@7eqeccl DOCUMENT FESUME 02557 - [A175-27851 The Work Measuremerit System of the Department Urban Development Has cf Housing and Potential but Needs Increase Its Reliability. Further Work to FPCD-77--53; B-183124. released June 20, 1977. June 15, 1977. 29 pp. + 4 appendices (8 pp.). Report to Sen. William Proxmire, Chairman, Appropriations: HUD-Independent Senate Committee on B. Staats, Comptroller Agencies Subcommittee; General. by Elmer Issue Area: Personnol Management and Compensation National Productivity (300); (2900). Contact: Federal Personnel Budget Function: General and Compensation Div. Government: Central Management (805). Personnel Organization Concerned: Department of Housing and Urban Development. Congressional Relevance: Senate Committee on HUD-Independent Agencies Appropriation: Subcommittee. The Department of Housing began developing a system and Urban Development (HUD) 1972, in order to establishfor work measurement standards in a basis for manpower for budget submissions requirements and allocat.on of perscnnel. Findings/Conclusions: HUD's standards were not justified,original claims of extensive standards were used as revised to develop estimates statements showed that staff requirements. for nly about 42 of The reliability of standards of weaknesses such as: varied because (1) achieving work efficiency, lack of studies on methods for questionnaire/interview (2) variation in data procedures, (3) insufficient produced by the of tasks, (4) use of definition subjective judgments, documentation, and (6) (5) lack of lack of procedures to standards. Discrepancies review and update were ncted in workload some appearing excessive forecasts with and some being understated compared with prior when years' accomplishments. process seemed to inhibit The budgeting reliable staffing estimates use of contract personnel. and led to Recommendations: HUD practices for developing should improve performing methods studies work measurement standards by (1) on task efficiency, data collection and (2) improving analysis, (3) defining detail, (4) assuring tasks independence of individuals in greater standards, (5) improving setting documentation, (6) formalizing process for reviewing a and updating standards, reevaluating staff resources and (7) The Subcommittee on to develop and maintain HUD-Independent Agencies the system. HUD to develop a more should encourage objective and reliable system, and require ork measurement that the budget submission comprehensive plan and statement on the progressinclude a *lstem's development. made in the (HTW) y\ '~' ;IESTRICTIItO _ Aeco0ing cOffe exep plot to e t. tp eneral oo ,fspec peic approval by the olfie of 4 i4ressionoa e o,aov V.s, REPORT OF THE 34. COMPTROLLER GENERAL t,.o N'.~ OF THE UNITED STATES The Work Measurement System Of The Department Of Housing And Urban Development Has Potential But Needs Further Work To Increase Its Reliability The budget staffing requests of the Depart- ment of Housing and Urban Development should be based on techniques Ivhich are re- liable and useful in the budget process. concept of work measurement offers The the potential to do this. The Department has made progress in devel- oping work measurement standards and some are more reliable and useful than others. How- ever, tne standards should not yet be accepted at face value for estimating the personnel required to efficiently perform the work of the Department. The Department shoula be encouraged to continue to develop its work measurement system and to increase the reliability of its work standards, where practical. First, the De- partment needs to improve its ratices for developing standards. FPCD-77-53 JUNE 15, 1977 Face0 I:I COMPTROLLER GENERAL OF THE UNITED STAT WASHINGTON, D.C. 08a B-183124 The Honorable William Proxmire Chairman, Subcommittee on HUD-Independent Agencies Appropriatinns Committee United States Senate Dear Mr. Chairman: This report describes the extent to which of Housing and Urban Development has developed the Department objective, reliable work measurement standards for personnel budgeting at the national level. Our review was made pursuant to your request of July 29, 1976, and subsequent discussions with your office. quested by your office, we did not obtain formal As re- comments from the Department. However, the results of our cussed with various agency personnel, review were dis- including the Secretary of Housing and Urban Development, and their comments have been considered in preparing the report. The report contains recommendations to the Housing and Uroan Development which are set Secretary of As you know, section 236 of the Legislative forth on page 28. Act of 1970 eouires the head of a Federal Reorganization agency to submit a written statement on actions taken on our recommendations to the House Committee on Government Operations Committee on Governmental Affairs not later and the Senate after the date of the report and to the Housethan 60 days mittees on Appropriations with the agency's and Senate Com- first request fcr appropriations made more than 60 days the report. we will be in touch with your after the date of office in the near future to arrange for release of the report auirements of section 236 can be set in so that the re- motion. Si y you Comptroller General of the United States REPORT OF THE THE WORK MEASUREMENT SYSTEM COqIPTROLLER GENERAL OF THE DEPARTMENT OF HOUSING Of THE UNITED STATES AND URBAN DEVELOPMENT HAS POTENTIAL BUT NEEDS FURTHER WORK TO INCREASE ITS RELI- ABILITY DIGEST Developing reliable work measurement standards takes time and is not a simple effort. Various tschniques are available to develop standards-- engineered standards, technical estimates, his- torical standards--to mention the more common. The method or combination of methods used will depend to a great extent on the activity to be measured. Consideration must be given to the long-term benefits and costs of detailed mea- surement methods versus less precise methods. Tne Department began developing its work mea- surement system in 1972. Progress has been made and work continues on the system. In testimony before the Congress and in recent budget documents, the Department referred to an extensive system of work measurement stand- ards directly relating staffing needs to work- loads. In March 1976 a Department official testified that 63 percent of its staffing was covered by "detailed work measurement standards." The Department's work measurement standards are not as extensive as purported. During testimony on the fiscal year 1978 budget, a HUD official clarified that standards are used to develop estimates for only about 42 percent of the Department's staff require- ments. Most central and regional office m- ployees, as well as support staff in area and insuring offices, are not covered by work measurement standards. (See p. 15.) While some standards are more reliable than others, they should not yet be accepted at FPCD-77-53 TAl Sht. Upon removal. the report cover date should be noted heron. face value for use in estimating the person- nel required to efficiently perform the work of the Department. There are weaknesses in the Department's practices for developing them: -- Methods studies on how to work efficiently and to eliminate nonessential and duplicate operations were not a part of standards development; hence, standards incorporated whatever inefficiencies existed in the way work was done. (See p. 7.) -- The questionnaire/interview procedures re- sulted in data that varied widely, making it virtually impossible to develop vaid stand- ards from such data. (See p. 8.) -- Tasks for some standards were no: defined in sufficient detail. This resulted in large task times and greater margins of error in the data. (See p. 9.) -- Data was discarded and personal judgments used to develop some standards, making them subjective estimates rather than true work measurement standards,. (See p. 10.) -- Program managers were involved in setting some of theiL own standards and may have influenced them on the high side. (See p. 11.) --Documentation was not available to support some standards, particularly the reasons why data was adjusted. (See p. 13.) -- Sample offices from which some standards were developed are probably not represen- tative of the Department as a whole. (See a. 13.) -- No formalized procedure exists to assure that standards are reviewed and updated when organizational or procedural changes are made to improve efficiency. Withoui this proce- dure, standards may soon be outdated. (See p. 14.) ii Because the standards cannot yet be accepted at face value, however, does not necessarily mean that they overstate personnel requirements. While it is likely that some do, it is also likely that some understate them. The Department may not have sufficient staff to adequately develop and maintain its work measurement system. (See p. 14.) Accurate workload forecasts are a key factor in aiving at personnel requirements. GAO ioted several discrepancies: workload fore- casts that appeaced excessive when compared to prior year accomplishments, forecasts that appeared understated when compared to the past or to procedure manual requirements, and workload which was not forecast at all. (See p. 24.) The budgeting process, which is the exclusive responsibility of the Department's central office, seems to inhibit the development of accurate workload forecasts and reliable staff- ing estimates. The central office decides how much of the total workload should go to each regional office, although changes can and do occur after the overall budget is approved. (See p. 24.) To compensate for perceived staff shortages in subsidized and other housing programs, the Department has been relying heavily on fee personnel (contract personnel hired to perform services such as appraisals) in the insured single family program. Fee personnel are paid from a separate mutual insurance fund. (See p. 23.) The extent to which fee personnel are used to supplement Departmental in-house staff was re- flected inaccurately in the fiscal year 1977 budget, and may not be accurately reflected in the fiscal year 1978 budget. After the fiscal year 1977 budget was approved, the Department increased the percent of appraisals estimated to be done on a fee basis from 15 to 75 percent. This frees for other work 200 staff-years ini- tially budgeted for in-house appraisals. In Tear Shee iii spite of this 75-percent estimated fee use in fiscal year 1977, the fiscal year 1978 budget reflects only a 25-percent fee appraiser e. Each -percent increase in fee use over 25 per- cent will free the equivalent of 4.69 F'aff- years to do other work. (See p. 24.) Accurate information relating time expended (in- put) and number of detailed tasks done (output) is needed to improve the Department's work measurement system. Such information, not now routinely generated, will permit management to better evaluate performance, assess organiza- t onal efficiency, and develop more accurate -ecasts of workload and staffing needs. HUD has established a time reporting system, but its accuracy is questionable because of re- porting deficiencies at the employee level. Moreover, the system is not integrated with the payroll system or the workload reporting sys- tem; the latter is in its early stages of development. (See p. 19.) GAO recommends that the Secretary of Housing and Urban Development improve the Department's practices for developing work measurement standards with the goal of achieving more ob- jective and reliable bases for estimating its personnel requirements. GAO recommends ways to improve HUD'- practices. (See p. 26.) In addition, GAO recommends that the Subcom- mittee: -- Encourage the Department to develop its work measurement system toward the goal of more objectively and reliably determining personnel requirements. -- Direct the Secretary to present with the fis- cal year 1979 budget submission (1) a compre- hensive plan for proceeding with the Depart- ment's work measurement system development and (2) a statement on the progress it has made in dealing with the issues discussed in this report. iv As requested by the Subcommittee's office, written comments were not obtained from the Department. Hwever, GAO discussed the report with Department officials and considered their views. Tear Sheet Contents Page DIGEST i CHAPTER 1 INTRODUCTION 1 HUD organization ard staffing 1 Work measurement standards in HUD 2 Techniques used in establishing standards 2 Review objectives and scope 4 2 RELiAILITY OF HUD'S WORK MEASUREMENT STANDARDS AND COVERAGE 6 Reliability of standards 6 Methods studies not performed 7 Highly subjective data collec- tion procedures 8 inadequate task definition 9 Adjustments to data 10 Objectivity of developed standards 11 Lack of documentation 13 Offices selected for data collec- tion may not represent HUD as a whole 13 No mechanism for updating stand- ards as efficiency improves 14 Staff resources for work measure- ment system development 14 Extent of coverage 15 3 OTHER ISSUES RELATED T HUD'S WORK MEASUREMENt SYSTEM 19 Integration of work measurement with time, workload, and cost reporting systems 19 Compatibility of data 19 Accuracy of time reporting 20 Accuracy of task reporting 21 HUD awareness of problem 22 Use of fee personnel 23 Importance of workload forecasting 24 4 CONCLUSIONS AND RECOMMENDATIONS 27 Conclusions 27 Recommendations to HUD 28 Recommendations to the Subcommittee 29 Agency comments 29 Page APPENDIX I Letter from Chairman, HUD-Independent Agencies, Senate Appropriations Com- mittee 30 II Analysis of HUD fiscal year 1978 budget request for personnel 31 III Summary of staff-years nd full-time em- ployment in permanent positions in HUD, fiscal years 1977-1978 32 IV Practical experiences with work. measurement systems 35 ABBREVIATIONS GAO 7eneral Accounting Office HUD Department of Housing and Urban Development OMB Office of Management and Budget CHAPTER 1 INTRODUCTION During thQ fiscal year 1977 budget hearings, the Secre- tary of the Department of Housing and Urban Development (HUD) testified that HUD had an extensive system of work measure- ment standards from which it could reliably estimate personnel requirements. The Chairman of the Subcommittee on HUD-Independent Agencies, Senate Committee on Appropriations, requested that we review HUD's progress in developing work measurement stand- 3rds and evaluate the extent to which existing standards are valid for determining personnel requirements. (See app. I.) HUD ORGANIZATION AND STAFFING HUD was created in 1965 to consolidate arious Federal housing activities into a Cabinet-level department. It is the principal Federal agency responsible for programs con- cerned with housing needs and improving and developing the Nation's communities. It has responsibility for the admin- istration of Federal assistance programs involving housing, community development, and related activities. HUD consists of three organizational levels: the cen- tral office, 10 regional offices, and 76 area and insuring offices. Determination of staffing requirements and budget formulation are functions of the central office, which also allocates staff and workload to the regional office level. Each regional office, in turn, allocates staff and workload to its own programs and to the area and insuring offices within its jurisdiction. Staff distribution between the cen- tral office and field offices is as follows: Staff ceiling Approved Budget request Fiscal year 1977 Fiscal year 1978 Number Pe:cent Number Percent Central office 3,867 24.8 3,959 23.6 Field offices 11,703 75.2 12,S31 76.4 Total 15,570 100.0 16,790 100.0 Note: A further breakdown of personnel by division is shown in app. III. I HUD's budget for salaries and expenses $454 million and $489 i estimated at million for fiscal years 1977 and 1978, respectively. WORK MEASUREMENT STANDARDS IN HUD In 1972, HUD established a task force to determine the feasibility of developing work measurement field offices. From this effort, standards standards for the were developed and published covering major field activities. that these standards were unrefined, HUD recognized but believed they were usable for determining manpower requirements. In 1974, a Manpower Resources Division Staff Resources Division) was established. (later changed to charged with the responsibility for developingThis Division was and maintain- ing manpower management systems, including productivity analysis, cost analysis, work measurement, systems. and manpower allocation The work measurement system and work standards are in- tended to provide an objective, reliable ing personnel requirements for budget basis for determin- submissions, allocating personnel t operating elements of the agency, and measuring efficiency ad productivity. Office of Management and Budget requires that work measurement, unit (OMB) Circular A-11 indexes be used to the maximum extent costs, and productivity practical by Federal agencies in justifying personnel requirements workload. The Director, OMB, emphasized for measurable and work measurement should be extended, that productivity to all areas and the data integrated whenever practical, into a mnagement con- trol and evaluation system. TECHNIQUES USED IN ESTABLISHING STANDARDS Various engineering and nonengineering be used for developing standards. techniques can Engineered standards are based on ment of the time a task should take to analysis and measure- produce acceptable quality under proper working conditions. They are generally developed using formal analytical techniques study, work sampling, standard data and such as time systems. Such standards are most useful predetermined time activities where detailed planning and for high-volume control are desired, and are frequently used in an industrial setting. 2 Nonengineered standards are those developed without using engineering techniques, and tend to be less reliable than en- gineered standards. Historical and technical estimates the methods most commonly used to develop nonengineered are stand- ards. Historical estimates are based on data relating time expended to the work produced. A drawback to this tech- nique is that it assumes that what has happened in the good practice and that fture conditions will be the past is same. With such standards, it is difficult to accurately assign a reason to significant deviations. They can, however, quickly applied to provide extensive standards coverage. be Also, the standards so developed can be updated over time the use of work sampling or methods studies and adjusted by more appropriately reflect the time it should take to to do the work. One secific historical technique is based on obtaining data for work performed under controlled work 4 ng conditions. After receiving detailed instructions on how to use the vey forms and the definitions of the work units, employeessur- record how they have spent their time over a short period. Technical estimates are derived by breaking jobs work units or stages and having a technically qualifiedinto son estimate how long each of the stages should take. per- Es- timates can be developed by a panel of knowledgeable who estimate time requirements and through people discussion reach a consensus of opinion. The work unit estimates are summed to obtain the standard time. A disadvantage ofthen this technique is that it relies considerably on the jdgment the persons making the estimates, and as such may vary of from the actual time it takes to do the job. This makes greatly difficult to assign causes to deviations from the standards.it An advantage of technical estimates, however, is the low cost of using them to develop standards. It may also be the only technique available to develop standards for highly technical or irregular work such as research or technical projects. The techniques we ave described are not Basically, the method or combination of methodsallused inclusive. will depend to a great extent on the activity to be measured. standard setter must consider the long-ternm benefits The and costs usually associated with detailed measurement methods against the drawbacks and economies of less precise methods. 3 REVIEW OBJECTIVES AND SCOPE The specific objectives of our review were to ascertain the extent to wnich the fiscal year 1978 budget estimates for personnel were computed by applying work measurement standards to workload forecasts, and evaluate the reliabil- ity of existing work measurement standards for national budgeting purposes. Our review was performed at HUD's central office in Washington, D.C., region V in Chicago, and region IX in San Francisco. At the central office, we ascertained the extent to which the fiscal year 1978 budget request for personnel is based on work measurement standards. At the central office and in region IX, we evaluated the processes used to develop and apply work measurement standards and consulted with technical advisors on the standards' reli- ability in relation to budgeting at the national level. Our review focused on the work measurement standards in relation to HUD's four principal program divisions: Housing Production and Mortgage Credit, Housing Management, Commun- ity Planning and Develgpment, and Fair Housing and Equal Opportunity. 1/ Within the four divisions, our major focus was on the larger programs and new or expanded programs requiring added staffing in fiscal years 1977 and 1978. At the regional offices we compared time charge reports to workload output reports and ascertained the methods used to allocate both workload and staff to the area and insuring offices. We also visited the Chicago and San Francisco area offices and the Cincinnati insuring office to evaluate the integrity of HUD's time reporting system. To compare the use of work measurement by HUD and non- Federal organizations, we contacted several trade associations and private mortgage credit institutions and inquired about their use of work measurement standards. We also contacted OMB to ascertain what technical assistance is available to agencies planning work measurement systems. (See app. IV.) l/The two housing dvisions have since merged into the Hous- ing Programs Division. 4 Our review did not deal with the extensiveness liability of HUD's producti;ity measurement system or re- or its financial management system. 5 CHAPTER 2 RELIABILITY OF HUD'S WORK MEASUREMENT STANDARDS AND COVERAGE The Subcommittee, in reviewing HUD's fiscal year 1978 budget, should not accept HUD's work measurement standards at face value for estimating the personnel required to effi- ciently perform the work of the Department. Further, the standards are not as extensive as purported in budget docu- ments and in testimony before the Congress. *RELIABILITY OF STANDARDS HUD, in developing work measurement standards for field activities of its major program divisions, did not follow acceptable practices for developing reliable work measurement standards. Specifically, the defects were these: -- Methods studies on how to work efficiently and to eliminate nonessential and duplicate operations were not a part of standards development. As a result, the standards which were developed or up- dated incorporated whatever inefficiencies. existed in the way work was done. -- The questionnaire/interview procedures resulted in data that varied widely, making it virtually im- possible to develcp valid standards from such data. -- Tasks for some standards were not defined in sufficient detail, resulting in larger task times and greater margins of error in the data. -- Data was discarded and personal judgments were used to develop some standards, making them subjective esti- mates rather than true work measurement standards. -- Program managers were involved in setting some of their own standards and seem to have influenced them on the high side. -- rocumentation was not available to support some standards, particularly the reasons why data was adjusted. 6 -- Sample offices from which some standards were developed may not be representative HUD-wide. --No formalized procedure exists to assure that standards are reviewed and updated when organizational or proce- dural changes are made to improve efficiency. Without this system, standards may soon be outdated. As a result of these weaknesses, we do not believe that existing work measurement standards should be accepted at face value for estimating personnel requirements in the fiscal year 1978 budget. This does not, however, mean that personnel requirements are overstated. While it is likely that some are, it is also likely that some are understated. However, because the stand- ards were developed by means of defective practices, they cannot be regarded as reliable indicators of how long it takes to perform tasks efficiently. The effect of using inaccurate standards for budgeting can be great. To illustrate, the fiscal year 1977 forecast for conditional commitments under the insured single family proposed construction program is 140,000 units. If the existing standard of 5.72 hours for processing these appli- cations is off by only 1 hour (less than 20 percent) it means an overstaffing or understaffing of 78 positions. Methods studies not performed Work measurement standards should be based on the most efficient and economical ways for performing given tasks. If the standards are based on existing procedures they will incorporate whatever inefficiencies exist in those procedures and perpetuate them. Therefore, methods studies to identify the most efficient way to work and highlight nonessential and duplicate operations should, where practical, be done before standards development. Yet, because of the time involved in making methods studies, it may be more practical to establish standards using historical data or technical estimates without first making methods studies. However, in these cases, meth- ods studies should be subsequently made, as soon as possible, and the standards revised to reflect "should take" time, especially where the initially collected data has high varia- bility. In developing its standards, HUD did not analyze the way its work was being done to identify inefficiencies prior or subsequent to developing them. Rather, it based its standards by and large on the existing ways of performing 7 tasks. Whatever inefficiencies existed in the way work was being done were incorporated into the standards. A number of recent studies have pointed out inefficien- cies in the way HUD's work was being performed. It has beer noted that because of differences in the types and degree of authority delegated to field offices there had been duplication of effort at all three levels of HUD's organization. To deal with these problems, HUD began a series of management initiatives, one of which was a study started in 1975 known as Process Analysis. This consisted of several detailed on-site analyses of the processes field offices used to administer HUD's programs to the public. In essence, Process Analysis was an attempt to disc:over, through dis- cussions with field managers, whether there were better, simpler, more realistic, and more effective ways to perform HUD's work. All four major program areas in the field, as well as administrative support functions and operations, were covered. The results of these process analyses were not used in developing new or rvising existing work measurement stand- ards, but: hey should have been, and future results should be also. HUD said in March 1977 that an analysis of the way work is done will be incorporated in future standards develop- rment efforts. Highly subjective data collection procedures In collecting data to develop standards, particular care must be taken to assure valid data. Time studies in whicn work performance is directly observed and recorded are generally reliable. On the other hand, interview or questionnaire methods by which individuals are asked to estimate the time they needed to do certain tasks are less reliable since they depend to a large degree on memory and can be influenced by personal bias. Particular care must be exercised when-the questionnaire/interview approach is used. A major step in HUD's approach to developing its work measurement standards was to solicit the estimates of super- visors and of employees doing the work on how long it takes to complete various tasks. In some instances, employees were asked to recall time spent doing tasks over a 1-year period. In all four program areas the frequent result of this approach was widely varied responses. In our opinion, 6 it is virtually impossible to arrive at a valid standard from data that vary widely. Further, when wide variance appears in data, it is a generally accepted practice to further analyze it and find out why. There was no evidence that HUD analysts attempted to do this. As discussed later, in many cases data was simply discarded. Inadequate task definition We believe another cause of unreliable data and ques- tionable standards is that in some cases the tasks for which data was requested were not defined in sufficient detail so that accurate measurements could be taken. We reviewed the breakdown of tasks for four activities in the Housing Production and Mortyage Credit area, as shown in the following table: Activities Single -ily Proposed Existing Proposed Multi- construc- construc- sub- family Category tion tion division insured Number of detailed tasks 25 19 9 35 Percent of tasks 1 hour or more 36 42 100 100 Average task time (hours per task) 1.0 1.2 12.6 24.7 The average task time and percent of tasks greater than 1 hour are highest for the proposed subdivision and multi- family standards because these standards were broken down into broad processing phases, not detailed tasks and sub- tasks. Conceiving tasks as long, general phases rather than short, detailed steps affects the ability of individuals to accurately estimate how much time a given task takes. This is illustrated by the following table, which was based on two series of estimates made by the HUD San Francisco region in revising single and multifamily standards. 9 Activity Single Family Proposed Existing Proposed Muiti- construc- construc- sub- family Category tion tion division insured Average task time (hours per task) 1.0 1.2 12.6 24.7 Average percent devia- tion in estimates between 1976 task time and 1975 task time 232 i4., 96.3 61.3 This table suggests a significant in ability to ac- curately estimate ime for given t:asks when the average task time is large. As to why tasks were not defined in ore detail, the HUD study group said that task sheets and descriptions used to develop the Housing Production and urtgage Credit stand- ards were based on HUD handbook procedures, supplemented by program knowledge. These officials described the handbook procedures for the four selected standards as (1) detailed for both single family existing and proposed construction, (2) not adequately detailed for single family proposed sub- division, and (3) very vague for multifamily insured projects. Thus, standards development was begun without clear defini- tions for all activities to be measured. Similar conditions existed regarding the standards developed in the Housing Management program area. Although the work measured by the Housing Management standards in- cludes numerous processing steps, the standards do not re- flect it. For instance, the task of inspecting property, listing property for sale, and executing a contract for sale are all covered by the single family property disposition standard. Adjustments to data Faced with the wide variability in sample data, HUD analysts frequently dscar. ~d data as being not representa- tive. For example: --In Housing Management sample data was eliminated to achieve the high statistical correlation HUD analysts believed as necessary to establish a standard. Data collected from 11 of 19 offices surveyed was eliminated in developing the standard for the management of insured multifamily projects. 10 -- In developing the standards for field activities of the Community Planning and Development Division, extreme data points were also discarded, usually one high one for each low one. The number discarded varied from standard to standard. In one instance, 36 employees were interviewed concerning the time spent in program management activities, and all but 13 sample results were discarded in developing the standards. Selective use of data is acceptable when a distorting factor is known to be present, and then the data may no longer be representative of the universe drawn. HUD personnel did not analyze the from which it was reasons for the wide variations in times reported by interviewees. Such analyses should have been made. Objectivity of developed standards Individuals selected to develop work measurement standards should be, to the extent practical, independent of the organizational element to which the standards ap- ply. If such individuals have proprietar,. interests in the programs for which the standards were developed, they might influence the results. They may be relied on for input where technical estimates are used, but the persons responsible for setting the standards should review these estimates and resolve significant differences. We examined the groups responsible for developing selected standards in the Housing Production and Mortgage Credit area. We noted that the individuals responsible for setting the standards were predominantly Housing Pro- duction and Mortgage Credit officials. In effect, they were setting standards for their own operations. The study participants who set the standards by division are shown in the following table. office and 11 1975 and 1976 Study Participants 1975 1976 1975 Negotia- 1976 Negotia- Standards tions Standards tions Location/ study task study task division committee force committee force HPMC with program responsibility 6 (60%) 8 (67%) 5 (71%) 14 (64%) Central office or region adminis- tration respon- sibilities 4 (40%) 4 (33%) 2 (29%) 8 (36%) Total participants 10 (100%) 12 (100%) 7 (100%) 22 (100%) We have no way of knowing if or to what extent these individuals influenced the standards. We have evidence, how- ever, to suggest that when section chiefs were allowed to set task times for their programs, as was the case in developing the Housing Production and Mortgage Credit standards, they tended to opt for the larger times, as shown in the following table. Section Chief Judgment Task processing time in hours De9ference Task times Final between obtained Average task average by task time and Activity/task interview time reported final Existing construction: Inner city 3.20 appraisal 3.00 3.10 3.20 +.10 Suburban 2.00 appraisal 2.67 2.34 2.67 +.33 CRV appraisal 1.00 1.00 1.00 1.00 - 2-3-4-unit 4.50 appraisal 5.00 4.75 4.75 - Proposed construction: Appraisal 3.0O 4.00 3.50 4.00 +.50 Valuation- 2.50 reprocessing 2.75 ".63 2.75 +.12 Valuation- .25 adjustment .25 .25 .75 +.50 Architect- 1.00 reprocessing .50 .75 1.10 +.35 Arcnitect- .30 adjustment .25 .28 .25 -.03 12 In most instances, the final field office task times selected by the section chiefs were the higher of the estimates ob- tained by interview. It appears to us that program managers have influenced some work measurement standards and, as noted in the preceding table, have tended to influence them on the high side. Lack of documentation We could not trace the process used in developing some standards because of incomplete documentation and the ina- bility of people to remember. For example, we were not able to review the procedures used to develop section 8 program standards, which account for 924 staff-years in the fiscal year 1978 budget, because HUD personnel could not locate the supporting documentation and the individual who developed the standards had retired. In addition, the reasons were not documented for some of the judgments made in adjusting the data to arrive at the standards. As a result, we were unable to completely evaluate why data was discarded, nor could . determine the basis for judgments. We believe it is important in developing, reviewing, and updating standards that the bases for judgments be documented and retained so that they may be examined in cle course of any future review or revalidation process. Offices selected for data collection may not represent HUD as a whole Since almost all HUD standards are national in scope, applying to all regions, it is important that they be rep- resentative of HUD as a whole. If they are not, some offices will be understaffed and others overstaffed, and the standards will be defective. In developing some of its Housing Production and Mort- gage Credit standards, HUD appears to have se. cted the locations for data collection more cn the basis of conven- ience or the initiative of a given region rather than the reprtsentativeness of the offices. For example, we were told that the San Francisco regional office took the initia- tive in revising the standards for two large programs in the Housing Production and Mortgage Credit area. In 1975, the San Francisco region revised standards for both the insured 13 single family and multifamily programs; these support approxi- mately 12 percent of HUD's staff-years. They were subse- quently updated in 1976, again based on work in the San Francisco region. According to HUD staff in San Francisco, these studies were initiated by the region to justify their regional staffing needs on the basis of their workload. The HUD central office applied these standards to all regional offices except the Boston, New York, Philadelphia, and Chicago regions, where they increased some single family standards by as much as 50 percent. There was no conclusive data to sup- port the increase in the standards for these regions. Regional and field office officials have questioned the appropriateness of national standards for many of the hous- ing and Community Planning and Development programs. They cited varying travel requirements, workload mix, and other differences which cannot be factored into a national stand- ard. HUD plans to study the feasibility of developing re-- gional standards. No mechanism for updating standards as efficiency improves Once standards are developed, a system is needed to assure that changes in procedures which influence standards are fed back into the standards review process. This is necessary to assure that standards are updated as more ef- ficient procedures are instituted. Yet in HUD, as organizational changes have taken place to increase efficiency, the results have not been used to revise existing standards. For example, standards did not change when HUD merged the two housing program divisions, a move designed to achieve greater efficiency and effective- ness. HUD standards have been revised over the years, but the revisions were not prompted by procedural or organiza- tional changes. Unless HUD institutionalizes a process for reflecting organizational efficiency improvements in work measurement standards, its work measurement system will continue to have defects and reflect greater staffing and costs than are actually required. Staff resources for work measurement system development The organizational unit in HUD responsible for iii- tiating and coordinating standards development is the Staff 14 Resources Division. The Division staff is made up of 15 professionals and 7 clerical personnel. We were told that the professional staff devotes approximately 5 staff-years to work measurement development and productivity measurement. In recent years the Staff Resources Division has taken on additional workload such as preparing regional operating plans and assisting in HUD budget preparation without cor- responding increases in staffing levels. These additional two responsibilities take up the majority of the Division's staff time between mid-May and mid-January. Only about 4 months per year are devoted to developing and validating work measurement standards; this is between January and May. We noted that due to other work, validating work mea- surement standards has not been done as planned. For ex- ample, the assisted housing standards are recognized to be inadequate because no experience data was available for the section 8 program when tandards were developed. The Staff Resources Division sees the need for establishing valid section 8 standards, but no firm -ecision has been made as to when and where these standards will be redevel- oped. A study, "Improving Work Measurement Systems in the Federal Goiernment," prepared by the U.S. Army Management Engineer Training Agency, recommended that several defense organizations staff their work measurement groups with one analyst for each 400 employees when using historical or technical estimates techniques. Applying ,this criterion to HUD's staffing levels indicates the Staff Resources Division should have about 44 qualified personnel for stand- ards development. While we do not advocate this s the ap- propriate ratio, it does suggest that HUD may not have the staff to adequately develop and maintain its work measurement system on a full-time basis. EXTENT OF COVERAGE During the Senate Appropriations Subcommittee hearing in 1976, the Secretary made the following statements about the work measurement system: "I think that our work measurement standards have become quite refined, probably better than in any of the other departments of government. I like to think they are among the best." 15 "* * * we are now able to tell you with a high degree of reliability the staffing impact of changes in program levels * "* * we have a staffing budget which is demonstrably related to workload estimates." In March 1976, a HUD official testifying before the House Subcommi:tee on Housing and Community Development, Committee on Fa-king, Currency and Housing, stated that 63 percent of tt Department's staffing was covered by detailed work measurement standards, and he furnished the following table: Percent of total HUD staff Covered by standards: 63 Field: Housing Production and Mortgage Credit 29 Housing Management 20 Equal Opportunity 3 pirection and administration 6 Headquarters: Office of Finance and Accounting 5 Covered by other criteria: 12 Field: Community Planning and Development 8 Direction and administration 4 Not covered: 25 Field: Legai, Federal Insurance Adminis- tration, other 3 Other headquarters (note a) 22 Total 100 a/Program staff offices, operating divisions, Inspector Gen- eral, and administration. In its fiscal year 1978 budget submission, HUD as- serted that the "* * * work measurement system has substantially im- proved the accuracy of the staffing estimates for the major program areas of Housing, Community Planning and Development, and Fair Housing and Equal Opportunity." 16 Further, the effort invested in the work measurement system was said to have resulted in realistic personnel estimates. The Secretary stated, "* * * we are handing over a strong and shiny piece of equipment to the new administration. Pro- grams, work measurement standards and tracking mechanisms are firmly in place." More recently (Ma-. 1977), and subsequent to our review and briefing to the Secretary on our review's results, HUD officials substantially evised their estimates of staff covered by work measurement standards. Instead of the earlier 63-percent figure, HUD indicated in testimony before the Subcommittee on HUD-Independent Agencies that only 45 percent of its staffing is currently covered by work mea- surement standards. HUD personnel later said that overhead allowances included in the standards for one operating di- vision should not have been. Without these allowances, only 42 percent of the Department's staffing for fiscal year 1978 is covered by work measurement standards. This figure is largely the result of the fact that standards were not used in developing staffing estimates for most activities of the central and regional offices and support activities in area and insuring offices. The following table shows HUD's 42-percent work mea- surement standards coverage in the fiscal year 1978 budget. Staff-years Covered Not covered Percent by work by work covered measurement measurement by standards standards Total standards Field: Housing Programs 5,917 2,237 8,154 73 Community Planning and Development 733 637 1,370 54 Fair Housing and Equal Oppor- tunity 66 355 421 16 All other - 3,448 3,448 - Central office: Office of Finance and Accounting 896 54 950 94 All other - 3,616 3,616 - Total 7,612 10,347 17,959 42 A more detailed breakdown of the fiscal year 1978 budget is shown in appendix II. 17 We believe that in prior years HUD oversold the exten- siveness of its system. According to HUD officials, the Department was under congressional pressure to develop a system which would justify to the Congress the basis of HUD staffing estimates. This year, HUD seems to be more realis- tically presenting the status of its system. 18 CHAPTER 3 OTHER ISSUES RELATED TO HUD'S WORK MEASUREMENT SYSTEM In addition to reviewing the coverage HUD's work measurement system, we examined and reliability of other issues which relate to HUD's ability to achieve the maximum potential from its system. Specifically, we reviewed -- the integration of the work measurement system the time, workload, and cost reporting systems with achieve measures of organizational efficiency, to -- the use of fee personnel ad its relationship personnel staffing, and to -- the validity of the workload forecasting system. INTEGRATION OF WORK MEASUREMENT WITH TIME, WORKLOAD, AND COST REPORTING SYSTEMS In addition to its use in budgeting for personnel, work measurement system was intended to provide HUD's determining and comparing the efficiency of a basis for its Essential to achieving this objective, however, operations. methods for reporting time and workload. are accurate These methods should be fully integrated with the work measurement HUD is not fulfilling this intent because system. -- data collected from its time and workload system is not always consistent with work reporting ment standards data requirements, measure- --the accuracy of time reporting data is questionable, and -- the accuracy of reported work accomplished is ques- tionable. Compatibility of data When a work measurement system is integrated with an accurate time and cost reporting system, tions is measurable. The work measurementefficiency of opera- measures of individual or group performance system establishes or output; the time and cost reporting system measures how long work takes and what it costs to do it (input). Efficiency can then be 19 expressed as the relationship between input and output. If the same mathematical units are used to describe both input and output, and the data is accurate, an organization can measure its efficiency. In reviewing HUD's time reporting system in efect during fiscal year 1976, we sometimes had difficulty cor- relating data. HUD analysts had similar difficulties. In October 1976 HUD revised its time reporting system, making it more compatible with the work measurement system. Ev(n with these revisions, the system does not provide data that expresses the actual work performed. For example, under the low income housing program, the following count- able work units are covered by the same time charge code: -- Management reviews. -- Engineering surveys. -- Utility surveys. Another problem exists with Housing Production and Mortgage Credit standards. Specific time codes apply to the work units covered by the standards, but other time expended is also allocated to the activities on a pro rata basis. Therefore, there is no assurance th. ' the time charges reflect the amount actually expended on the program. To achieve compatibility with the work measurement standards, HUD needs to accumulate time charges at a lower work level than it dols now. Accuracy of time reporting HUD's Departmental Time and Cost Reporting System requires field office employees to record daily time use information by program activity. Time charges collected through this system become the source data for work measure- ment and productivity analyses. If the time reporting data is not current, its accuracy and the validity of the re- sulting analysis become questionable. This is particularly true when individuals record time spent on numerous activi- ties with several different codes. Our review of daily time reports and interviews with staff, conducted well into the month when the reports should have contained several entries, showed that 70 percent of the staff had noncurrent reports; most of them also charged 20 time to two or more codes. In one case, a.n individual with a noncurrent report told us he usually ctrges time to more than 30 different codes. The results of our observations at five locations in regions V and IX are shown below. Number of Number of daily time reports staff Not Location interviewed Current Percent current Percent Region V 56 12 21.4 44 78.6 Chicago area office 94 30 31.9 64 68.1 Cincinnati insur- ing office 55 40 72.7 15 Region IX 27.3 111 27 24.3 84 75.7 san Francisco area office 154 31 20.1 123 79.9 Total 470 140 29.8 330 70.2 An area office official told us 'a did not want his staff spending time filling out time forms every day. He thought the reports were merely an exercise and had no credi- bility as a management tool in the area office. To him, there was no relevance between the work of his staff and the Departmental Time and Cost Reporting System. HUD's internal auditors, in reviewing daily time reports in the New York region, also found numerous deficiencies which raised ques- tions about the integrity of 'he system. Some of these were: -- Employees prepared the forms only at the end of each month. -- Employees prepare two to four pro forma forms in advance and submit them in successive months on a rotation basis. -- Program or activity codes are charged according to supervisory instructions which are based on remain- ing budgeted time or funding availability. We were told that HUD is nearing completion of an exten- sive study of the quality of time-reporting data. Accuracy of task reporting Time expended is one basic element of work measurement systems; the other is work accomplished, or output. 21 Some work covered by work measurement standards is not counted. For instance, although HUD has developed standards to cover environmental clearances, such clearances are not as yet counted when performed. In addition, in the Housing Management program and in the Community Planning and Develop- ment program areas there are no counts of work accomplished under several standards. When workload is ounted, its accurac]/ in some instances is suspect due to inconsistencies in definition of tasks. For example, central office data showed region V resolved 24,829 section 518(b) cases during fiscal year 1976; the region reported a total of 29,193. We were told that a major effort is underway to refine the definition of line items included in HUD's Regional Operating Plan and to identify specific sources for each so that all field and central office personnel will follow uniform instructions. HUD awareness of problem Housing program staff in the central office are also aware of and concerned about the lack of feedback on pro- gram performance. One problem they identified was that data has to b tracked to many sources and manually analyzed and manipulated to be usable. Because of this and other problems, the eficiency of program activities was not being routinely measured. Similarly, payroll costs and task accomplishment data were not being combined to provide pro- ductivity or unit cost information. The housing program staff has developed a Housing Opera- tions Plan System to attempt to remedy these problems. The system is designed to assist managers in carrying out their responsibilities for goal setting, performance monitoring, and other analyses. The basic data, as planned, should com- bine time and task reporting data into an accessible, common dcta file to facilitate workload forecasting and development of work measurement standards. The development of the ousing Operations Plan System will follow a modular approach by program area. The single family application processing module is the first one being developed and is scheduled for completion about October 1977. HUD is currently conducting a data quality assessment and has identified problems with the procedures that are used to collect and process the single family workload in- formation. The Department also recognizes that a system for 22 insuring uniform procedures on data collection from region to region is lacking. and analysis HUD believes the program-by-program approach sary because most of the workload reporting is neces- which data will be obtained are either systems from inaccurate or incom- plete. Long-range plans are to expand housing programs, and consideration willte system to other panding it to a Department-wide system. be given to ex- USE OF FEE PERSONNEL HUD's budget submission sibly 1978 does not accuratelyforreflect fiscal years 1977 and pos- sonnel are used to supplement in-house the extent fee per- staff in the insured single family program. Fee personnel arate mutual insurance fund. are paid from a sep- In fiscal year 1976, HUD paid about $9.4 million to fee appraisers, to inspectors and $.2 million to mortgage $1 million The following chart shows the extent credit examint=s. and purpose of fee use during fiscal year 1976 and for the 5 ber 30, 1976. months ending Novem- Appraisals Inspections Credit checks Number Amounts Number Amount Number Amount 1976 189,342 $9,42,350 88,654 $1,034,676 37,400 $187,275 July- Nov. 1976 96,467 4,798,096 53,380 621,700 14,430 71,965 Total 285,809 $14,222,446 142,034 i1,656,376 51,830 $259,240 Historically, fee personnel have been HUD staff during seasonal peaks and in used to supplement fiscal year 1977 budget indicated that remote locations. The 54,000 single family appraisals, 15 percent of the total, would basis. However, the use of fee personnel be done on a fee has become much more extensive. In March 1977, HUD estimated that 271,000 appraisals--75 percent, not 15 percent--would fee basis during fiscal year 1977, the be done on a equivalent of 335 staff-years. In the Chicago region, HUD appraisers to the section 518(b) program which was are being diverted Eiscal year 1977 budget. An area office not included in the region has reassigned in the appraisers to the section San Francisco 8 subsidized housing programs, which in their view due to poor workload forecasting. The have been understaffed new Secretary of HUD supports the increasing use of fee personnel. 23 Notwithstanding the estimated 75-percent fiscal year 1977 and the inclination fee use in to make greater use of fee personnel, the fiscal year 1978 budget submission re- flects only a 2 5-percent fee appraiser use and no use of fee inspectors and mortgage credit examiners. HUD's budgeted fee use of 25 percent of 352 staff-years for appraisal services. is the equivalent sonnel will be freed to perform other In-house per- duties as HUD increases the fee use above 25 percent. The over the 2 5-percent level result in extent to which increases increased staff-years tor the appraisal function or provide staff for other purposes is reflected below. Percent fee Resulting staff-year appraisals availability 50 117 75 234 100 352 Note: Based on HUD standard of and HUD estimated workload2.23 hours per appraisal of 379,100 appraisals, of which 25 percent will be done by fee personnel. In effect, each -percent increase in fee appraisal use over the 2 5-percent budgeted results in an overstaffing of 4.69 staff-years. The use of fee personnel to supplement also adversely affect the quality of HUD staff may HUD work. Several studies have found that: -- The $50 appraisal fee, plus clerical costs, was at least 20 percent more expensive than a HUD staff appraisal. -- Fee appraisers make more mistakes because of unfamiliarity with HUD than HUD appraisers regulations. IMPORTANCE OF WORKLOAD FORECASTING A reliable workload forecasting system is as important as reliable work measurement standards. Forecasts plied by standards in arriving at personnel are multi- requirements for national budgeting purposes. key determinant of the appropriate Accurate forecasts are a staffing levels to meet HUD's work objectives. We identified several factors casting doubt on the accuracy of forecasts in the fiscal year 1978 budget. 24 Budget formulation is the exclusive responsibility of HUD's central office. Yet some experts on work measurement recom- mend that agencies formulate budgets from the bottom up. The fiscal year 1977 budget formulated by the HUD central office did not show any workload for the section 518(b) and (d) pro- grams. After a preliminary allocation of workload, however, the central office was informed by its regional offices that work was expected to be performed under the section 518(b) program during the budget year. HUD now expects a combined workload of 37,500 cases under the programs, and the use 120 staff-years. If the regions had been consulted as partof of the budget formulation process, this situation may have been avoided. We also noted workload forecasts and 1978 that appeared excessive when for fiscal years 1977 compared to fiscal year 1976 accomplishments. For instance, in fiscal year 1976, HUD did not come close to its goal of approving 50,000 units under the section 235 insured single family program. In- deed, as of December 31, 1976, or after 1 year of operations, the program had firm nationwide obligations for proposed construction of only 3,319 units, with preliminary reserva- tions for an added 14,888 units. A variety of problems were cited as reasons for the limited activity of the program: -- High land and builaing costs making the mortgage limits totally infeasible in eastern metropolitan areas. -- Reluctance of mortgagees and builders to participate in the program. -- Lack of eligible borrowers able to make the required down payments. Nevertheless, HUD forecasts that preliminary reserva- tions under the program will be 100,000 a year in both fiscal years 1977 and 1978. The fiscal year 1977 estimate of 462 staff-years for the section 235 program was determined by applying the 100,000-unit forecast to the work measurement standards. If only half of the workload is accomplished this year, it would mean an overstaffing in this area of 231 staff-years. A HUD official stated that in the current fiscal year (1977) a probable shortfall in the section 235 program will be more than offset by an overage above the budget estimate for the section 23(b) program. Both of these programs, we were told, are single family and utilize the same skills and staffs in the field offices. 25 Some important work required by administrative procedures rather than by statute is not being accomplished. For ex- ample: -- Our August 1916 report 1/ indicated that realty specialists of the Chicago area office had not been properly supervising the activities of area mnanagers. An official of the area office stated he had only 35 authorized positions to d 48 staff-years of work. -- The April 1976 HUD central office evaluation of HUD region V reported staff shortages in 10 housing man- agement areas, noting that physical inspections of multifamily projects were being done on an emergency basis and that there was a tremendous backlog of unreviewed financial statements. We did not review in detail HUD's entire workload fore- casting system. The examples presented above may not be representative of t total system and should not be so in- terpreted. A HUD central office official stated that several years' experience with field participation in budget formulation has clearly demonstrated that field input exaggerates the demand fcr HUD programs and thus tends to make workload forecasting unreliable. l/"Protecting and Disposing of Single-Family Properties Ac- quired by the Department of Housing and Urban Development'," Aug. 31, 1976 (CED-76-141). 26 CHAPTER 4 CONCLUSIONS AND RECOMMENDATIONS CONCLUSIONS We believe that a need exists in HUD and overnment-wide for objective, systematic ways to estimate personnel require- ments. The Congress, too, needs budget requests that are based on reliable personnel requirement estimating techniques. The concept of work measurement offers potential for yielding more objective anc reliable personnel requirements estimates. Yet, the nature of much of the work in civilian govern- ment agencies, including HUD, is such that engineered ap- proaches to establishing work measurement standards ,hould be very costly, and in many cases may not be cost effective. This is because much of the work is seasonably variaDle, non- routine, or has long cycle times. In these cases, work mea- surement standards based on historical experience or properly conducted technical estimates may be the most cost effective method for developing the time to perform a given eration. task or op- These types of standards can be established quickly and at significantly lower cost than engineered standards. The inherent disadvantage, however, is that they are initi- ally less reliable since they tend to be based on past ex- perience and therefore incorporate whatever inefficiencies exist in the way work is done. However, the standards so developed can be updated over time by the use of work samp- ling or methods studies arid adjusted to more appropriately re- flect the time it should take to do the work. HUD's work measurement standards have not progressed to the point where they can be accepted at face value for es- timating personal requirements. In part, this is because of weaknesses in the practices used. In spite of this, HUD should be encouraged by the Subcommittee to develop its measurement system. It would, in our view, be a mistake work to, as one HUD official put it, "throw out the baby with the water." However, in moving forward, HUD needs to be more bath cerned with assuring that the standards developed are basedcon- on sound work measurement practices. The method or combination of methods should be used that is cost effective. We identified weaknesses in HUD's procedures for stand- ards development. These must be corrected before standards are further developed. 27 HUD's work measurement system is also not integrated with time, payroll., and workload reporting systems. Although HUD has a time reporting system, its integrity is questionable because of deficiencies in initial reporting at the employee level. In addition, HUD's workload reporting system is in an early stage of development, and steps have been taken to get better feedback on actual performance to compare with work measurement standards. To compensat- for perceived staff shortages in the subsidized and o,er housing programs, HUD has been relying on fee personnel in th: insured single family program. The extent HUD is using fee personnel to supplement its staff has not been ully disclosed to the Congress. Faulty workload forecasts may be contributing to unreli- able estimates of personnel requirements. If the goal of obtaining reliable estimates of personnel requirements is to be realized, HUD must develop accurate forecasts of workload. RECOMMENDATIONS TO HUD We recommend that the Secretary of Housing and Urban De- velopment improve the Department's practices for developing its work measurement standards. To do this, the Secretary should: -- Perform methods studies or other similar studies to identify the most efficient and effective way to do the tasks beina measured. Standards should, to the degree practical, be based on the time it should take to efficiently perform the given task. -- Exercise more care in using the questionnaire/interview approach to data collection and evaluate other methods of aata collection. The costs and benefits should be considered in deciding what approach or approaches to use. -- Define in further detail the tasks for which stand- ards will be developed, thereby increasing the re- liability of the data collected. -- Examine data which seems out of line and ascertain the reasons for it. -- Assure that the individuals responsible for setting standards are independent of the organization being measured and that they carefully review technical estimates of program managers for accuracy and relia- bility. 28 -- Better document the procss of and reasons for judg- ments used in developing work standards. -- Assure, to the, extent feasible, selected for data collection to that sample sites develop standards are representative of the whole Department. -- Formalize a process for reviewing and updating stand- ards as organizational or procedural changes occur. -- Reevaluate the sufficiency of staft rsourcen to de- velop and maintain the work measurement system. -- Integrate the work measurement system with the Depart- ment's time, payroll, and workload reporting systems. -- Insist upon top management's recommitment measurement system effort and communicate to the work HUD staff involved. it to all RECOMMENDATIONS TO THE SUBCOMMITTEE The Subcommittee should encourage work measurement system toward HUD to develop its the goal of more objectively and reliably determining staffing requirements. The Subcommittee should direct the Secretary to present with HUD's fiscal year 1979 budact hensive plan for proceeding witi, submission (1) a compre- its work measurement systems development, and (2) a statement ment has made in dealing with the on the progress the Depart- report. issues presented in this AGENCY COMMENTS As requested, written comments Department. However, we discussed were not obtained from the our report with Department officials and their views have been considered. 29 APPENDIX I APPENDIX I JMw L MCCL..I ANK.., UAIM W*1RbII. e, MAN . WASH. MILiO N. "UW, N. OAK. JoE C. Irrm·. MiuE. WOAN L UUICA, I. JOIN 0. PSIITORS, n.I. CUIO P. Ch iJ. "rm C. MY". W VA. HInAM L o. Hi.W.II Ur. WM. "CDW. WA DWAID W.oOKE. . iMA·fSS. MINE MANWI$LD, MOT. MARK 0. IMTI1·I.D, 0RE4i, WILLIAM CIOXII. JOGM" n. MorMo, N. Mo. E SEVlwS. ALA CHARLES MC C. MATHIAS, N.MD. en DOAL K. *INOUM WAII RICHAD ·. SCHWIK. PA. COMMITTrr ON APPROPRIATIONS 1NNSY wU H. I.C. H Y ICLLMON, OKLA. *Irm"H AY, I. WAHINGTON. D.C. naAS P. EABIN, 2051S Mo. LAWTON CHILES, FLA. J. SmrNrNJOT LA. WALTE U. WDmIUOT. KY. maw"11CI -A-O--AMW .cAwL, STAMrTOr July 29, 1976 Honorable Elmer B. Staats Comptroller General General Accounting Office Washington, D. C. 20548 Dear Elmer: For some time now I have been concerned over the increasing demands for staffing at the Department of Housing and Urban Development (HOD), and the extent to which these staffing levels are supported by valid, objective work measurement standards. In recent testimony, HUD officials again requested additional staffing and stated that such standards now exist for a substantial percentage of their personnel. I would therefore like to request that you undertake an independent study for this Subcommittee directed at assessing the progress that HUD has made in recent years in develop- ing these work measurement standards and the extent to which the standards that now exist are valid for determining personnel staffing requirements. As a miniun I would appreciate your being in a position to brief the Subcommittee on yourtentative conclusions by April 1977 with a final report to follow at a later date. HUD-I ndent Agencies Senate Appropriations Committee 30 APPENDIX II APPENDIX II ANALYSIS OF HUD FISCAL YEAR 1978 BUDGET REQUEST FOR PERSONNEL Percent of total Staff-years HUD staff Covered by standards Housing Programs: Insured single family existing and proposed construction 1,269 7.1 Section 235 single family housing 507 2.8 Section 8 subsidized housing 924 5.1 Insured multifamily 323 1.8 Public housing construction- traditional and Indian 285 1.6 Single family property disposition 640 3.6 Multifamily property disposition 158 .9 Management of public housing 500 2.8 Management of multifamily loans 613 3.4 Ail other programs 698 3.9 5f917 33.0 Community Planning and Development: Block grants 457 2.5 Comprehensive planning 73 .4 Categorical programs 97 .5 All others 106 .6 733 4.0 Fair Housing and Equal Opportunity 66 .4 Office of Finance and Accounting 896 5.0 Total 7,612 42.4 Covered by staffing allocations, other criteria, or not covered Central office functions other than the Office of Finance and Accounting 3,670 20.4 Field. Housing Programs 2,237 12.5 Community Planning and Develop- ment 637 3.5 Fair Hcusing and Equal Opportunity 355 2.0 Administration and direction 2,08S li.6 Reimbursable disaster 70 .4 All other 1,289 7.2 10,347 57.6 Total staff 17,959 100.0 31 APPENDIX III APPENDIX III SUMMARY OF STAFF-YEARS AND FULL-TIME EMPLOYMENT IN PERMANENT POSITIONS IN HUD FISCAL YEARS 1977-1978 Fiscal year 1977 Fiscal year 1978 Employ- Staff- Employ- Staff- ees years ees years Housing programs: Washington 705 793.0 685 768.0 Field 7,223 7,580.2 7,976 8,153.6 Total 7,928 8,373.2 8,661 8,921.6 Government National Mortgage Associa- tion: Washington 41 40.8 43 44.7 Community Planning and Development: Washington 273 296.8 283 309.8 Field 1,189 1,251.1 1,377 1,370.0 Total 1,462 1,547.9 1,660 1,679.8 New Communities Administration: Washington 75 87.6 75 38.0 Federal Insurance Programs: Washington 191 207.0 215 237.0 Field 107 97.0 184 161.9 Tctal 298 304.0 399 398.9 Consumer Affairs and Regulatory Functions: Washington 189 188.7 189 201.0 Field 180 185.0 180 185.0 Total 369 373.7 369 386.0 32 APPENDIX III APPENDIX III Fiscal year 1977 Fiscal year 1978 Employ- Staff- Employ- Staff- aes years ees years Policy Development and Research: Washington 204 227.0 208 229.0 Field - 17.0 - 15.0 Total 204 244.0 208 244.0 Fair Housing and Equal Opportunity: Washington 94 91.7 161 162.8 Field , 363 398.0 388 421.0 Total 457 489.7 549 583.8 Federal Disaster Assistance Admini- stration: Washington 68 82.4 68 83.4 Field 100 155.2 100 173.7 Total 168 237.6 168 257.1 Departmental Manage- ment: Washington 162 173.3 164 175.3 Office of General Counsel: Washington 200 217.1 200 220.2 Field Legal Services: Field 315 332.0 325 341.7 Office of Inspector General: Washington 87 92.2 90 94.2 Field 404 405.9 408 412.5 Total 491 498.1 498 506.7 33 APPENDIX III APPENDIX III Fiscal year 1977 Fiscal year 1978 Employ- Staff- Employ- Staff- ees years ees years Administration: Washington 1,578 1 992.5 1 578 1,952.8 Field direction and operational support: Field 668 732.1 668 737.0 Field administration: Field 1,154 1,279.0 1,225 1,351.7 All activities: Washington 3,867 4,490.1 3,959 4,566.2 Field: Regions 11,052 a/11,734.4 12,099 12,535.0 Other 651 698.1 732 788.1 Subtotal, field 11,703 12,432.5 12,831 13,323.1 Total permanent full-time 15,570 16,922.6 16,790 17,889.3 Reimbursable disaster staff-years 149.7 70.0 Total staff-years 17,072.3 17,959.3 a/Includes 1,725 positions authorized 10 LO. regional offices and 9,228 positions authorized for 76 urea and insuring of- fices. 34 APPENDIX IV APPENDIX IV PRACTICAL EXPERIENCES WITH WORK MEASUREMENT SYSTEMS Work measurement systems have received increasing tion in recent years, but some studies conclude atten- agencies are not getting the maximum benefits that Federal work measurement. The following are examples possible from of some of the experiences public and private institutions have had with work measurement systems. Study of work measurement in Fed eal agencies The U.S. Army Management Engineering Training which has been providing work measurement training Agency, sulting services to Federal agencies since 1952, and con- conducted a study of the work measurement systems of 11 Federal cies. In a report issued in 1973, it concluded agen- few exceptions the initial attempts to establish that with work surement systems had been only partially successful. mea- benefit was not being derived from many of the existingMaximum measurement systems, in spite of their use in the work Government for over 20 years. Federal Department of Defense In August 1976, we reported that the full potential the Department of Defense's work measurement of efforts was not being realized. 1/ We expressed the belief that savings could have been considerably greater than the $121 million if the work measurement potential had been fully reported realized. In a review of work measurement systems in the services' real property maintenance operations, military 2/ we noted that standards used were largely outdated and that ices did not have adequate work measurement the serv- system reporting. As a result, the services fell short of their potential. We also reviewed operations at a Government-owned, contractor operated ammunition plant and found that the work l/"Improvements Needed In Defense's Effort to Use Work Mea- surement" (LCD-76-401, Aug. 31, 1976). 2/"Major Cost Savings Can Be Achieved By Increasing Produc- tivity In Real Property Management" (LCD-76-320, 1976). Aug. 19, 35 APPENDIX IV APPENDIX IV measurement program was partially successful, but did not fully meet the objectives intended. 1/ Although the plant had made notable progress in its wore measurement program, it was not achieving intended results because of deficiencies in the direct labor standards, and was not achieving attain- able productivitiy. Department of Health, Education, and Welfare Like HUD, the Department of Health, Education, and Wel- fare has a staff management system designed for estimating staff requirements and for fostering efficient management of personnel resources. As part of its efforts, the Department has developed, with the help of outside consultants, manuals describing procedures for gathering data and implementing manpower management systems. One of the Department's component agencies, the Public Health Service, has conducted work measurement and manpower utilization studies to broaden the coverage of its work mea- surement system. Farmers Home Administration The Farmers Home Administration is in the second year of a 3-year project to establish a work measurement system for its 1,750 county offices. With assistance from the Joint Financial Management Improvement Program, the Civil Service Commission, the General Services Administration, and GAO, it developed a plan to survey 75 percent of its offices through a process of detailed employee time reporting. Upon completion of the project, work measurement standards rep- resenting local conditions for each State are to be developed. Currently, data generated by its work measurement system is used to estimate national staffing needs, subject to ad- justment where the data is not reliable. Local government Work measurement standards were successfully used by a city water department to increase its productivity. With the help of consultants, the department developed work l/Letter report to the Secretary of the Army on e-xults of review at Lone Star Army Ammunition Plant, 1Txurkana, Texas (B-175462, Oct. 26, 1976). 36 APPENDIX IV APPENDIX IV measurement standards and systems to report time and produc- tion data. As a result of a methods improvement study and work measurement, the department reduced its staff 25 per- cent while workload and service levels stayed the same. Private mortgage banking and insurance industries Based on contacts with six companies and a number of industry associations, use of formal work measurement sys- tems dces not appear to be widespread in the mortgage bank- ing or mortgage insurance industries. Standards and work measurement systems were generally considered impractical because of small staffs, variety in the work performed, and nonuniform work units. Staffing requirements were not based on scientific methods, but instead on overall workload volume indicators (such as the number of loans or dollar volume per employee), subject to management judgment and discretion. 37
The Work Measurement System of the Department of Housing and Urban Development Has Potential but Needs Further Work to Increase Its Reliability
Published by the Government Accountability Office on 1977-06-15.
Below is a raw (and likely hideous) rendition of the original report. (PDF)