What GAO Looks for

Published by the Government Accountability Office on 1977-06-08.

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

                               DOCUftENT   RESUfilE

     O ~ 6 39   - (A1852811)

     What GAO Looks Tor. June 8, 1911. 1 pp.
     Speech by Donald Eirich, Associate Director, Logistics and
     communications Div.
     Issue Area: Federal Procurement of Goods and Services ,1900,.
     Contact: Logistics and Communications Div.
     Budg e t Function: General Govern.ent: Other General Government
            (806) •

               Most of the procurement audit work being done by GAO is
     being done prior to procurement. congress wants to know whether
     t~e procurements are competitive, n~c~ssary, and cost effective
     when sy~tems are in the conceptual stage or as far in advance of
     the proc'lrement process as is possible. Audit criteria involve
     agency compliance with Federal directives and the agency's own
     implementing criteria. A 1977 report on problems founa with
     G~vern.ent acquisition ana use of computers in procurement
     audits identified general proble m areas: not adequately
     deter.ining the scope of work to be performed; no cost-benefit
     analysis of alternative ways to meet needs; inaccurate
     assessment of current utilization; acquiring equipment sooner
     than necessary; procurement without competition; avoidanc~ of
     authority to delegate procurement resp~nsibility; communications
     applications n0t identified; and security requ i ra.ents not
     consider~d. Areas being emphasized in procurement involve:
     competition, need, expected benefits, costs, privacy protection,
     and meeting agency needs. (RRS)


                                WHAT GAO LOOKS FOF

             ' We were originally asked to talk about post procurement
        a'u dits.   While this used to be the standarci fare of auditors,
        including GAO, we haven't done much of this in" the past few
        years so I have shifted the emphasis. Most of the work we have
        been doing is at the request of Congressmen or congressional
        committees and most is pre-procurement. However, the audit
        objectives are essentially tne SJme for both timeframes.
              The primary concer,'lS of the Congress, as we perceive them,
        are: Is the procurement    co~~titive,   ia the new system needed,
        is it cost effective, can it protect personal privacy and does
        it meet the agencies needs. So that prerogatives can be exer-
        cised, the Congressmen want to know these things early--when
    ~   the systems are in the conceptual stage, or at least as far in
        advance of the procurement action as they       beco~e   aware of it.
        In some case, unfortunately, this has been after the RFP has
        been released or even after the    benc~~ark    has been completed.
              Actually, our approach is straiqhtforwacd.         Being auditors,
        we examine the manner of agency compliance with FMC 74-5, those
        other directives you have discussed during the past few days,
        pertinent parts of FPMR 101 and the agency's own i mplement j,ng
        directives.     These are the audit criteria.


              On occasion, we have questioned or taken exception to
       the criteria.      Fo~    example, we are presently examining into
       the reasonableness of the 24.7 percent cost of Federal
       employee retirement benefits, directed for use in OMS Circular
       A-76 cost comparisons.         As another example, we found at one
       aqency that charges for computer time provided on a reimburs-
       able basis to other agencies did not include depreciation or
       overhead, and we recommended that its policy be changed.            !I
              For the most part, thotJqh, we rely on the Government-
       wide guidance issued by Executive agencies having central ADP
       management responsibilities and the implementing directives
       of the agencies.         These are the criteria we au1it against.        o
              In March 1977, GAO issued a report on ·Problems Found
       With Government Acquisition and Use of Computers from
       November 1965 to December 1976" (FGMS-77-14).         This   doc~ment

       contained a complete list of the 175 GAO reports issued
       during that period, classified into 14 general problem areas,
       with numerous subsets.         Examples of some of the topics related
       to ADP procurement were:

       !/   REF:   (1)   GAO Report "Designation of Lawrence Berkel~y
                         Laboratory Co~puter Facility as a Federal
                         Scitntific Dat~ Processing Center Could Sav~
                         Millions" (LCD-76-112), 12/30/76
                   ( 2 ) CG Decision B-136318, 1/~1/77



             • Not adequately determining the scope of wock
               to be performed
             • No cost/benefit analysis of alternative ways
               to meet needs
             • Inaccurate assessment of current utilization
             • Acquiring equipment sooner than necessary
             • Procurement without competition
             • Avoidance of GSA authority to delegate procurement
               'responsibil ity
             • Communications applications not identified
             • Security requirements not   con~idered

             Let me comment on a few of the areas that we are presently
)    emphasizing.

     COMPETITION -      You are all aware of   a~   extremely difficult
     problem in pursuin9 full and free competition--how
     to handle conversion costs in the procurement process.
             Attempts by agencies to avoid conversion costs has
     resulted in a large number of so-called interim         u~grades--

     sole source or limited competition procurements--which have
     been of considerable concern to the Congress.          The Chairman
     of the House Committee on Government Operations elaborated
     on a recommendation, in its October 1976 report,        !/   in a
     December 1976 letter to the Administrator, GSA.          It said that
     until    ~n   aqency's software has heen convected to standardized

     !/   HR 94-1746


  higher level lanquages, no conversion costs should be considered
  in evaluating hardware bids.    After conversion to higher level
  languages, incidental out-of-pocket costs may be considered.
       In attempting to satisfy this requirement, GAO worked
  with the Depart,ent of Agriculture with GSA's concurrence, to
  develop a new F:ocurement technique and test it in two pending
  procurements.   t t is a two step process.    Some of the key
  elements in th i; case are:
       •. Conversi ;'n will be a mandatory option in the RFP i.e .. ,
          mandatocy for the vendor to quote a separate price
          and optional that the Government accept it. Only
          the dollar amount ~roposed oy the vendor will be
          considered in the evaluatio~ of the conversion
          aspects o f. the ~roposals.
       • Equipment vendors will be required, under a man-
         datory option, to offer two persons as conversion
         monitors, regardless of who receives the conversion
       • After the ~quipment is selecteo, a second solici-
         tation will be made to software firms. At this
         time, the successful equipment vendor will be
         given the opportunity to give his best and final
         offer for the convecsion task.
       • Programs consid&red must be in standard Fortran
         or Cobal. Where there is a Federal standard, it
         will apply. Programs !!'lust be running on existing
         equipment w r ~n the DPA is released.
       • Vendor uniq · = extensions will be permitted pro-
         vided that \. l · ~ benefits are established by
         trade-off an r lysis. Approved extensions will be
         fully documented.
       • Recesign and ~" esystemization must be considered
         in preference to conversion, such as programs
         originally ~repared for first and second genera-
         tion equipment.


              Admittedly, this means more work for the Government.
        We "believe, however, that the benefits of more thorough
        planning and redesign considerations, standardization and
        economies will compensate.
              GAO presently has a draft report out for comment by
        affected agencies which        discus~es   reasons for the high con-
        version costs and ways in which such costs can be reduced.
               I have spent a disproportionate amount of time on this
        concept because it is new and, I think, of int e rest to this
        assembly.!/       GAO expects to be involved in thi s question
        with other agencies, whose ADP procurements we are asked to

)       review.

        NEED -    w~   evaluate the composition of the projected workloads.
        We review the revalidation of workloads, anticipated growth
        and estimates for new applications by examining the nature
        of studies made and documentation down to the user level.
        We have found regression analysis useful in a variety of
        situations, both for validation of expected workload growth
        and for sizing the proposed hardware buy.            GAO is on the
        INFONET, and we use its standard regression analysis soft-
        ware package.       Probably other TS services provide the same
        feature.       We have found   proj~cted   workloads overstated,

        !/   This process is explained in detai~ ·      in   Department of
             of Agriculture RFP No.       dated


     using this technique.            Also, 1isk, core and CPU resources
     can often be extended.            We have found instances where
     ~1stem   enhancements, some ioentified by the agency itself,
     that would expand these capacities and improve the performance
     of the existing      co~puter      had not   ~een   made before initiating a
     new   procure~ent.     While recognizing that technology continues
     to change, GAO, in 1974 issued a report describing ways to
     improve computer operations in six areas.               1/

     EXPECTED   BENEFI~S    -   ,~a   look at these closely.      We have rejected,
     from an audit standpoint, broad estimates of increased produc-
     tivity not supported by detailed justifications.                We have
     accepted other estimates, when broken down to specific
     functions and the rationale is adequately described.                However,
     time and motion studies and pilot or protype tests are the
     best .demonstration of benefits.             These are particularly useful
     where terminals and networks are being planned.

     COSTS - We examine the estimates of the costs of alternative
     solutions that wece considered and,             sometim~s,   other alterna-
     tives that we consider feasible.              Believe it or not, 12 years
     afte~    the Bcooks Bill, we are s ' 11 examining the lease/
     purchase financing alternative.              In some cases, where the

     1/    Teo1s and Techniques foe Improv i ng the Efficiency
           of Federal Automatic Data Processing Operat1ons
           (8-115369) June 3, 1974


      purchase alternative proved advantageous, we found that GSA
      had sufficient capital in its ADP Fund and were able to direct
      the agency toward this alternative.

      PRIVACY PROTECTION - The best way we have found so far to
      evaluate compliance with this requirement is to examine the
      existing computer security environment, generally in accordance
      with the pr.inciples of FIPS PUB. 41.   We have found instances
      where available protections such as personnel and terminal
      profiles and even passwords have not been implemented in the
      existing system nor specifically provided for in planning
      the new system.
      MEETING AGENCIES NEEDS - We observe and 6iscuss the existing
      and proposed applications on site, at the user level.       We
      look for key indicators of user dissatisfaction--such as the
      number and nature of unique applications being employed to
      meet management needs, the existence of supplemental manual
      or automated systems, reports on backlogs and error corrections
      and the number and significance of software changes being made

             Time constraints have permitted only the broadest commen-
      tary on these audit objectives.    We should   recogniz~   that GAO's
      approaches vary considerably according to the scope and
      character of the ?acticular operation, and the time available
      for   ou~   rev iew.