FNS program data). ~/, 2/

~/All   the research on price-participation relationships that
    we found was either directly conducted, or sponsored, by
    USDA.  In each instance, the work indic,ted a high level
    of professional competence. The Comprehensive Study, how-
    ever, misreported the research findings by stating, as a
    general rule, that "paying students respond by reducing
    participation 3 to 6 percent for eVJry 10 percent increase
    in prices charged." Such a relationship (price elasticity
    in economic terms) is specific to a particular price level.
    At prices of 35 cents and above, the elasticity is markedly

2/USDA, FNS, Fiscal Year 1973 Statistics and Historical
-   !~~!~,   Wash~~-5~c~~-r974----------------------------

3/USDA, FNS, "Estimates of Needy Children in National School
- Lunch Program Schools Eligible for and Reached with Free
  or Reduced Price Lunches," Survey Reports for Oct. 1972
  and Mar. 1973.

                                      FIGURE 1.1

  ....     .'   ~uNiTED:STAl'ES FlsCAL'UAR 1973
   -   '   -       • l<

                                             . ....- - - - - - - - - - - ,

                             PARTICIPATION 9.4.1. _ 1.079 (PRICE) a


  O....._ ~ _....._       ....      L-_~_          ....._   ......_ - - ' ' - _......_   ...

                                 PRICE OF LUNCH, CENTS

  a The price coefficient (-1.079) has a standard error equal to 0.193, indicating
    that the coefficient is statistically highly significant.
    The value of the squared correlation coefficient ,R2, is 0.389, indicating that
    price olone explains about 39 percent of the variation in the participation rates.

Although differences in methodology preclude precise
comparisons, the price-participation relatlonships·shown on the
previous page appear to be in substantial agreement with the find-
ings ot three studies 11,2/,31 cited in USDA's 1974 report. The
1 ine has a downward slope- (determined by the equation' s p.r ice
coefficient, -1.079) and indicates that a 5-cent price in-
crease is associated with an "expected" 41 de~line of ~b6ut
5.4 percentage points in the daily participation rate. Price,
however, explains only 38.9 percent of the variation in State
participation levels (i.e., an interpretation of the squared
correlation coefficient). In a statistical sense these
findings indi~ate that price increases tend to depress par-
ticipation rates, but that other variables--accounting for
61.1 percent of the variation in participation--have a
dominant influence on the magnitude and direction of such
     The implications of these findings are clear. Price-
participation forecasts rest on the assumption that "all other
things remain constant."  This assumption is tenuous, as
evidenced by the increases in both price and participation
rates which occurred between fiscal years 1972 and 1973.
      In view of the above discussion, and because participa-
tion studies have concentrated on the impact of "price,"
decisionmakers have very little quantitative support for
estimating the participation impacts of variouS policy al-
ternatives. These limitations can be summarized as follows:

!/see footnote I, p. 93.

2/USDA, Economic Research Service, "Factors Affecting
- Participation in the School Lunch Program," Unpublished
  Working Paper, June 1971.
3/Braley, G. A. and P. E. Nelson, "Effect of a Controlled
- Price Increase on School Lunch Participation: Pittsbu.rgh
  1973," ~!!!~E.!£~~_~£~E.~al_£L~!iE.!£.!:!lt~r.al_~££~ic!!.,F eb.

4/The standard error of the price coefficient introduces
- variation about the "expected value," with a 99-percent
  likelihood that declines will be in the range of from
  3.1 to 7.6 percentage points.

     --price-participation provides an extremely weak fore-
       casting tool.
     -~The relative importance (rank) of the individual fac-
       tors affecting participation has not been fully deter-
       mined   a

     ~-The "recognized" factors affecting participation have
       not been shown to account for the major variations in
       program participation.

     Since};?articipation rates are a primary measure of pro-
gram performance, we believe that methods of improving and
better integrating the resultS of participation studies need
to be examined closely. By combining several of the factors
affecting participation, it may be possible to develop a
cause-effect relationship which would "explain" most of the
major reasons why children do or do not participate in NSLP.
      To illustrate the basic features of a multifactor rela-
tionship, we added one additonal factor, "per capita income,"
to the price-participation relationship described earlier. 1/
The result is shown as a mathematical eguation in table 7.3-
on tl ~ following page.

l/Although "per capita income" serves to illustrate the de-
- velopment of a multi factor relationship, it is very sus-
  pect in an economic context. Specifically, average per
  capita income for each State includes the incomes of house-
  holds without children, with children who pay fully or re-
  ceive free/reduced-price lunches, retirees, etc. Thus,
  even though the relationship shows a high level of statis-
  tical importance, it should not be given too much emphasis
  in an economic sense. A more meaningful factor would be
  an income series which is more closely representative of
  the "household income" of regular-price students; however,
  we have yet to find such data. The results illustrated
  by this price-per-capita-income example show a need for
  further research in developing such a variable. ERS has
  stated that it joins us in recognizing the importance of
  a well-specified income variable for improving forecast

                              Table 7.3

       Regular-PI." ice Participation as a Func,tion of
                Per Capita Income and Price,
               United States, Fiscal Year 1973

 Regression equation:
              r   = 121.8 - O.5862p - 10.121 + e
    where:    r = daily participation rate expressed as a
                      percent of regular-price enrollment

              p = average regular price charged for the NSLP
                  lunch (cents)

              i = per capita income (in thousands of dollars)

             e    =   error term
Notes: The value of the squared correlation coefficient, R2, is
       0.557, indicating that the combination of price and income
       explains about 55.7 percent of the variation between State
       participation rates, or almost 17 percent more than price
       The coefficients for price        and income have standard errors
       equal to 0.2022 and 2.366,        respectively. Each indicates a
       high degree of statistical        significance. The value indi-
       cating correlation between        price and income is 0.5695.

Both price and per capita income are statistically important
factors in lIexplainingll variations in participation. Their
coefficients are negative (-0.5862 and -10.12, respectively),
indicating that an increase in either would tend to lower
participaton rates.    Furthermore:
     --If price is held constant, a $1,000 increase in per
       capita income would be expected to lower the rate of
       participation by about 10.1 percentage points.
       Alternatively, if per capita income is held constant,
       a 5-cent price increase would be expected to lower the
       rate of participation by about 2.9 percentage points.
       In terms of their impact on participation, a $1,000

       in'c·rease in per capita income is equivalent to a
       l7.3-cent increase in the price charged for an NSLP
       lunch. 1/
     --P.rice a'nd per capita income, taken together, account
       for. about 55.7 percent of the variation in State
       participation rates. 2/ Although this value is low for
       forecasting purposes,-it is considerably higher than
       t6e value obtained in the analysis of price relation-
     Of course, price and per capita income are not the only
factors that affect participation. The relationship could be
expanded to include the influence of a number of factors--
both quantitative and qualitative. 3/ Appropriately coordi-
nated, we believe that the use of multifactor relqtionships
in future studies of NSLP participation offers the best
possibility for developing a functional undetstanding of the
nature and importance of the major factors affecting partic-
ipati~1 and, quite possibly, a reliable basis for estimating
the participation impacts of various program modifications.

l/Changes in price and per capita income, measured in 1973
- constant dollars (i.e., annual increases in the cost of
  living do not automatically increase per capita income as
   used in the equation).
2/Taken independently, price explained 38.9 percent of the
- variation in participation and per capita income explained
  48.0 ,ercent. The combined relationship, as indicated by
  the 55.7 percent reported above, is not usually cumulative.
  This indicates the "Sharing" 'jf a certain participation
  influence between price and per capita income.
3/USDA's 1971 study (see footnote 2, p. 98) combined "price" with
- several qualitative factors to determine their effects on
  NSLP participation. Although the study did not develop a
  reliable forecasting relationship, it did determine that
  the following qualitative factors affected participation:
  (1) the presence of a la carte foods, (2) the presence of
  vending machines, and (3) whether or not students were
  permitted to leave ca~pus at lunchtime.

      In describing price-participation relationships, we
indicated that about 39 percent of the variation between
State participation levels was explained by differences in
the price of the NSLP lunch. "Per capita income," however,
accounted for 48 percent of the fluctuation in regular-price
participation. In every respect, per capita income was stat-
istically a more important variable than price in regard to
NSLP participation.

     As a normal condition, the participation of regular-
price students would be expected to increase with income.
This was not the case. Our analysis revealed that as per
capita income advanced, student participation declined. In
economic terms, the NSLP lunch exhibited the characteristics
of an "inferior good. 11
     The income-participation relationship, because it is de-
rived from statewide totals and because per capita income may
not be representative of the "household income" of regular-
price students, should not be considered in formulating NSLP
policies unless substantiated by further research.   (See
footnote 1, p. 99.) However, it does suggest the possibility
that as "real income" increases children are provided greater
funds and therefore select their noon meal from a broader list
of alternatives--including some which may cost more than the
type A lunch. If this is true, NSLP is n~t "pricing-out"
students; it is merely losing the ability to use lower prices
as an incentive for participation. Our I'example" would pro-
vide two very important conclusions regarding program partici-
pation, namely that:

    1.   Increases in "real income" (standard of living)
         will tend to reduce the participation of regular-
         price students even if the "real price" Of the NSLP
         lunch is held constant.  For example, the practice
         of increasing Federal subsidies to compensate for
         advances in the Consumer Price Index should not be
         expected to sustain the current participation rates
         of regular-price students.

    2.   Attempts to sustain participation rates in the reg-
         ular-price program should concentrate on noneconomic
         aspects of the school lunch program (e.g., social
         climate of the cafeteria and the selection, prepara-
         tion, and presentation of program lunches).

~!'.!.ching   resou~~~ need~

       A discussion of NSLP participation should also consider
       --the p'rogram's objective is to safeg'lard health,
       --children have differing needs for nutritional
          assistance, and
       --program resoUrces- are limited-.
The trade-off between differentially subsidizing the lunches
of economically needy children and applying the same resources
tow.ard an increased subsidy for regular-price lunches is ob-
vious. The need to provide nutritional assistance for eco-
nomically needy children (e.g., free and reduced-price
lunches) is considered to be more important than the need to
increase regular-price participation. In effect, two things
should be noted:
       --Program effectiveness depends more on satisfying re-
          cognized needs than on ~total" sales.
       --Only as requirements for nutritional assistance are
         identified can the effectiveness of participation
         policies be judged and available resources matched
         w'i th needs.
     By considering the NSLP enrollment as being grouped into
four participation cate90ri~s--nonparticipatingstudents and
regular, reduced-price, and free lunch participants--it is ap-
parent that a participation change in one category necessarily
affects one or more of the others. Present studies of NSLP
participation, by confining the!ir observations to a single
participation category, generally exclude these impacts. They
do not, in general, identify the full impact of a participa-
tion change.
     A Pittsburgh study 11 analyzed the effects of a price in-
crease on student movements between participation categories.
The study reported that
       --about one-fifth of the l'dropouts " from the tegular-
         price program joined the free lunch program (implying
          that at lower prices some of those eligible for free

!/See footnote 3. p. Y8.

       lunches preferred to participate' as regular-price
       students); and

     --four-fifths of the students ceasing to purchase type A
       lunches did not transfer into the free lunch category
       (implying that many children shifted to alternative
       food sources or went without lunch). !/

The increase in free lunch participation is an important cost
consideration in that Federal reimbursements may likewise in-
crease (refer to eh. 10). However, in regard to the program's
objective of safeguarding health, a more crucial question is:
what happens to students that leave NSLP? For example:

     --What do program "dropouts" substitute for the NSLP
       lunch? And how does the alternative compare with the
       price and nutritional content of the NGLP lunch?

     --Do the reasons which precipitate a decline in participa-
       tion affect the selection of alternative food sources
        (e.g., does a participation decline due to the availa-
       bility of competitive foods have the same nutritional
       impact as a decline caused by price increases)?
     --Is the process of a decline in participation revers-
       ible? That is, by reversing the conditions causing
       a decline, would NSLP reattract the s&me I'dropouts"
       into the program?

     Further research is needed to develop a "unified expla-
nation" for the causes ~nn impacts of changes in NSLP partici-
pation rates. We belie     . lt such research, properly coor-
dinated, Should enable L.JA to better estimate and prepare
for the impact that various program changes would have, and
to improve the direction and effectiveness of outreach

     Further effort is needed to develop a "unified explana-
tion" for the causes and impacts of .:hanges in the program's

l/Since Pittsburgh schools did not offer a reduced-price
- lunch prior to the price increase, the effects on-that
  category of participation were not analyzed.

participation rates.   We recommend that the Secretary of
     --Improve the accuracy of participation forecasts and
        determine the relative importance of individual factors
        (including price) which affect participation.
     --Determine how changes in school lunch program partici-
       patiOn affect the magnitude and characteristics of un-
       met nutritional needs in the nonparticipant population.
     USDA agreed that there is a r,eed to prioritize the factors
affecting participation and to determine the extent to which
they individually and collectively influence participation.
(See app. I.) It indicated that such work has been an on-
going objective of the Food and Nutrition Service.
     USDA did not address our recommendation about determin-
ing the influence of participation changes on the unmet nutri-
tional needs of the nonparticipant population.

                             PART IV
                          PROGRAM COSTS
      In 1975 the School Lunch Program's operating expenses rose
to about $3.8 billion, a 73-percent increase over the $2.2 bil-
lion of fiscal year 1970. In this same period the Federal
share of program expense increased from about $0.6 billion to
$1. 7 billion.
     Although the Federal Government shoulders the largest
portion of program expense, Stat" and local governments have
the greatest control over operating efficiencies. As pointed
out in an Urban Institute report, 1/ the State governments
are in a strategic, and in some ways u~ique, position to in-
fluence some of the factors that increase cost growth within
NSLP. State government is especially able to improve program
performance in areas where either the Federal Government or
the localities cannot or will not act. States have the power
to prescribe regulations, implement program incentives, and
to some extent, provide services and resources for local use.
Although the State's domain is circumscribed by Federal and
local prerogatives, State governments, by their actions or
inaction, have an important impact on program costs through

     --administering local functions such as education and
       welfare services and
     --disbursing NSLP funds in a differential manner.
     Because of the rapid increase in NSLP expenses and because
an increasing proportion of this expense is supported by Federal
funds, we thought a review of selected program areas might pro-
vide insights for future savings. Chapters 8 through 10 de-
scribe, in turn, the impact on Federal funding as influenced
by USDA food distributions; the dominant factors in cost
growth; and trends in Federal, State, and local financing.

                              CHAPTER 8


     The National School Lunch Program has provided an im-
portant outlet for foods acquired under USDA price stabili-
zation and surplus removal actions (see table 8.1 below).
Federal funds, principally from agricultural programs, have
paid for the purchase of commodities and the cost of trans-
porting them to the States. Once at the designated warehouse
or car-side location, the State or recipient agency has borne
the final distribution expense.

                              Table 8.1

               USDA Commodity Distributions to NSLP,
                   Fiscal Year 1971-75 (note a)

                 Value of commodity distributions          Share of
Fiscal                       (note b)                        NSLP
 year            Sec. 6   Sees. 32 & 416    Total         food costs


1971             $64.3          $213 .0       $277.3          19.7%
1972              64.0           248.0      c/3l2.l           20.0
1973              59.5           200.7      .'Y 260 • 2       15.6
1974              67.3           248.8          316.1         16.4
1975 (est. )      63.7           357.6      e/42l.3           18.9

£/Source:      FNS/Program Reporting Staff publications.

b/Value is cost to Federal Government. Commodities are
- obtai~~d by the Secretary of Agriculture with funds
  appropriated by the National School Lunch Act (section
  6 funds); with funds arising from tariffs on imports
  (section 32 funds); and from the Commodity Credit
  Corporation's purchase of surplus foods (section 416
  funds) •

£/Difference in total due to rounding.

d/Excludes $70.8 million cash in lieu of commodities to
- schools.

e/Excludes $5.2 million cash in lieu of commodities
- authorized for NSLP schools in the State of Kansas.

      NSLP, in 1970, accounted for less than half of USDA's
food distribution effort. But as the Nation's agricultural
policies reduced the need for surplus food outlets, alterna-
tive means were employed to support other Federal programs
(e. g., food distribution to needy families was replaced by the
Food Stamp Program) and NSLP became nearly the total focus of
USDA's food distribution activities. The Congress increased'
the level of commodity support for NSLP and authorized the
Secretary of Agriculture, in some instances, to purchase non~
surplus foods for distribution to schools.   In 1974, USDA re-
ported to the Congress that

     "* * * the Department will also be looking into whether
     the continuation of * * * [the current food distribution
     program] * * * is either feasible or necessary in view
     of the shifts in U.S. farm policy, the phaseout of the
     food distribution program for needy families, and the
     fact that most of the foOd for the child nutrition pro-
     grams is already being purchased locally. The Depart-
     ment believes that a single cash payment, increased to
     reflect past commodity support, may be preferable." 1c!

This statement raised two major issues. The first, an issue
involvina future agricultural policies and the abandonment of
a proven oommodity outlet, is beyond the scope of this report.
The second issue, cash versus commodity support, is addressed
in the balance of this chapter.

     In present program reporting, commodities are valued as
the sum of procurement cost and the cost of delivery to the
States. For comparing commodity distributions against a cash
alternative, it is necessary to consider the cost of food as
delivered to a school--a cost which includes administrative
and intrastate distribution expenses. ~!

!!USDA, fomp~he~~ive Study of th~ Child Nutrition Programs.
   Committee Print of the Committee on Agriculture and------
   Forestry, U.S. Senate, Wash., D.C., Sept. 1974.

-2!Report   of the Commission on Government Procurement, Vol.
  ~ Dec~-1972:-------------------------------------

Cost of USDA foods increased.by
Intrastate d1str16UtI~~ expenses

      A study to determine the average cost for distributing
com~odities    at the State and local level and the apportion-
meilt of.those costs among participating agencies was completed
!.n Apr i l of 1974. !/
     This study, using 1973 data, determined that:

     --Average cost for intrastate distribution of commodi-
       ties was 53 hundredths of a cent per school lunch--
       in aggregate, about 6 percent of the cost of pro-
       viding commodities.

     --If food distribution to needy families was eliminated
       with an assumed reduction in total administration
       costs of 25 percent, intrastate distribution costs
       would have increased to 57/100 of a cent per lunch--
       an 8% gain.

     --The major components of intrastate distribution expense
       were:  transportation (56.6%), warehousing (22.6%), and
       administration (20.8%).

     --Schools paid 61 percent of local distribution costs;
       States contributed the remaining 39 percent. Twenty-
       eight States made an assessment against recipient

USOA foods less exeensive than
~£~oo~~ oe~~-ma£~~~:E~rchaS~~
      In February 1975, USDA's Economic Research Service (ERS)
published "Costs of Foo:ls Purchased by USDA and Local School
Systems, 1973/74." 2/ This study determined the cost for a
uniform "market basiet" of 15 foods as purchased during
1973-74 school year at prices paid by:


2/USDA, Economic Research Service, Costs of Foods Purchased
- £~_~~Q~~~~LOC~l-~ch~l-~~st~,-f973774~~S-S92;-----­
  Wash., D.C., Feb. 1975.

       --the   USDA's commodity program,
       --the   largest school systems (over 25,000 students),
       --the   smallest school systems (under 2,500 students), and
       --all   school systems combined.

     The st.udy's "market basket consisted of foods with sub-

stantial usage which had been purchased between July 1973
and April 1974 by school systems and USDA. On the open mar-
ket, schools purchased these foods through the following
types of sellers: processors (canners, freezers, packers);
wholesalers; county or State purchasing agencies (which buy
all items for the schools); and retail merchants. The simple
average price paid for each pound of "market basket" foods
purchased from these sources is shown below:
                            Table 8.2
                            - ---

                            Proc-   Whole            ing
Food                        essor   saler               Retailer
                            -----          ---!!9.~l          -----
                            ----------(dollars per pound)--------

Turkey                  0.846              0.762              0.639
Chicken                  .563               .614      0.480    .568
Frankfurters (all meat)  .850               .889       .437    .902
Ground beef (20% fat)    .998              1.009       .861   1.006
Cheese, processed        .990              1.007       .954   1.078
Flour, all purpose       .165               .145               .195
Margar ine                                  .410               .676
Rice                                        .356               .430
Corn, canned                 .179           .192       .159    .207
Tomatoes, canned             .229           .214       .237    .217
Peas, canned                 .172           .190       .202    .216
Peaches, canned                             .252       .363    .266
Pears, canned                .236           .274       .238    .287
Pineapple, canned            .248           .251       .276    .246
Potatoes, frozen french                     .239       .207    .199

a/Source: Costs of Foods Purchased bl USDA and Local School
- ~lst~~~L 1973Z2i;-ITSDA-Pub:~RS=592.---------       ----

      The county/State purchasing agency had the lowest price
fo'r fivefo"ods, with a marked advantage for ch icken, frank-
furters, and ground beef. Compared with the most expensive
source, ~heir price was 13.4 cents a pound lower for chicken,
14.8 cents lower for ground beef, and 46.5 cents lower for
frankfurters. Since these foods are relatively high in cost
and are excellent sources of protein, the ?rice differences
are especially noteworthy.
      Prices showed regional differences throughout the Nation
at all levels of distribution--retail, wholesale, and proc-
essor.    However, for each of the 15 food items, the average
price per pound paid by the largest school systems (25,000 or
more ,students) was consistently lower than the price paid by
the smallest school systems (fewer than 2,500 students).
     During the study period, USDA purchased about 441.5 mil-
lion Dounds of the 15 "market basket" foods. By weight, this
amounted to about 40.3 percent of all USDA foods donated to
schools in fiscal year 1974. The overall procurement cost
of these foods was eas-ily determined.   However, since schools'
open market purchases included delivery to the feeding site
and administrative expenses, a comparable cost for USDA foods
required that USDA's procurement CJsts be adjusted to include
a pro rata share of Federal administrative and intrastate ex-
penses (obtained from the A. T. Kearney study described
earlier). The resulting USDA food costs, expressed on a oec
lunch basis, are shown on the following page:

                          Table 8.3
                    USDA Food Cost Per Lunch--

                                               Food cost per lunch
Food procurement                                      -7:"65
Federal administrative expense                          .15
Intrastate costs:
   War~housing                                          .12
   Transportation                                       .30
   Administrative                                       .11

         Total                                         8.33
a/According to the Kearney study, food cost per NSLP lunch
- would rise to 8.37 cents if the Needy Persons Program
  were eliminated. Since that program is being phased out,
  the higher figure will be used throughout subsequent
  sections of this chapt~r for comparative purpose.
  Source: IC~is~7s~Of_FOO~~-f~~~~-~yUSE~~~~_~££~! Sc~£ol
  ~y~te~_~_l-!i, USDA Pub. ER~-59~

     Cost comparisons between foods purchased by USDA (ac-
cording to its published specifications) and the cost of the
same foods (matched as closely as possible) purchased by 150
school systems are shown below: !/

                          Table 8.4
                    ££~E~~i~£~_£f_E£od Co~~~:
                                Food cost per lunch
                                      (~nts)                   Index
USDA-purchased foods                   8.37                    100.0
Open Market Purchases:
   Average U.S. public school          8.93                    106.7
   Largest schools                     8.35                     99.8
   Smallest schools                    9.86                    117.8

liThe "as closely as possible concept" still leaves a sub-
- stantial margin for differences. For example, USDA speci-
  fications call for net drained wL.ghts for canned foods
  and for can or container size. In the open market, few
  school systems specify drained weights. To the extent
  that USDA drained weights vary from market practices, the
  respective prices will probably reflect this difference.

Thus, ERS reported:
     "The. size of the differences paid for food by the school
     systems and USDA and their statistical significance pro-
     vides some quantitative support for making program
     policy decisions. Estimates for the total 1973/74
     school lunch program may be made in terms of food costs
     per lunch • • * with the recognition that the results
     for the 15 foods are assumed typical of results that
     would be found if all commodities were studied."
     Prices reported by school systems with 25,000 .students
or more are essentially the same as those paid by USDA. HOw-
ever, the smallest school systems ~aid prices averaging 17.8
percent higher than those paid by USDA; and the average U.S.
public school paid an average price 6.7 percent higher. As-
suming these proportions are representative of 1975's com-
modity program, equivalent cash support would have increased
program expenses by $23.2 million.
OF COMMODITIES        ----------
     In January 1975, the Senate Select Committee on Nutrition
and Human Needs asked the Nation's State School Food Service
Directors for their preferences between cash and commodi-
ties. 1/ Of the 36 respondents (1 undecided), 22 preferred
commOdIties, 8 preferred cash, and 5 wanted an option for
individual school districts. Those favoring commodities
frequently noted inconveniences, but felt the cost savings of
USDA foods outweighed all other considerations. Their com-
ments included:
     "Discontinuing the commodity program is not a reasonable
     alternative * * * food service directors should have some
     input into types and kinds of food. Improvements are
     needed in information regarding delivery and also in
     frequency and quantity of delivery."
     "If the commodity program were to end, at least 15 cents
     per meal would be needed to offset the loss • • • The use
     of bids with standards of quality presently found in USDP

l/U.S. Senate, Select Committee on Nutrition and Human
- Needs, School-K££2-fI0gr~~~2~-1~75,Government
  Printing orrIce, Wash., D.C., 1975.

    donated foods is impossible in 95% of South Dakota

    "We heartily endorse the continuance of the commodity
    program * * * we are now geared to effectively adminis-
    ter the program with warehouses and trucking equipmant."

        the commodity program were discontinued * * * no
    school program, other than some of the major cities'
    programs, have the technical know-how or volume to ob-
    tain the quality and quantity per dollar that the USD~
     On the other hand, Kansas deactivated its intrastate
commodity distribution network and in fiscal year 1975 began
a program of full cash support in lieu'of commodities. ~fter
a half-year's experience Kansas' Director of School Food
service indicated a strong preference for cash in lieu of
commodities. Iler reasons included:
    1.      Less paperwork and time are involved in inventory
     2.     More variety in menus is possible.
     3.     There is better planning for utilization of facili-
            ties, particularly storage facilities.  Purchases
            and deliveries are scheduled for convenience and

     4.     There is a savings of actual cash formerly spent for
            freight and storage charges on commodities.

     5.     More food money is available to negotiate good buys
            on food items.

     6.     Cash does not have 'price support' effect which
            raises prices to other consumers. !/

     ~lthough early experience in Kansas shows the convenience
of cash, we do not know of any study which has evaluated the

~/George, r., Testimony before the Senate Subcommittee on
  ~gricultural Research and General Legislation, ~pr. 1975.
  (Adverse experiences with the commodity distribution pro-
  gram are noted by Kansas school food service directors in
  ch. 5 of this report.)

impact of that State's changeover, either in terms of student
participation or In terms of total program costs. We believe
that a study of the Kansas changeover can provide much useful
info~matiqn cbncerning the commodity program's influence on
various aspects of NSLP. However~ because of Kansas' accessi-
bility to the Nation's agricultural producers, we are not
sure that a comoarison of food costs i- that State will be re-
presentative of- the Nation as a whole. In our opinion,
USDA's "market basket" comparison (described earlier in this
chapter) provides a better technique for appraising the cost
trade-offs between cash and commodity support for NSLP over-
     Based on the facts and impressions described in this
chapter. ending commodity distributions to MSLP schools may
have the following impacts:
     --The possible disestablishment of a proven commodity
       outlet, including the State facilities and administra-
       tive organization essential to its o~eration.
     --Increased NSLP food costs. Comparable foods, purchased
       on the ODen market, are estimated bv ERS to cost about
        6.7 percent more than USDA commoditles--an increase
       which would have amounted to $28.2 million in FY 1975.
       While this "6.7 percent U value represents the averaqe
        increase, table 8.4 shows sizeable differences between
       the procurement economies of large and small school
       systems. Prices reported in large districts are es-
       sentially the same as those paid by USD~. The small-
        est systems paid prices 17.8 percent higher.   In
       this context, commodity distributions are shown to
       provide greatest benefits to schools with high food
       costs. A fixed-rate cash assistance program inverts
       these benefits--schools with high food costs receive
       the least assistance.
The precise extent of program impact may depend on the actions
of the individual State governments. ~s concluded by the ERS
     "Given the size of food purchases involved (about $1.8
     billion annually), a potential exists for saving the

        Nation's schools several hundred million dollars by more
        effective procurement practices. !/

State agencies, by providing services such as volume pur-
chasing of foods and equipment and regional contracting for
storage and transportation services, might be able to im-
prove the food service economics of small and medium-sized
school districts. These actions could, at once, provide
significant savings in program food costs ar.d enhance a USDA
interface for intermittent sales of agricultural products.

~/See   footnote 2, p. 10,

                          CHAPTER 9
     A 1974 USDA study of regional cost variations in the
National Student Lunch Program reported that !/
    --The Northwestern and Western States had the highest per-
      lunch costs; the Southeastern States had the lowest.
    --For total food costs, the Northeast was highest, the
      Southwest lowest.
     -~Labor costs were highest in the Northeast and lowest
       in the Southeast.
     --Cash outlays for other than food and labor showed the
       greatest variation, with the West-Central and Western
       Regions highest, the Northeastern the lowest.
     --Synthetic cost analysis of data from 30 cities indi-
       cated food costs varied little between major cities
       except for Anchorage and Honolulu, where costs were
       about one-fifth higher.
     --Some va'iation in costs among States and regions may
        be due to differences in accounting by the State
        reporting agency.
    --Characteristics of the program are heavily influenced
      by State operating policy. For example, management
      analysis of Hawaii for fiscal year 1973 showed program
      management was centralized, with good accounting, some
      centralized purchasing, and high overall efficiency.
     While the study found sizeable cost variations between
States and regions, differences in accounting methods prevented
it from determining any meaningful association between costs
and NSLP's operating efficiency. Since that time USDA has
tested a uniform accounting manual for school food service
systems and has encouraged the adoption of this manual
throughout NSLP. If USDA's manual is accorded widespread

!/USDA, ~2~E£~~~!~~_Sfi!~~l_2!-!~~~~!!~-~~!~!!!~-E£2~~~~~,
   Committee Prlnt of t e Commlttee on Agrlculture and
  Forestry, U.S. Senate, Wash., D.C., Sept. 1974.

use, a more uniform data base will be available for program
accountability, planning, and evaluation.


     Cost growth for NSLP overall is shown in table 9.1 below.
~or fiscal years 1970 through 1975, the table shows the num-
ber of meals served and the total program cost--including
commodity donations. On the right-hand side of the table,
these costs are deflated to 1970 dollars and then compared as
an I'adjusted cost per lunch. 11

                              Table 9.1

               NSLP Cost Growth--Fiscal Year 1970-75

                                                     Adjusted cost
                                    CPI food-      (Fiscal year 1970
                     Total cost     away-from-     constant dollars)
Fiscal        Meals    actual       home index     Total     Cost per
 Year         served   (note a)      (note b)      cost       lunch
              ---(millions)----                  (millions)   (cents)

1970          3,565.1   $2,208.0         100.0   $2,208.0      61.93
1971          3,848.3    2,427.9         105.9    2,292.6      59.57
1972          3,972.1    2,712.1         110.5    2,454.4      61. 79
1973          4,008.8    2,984.5         115.9    2,575.1      64.24
1974          3,998.9    3,347.6         130.4    2,567.2      64.20
1975(est. )   4,076.8    3,751.0         144.9    2,588.7      63.50

~/Includes    value of donated commodities.

b/Adjusted to fiscal year 1970 (Sept.-June) = 100.0.

During this period, the number of meals served increased by
approximately 14 percent; program costs (including commodity
donations) increased by 70 percent. When deflated by the
Consumer Price Index (CPI) (the escalator used to determine
Federal reimbursement), the adjusted cost of producing .an
NSLP lunch increased from 61.93 to about 63.50 cents.

Component costs

     Table 9.2 presents the cost of the various components
of a school lunch, for fiscal years 1970 to 1974:


            f£!.Eonent_ Costs of NSLP Lun~E==~1§cal-!ea!_1970:~
             Local          Federally                   Other     Donated     Total
Fiscal       food            donated                    cost      goods &    per-lunch
 Y!~       E.!!.!.£hase~   commodities    Labor    expenditur~~   services    cost
                           ------                                 -----

1970         28.36            7.44        19.68         4.45        2.00       61. 9'3
1971         29.43            7.21        20.42         4.63        1.40       63cD9
1972         31. 49           7.86        22.09         5.56        1. 28      68'.28
1973         35.13            6.49        24.29         6.64        1. 90      74.45
1974         40.39            7.90        26.26         6.92        2.24       e3.71

rate   ,        9.2           1.5          7.5         11. 7        2.9         7.8

The average annual growth of the total "oer-lunch" cost, at
7.8 perc€nt, outpaced the CPI (food away from home) for the
1970-7~ oeriod--the latter had an annual orowth rate of about
6.9 L>ercent.  The larges-t. annual increase -was for "other cash
expenditures" (11.7 percent)     followed by increasps in local
f00d purchases (9.2 percent) and labor (7.5 gercent).      Donated
gf)ods and services and federally donated commodities increased
o~ly slightly--2.9 percent and 1.5 oercent, respectively.

      ~o determine the dopinant factors contributing to cost
growth in NSLP, we increased fiscal yeer 1970 meal costs to
reflect changes in the CPI and then compared these "adjusted
costs ll with the actual costs incurred in fiscal year 1974,
as shown in the table on the next page.

                                             Table 9.3
             Highlighting the Factors Affecting Cost Growth of NSLP Lunch
     (Meal cost for Fiscal Year 1970 adJusted to Fiscal Year 1974 prices and compared
                     with actual Fiscal Year 1974 costs) (note a)
                                Local         "ederally               Other        Donilted      Total
                                food           donated                cash         goods So    per-lunch
                              purchases      commodities   ~       exp~nditures    services      cost
                                - - - . - - - - - - - -·--(cents)---------·--·-·-- ..-
  Actual cost                    40.39           7.90      26.26      6.92              2.24    83.71
  Adjusted cost
    (note h'                    li.:. 'l:~       9.70      25.67      ~                 2.61    80.76
      Difference (note   c,       3.41          -1. 80      0.59      ~                -0.37     2.95

  a/SinCe the CPI's ~tood away fcom home" component medsures changes in consumer
  - purchasing power rather than changes in the individual component costs, the
    differences shown are not precise measures of cost growth. In relative size,
    however, we believe these differences serve to highlight the cost categories
    most responsible for escalations in meal cost.
  b/FY 1970 meal cost escalated by CPI {food away from hamel to fiscal year 1974
  - prices.
  c/Each I-cent difference in cost represents about $40 million in program costs
  -(based on the nearly 4 billion meals served in fiscal year 1974).

      Between fiscal years 1970 and 1974, the cost of preoar-
ing an NSLP lunch increased by about 2.95 cents more than the
amount exolained by changes in the CPl. With nearly 4 bil-
lion meals served in fiscal year 1974, this increase added
$118 million to program costs (e.q., 4 billion meals times
2.95 cents).   The primary source of Qrograro cost increases
was in the cost category "local food ?urchases (3.41 cents) ~                     ll

the secondary source was in the category "other cash excend-
itures" (1.12 cents).   Part of the increase in local food
Durchases compensate~ for lower levels of commodity sup-
oort,ll but s~ch ourchBses accounted for less than-halt of
the increase observed in this category.

     We do not know of any studies which have investigated
the precise cause of these increases, but the relative
change in local food purchases 3nd "other cash expenditures"
suggests that schools increased the use of convenience foods
and labor sRving devices (such as disposable utensils).

l/Legislative orovisions, effective at the beginning of
  fiscal year 1975, authorized a minimum level of commodity
  assistance at 10 cents ~er lunch, or cash payments in lieu
  thereof, with provisions that the rate be adjusted on an
  annual basis to compensate for changes in the CPI for food
  away from home.

Although these items at times may provide economic advantages,
a North Carolina study II has pointed out that this is not
always the case. Labor-usage may be established by State
guidelines which specify in a policy formula the number of
workers to be used for the number of meals served at a par-
ticular school. In the presence of such guidelines. there
may not be an opportunity to reduce labor costs. The intro-
ducti-Qn of labor-saving features could be an unnecessary
    Alternatively, USDA's comments on this report make the
important observation that
     lloverall program costs have just about kept pace
     with inflation. Hence differentials in rates of in-
     crease in purchased foods compared with other cost
     components may reflect a deliberate effort to mini-
     mize labor costs. Wage rates for cafeteria workers
     have risen and continue to rise at a relatively ra-
     pid pace. Other purchased inputs in part may be
     substituted for labor. The increased purchases of
     'preformed beef patties' in lieu of bulk ground
     beef and of individual portion pizzas are illustra-
     tive. Thus, disproportionate increases in food
     purchase expenditures need not reflect ineffective

Cost variations due to
     Another factor influencing cost growth is the decline
in daily participation levels. During the fiscal year 1970-
74 period, the average daily participation per school de-
clined from 276 to 263 students, a 4.7 percent reduction,
as shown in the table on the following page.

!/Nicholson, R. H., So~~££~£~!~-~~P7ct~-£f-~~~_~~~!on~!
  ~~hoo!~~~£~~££g~~~_!~_~££~_~~~!ln~,Eco~omlc~ Informa-
  tion Report No. 12, North Carolina State Unlversity,
  Raleigh, N.C., July 1973.

                                Table 9.4
                     Served Per NSLP School

Fiscal                                              Oaily lunches
                              Enrollment               served
 l~~E                         ---------                -----
1970                             541                      276
1971                             540                      279
1972                             528                      275
1973                             507                      268
1974                             515                      263

~/FNS/Program       Reporting Staff publications.

        In general, there is an inverse relationship between the
number of daily meals served in a school and the unit cost of
preparing a program lunch. Low unit costs are associated
with high participation levels, and high unit costs are
associated with low participation levels. For example:

       --By analyzing fiscal year 1972 data from 160 schools
         across the Nation, a USOA study estimated that for
         each additional 100 meals, the unit cost of preparing
         the type A lunch declines by approximately 3 cents. !/

        --A North Carolina study reported economies of scale
          in labor costs.  The study analyzed scale economies
          for elementary, junior high, and senior high schools
          separately.  It did not find evidence of scale econ-
          omies in high school food service operations.     However,
          in elementary and junior high schools, the study
          reported that labor costs per plate dropped 2 percent
          for every IO-percent increase in the number of meals
          served.    ~/

l/USDA, Economic Research Service, ~£~!_~!ru£!~~_£!_~£~££~
- ~~~~, Unpublished rept., 1973.

~/See    footnote, p. 121.

       Given the large volume of lunches served each year, and
the large quantities of foods used, we believe that efforts
to reduce food costs provide the greatest opportunity for
program cost savings. At today's participation levels, each
pe,nr\y saved' in meal costs would reduce NSLP expenditures by
about $40 million. Some of the areas in which we believe
Federal and State governments can act to lower food costs
without adversely affecting the program's nutritional
standards can be summarized as follows:

     1.   ~~~~!.ngUSDA's me2.L~u!.~~!.~_to'emE!:'.~~ize~
          ~tr!tio~~l standard ratner_~ha~he-lX~-~-~!.
          pattern. There are a number 01 lndicatlons that
          NSLP's type A meal pattern may'increase the cost
          of program lunches.  (See ch. 5.) Some authorities
          believe that revising USDA's regulations to focus
          on a nutritional standard would provide lower cost
          meals, less plate waste, and higher levels of
          student participation.

     2.   ~£~~ring_th~_eE~te!.~_£Sgd~~~~~ts-~-~!:'.~~£hOO!
          lunch. Based on the stu les we revlewed {see cha
          JT-,-the Nation's schoolchildren have mean protein
          intakes well in excess of RDA standards.   In this
          regard, it appears that NSLP's nutritional stand-
          ards place undue emphasis on protein, usually the
          most expensive component of the NSLP lunch.

     3.   !~ef~~i~~-!~-food P£££~~~~~~£~~mi~_£f_~~!!
          ana medlum~slzed school ~ystems. USDA and large
          scnoor-systemS-purcnase-rooaS-at prices consider-
          ably lower than the vrices paid by small and medium-
          sized school systems.  (See ch. 8.) It appears that
          sizeable reductions in program costs might be
          achieved by: .!/

l/According to ERS estimates (see table 8.4), the average U.S.
- school pays prices 6.7 percent higher than those paid by
  USDA or large school systems for comparable food purchases.
  In fiscal year 1975, approximately $1.808 billion of NSLP
  foods were bought by schools in local markets.  A potential
  exists to reduce NSLP food costs by more than $100 million
  per year by improving the food procurement economies of
  small and medium-sized school systems.

              --States consolidating the food purchasing
                operations of small and medium-sized school
                systems to take advantage of volume pur-
                chasing economies, !/ and
              --USDA directing proportionately greater com-
                 modity support to those school systems which
                 pay the highest prices for local market food
      State governments can also act to facilitate productiv-
ity increases in school food service operations. Although
productivity increases in the food service industry have
historically been low, some authorities believe the pressure
of rising labor costs will necessitate greater improvements
in the future.   It is not expected, however, that such pro-
ductivity increases will offset increasing labor costs.
Investigation into labor savings by FNS is a continuing
effort.   Its success will depend largely upon each State's
willingness to employ FNS' recommendations and equipment
support to offset labor expenses. The introduction of
convenience foods and labor-saving equipment in those States
with policies requiring a fixed number of workers per meal
will, in general, aggravate cost growth.

     In light of the potential for cost savings in the food
procurement area, we recommend that the Secretary of Agri-
culture examine approaches and implement procedures for im-
proving the food procurement economies of small and medium-
slzed school systems.

     USDA stated that actions related to our recommendation
are currently underway.   (See app. I.) A report dealing with
the food procurement economies of small and medium-sized school
systems is scheduled for completion in this fiscal year.

IjUSDA suggests that when a State is composed of both large
- and small school systems, the development of a centralized
   purchasing system should be used on a voluntary basis. A
   compulsory participation arrangement, while helping smaller
   systems, could penalize the larger ones as they could not
   capture the savings from advantageous local bids whenever
   they become available.

                         CHAPTER 10

     The Federal Government provides assistance to States for
serving lunches to schoolchildren. The assistance is pro-
vided on a meals-served basis and includes the following:

     --£~~~_~~~~.    Section 4 of the National School Lunch
       Act provIdes a cash reimbursement rate (general cash-
       for-food assistance) for all meals served to school-
       children. Section 11 of the act provides an addition-
       al cash reimbursement (special cash assistance) for
       meals served free or at a reduced price to children
       from poor or near-poor families. 11 These rates are
       adjusted on a semiannual basis (Jan. and July) to
       reflect changes in the series for food away from
       home of the Consumer Price Index.

     --Commodi~ assistance. Section 6 of the National
       Schoor-Lunch-Act-reguires that

          1'* * * the national average value of donated foods,
          or cash payments in lieu thereof, shall not be less
          than 10 cents p~r lunch, and that amount shall be
          adjusted on an annual basis each fiscal year after
          June 30, 1975, to reflect changes in the series for
          food away from home of the Consumer Price
          Index * * *.11

       During fiscal year 1974, $316 million worth of agri-
       cultural commodities and other foods were supplied
       to States, of which $67 million represented foods
       purchased eXFressly for NSLP (section 6 funds) and
       $249 million represented commodities contributed
       through the Federal price support programs and pro-
       grams for strengthening markets, income, and supply.

      Federal assistance rates for the school lunch (effective
Jan.-June 1976) are shown on the following page:

I/Special cash assistance for a reduced-price lunch is 10 cents
- less than the spe~ial cash assistance for a free lunch.

                           Table 10.1
              School Lunch Federal Assistance Rates
              ------------- ... _-----------------

                              Regu ar-       Reuucea-
                          ££!.~_i.~!:~_~)        ££~_(note_~)    rree
                                - - - - - ( cents )-)- - - - - - -

General cash-far-food
  assistance (sec. 4)           12.50              12.50        12.50
Special cash assistance
  (sec. 11)                                        46.75        56.75
Commodities or cash in lieu     11.00              11.00
                                                   ----         --
     Total                      23.50              70.25        80.25
~/Eli9ibility   guidelines for free and reduced-price lunches:
    --Poverty guideline for fiscal year 1976: $5,010 for a fam-
      ily of four. Any child from such a family is entitled to
      a free lunch.
    --States have the option of increasing their free lunch
      guidelines up to 125 percent ($6,260) of the Secretary's
    --Children from households with an annual income level
      which falls between the State's guidelines for free
      lunches and 95 percent above the Secretary's poverty
      guideline ($9,770) are to be served reduced-price lunches
      at a price not to exceed 20 cents.

For the 1976-77 school year, the average level of Federal
assistance for regular, reduced-price, and free meals is ex-
pected to rise to 25.6, 77.3, and 87.3 cents, respectively. !I

     States must match the Federal cash grant for regular-
price lunches from sources within the State at a 3-to-l

liThe Budget of the United States Government, Fiscal Year
-   I~Z2~~E~ndI~; Governmenr-printTng-6flice~-wash.,-57c7,
    1976, p. 173.

ratio and 8 percent of the matching funds must come from
State appropr ia ted funds. (For States wi th below-aver age
per capita incomes, the ratio m3y be reduced.) 8etween
fiscal years 1974 and 1975, annual contributions from sources
within the States increased from $1.97 billion to an estimat-
ed $2.14 billion, about 60 percent of which came from
students' payments. (Refer Co table below.) Traditionally,
States have exceeded the matching r~quirements prescribed in

                                               T<Jble 10.2

                              Sources of National School Lunch Program Funding
                                       Fiscal Year 1910-75 (note a)

               Federal                    Children's               State " local
 riseal   contribution                     payments                contribution                  Tota I
  year     (mi 11 ions I        Percent   (,dllions)     Percent    (millions)     Percent     I note b!

 1970           S     565.5      25.5      $1,105.0      H.B         5546.6         24.7       $2,211.1
 1911                 809.5      32.5       1,090.2      4].7         593.3         2] .8       2,493.0
 1972               1,050.8      38.5       1,080.4      39.5         59~.O         21.9     !?12,130.3
 1973               1,142.4      38.6       1,123.7       38.0        692.7         23 ...      2,958.8
 1914               1,401.4      41.6       1,174.2       34.8        796.8         23.6        3,312.4
 1975(est. )        1,702.0      44.3       1,290.0       )).6        SSO.O         22.1        ).8~2.D

 !/FNS/Program Reporting Staff publications.

 ~lThe pro9ra~ operates on a nonprofit basis. Variations between funding and                 co~ts   are
   carried forwlrd as a surplus/deCicit to the succeeding year's oper"tion.
 £/Di[Cerences due to rounding_

     Between fiscal years 1973 and 1974 the number of regular-
orice meals declined by 90.3 million and the number of free
and reduced-price lunches increased by 80.4 mil1ion--an over-
all decrease of 9.9 million meals served. However, ~esDite
the fact that fewer meals were served, the shift to free and
reduced-9rice lunches actually increased the Federal Govern-
ment's share of program costs:

                                 Table 10.3

                 Millions of meals
Fiscal    Regular=--Reaucea=------
 :i~£         eEic~      e£l~~       Free

1972           2,686.8    78.6     1,206.7    67.6   2.0   30.4
1973           2,606.4    38.5     1,363.9    65.0   1.0   34.0
1974           2,516.1    45.5     1,437.3    62.9   1.1   36.0
1975(est. )    2,451.2    89.4     1,536.2    60.1   2.2   37.7

-a/FNS/Program Reporting Staff publications.
     In fiscal year 1976 the serving of reduced-price lunches
was changed from a local option to a mandatory requirement
for all NSLP schools. Since that time the number of meals
served at reduced prices has increased significantly--and so
has participation in the free lunch program. The Federal
share of program costs has continued to grow.

     Because of the strong possibility for further increases
in the free and reduced-price program, several observations
on the funding structure should be noted.  These are:

     --The reduced-price lunch, combined with a 20-cent
       student charge, provides the least-cost alternative
       in regard to State and local contributions. The
       free lunch is, in general, the most expensive alter-
       native.  (See table 10.4.)

     --The special cash assistance subsidy for a reduced-
       price lunch, at 10 cents less than the rate provided
       for a free lunch, increases faster than changes in the
       CPI (e.g., CPI adjustments are based on the free
       lunch rate rather than on the rate provided for a
       reduced-price lunch). This feature is expected to add
       more than $2 million to Federal program costs in
       fiscal year 1977.

     --The 20-cent ceiling established as the maximum child's
       payment for a reduced-price lunch does not provide for
       the absorption of cost increases due to advances in
       the cpr. Overadjustment of the Federal subsidy

      (described in the preceding paragraph) compensates
      for a portion of the increase, but the contribu"ions
      of State and local governments also need to increase
      faster than advances in the CPl.

     --Current law requires that States must match the Feder-
       al cash gcant for regular-price lunches from sources
       within the State at a 3-to-l ratio and that 8 percent
       of the matching funds must come from State appropri-
       ations.   (See table 10.1.) Since there is no require-
       ment for State appropriations to complement the
       Federal subsidies for free and reduced-price lunches,
       the continued shift of students to those programs may
       place an inordinate burden on the resources of some
       local governments.

                           Table 10.4
                Fiscal Year 1974 School Lunch

                                               T:t.ee of lunch
                                        Regular= -Reduced
                                         E£!£~        E£!~            Free
                                         - - - - - - { ( cen ts ) } - - - - -

Total cost per lunch                     83.7           83.7          83.7
  Deduct:  Federal contribution
             (note a)                   18.2            54.0          64.0
                                        ---             - -           ---
Net cost to States                       65.5           29.7          19.7
  Deduct: children's payments
              (note b)                  46.3            20.0
                                        ---             ----
Required State & local
  contribution (note c)                 19.2              9.7         19.7
~/Includes   commodities distributed to States.

~/Assumes   20-cent charge for all reduced-price lunches.

c/State and local contributions for FY 1974 were in excess of
  this requirement and are assumed to be applied to previous/
  subsequent deficits.

      At Dresent, we believe the Federal subsidy structure has
facilitated a cost-effective increase in t:SLP's participation
levels (i.e., if the funds eKoended for the reduced-or ice
program were used to increase· the Federal contribution to the
regular-price proqram, the increase in participation would be
less than t~at achieved by the reduced-price program).
However, it should also be n.,ted that any increases in the
cost of preparina an NSLP lunch (in eKce~s of CPI adjustments)
must be borne by State and local contributions. The large
increases in free and reduced-price participation, combined
with a requirement to increase the subsidies for reduced-
price lunches faster than advances in the CPI, 9lace a oremium
on efficient ~rograrn o~erations. Any program cutbacks by
State or local officials would likely be reflected in the
child's payment for a regular-price lunch.

                            PART V
     The Office of Planning and Evaluation, USDA, has defined
program evaluation as
     "* * * the formal systematic assessment of the actual per-
     formance of ongoing programs in meeting the goals of USDA
     missions, achieving program objectives, and serving specified
     target groups. It is concerned with measuring the effects
     and benefits flowing from programs and their costs. It ex-
     amines the extent to which program activities have been
     carried out in relation to the opportunities that have the
     most favorable benefit/cost ratios or otherwise maximize
     the beneficial effects in relation to cost."
     Chapter 11 addresses the current state of evaluation of
NSLP. Unresolved issues and suggested corrective measures are
presented for committee consideration.

                          CHAPTER 11

                     EVALUATION RESEARCH'

     Program evaluation reoresents one of the most effective
tools available for closing the gap between policy formula-
tion and responsive program administration. A well-directed
evaluation provides objective evidence on what a proqram
accomplishes, ho~ these accomplishments compare with intended
objectives, and how effectively program resources are managed.
For Federal 9roarams, good evaluation studies not only help
to measure program results but also provide an analytical
tool to assist the Congress in apportioninq scarce budget
resources, in considering revisions to an existing program,
and in overseeing progLam administration.

     While t~is report brings together a great deal of
information about various aspects of NSLP, its most import-
ant findinqs are that fundamental issues about the program's
im~act remain unresolved.  Four questions, which we believe
are the basic logic ste9s for evaluatinq NSLP's effective-
ness, have not been satisfactorily answered.

     1.   What is the program's impact on the participants?
          Does the program, nationally, safequard children's

     2.   What is the program's impact on the consumption of
          agricultural commodities? Do children consume more
          eqricultural products under NSLP than if it did not
          exist? And how does the change in consumption, if
          any, affect the Nation's agricultural economy?

     3.   Is the program reaching the defined target popu-
          lation? To what extent are nutritionally needy
          children participating in NSLP and what are the
          health conditions and dietary habits of those who
          do not participate?

     4.   To what extent are the specified services provided?
          And, in relation to alternative ways of providing
          these services, are program services provided in
          the most cost-effective manner?

     The National School Lunch Act of 1946 established two
major objectives: (1) to safeguard health through a program
of nutrition intervention and (2) to supplement farm income
by increasing food demand. Over the ensuing years, national
priorities· changed; NSLP has become primarily focused on one
objective--safeguarding schoolchildren's health.
     To help meet this objective, the Secretary of Agriculture
reguires that meals served under the program be designed ac-
cording to a specified (type A) food pattern which should
provide, on ~he average, one-third of each participant's
recommended dietary allowance. This lunch--as designed,
served, and eaten--is, in our opinion, one of the most
crucial factors affecting program effectiveness. The quantity
and type of food included in the lunch largely determine its
cost and the amount of agricultural commodities consumed.
The price and presentation of the lunch determine how well
the program reaches the Nation's schoolchildren. And, the
nutritional qualities of the lunch determine how well the
program safeguards health.

-Health   considerations
     Although education in nutrition is regarded as a major
strategic method for safeguarding public health, it appears
that State and local programs of nutrition education have
not been completely successful in developing good food habits.
Nutrition, the lack or the excess or the quality of it, ap-
pears to be a problem for millions of the Nation's school-
children. The threat is not overt, as in deficiency diseases
such as beriberi or scurvy. It is more complex, often without
visible signs, and usually associated with one or more of the
      --Deficiencies in RDA nutriture, which may impair growth,
        development~-and the-ablflty to withstand infectious

     --Excessive intakes of calories, which may contribute to
       the-devefopment-of-heart-and allied diseases.
     --Poor choices in the nonnutrient eart of diet, which
       may-contrlbute-ro-the-aeveiopment-of-dlseases such as
       tooth decay and, ir the opinion of some authorities, hyper-
       tension and bowel cancer.

     While these problems suggest a need to place greater
emphasis on the subject of nutrition education, it should be
recognized that such actions are traditionally the preroga-
tive of State and local governments. NSLP's authorizing
legislation expressly prohibits the program from imposing
any requirement relative to the teaching of nutrition to
schoolchildren. The program's health impact, therefore, is
directly dependent on the benefits of eating a program lunch.

     NSLP's requirement for each lunch to approximate one-
third RDA should not obscure the fact that the lunch is but
a supplement to a childls home diet.  Its effectiveness lies
not in its nutritional content alone, but rather in how well
it complements the home diet in providing optimum nutrition
for the individual.

     In our opinion, the design of the NSLP lunch needs to
be reassessed. Not only does the program's single meal
pattern appear Iiout of phase " with the needs of school-
children, it also has an inherent capability for producing
undesired side effects. As set forth in this report,
indications are that the current lunch

     --provides a valuable source of nourishment for some
     --may contribute to obesity in others; and
     --is relatively ineffective in improving iron nutriture
        (the most prevalent deficiency reported for school~

     The issue as to whether or not NSLP provides a net
health benefit is complex and riddled with uncertainty.
There is some evidence that the school lunch, if paired with
a nutritional supplement or with the school breakfast, can
improve the nutritional levels of schoolchildren. But, the
overall health impact of the NSLP lunch itself is presently

     Few studies have attempted to evaluate NSLP's effective-
ness as an agricultural program. We did find consistent
indications that NSLP participants consume a greater quantity
and variety of commodities at lunch than nonparticipants,
but there was no conclusive evidence that this represented
and overall increase (at home and at schoOl) in consumption
or that such increases were caused by NSLP. Nevertheless,

  we believe that NSLP, through substantial purchases of foods
  in local markets and as an outlet for foods acquired under
. USDA .price stabilization and surplus removal actions, has
  probably strengthened the overall demand for farm products.
     There is presently some controversy among school food
service directors as to the influence of the type A meal
pattern and/or USDA's commodity distributions on NSLP's
effectiveness as a nutrition program. The areas of contro-
versy can be summarized as follows:
      --Type A meal pattern. School food service personnel
        appear to be almost evenly divided in their opinions
        of USDA's type A meal pattern: half believe the
        pattern is needed to safeguard the program's nutri-
        tional standards, and half believe that the pattern
        inhibits student participation. The latter group
        emphasizes that one-third RDA can be met in many forms
        and that the inflexibilities of USDA's food pattern
        contribute to higher costs, food waste, and a meal
        design which is not representative of today's eating
      --USDA'S commodity distrlbutions. Current legislation
        mandates a guaranteed level of commodity assistance
         which, except in special circumstances, is provided
         in the form of foods acquired under USDA price
         stabilization and surplus removal actions. In essence,
         a sizeable share of NSLP foods is provided without
         regard to the menu planner's desires. Many school
         food service directors believe that USDA's commodity
        distributions provide high quality foods at substantial
         cost savings which, by keeping meal prices low, encour-
         dge higher levels of student participation. There
         are, however, many complaints that administrative
         problems in the timing and quantity of commodity
        deliveries interfere with menu planning and student
         acceptance of the NSLP lunch.
In each instance, the points of disagreement appear to be a
result of adminIstrative practices rather than legislative
requirements. And each of the opposing viewpoints is worthy
of   consideration~

     Present conditions in the Nation's agricultural economy
are considerably different than when the program's agricul-
tural objective was enacted (i.e., while significant mar~et
imbalances still occur, the agricultural economy is no longer
characterized by seemingly permanent excess supply; concern
has tended to shift to the problem of shortages and away from
the problem of surpluses).
Because of this and because the agricultural objective pro-
claims that a major purpose of the program is to increase
food demand (thereby increasing food prices), the objective
itself may no longer be desirable.

E2£!i£lEation   considera~ion~

      Between 1971 and 1975, an expanded .free/reduced-price
program substantially increased the participation of low-
income children: but, because much of the increase was offset
by declines in the participation of regular-price students,
overall participation levels tended to remain constant. The
shift toward low-income children (the popula.tion group with
the greatest prevalence of nutritional problems) probably
increased NSLP's overall effectiveness as a nutrition program.
On the other hand, NSLP became less effective in reaching the
regular-price student (a population group containing several
times as many nutritionally needy children).
     In fiscal year 1975, 44.8 million students Cabout 88
percent of the Nation's total) were enrolled in NSLP schools.
Roughly one-fourth were eligible for free or reduced-price
lunches; the remainder had to pay the "regular" price. Of
those eligible in each group
     --86 percent participated in the free/reduced-price
       program and
     --47 percent participated in the regular-price program.
Of all U.S. schoolchildren who did not eat the NSLP lunch,
about 76 percent were "nonparticipants in NSLP schools.11

     It appears that the NSLP enrollment itself presents the
greatest opportunity for further increases in program
     Although many authorities have expressed a desire to
improve NSLP participation levels, the question remains as
to how this can best be accomplished. One method for improv-
ing participation would be to lower the price of the NSLP
lunch. However, price is not the sole factor influencing
participation; daily participation levels are also affected
by noneconomic factors such as

    --the presence of competitive food sources,
    --attitudes of school administrators, and
    --menu choice and food preparation.
     Available studies, though beneficial in identifying
some of the "factors" affecting participation, provide very
little quantitative support to assist NSLP decisionmakers
in estimating the participation impacts of various pOlicy
alternatives. Our research indicates that
     --price-participation relationships provide un ex-
       tremely weak forecasting tool,
     --the relative importance (rank) of the individual
       factors affecting participation has not been fully
       determined, and
     --the IIrecognized factors" have not been shown to be the
       major cause(s) for variations in NSLP participation.
Moreover, there is a lack of information about how a change
in NSLP participation affects the nonparticipant population--
information which is needed to assess the full impact of a
participation change and to target the program toward those
children in greatest need.

Cost considerations
     While it is true that NSLP operating expenses increased
rapidly over the 1973-75 period, the cost increases appear
to be due primarily to inflation. Discounting the effects
of inflation, the cost of producing an NSLP lunch actually
     On the other hand, we have some doubt as to whether or
not program services are provided in the most cost-effective
manner. As set forth in this report, it may be possible to
reduce NSLP food costs by more than $100 million per year
without sacrificing the program's nutritional impact if
Federal and/or State governments act to:
     --Revise the program's regulations to emphasize a
       nutritional standard rather than the type A pattern.
     --Lower the protein requirements for the school lunch.
     --Improve the food procurement economies of small and
        medium-sized school systems.

APPENDIX I                                                           APPENDIX I

                                   FOOD AND NUTRITION SERVICE
                                      WASHINGTON, C.C. 20250

 :r-rr. Henry Eschw:ge
 Director, Cow~uriity and                                               i~PH 2 '! i~11
   Economic Develoyment Division
 United States General Accounting Office
 Washin€Go~, D.C.   20548

 Dear I1r. BGch'tlege:

 The enclos'i.ll'e to this letter responds to ths General A,~C;Ol.L'>j,ting Office's
 draft report titled, t1Lllpact 2..:."1d Effsctiveness of Sob.:::.: IlilllCh Prog.r207J:
 A S;y"nthesis c.f E\raluation Studies. II \oi1:ile the enclomtI."n ~.s o ffe:·:lC. -:lS
 the response of the Food. and Nntrition Serrice, 'ile h£;78- il1coJ.'pora.tGo.
 the resp(;D.se of the Economic P..et-i2arc!1 Serlice ",hich was sE:nt to ~YCl).
 UDc.cr SelKl.:::'ltc cove:!'. ~.'he Ag1.'icultu:!."'al Stabilization Dnd ConBe:t"taticl1
 Service and the ncpartt::ent r s Officl:l ()f r'ianagement and Fhlance offe::'EHl
 no fonnal cOr."!!1ents for i.nolusion in our report. Beyond the specifi,J
 points covered in the enclosure "Ie su.gges-G that GAO subr:d.t the PO'1't~_0n
 of the report (~ld sw~~aries) dealin~ with nutritional espects of the
 National School Lunch Progran to a panel of nutritioniG';;s appointed
 by a technical advisor;'f e;J'oup such as the Food and Nutrition Boar·i
 of the 1Taticnal Acad.e~:1Y of Sciences for revi~w prior to its subrr(lssiol1
 to the Congress. 1/

 GAO's reccrJIJendation regarding the effect of commodity distributi.()J1
 surges on the National School Lunch Program's nutrit:1.onal objecti-ve
 is not covered in the attached response. The DepartT2s;1i;' E: posi til)n
 on this issue is :overed in our response to GAO's report CE~·77-32
 dated January 1, 1977. USDA's statement of actions taken on the
 recommendations in that report has also been sent to GAO.

 We hope you find our comments and suggestions usellll as you prepare
 your final report to the Con~ess.

 S'" ........""lv
  -7 7) "           .
                         'I   ,/

 Enclosure              v._

 l/GAO does not believe this is necessary since nutritional
 - experts in and out of Government were consulted during the
    study, and the work reviewed by HEW.
 APPENDIX I                                                                 APPENDIX I

Otlanlz8tlon   o~   Response

This T@port follows the same organ128tlonal pattern as the major GAO report

(i.e. not the summary).        FNS's comments on each of the five parts of the GAO

report are covered under similarly titled sections of this report.                       Where

p08.1ble~   we have referenced the specific page(s) of the GAO report to

which the comments apply.
                       ISee G.\O note 4, p. 165.J
~~al    Comments

An overriding theme of the GAO report 1s the need for          tI   •   •   •    a comprehensive

evaluation of the program's effectiveness in meeting its stated legislative

objectives,"   The need for such an evaluation is recognized in FNS.

An evaluation plan projecting FN5' research plans over the next five years

has been drafted and is currently under review.           This plan calls for develop-

ment of a methodology for assessing program performance in light                    ~f   the

nutritional objectives contained in      OUT   enabling legislation.              Because the

plan is under review. and therefore subject to change. it is not available

for release.   We will. however. forward a copy of the plan to GAO as soon
as it is available.     The Department conducted     8   comprehensive review of

studies of the program in 1974 and submitted it to the Congress in response

to Public Law 93-150.     This report pulled together all of the existing

information on the subject that vas available at that time.                     We note that

the GAO "mace extensive use of this report in its critique of food program
evaluations.   Our evaluation plans call for another major compilation of

existing information in 1981.

 APPENDIX I                                                    APPENDIX I
As pointed out in the Economic Research Service's comments, the literature

review in GAO's report 1s reasonably complete, but does not indicate to

the reader that almost all of the research reported was directly conducted

or funded by USDA with professional staff responsible for research oversight.

This 1s a significant omission.

Finally, there are two summaries of the overall report.    The first is the

separate summary titled liThe School Lunch Program -- Is It Working?" The

second is PART V of the main report.    Consistency between the two summaries

and the main report is essential.   Thus, any of our comments or recommended

changes to the main report which GAO decides to accept, should also be

reflected In the two summaries.


- Program Philosophy and Basis

  References:   Page 20, Main Report; Pages 5, 6 & 8 Summary

It has never been the philosophy of the Department that the basis for the

NSLP is to serve as a nutrition intervention program to prevent a state of

disease.   Adequate nutrition will allow for maximum dividends in the main-

tenance and promotion of health, but cannot guarantee total well-being and

absence of disease.

The Department has not evaluated the impact of NSLP on safeguarding the .health

of the Nation's children.   To assess the impact of the NSLP on the overall

nutritional well-being of its participants would be difficult.   The program
is designed to provide a maximum of five meals per week.   Assuming that

an average of 1/3 RDA is provided through the lunch over the five day period

  APPENDIX I                                                          APPENDIX I

this would be only 20% of the- child's total nutritional requirements for
that .period.   It would be most difficult to demonstrate any significant

changes among large groups of people as a result of this small amount of

food.   The complexities associated with determining nutritional status and the
lack of standards that contribute to good health and adequate          nutritio~

(as presented 10 Chapter 3) further complicate          the evaluation proceS$

being suggested in the report.    Because of the relatively small proportion

of the total nutritional requirements the NSLP is expected to provide and

the complexities associated with determining nutritional status, it is

~uestionable    that the study would   b~.?ucces$ful    in   accompli5hin~   its objectives.
             [See GAO comments on p. 45 of this report.l
  Meal Pattern/Plate Waste

  References:    Pages 39 & 108, Main Report; Pages 4, 5, 6, 7, 10 & 14,          S~ary

The nutritional goal for the NSLP is to provide approximately 1/3 or more

of the RDA for children of various ages.       The RDA's are estimates of the

average known nutritional needs of population groups and are not recommended

intakes for individuals.    Establishing a simple nutritional standard per se

would be a misuse of the ROA's.    The RnA's are not to be confused with

requirements.    They are estimates that exceed the requirements of most

individuals and thereby ensure that the known nutritional needs of

practically all healthy persons are met.       The basis for estimation of

RnA is such that, even if a person consumes less than the recommended

amounts of some nutrients, his diet is not necessarily inadequate for

those nutrients.

  APPENDIX I                                                          APPENDIX I
 On   page seven of the s\unmary report. GAO refers to lithe program's sioble

 meal pattern."       The program does not have a single meal pattern.       The

 Type A Pattern is based on minimum requirements designed for the lO-to-12-

 year old    ~hild.   V~ried   portion sizes are recommended for older and younger

 children to meet their     s~ecific   nutritional needs.     The Type A Pattern has

 been reviewed and revised periodically since development in 1946 to reflect

 current nutrition knowledge and food consumption habits.           While the report

 criticized the Pattern for its limitations and its appearance as being un-

 responsive to today's eating habits t it reported that butter/fortified

 margarine was part of the Pattern.       This requirement was deleted in

 June, 1976 and the report should definitely be modified to reflect this

 fact.   Previously in 1969, the amount of butter/fortified margarine had

 been reduced from two to one teaspoons.         Eliminating this requirement is

 consistent with the knowledge of possible undesirable side effects of

 large amounts of fat in the diet.       Other changes that have been implemented

 include:     1) In 1973, the definition of milk was expanded from whole fluid

 milk to include fluid forms of whole, lowfat, skim, cultured buttermilk

 and flavored forms of these milks.       2) In 1974, the definition of bread

was expanded to include crackers, taco         ~hells,   pizza crust, etc.   These

 changes have increased menu planning flexibility within the Pattern while

 maintaining its nutritional integrity.
[The report has been revised to show the current type A pattern.
 See pp. 6d and 6~.J
 The kinds and amounts of foods as specified in the Type A Pattern are based

on the four food groups from the Daily Food Guide.            These four food groups

have served as the basic framework for menu planning by nutritionists for

years.      We agree with Dr. Mayer's statement from page 5 of the summary that

 eating habits have changed.       However. the flexibility provided by the Tvoe A
  APPENDIX I                                                       APPENDIX I
Pattern enables menu planners to offer meals which respond to these changes.

Dr. Mayer'. example of    ~   typical lunch supports the Type A Pattern with the

exception of one 1e8& fruit or vegetable.       Other menu planning approaches

will not likely provide greater flexibility than is provided through the

Type A Pattern 1f the nutritional goal is to be maintained.       The types and

quantities of foods specified by the pattern are those required to satisfy

the goal.    Consequently, an expression of this goal as a requirement would

result in the utilization of approximately the same types and quantities of

foods.      [See GAO comments on P9. 6Y ana 70 of this report.]

Three studies have been conducted to compare nutrient standard methods of

menu planning with Type A menu planning 1n school lunch.       The study con-

ducted by Colorado State University comparing the manual nutrient standard

with Type A did report that no significant differences     ~~re   found in either

student's rating or consumption of menu items between the two methods.         How-

ever, 1n the study conducted in the Dade County (Florida) Public School

System, consumption and acceptability rates were higher in schools serving

Type A menus· than in schools serving menus planned by a nutrient standard.

Additionally, a similar study in Memphis City (Tennessee) School System also

found that students consumed more of the Type A meals than of the meals planned

by a nutrient standard.       [The report has been revised to acknowleoge
                               FNS' comment.  See ?   76. )
A frequent criticism has been that the Type A Pattern 1s not responsive to
the protein contributions of food components in the pattern, because credit

is not given for the protein in both meat and milk.      The pattern does not

require meat and milk for protein alone, but for all other nutrients      8S

well.    Using iron and thiamin as examples, the goal of 1/3 of the RDA could

      APPENDIX I                                                    APPENDIX I
not be Met without the specified quantities of both meat or alternate and

milk.    Consequently, expressing the requirements on a nutrient basis would

not   l~r    the cost of   ~~~.

On page 30 of the main report, GAO states     iron-t~-calorie   ratios were   lo~cr

for school lunches than children's home diet (based on Ten State Nutrition

Survey data).      A recent nutrient calculation of the Typ'c"A Pattern, based on     fOf

representative of frequency of service to 60 test groups over a four week

period shows that the Pattern furnishes approximately 8 mg. iron per 1,000

calories.    This amount is well over 6 mg. iron per 1,000 calories which is

the awount   e~~ected   from a varied, well-balanced diet as specified by the

RDA's.   lSee GAO COiTlillents on PP.   71 and 72 of this report.)

Accumulating evidence demonstrates that the amount of iron potentially

available from foods depends not only upon the amount of iron supplied but

the nature of that iron and the composition of the meal with which it is

consumed.    This fact is demonstrated in the RnA for iron, which bases its

requirement on the assumption of an average availability of 10 percent of

the food iron.     Furthermore, there has been consideration of expressing the

requirement by a different method.      The total iron content of the diet is

thus a relatively poor indicator of the adequacy of the diet with regard to

iron.    Two of the factors known to affect iron absorption are. the source

of iron in the diet and other foods consumed with the supply of iron.         The

mp.als served in NSLP have a positive effect on the availability of iron.

Listed below are factors that enable the NSLP to positively affect iron

nutriture:    1) most meals contain heme iron from meat, poultry and fish -

this form of iron is most readily absorbed; 2) the meat/meat alternate

  APPENDIX I                                                           APPENDIX I
component of the NSLP    erbances the absorption of iron from other sources; and

3) the NSLP provides a high level of vitamin C which enhances iron

The statement on page 10 of the summory report (attributed to half of the

school food service personnel) that states, "1/3 RDA can be met in many

forms and that the inflexibilities of USDA's food pattern contribute to:

higher costs, food waste, and a meal design which         i~   not representative (If

today's eating styles", appears unfounded.      Based on the foregoing discussion

the statement reflects a lack of understanding       Ot   the Type A Ptittern and the

intent of the RnA's.
       [See [>[>. 74 ana 75 ana table 5.7 in this re[>ort. ]
A recent review of the Type A Pattern, based on the 1974 revision of the

RnA has led to the development of recommended revisions which are under

consideration b) the Department.     In an effort to reduce plate waste while

maintaining the nutrioonal goal of the lunch program, the revisions would

specify minimum meal requirements by age/grade groups, thereby allowing

significantly smaller portion sizes for elementary school students while

more accurately meeting the nutritional     need~   of children of all ages.

   FNS Suggestions for Improving Participation and Minimizing Plate Waste

On page 6 of the summary report, GAO states that, " .. • • the Type A lunch

is often presented in a form which discourages student participation and

contributes to plate waste. II   The Food and Nutrition Service shares GAO's

concern about program participation and plate waste.           However, there are

ways of addressing these concerns short of abandoning nationally established

meal standards..   The following list of activities is suggested as a means

to positively effect participation and help reduce plate waste .

APPENDIX I                                                 APPENDIX I
  Provide Quality Food through Effective For1service Management

  1.   Increase understanding of school foodservice personnel in

  subjects of good menu planning, quality food production, and

  imaginative techniques for merchandizing school lunches.       Menu

  planners must be aware of the proper techniques for using the Type

  A Pattern to provide nutritionally adequate meals using a      v~rlety

  of foods of the kinds and amount9 children will enjoy and consume.

  2.   Develop materials to assist school foodscrvice personnel in

  planning menus and writing specifications f9r pre-prepared foods

  as well   8S   food handling techniques and serving methods for these

  foods.    Monitoring guidelines should be provided for State and

  local school foodservice personnel.

  3.   Implement the recommended revision of the Type A Pattern to

  provide more flexibility in portion size adjustment according to

  age groups and to permit use of more conventional foods.

 Develop an Awareness of the Importance of Nutrition to       Heal~h

  4.   Direct nutrition education activities toward the emphasis of

  foods, the development of good eating habits and their relationship

  to health, growth and development.       Nutrition education activities

 should utilize modern teaching techniques that relate nutrition to

 day-to-day activities in both the classroom and the lunchroom.

 5.    Encourage more schools to involve students 1n the lunch program

 through activities such as menu planning, cafeteria decoration, and

 building a student awareness of nutrition and the importance of

 minimizing plate waste.
 APPENDIX I                                                     APPENDIX I
     Enllst the Support of School Administration and Program Cooperators

     6.   Emphasize the importance of a complete lunch program 1n which all

     persons involved contribute to its effectiveness.    Encourage school

     administrators and teachers to assist the school food service manager

     and students 1n developing an effective program.    Encourage school

     administrator and teachers to eat with the students and to schedule

     lunch periods that minimize length of serving lines and provide

     adequate time for eating.   Encourage all school personnel to establish

     effective communications with parents and the community.

     7.   Encourage schools to provide choices for elementary and

     secondary students within each component of the lunch.

     8.   Encourage schools to eliminate the sale of "snack" foods during

     the lunch period.

On Page 39 of the main report, GAO indicates that NSLP lunches should be

designed to better supplement the school child's home diet.     Based on

previous discussion 1n this paper, there is no way to determine and evaluate

each participant's specific nutritional needs on a daily basis, let alone

produce and serve meals to meet these needs.   USDA has consistently

encouraged schools to offer a choice in Type A meals, including a salad or

diet lunch. 1n an effort to meet the nutritional needs of various segments

of the student population.   However. even when varied meals are offered,

there is no way to ensure that each student will select the meal most

applicable to his/her nutritional needs.

  .4PPENDIX I                                                                  APPENDIX I

   Non-nutrient Diet            Diseases

   Reference:     Pagt        Summary

The statement on i.          6 of the SUI.1T:'lary repoTt~ that, "Poor choicr:.~; in the.'

non-nutrient part (lr diet which n:ay contribute to • • • tooth decay

hypertension and bOHel cancer", is not sufficiently qualified.                    The absolute

causes as "'ell as the dietary and health practices related to these abnornal-

ities arc not kno",,"l1.     lSee technical note on p.             OJ   oE tnis report.J

   Caloric   Intal~e   and   Developnl~nt   of Heart Clnd Allied Diseases

   Reference:     Pages 6&7      Summary

On page 7 of the    summ~ry     report CAD states that, "Indications are that the

current lunch -- may contribute to phesiey                    "
                15ee p!=>.    45 anu 4b oE tnis report.i
The nutritional goal of school lunches is approximately 1/3 of the RDA or

more (over a period of one week)           fo~   nutrients other than energy.        Foods

specified in the Pattern will not generally result in a                 m~al   containing 1/3

RnA for energy.     It is believed to be desirable for lunches to furnish less

food energy as a percent of the RnA than for various other nutrients.                    }~ny

children do not need a comparable high level of food energy at lunch time

because food eaten at other meals and snacks frequently provide more than

2/3 of their daily energy requirements.                Furthermore, the level of energy

is only one of the two important factors contributing to obesity.                   Exercise

is equally important.

   APPENDIX I                                                     APPENDIX I
  In it.s respol1se to GAO, the Economic Research SPTvice discussed two basic

  methods of evaluating the NSLP's impact on the nationts economy.     Along with

  their response, ERS transmitted two reports which assessed the NSLP's
  economic impact using each of the methods.     The ERS summarized findings

  from the more recent of the two reports which was released in September 1976.
[The studies provided by ERS are summarized in ch. 5 of this report.]
  In addition, the Department is conducting a survey to determine the kinds

  and amounts of food used in the nation's schools.     This national probability

  sample of food use will provide information to further evaluate the impacts

  of the NSLP (and the School Breakfast Program) on the demand for agricultural

  products.    An outside contract for the conduct of this study has been under-

  way for some time.    Data collection has been completed and analysis is in

  progress.    At the time of GAO's review, the most recent report on the NSLp 1 s

  .effectiveness in meeting its agricultural objectives was based on data from

  the early 1960's.    However, as mentioned above, since that time a formal

  evaluation has been released and another is currently underway.
  [See GAO note on p. 62 of this report.]
                           [See GAO note 1, p. 153.J

  The GAO recommendation (page 90) to improve the reliability of participation

  projections and determine the relative importance of factors which affect

  participation, has been an on-going objective of the Food and Nutrition

  Service.    National projections of program participation and costs are

  updated by the Service on a quarterly basis for internal management and

 APPENDIX I                                                       APPENDIX I
budget purposes.     The GAO appears to be under the misconception that the

Department u.e. primarily prices paid by the paying child in its projections
of participation.     While that Is an important variable in assessing the

impacts of alternative legislative proposals. trends in past performance,

enrollment, free and reduced price eligibility levels and several other

nonprice factors are taken Into consideration In the development of each

national projection.     New legislative developments have often been   do~i­

oating factors.     A special study, "The USDA study on High School Partici-
pation in Child Nutrition Programs", cited on page 76 of the GAO report.

explicitly pointed to a large number of nonprice factors affecting parti-

cipation in high schools.     That study was completed in 1975.    FNS agrees

that there is a need to prioritize these factors and determine the extent

to which they individually and collectively influence participation.

The ERS response discussed several reasons why per capita income is a

poor series to adopt as a proxy for household income in evaluating parti-

cipation.   FNS is in accord with the concerns expressed in the ERS response

and has similar objections regarding aggregation of prices at the State

level to determine the percent of variation 1n participation explained

by price.   State agencies do not establish lunch prices for schools within

their jurisdiction; price setting is a local function.       States report prices

which are average prices within States and include individual schools      an~

districts with widely varying prices.       Thus. the State is not the appro-

priate sampling unit to determine the percent of participation variation

attributable to price.      [See GAO note 2, p. 153.)

FNS also joins with ERS 1n taking exception to the statement on page 81

 APPENDIX I                                                                   APPENDIX I
that," • • • price-participation relationships reported in the USDA's

study would be technically correct for NSLP lunches priced in the 20 to
35 cents range."   In USDA I   5    comprehensive study, data from t.he Pittsburgh
8tudy were weighted heavily 1n projecting participation rates for a broad

range of program options.          The Pittsburgh study included observations of

participation behavior at two prices (20t and average of 46.67 cents).

Since the Pittsburgh study, FNS conducted an informal study In Fairfax

County, Virginia, with price observations up to            6oi,   which lends further

support to the projections In the Comprehensive Study.                  Tilt.. .::., we object

to GAO's contention that USDA's analysis was lIflawed" because prices

in fiscal year 1974 were above the 20-35 cent' range.
                   [See GAO note 3, p. 153.J
In its response, ERS generally covered the food procurement aspects of

program costs, and indicated that increasing labor costs may force sub-

stitution of foods which require less labor and are thus more expensive.

We agree that disproportionate increases in food purchases need not re-

fleet ineffective management.

The GAO report in assessing the regional cost variations in operating

the National School Lunch Program as reported in the Department's com-

prehensive study of the child nutrition programs, indicated that dif-

ferences in accounting methods among the States prevented it from

determining any meaningful association between costs and program operating

efficiencies (page 102).       The Department recognizes that the cost data

available for that analysis were not sufficient for fully answering the

questions posed. partly because of accounting problems.                  The Food and

Nutrition Service has developed        al"'r:.ountin~ <~·n·structions   which when fully
.APPENDIX I                                                   APPENDIX I
implemented will ensure more standardization in accounting practices

throughout the NSLP.     The implementation of these instructions has been

a major FNS objective for the past several years.     With implementation

of cost-based accounting, FNS will be 1n a position to better assess

regional cost variations.

The GAO recommendation (page 109) to examine approaches and implement

procedures for improving the food procurement economies of small and

medium sized school systems is already underway.     An outside contract

with A.T. Kearney and Company has been underway since last summer to

accomplish this objective.     A report is expected before the end of

fiscal year 1977.


As mentioned earlier, this section of GAOls report hinges on the pre-

ceding sections.     Thus, any suggestions previously covered should also

be considered in terms of their impact on PART V (as well as the Summary).

In its assessment of evaluation of the program, the GAO did not appear to

be aware of the study of the effectiveness of the program conducted for

the State of Washington by Washington State University.     That study

assessed the importance of the program both for low income children

receiving free or reduced price lunches and higher income children paying

for their lunches.     The impact of the Program on the overall diet of

recipients was measured, and the nutritional status of participants

versus nonparticipants was assessed using biochemical measurements.

Food intake data were obtained both from the children at school and

from the parents of the children through hone visits.     Statistical

         APPENDIX I                                                  APPENDIX I
           assessment was made of the net additional amounts of food obtained

           through the Program upon the total food availability to the household.

           The report of this study is available on loan from the Information

           and EducatiOnal Materials Center, National Agricultural Library,
          Beltsville, Maryland.

    GAO notes:   1. Comments have been deleted because of changes
                    to the final report.

                 2. The final report has been changed to reflect ERS'
                    comm~ntf; regarding the use of "per capita income"
                    as a pr::>xy for household income. (See p. 'J'.)
                    FHS' objection to the aggregation of prices at the
                    State level is a different matter, and was not ad-
                    dressed by ERS.
                      All cross-sectional regressions implicitly have an
                      identity problem, that is whether the stuaents in-
                      cluded in each school/State of the cross-sectional
                      survey can be treated as being part of the demand
                      curve. Our work, as presented in this report, is
                      based on the averag' price charged (on a per lunch
                      basis) in each State. The findings are consistent
                      with, and supported by, USDA's own findings in cross-
                      sectional surveys of individual States and in a before
                      and-after study on the effects of a price increase in
                      Pittsburgh public schools.   (See footnotes 2, I, 2,
                      and 3 on pages 92, 93, and 98, respectively.)

                 3.   FNS reported in its CO~Erehensive St~z-£f_!~~Ch~lo
                      Nutrition Programs that a 10 percent Increase in priCE
                      would cause-a·3=-ro 6-percent decline in paying studer
                      participation. Our report states that such a relatior
                      stlip is technically corr~ct only for NSLP lunches prie
                      in the 20 to 35 cents range. E~S' comments reinforce
                      our conclusion by noting that 11* * * for every I per-
                      cent they raise the lunch price above 35 cents they
                      will average a loss in student participation of l.dd
                      percent * * *.11 In other words, at a lunch price of
                      35 cents, a IO-percent increase in price woula cause
                      about an lS.d-percent decline in paying stuuent

i                                           153
APPENDIX II                                                           APPENDIX II

                        ECONOMIC "ESE.ltCH . . .VICE
                            W"8HINGTOH. D.C:.   uno

                                                 March   I~.   1977

 Mr. Henry Eschwege
 Director, Community and
   Economic Development Division
 United States General Accounting Office
 Washington, D,C. 205~8

 Dear Mr. Eschwege;

 We have reviewed the draft GAO report on the Impact and Effectiveness
 of the School Lunch Program as requested in your letter of February 15.
 Our comments are confined to Parts I I-IV since nutritional science is
 outside our charge.

 Our detailed comments are enclosed.


   • I
 Acting Deputy Administrator


APPENDIX II                                                    APPENDI:< II
                         EcoNoMIC REtIoEA"CH SERVICE
                             WASHINGTON, D.C.   20210

                          COMMENTS ON GAO REPORT

     This report presents a synthesis of National School Lunch Program

evaluation studies.   Although the review of literature seems reasonably

complete, nowhere does it give the reader the idea that almost all of

the research reported either was directly conducted by the USDA, or funded

by it with USDA professional staff responsible for research oversight.

     The National School Lunch Program's economic impact can be assessed

at various levels within the economy.       For instance, identification of

the net increase in business receipts received by specified sectors

(e.g., agriculture; meat and poultry manufacturing; wholesale trade;

retai 1 trade; etc.), as a result of USDA's (a) cash transfers to the

States and (b) its purchase and distribution of commodities to schools

is one type of comparison.   The determination of the tonnage and/or dollar

value of increased sales of a particular commodity (e.g., carrots, lettuce~

mi lk) is another.

     The GAOls contention that it llcould not locate any formal attempt

to evaluate the NSLp1s impact on the Nation's economy ,II (p. 52) is not

surprising as published reports have been few in number.      The most recent

was released in September 1976.
                       [See GAO note 1, p. 162.)

APPENDIX II                                                        APPENDIX II

       Both this report and an earlier one may be of interest and are enclosed.

The September study presents the sector level comparison while the earlier

study illustrates the commodity-by-commodity approach.         The commodity

approach report is dated, but it illustrates the research prob-lems and

the kind of findings such an approach yields.

       The sector impact approach reports data for as recently as fiscal

year 1974.    I t compares operating results along with     simulatiol~s   of what

three alternatives would have yielded if they had been substituted.              In

1974, the cash transfer of about $1.1 bilJion resulted in a net increase

in business receipts of $573.2 million and in GNP of $397.5 million.

With respect to business receipts, some sectors gained while others would

have gu ned more if there had been no program.      Thus, agriculture, food

manufacturing, and the wholesale trade sectors gained a total of           $9~2.6

millionj whereas other sectors such as the retai 1 trade sector would have

gained $106. I million more without the program.        Schools buy primarily

from wholesalers and food manufacturers.

       Commodi ty distribution yielded analogous results.      In fiscal year

1974 when the USDA purchased $319.2 million of food products which were

distributed to the schools. the business receipts for        th~   Nation rose

by a   ~   $409.2 mi Ilion and GNP by $50.2 mi Ilion.    Agriculture, food

manufacturing, and the wholesale trade sectors gaineu        $556.~   million    In

business receipts whiJe retaiJ trade would have gained        $~1.5   miJJion more

in business receipts without it.     (For complete details, see the enclosed.

copy of AER-350.)

       The results of each of these studies supports the GAO's belief that

liThe NSlP has supported domestic demand for agricultural products l l (p. 64)--

APPENDIX II                                                         APPENDIX II

and, it should be added. the demand for the services of food manufacturing

and trade sectors responsible for moving food from farms to school children.

However. there were economic opportunity costs.        Other sectors would have

gained more business receipts without such an increase in the final demand

for food.     Even so. the net gain for the economy was greater with the

NSLP than it would have been without it.

     FNS administrative reports and studies long have documented that

agency1s concern for increasing and for accurately forecasting the numbers

of program participants.     Published and unpublished reports have identified

factors which have been viewed as being associated with participation

and as of being of use for forecasting numbers of participants.             These

inc Iude those de ta i led in the GAO report.   In cons j der J ng rese.1i--ch d i rec ted

toward identifying the relative importance of factors which explain

participation, priority was given to     f~ctors   over which the school had

substantial degrees of control, e.g., prices and costs of lunches, and

also to those which could be meaningfully quantified.          This did not include

the income variable.

     Specifically, per capita income for each State is a poor statistical

series to adopt as s proxy for household income because it does not yield

measurable associations which are statistically llclean        CUt-"    Average per

capita State income includes the incomes of households without children,

with children who pay fully, on free- or reduced-price lunch participation,

composed of old maids or bachelors, and households composed of retired

pecple.     The per capita State incomes for Florida and Arizona are affected

substantially by their numbers of households composed of retirees.

APPENDIX II                                                        APPENDIX II

      States have such great variations of income within them that the per

capita income statistic is unsuited for between-State comparisons.           If

each Statels within-State income was homogeneous so that the variance

within each State was small while differences between States were sub-

stantial   I   the income specification adopted by GAO would hold up--but

the present data suggest that this is not the case.

      In addition, States which have high proportions of free- and reduced-

price lunches would have State per capita income averages not representative

of households from which full-pay students        com~.   Thus, even though the

model's coefficients have high statistical significance, within an economic

context they are very suspect.

      The need for a cleanly specified income variable is emphasized when

the added factors cited from the literature by GAO are considered.

Elementary students usually are relatively captive lunch patrons.           They

must eat on school premises and typically the only alternative to the

Type A lunch provided is a bag lunch.         Secondary students   have varying

degrees of freedom ranging from use of only on-campus facilities or. with

both Type A and non Type A lunch, choice of such faci lities plus permission

to leave the school premises and buy from off-campus sites.         The true

income impact can be identified only if " a ll other variables are held

constant. 11     Use of an average per capi ta State income in a regression such

as GAO conducted could attribute influence of these other factors to

income.        In essence, the GAO report places too much emphasis upon a

crudely specified regression equation. [See GAO note 2, p. 162.1

      In discussing the price-participation relationships, GAO presents

data which do not have clear source identification (footnote I, p. 81,

     APPENDIX II                                                   APPENDIX II

     is missing).   GAO contends that, lithe USDA study is technically accurate

     only for the 20-35 cent range.   I   Actually, there were several studies

     conducted independently (Wa-   "ngton State, North Carolina, and two ERS

     studies) which yielded similar results within this range.      Of these.

     al I but one were cross-sectional in character and analogous to the one
     GAO has developed and reported in its figure 7. I, p. 82.     For these cross-

     sectional studies, the GAO statement is essentially correct.      It is only

     partially correct for the Pittsburg      study which was more analogous to

     an experimentally controlled design than it was to cross-section.         The

     Pittsburg   system which had maintained a price of 20 cents for an extended

     time, had a single price increase to an average price of 46.67 cents.           The

     Pittsburg   own-price elasticity at 20 cents however, was almost identical

     to the cross-sectional studies at 20 cents (-.47 versus -.50).      Also.

     when the Pittsburg   and cross-sectional results were placed on a common

     statistical basis for comparison. the results were similar.      Because of

     the similarity of results in their corresponding price range, the Pittsburg

     results for the range of 35 cents to 46.67 cents become particularly

     relevant as they go beyond the other studies.      At 31.2 cents the Pittsburg

     own elasticity was -1.0.   Above that price the elasticity rose rapidly.

     From 35-46.67 cents the average was -1.88.     At 46.67 cents its own-point

     elasticity was -2.95.   [See GAO note 3, p. 102.J

          No regression equation wi 11 necessari Iy yield a precise estimate

     beyond the range of data from which it was computed.      However, when

     results from experimentally controlled situations and cross-section data

     taken independently agree within the same range and observations of the

APPENDIX II                                                            APPENDIX II
former go beyond those of the cross section, the results from the former

can be used to make administratively meaningful decisions.               If a   ~chool

Board recognizes that for every          percent they raise the lunch price

above 35 cents they \'Iill average a loss in student participation of 1.88

percent, actually losing as much as 2.95 percent if they go as high as

46.67 cents, the message would appear to be loud and clear.

      In summary. ERS joins the GAO in its recognition of the importance

of a well specified income variable for improving forecast models.

Unfortunately, ERS has yet to find currently available income series

which are adequately specified for this purpose of forecasting school

lunch participation.       Consequently, until such an adequately specified

income variable becomes avai lable, it is analytically questionable to

claim that a State per capi ta income series is a more lIimportant variable"

than others, such as price.        lSee GAO note 2, p.162.l

      GAOls contention is based upon differences in coefficients which

they ob ta i ned.   Hmvever I un til these coeff i c i en ts have been s tanda rd i zed

even a statistical comparison of relative importance is not possible.

In this instance, even if standardized coefficients were different in

terms of    stati~tical   significance, analytically they would be meaningless

because of the crude identification of the income variable.

      Overal I program costs have }ust about kept pace with inflation.

Hence differentials in rates of increase in purchased foods compared with

other cost components may reflec, a deliberate effort to minimize labor

costs.     Wage rates for cafeteria workers have risen and continue to rise

   APPENDIX II                                                      APPENDIX II

   at a relatively rapid    pac~.   Other purchased inputs in part may be substituted

   for labor.     The increased purchases of lip reformed beef patties l' in lieu of

   bulk ground beef and c,f individual portion pizzas are illustrative.       Thus,

   disproportionate increases in food purchase expenditures need not reflect

    ineffective management.
[The report has been ~evised to acknowledge ERS I              comment. See p. 121.
        While smaller school systems typically could realize savings if they

   participated in a State- or county-wide buying arrangement, an additional

   caveat probably shculd have been added in the original ERS report.        When a

   State is composed of both large and small systems, the development of a

   centralized purchasing system should be used on a voluntary basis.        A

   compulsory participation arrangement, while helping smaller systems, could

   penalize the larger ones as they could n0t capture the savings from

   advantageous local bids whenever they become available.
[The report has been revised to acknowledge GRS' comment. See p. 124.)
        Many readers \'1i II not study this report beyond its summary.     Conse-

   quently, it is important that the summary be written clearly and accurately.

   Because of the information presented in the preceding comments and the two

   enclosures, GAO may wish to revise some of its statements, particularly on

   pages 120-124.     For example, the introductory sentence under the heading,

   Agricultural considerations, p. 120, needs revision to reflect the added

   information.     The statements at the bottom of p. 122 and the top of p. 12')

   are questionable.     IICurrent forecasts based on price participation studies

   are not reliable" is too strong a contention.       The use of elasticities of

   demand can yield workable estimates of what the impact of specified price

   changes for lunches will have in terms of changes in numbers of participants.
                         [See GAO note 4, p. 162.J

APPENDIX II                                                APPCNDIX I I

  GAO notes:

  1.   The final report has been revised to include a summary of
       the studies provided by CRS.  See "Impact on the agri-
       cultural economy,!1 ch. 5.
  2.   The final report has been revisea to better qualify the
       limitations of using "per capita income" as a 'regression
       variable.   See pp.   ~~    to 102 of this report.

  3.   CRS' comments pertain to the lack of clear source identi-
       fication in an earlier draft of this report.  The final
       report has been revised to clarify the point.  (See p. ~6.)
       In essence GAOls contention is         ~hat   USDAls Comprehensive
       ~~of the Child Nutr ition Programs misinformed the
       Congress by stating that "paYIng students respona by re-
       ducing participation 3 to 6 percent for every 10 percent
       increase in prices charged.      11
                                             Such a relationship, in
       GAO's opinion, would be true only for lunches priced in
       the 20 to 35 cents range.  CRS' comments support the
       contention by noting that ". • • for every 1 percent they
       raise the lunch price above 35 cents they will average a
       loss in student participation of 1.S8 percent • " . In
       other words, at a lunch price of 35 cents, a lO-percent
       increase in price would cause about an lS.S-percent de-
       cline in paying student participation.

  4.   The statement that "current forecasts based on price
       participation studies are not reliable" has been revised
       in the final report to read that "price-participation
       relationships provide an extremely weak forecasting tool."
       The interpretation of what constitutes a "wor kable esti-
       mate" in projecting participation levels is, of course,
       dependent upon the degree of precision required.            The
       limitations of current price-participation forecasts are
       shown on pp.  ~6 to 9i  of this report.

APPENDIX II I                                             APPENDIX I II

                          OFFICE OF THE SECRETARY
                              WASHINGTON. D.C.   20Z01

                                                         i' Dr. 1
                                                         .,. H, ,

  Mr. Gregory J. Ahart
  Director, Human Resources
  United States General
    Accounting Office
  Washington, D.C. 20548
  Dear Mr. Ahart:
  The Secretary asked that I respond to your request for our comments
  on Your draft report entitled, "Impact and Effectiveness of School
  Lunch Program: A SYnthesis of Evaluation Studies." The enclosed
  comments represent the tentative position of the Department and
  are subject to reevaluation when the final version of this report
  is received.
  We appreciate the opportunity to comment on this draft report before
  its publication.
                                    Sincerely yours,

                                    Thomas D. Morris
                                    Inspector General

APPENDIX III                                             APPENDIX III


That the Secretary of Agriculture:
     -- should require a formal, systematic evaluation of the NSLP's
     performance in meeting legislative objectives. The evaluation
     should be coordinated to utilize the expertise .and resources of the
     Department of Health, Education, and Welfare (HEW), in all matters
     pertaining to the health and nutritional status of school children;
     and to provide effective and timely reporting of information needed
     for Congressional oversight.
     -- with assistance from HEW, determine t~(, nutritional standards"
     needed for tJe NSLP to best safeguard sClloolchi1d health; and, if
     found desirable, revise the program's m~a1 regulations to reflect
     nutritional requirements that will prov'de menu planners with
     planning f1~xibi1ity, improve the program's cost-effectiveness;
     encourage higher levels of student participation; and reduce plate
While we concur with the GAO recommendations, we have concerns, listed
below, about the accuracy of some statements ir, the report and the
validity of some of its reasoning. We will, however, assist the Depart-
ment of Agriculture in carrying out the intent of GAO's recommendations.
FolloWing are some areas of the report which should be modified:
     1.   There are a variety of statements asserting specific relation-
          ships between nutrition and disease which, while intriguing
          hypotheses and which are now being studied, cannot be con-
          sidered as authoritative fact. For example:
                        [See GAO note 1, p. lb5.]
          a.   References to deficiencies in fiber in the diet causing
               increases in the risk of bowel cancer. (Page 10);

                        [See GAO note 2, p. 165.J

APPEN"D I X II I                                             APPENDIX I II

         2.    The report criticizes the regular Type A school lunch because
               it contributes to obesity in some children and has not been
               able to improve iron nutriture. Since the report elsewhere
               concluded that present studies of NSlP are inadequate to
               evaluate nutritional impact, it is premature to implicate the
               program on these grounds. This is particularly true since, as
               the report points out elsewhere, the school lunch provides
               only one-sixth of the meals of the participants and can,
               therefore, only be a supplement to home meals.

                            [See   GAO   note 3 below.]

    GilD      notes:

    1.        The report has been revised to agree with the comment.
              See "technical note 'l on p. 9.
    2.        Comments have been deleted because of changes to final
    3.        See   GAO   remarks on p. 45.
    4.        Page references in the agency comments in appenolxes
              I, II, and III refer to the draft report and/or summary,
              and may not correspond to the final report and summary.