Show me the money

Evidence is sorely lacking that the Bush Administration’s proposed A-76 rules for contracting will bring budget savings

by Max B. Sawicky

The fact of the matter is that DoD and OMB estimate that historical savings from competitive competitions have ranged from 20% to 30% no matter who wins. Those are unaudited numbers. They are OMB and DoD’s numbers. GAO has done work in this area, and we do believe that there are real cost savings. However, we can’t express an opinion as to whether or not that 20% to 30% range is reasonable, because the cost accounting systems are just not of a state that we can form an opinion on it.

— David Walker, Comptroller General, U.S. General Accounting Office (GAO), July 24, 2003

Soon after the Bush Administration assumed office in 2001, Director of the Office of Management and Budget (OMB) Mitchell Daniels promised a radical expansion of outsourcing. He claimed that soliciting bids from business firms for work done in the public sector would either save money or provide more service for less cost than if the work were performed by federal employees. To this end, the OMB issued a revised “Circular A-76,” which codifies detailed guidelines on so-called public-private competitions for work now done by federal agencies. However, further evaluation of the principal evidence put forward by the Bush Administration in support of their claim that the A-76 revisions for service contracting will save money shows that this evidence has a number of shortcomings.

The evidence put forward by the administration is flawed in the following ways:

• There is no evidence that the examples cited in the studies touted by the administration are representative of the functions that could conceivably be open to bid by private sector contractors. They are too few in number for any such inference. Even if the measurement of savings in these few cases are accepted at face value, the relevance of these cases to the prospective expansion of contracting still lacks justification.

• Selection of the cases made available for study was explicitly biased. The data employed in the studies were in great part provided by parties with an interest in how their management or implementation of contracts is viewed. The cases made available for study by the Department of Defense (DoD) are likely to appear more successful than cases chosen at random.

• The measurement of purported savings fails to account for important related factors, in particular the quality of services provided, but also nonlabor costs, and costs that are shifted to other federal agencies or the taxpayer. The savings rates are calculated by comparison to old ways of doing business, rather than arrangements for labor-management restructuring that might be preferred alternatives to the A-76 process.

• None of the evidence is derived from actual experience under the A-76 revisions established by the administration. The revisions are so new that it is too early to be able to draw conclusions based on experience with the A-76 rules.

• The federal government has yet to construct management information systems that would enable well-informed, economical decisions about outsourcing.

The Bush Administration’s privatization agenda has drawn bipartisan opposition in the Congress. Appropriations bills currently pending, such as S. 1589 for Transportation and Treasury Department funding, are expected to draw amendments that block funding of the administration’s new contracting guidelines, detailed in the revised Circular A-76. On September 9, 2003, legislation to this end offered by Rep. Christopher Van Hollen passed in the House of Representatives.

A recent administration brief on behalf of expanded contracting asserts the following:

Both the public and private sectors have conducted independent studies to document the effects of public-private competition. Each has reached the same conclusion: subjecting in-house operations to competition consistently generates cost savings—anywhere from 10 to 40% on average, regardless of whether the competition is won by a private contractor or the government (OMB 2003, p. 2).

A footnote to the cited passage refers to three studies that are examined in detail in the following sections. In testimony before Congress, the Administrator of Federal Procurement Policy in the Office of Management and Budget, Angela B. Styles, noted that “use of public-private competition has not taken hold outside of the DoD” (Styles 2003). This statement is misleading in a way that is basic to this paper and to the A-76 debate. The fact is that public-private competition is not widely used within the Department of Defense, relative to the volume of actual contracting in terms of number of contracts or dollar volume. When public-private competition is used, its workings are seldom documented sufficiently for the conduct of useful research. The evidence put forth by the administration bears this out, as described below.

In 1999, the Department of Defense resorted to the A-76 framework in less than 2% of service contract spending. On the civilian side, the comparable level was one-tenth of 1% (GAO 2001). From 1993 to 1997, fewer than 40,000 federal jobs were subjected to A-76 competition (Kettl 1998). From 1997 to 2001, the comparable figure for DoD positions subject to competition of some type was 38,016 (Commercial Activities Panel 2003).

A second key consideration is that the old A-76 regulations were focused, however imperfectly, on cost minimization. The new revised regulations explicitly abandon this priority in favor of what is called “best value.” Under best value, decision makers may reject the low bid in a competition for the sake of a “cost-technical tradeoff” that, at his or her discretion, the decision maker thinks is a better deal. The murkiness of comparing alternatives with multiple dimensions, cost and quality among them, renders evidence of cost savings under the old A-76 less germane to the debate over whether to fund competitions governed by the administration’s new “competitive sourcing” guidelines.

What would constitute evidence that public-private competition under the new A-76 regulations will save money?

The simplest fact of the matter is that there is no experience from which to learn under the A-76 revisions because they are new. However, this does not preclude any change in regulations aimed at improving procurement. It would be possible to organize rigorous experiments to test the effectiveness of new or old A-76 rules, as the federal government has done in the past in other areas. These experiments would cost money, but untested A-76 procedures would cost money as well.

One type of experiment is exemplified by the way studies of new drugs are done. To test the effectiveness of a new drug, the Food and Drug Administration sponsors studies where a group of afflicted persons are randomly assigned to the drug or a placebo. With a sufficient number of test subjects and some common statistical assumptions, the results of the test provide information on the effectiveness of the treatment.

Alternatively, some studies are based on collecting a sample of information from a larger population. Care must be taken in collecting the sample to ensure that it is sufficiently large and diverse to be representative of the population of interest. Otherwise, what is true of the sample is not necessarily true of the population.

In the case of contracting, such an approach would entail a sample of tasks performed by public employees and contractors and would require data on costs and productivity. Statis
tical analysis of such data would shed light on what sort of arrangements are most economical.

Nothing remotely resembling these basic research methods is employed in the studies cited by the Bush Administration and others on the purported benefits of contracting out or public-private “competitions.” In all cases, the research depends on a small number of case studies. No statistical tests are conducted in these studies, though in some cases claims are made as if a valid statistical exercise had been carried out.

In other words, no case study or small group of case studies can support the claim that more contracting out or competition of any sort would have any particular result, including budget savings. Each study is a data point of one, hence anecdotal in nature, as far as generalizations about the hypothetical effects of new policies are concerned. Arguing that outsourcing will save money by referencing a case study is tantamount to pointing to the successful treatment of an individual in our example of a drug experiment. But case studies, including the ones discussed here, can be valuable in other respects.

The three studies discussed in the following section provide original empirical evidence based on a total of 22 cases. Of these 22 cases, five resulted in keeping the work “in-house,” to be done by public employees. Between 1997 and 2001, the DoD implemented 364 competitions, 65% of which resulted in a choice of public employees (Commercial Activities Panel 2003). Even in the very limited experience of A-76 competitions, the case studies used as sources in the studies discussed here are starkly unreflective of the results of the A-76 process.

The administration and advocates claim that case studies provide evidence that competitions provide costs savings in the range of 30%. This is irrelevant because all that case studies can do is document the cases in question. For example, if we found 10 Chevrolets that turned out to be lemons, this would not be evidence that Chevrolets tended to be bad automobiles. If through case study we discovered what made each car function improperly, that would shed light on possible problems with Chevrolets. It would not mean, however, that Fords were better cars.

The case studies in question provide some insight into how savings may be achieved under competitions, when savings are in fact achieved in the first place. They provide no basis for expecting that other competitions will turn out well.

Other problems with the studies cited by the Bush Administration

None of the studies advanced as evidence for savings under competition deals with productivity in any systematic or analytically substantive way. In other words, savings are reported without regard to differences in output or outcomes. For example, a savings of 40% is treated as superior to one of 35%, regardless of the likelihood of differing performance. If one person spends $15,000 on an automobile, and another spends $25,000 for some other make and model car, we would not assume that the first saved $10,000, compared to the second.

All of these studies were prepared by contractors. Most of the data for these studies are derived from interviews with persons with a vested interest in the outcome of competitions. The bulk of the interviewees are in the managerial ranks of either the public agencies or of the contractors.

In addition to the possibility of bias, the actual development of data on savings is inherently difficult. Comptroller General David Walker of the U.S. General Accounting Office has testified as follows:

[C]alculating savings that accrue from A-76 competitions, as required by the new Circular, will be difficult or may be done inconsistently across agencies without additional guidance, which will contribute to uncertainties over savings (Walker 2003).

All of the studies discussed here, as well as the vast literature on the subject, take note of the inadequacy of existing government management information systems, for the sake of accurately ascertaining the costs of activities performed by either public employees or contractors. The construction of such a system would make possible, as a routine matter, informed judgment as to the merits of alternative methods of service provision. It would also facilitate the development of incentives to generate better performance by public employees or contractors. As economists in this field have noted, with well-designed incentives there is no reason why public employees should be any less effective than those working for a contractor.

A basic concern in a finding of savings is the point of comparison. Savings rates in the studies are calculated relative to arrangements prior to competition under A-76. In the competition process, the public sector competitor submits a Most Efficient Organization (MEO) plan for restructuring. There could turn out to be relatively little savings under contracting relative to the MEO cost.

A-76 advocates would argue that only under pressure of an A-76 contest could an MEO emerge that is competitive with private sector costs, but none of the studies provides any evidence for this claim, other than background interview statements. Sclar (1997, 2000) provides evidence that labor-management cooperation can produce cost savings without need for public-private competition.

At least in part, competitions are motivated by a desire for budget savings. All three studies discussed below document the tendency for agency savings to be reabsorbed into other agency or departmental budgets. Some of the money saved never returns to the federal budget, much less to the taxpayer. This happens for two reasons: first, because cost savings are subject to reprogramming. Reprogramming may enable another agency to provide more services, but at the same time it could run against the grain of budget priorities set by the Congress. If a given sum of money is appropriated with a given level of service in mind, and if that service can be provided at lower cost, it could be argued that disposition of the savings should be subject to the discretion of the Congress. Savings from competitions have a tendency to stick in the functional area, irrespective of the intent of legislation.

Second, the new A-76 regulations incorporate the principle of “best value.” Under best value, the winner of a competition need not be the low bidder. Subject to the discretion of the decision maker, a bidder who provides evidence of a higher level of quality relative to cost may be awarded a contract. Aside from whether the decisions in question are well founded, in effect the option of best value permits the decision maker to expand public services on his or her own initiative, rather than return cost savings to the Treasury. Best value implies that the performance promised exceeds what was originally requested in the Performance Work Statement (PWS). If best value does not guarantee performance beyond the Performance Work Statement, best value would automatically be the lowest bid.

Although obtaining more services at a lower cost is a good thing for the agency and its customers, it is not necessarily good for the federal budget. The decision on whether to expand services in a particular function properly rests with Congress. In principle, potential savings that result from competition should not necessarily be at the disposal of the agency in question.

Further evaluation of specific studies

The following sections provide further detail about the three studies cited by the Office of Management and Budget in support of their claims about likely savings from public-private competitions under the revised A-76 rules.

I. Long-Run Costs and Performance Effects of Competitive Sourcing , Frances Clark, Cheryl Rosenblum, Murrel Coast, and Elaina Smallwood, Center for Naval A
nalysis (CNA), February 2001.

For this report, 16 competitions were selected; 14 of these were won by contractors and two were won by public employees. The total cost of this work prior to competition in 1996 was $100 million, involving 2,800 jobs (in terms of “full-time equivalents,” or FTEs). At the time, 203,000 positions were classified as ripe for competition. This is only a tiny fraction of defense procurement and cannot be called a representative sample in any meaningful sense.

As shown in the statement below, the CNA does not claim its sample is representative.

We included only two in-house wins in our analysis because good documentation of the post-competition costs was very scarce. These two wins are probably a self-selecting group and because good documentation exists, we can probably assume that the functions are well managed. As a result, they may demonstrate higher savings than we would have found had we been able to analyze a larger sample of in-house wins. The contract wins we analyzed may also be self-selecting because we dropped 11 contracts from the analysis due to missing data or because the contracts were later combined with other contracts and it became impossible to track changes in cost (CNA 2001, p. 68).

The authors caution that their sample “should not be considered a representative sample of the population of competitions, [but] the competed functions do reflect the majority of competitions completed during the 1988 to 1996 period” (p. 11).

The problem here is that the population of interest is not competitions, but all potential competitions, or what are classified as “commercial activities.” A representative sample of competitions could provide information about the universe of all competitions, not of functions that have not been subject to the A-76 process. Second, the fact that the selected functions reflect a “majority” of competitions has no statistical import.

It is reasonable to suspect that functions would be chosen for competition with some view as to the prospects for success. There is no legal requirement to open any particular function to competition. It is also reasonable to suspect that those competitions made available for study by the Department of Defense provided better examples of the managerial skill of the DoD. It is also possible that those functions where savings were best documented are not representative of competitions. As noted by the authors of the CNA study, a competition that is well designed and managed is likely to be better documented as well. Conversely, a poorly managed function need not encourage its sponsors to record a detailed chronicle of their sub-par performance. So multiple factors make for bias in the selection of the objects of study. In competitions, the CNA itself advises the DoD to search for “low-hanging fruit.” As a management practice, this makes perfect sense, but insofar as it has been followed in generating the cases available for study, it makes analysis less credible and generalizations from experience less tenable.

The CNA reports that originally it had selected 49 competitions for study, but study of 33 of these proved to be infeasible because of insufficient documentation. Moreover, there was a systematic pattern in rejected competitions—most of the cases where public employees won the work were rejected for inclusion in the final 16 cases.

The CNA also excluded small jobs, defined as involving 20 or fewer workers, because they “typically generate smaller cost savings.” They also chose functions with a mixture of results, in terms of public and private wins. They also limited selection to competitions that had been conducted for at least 10 years. All of these factors underline the inappropriateness of generalizations from the cases under study.

On the whole, the fault here is not with the CNA. They can only study what the DoD makes it possible to study. Doubtless they would have been happy to work with a rich, comprehensive data set, had it existed.

Although DoD has collected some data on the results from earlier A-76 competitions, there are no centrally collected data that track actual cost and performance after the competition is completed. Also, the cost and performance data that do exist at the local level have not been consistently maintained because OMB, departmental, and service requirements have changed over the years. For these reasons, we have relatively little data on the longer-term impact on cost and performance (CNA 2001, p. 8).

As noted earlier by Walker, the determination of savings is a challenging task. The CNA study in this realm is conceptually sophisticated. It distinguishes between observed costs and costs that are adjusted up or down for, respectively, contraction or expansion of the work to be performed over the course of the contract. In other words, if a job is bid for $1,000, the observed cost is $2,000, and the contractor has been compelled to do twice as much by the federal government, what is called the “effective cost” is only $1,000. The idea is to factor productivity into cost so that actual costs are comparable to the bids that win the competitions. Observed costs exceeded bids in every case analyzed by the CNA.

While “effective cost” is indeed the concept of interest, it should be realized that this number requires counterfactual extrapolation from observed costs. It is not actually observed, but rather derived from a model of costs based on experience that is applied to a hypothetical situation. The sophistication of this sort of analysis goes back to the question of relying on interviewees for the data in this study.

Another drawback of “effective cost” is that it fails to reckon with undesirable expansions of activity. If an agency makes a deal with a contractor to do a certain service, and midway through the contract the agency determines that some other service is required, to some extent the new service will be an add-on. The contractor will expect the original arrangement to remain in force and supply extra services at additional cost. This inflexibility is concealed by the use of the effective cost concept. By contract, in a federal agency, workers and resources can be redeployed as circumstances warrant. The implicit contract with workers is to do whatever is asked of them, however that changes over time.

One question raised in the CNA results arises from their report that “effective costs” exceeded the bid in nine out of 14 cases. Why were contractors paid in excess of their bids, if not for providing additional services? Another question is suggested by the fact that the worst contract performance was in the area of grounds maintenance, arguably one of the easiest things to contract out.

Unlike the other studies, the CNA reports collected data on contract performance. Quality of output was rated in a scale from one to five, based on interviewee response. The middle response of three is described as “neither satisfied nor dissatisfied.” The meaning of this formulation is elusive. A response of “three” makes sense as halfway between dissatisfied and satisfied, though different respondents could conceivably gauge their choices on this simple numerical scale very differently.

In the report, “three” is described as “neutral.” In the summary of results, however, the neutrals are lumped together with the “satisfied” respondents. One could as easily say that a given percentage of respondents were either dissatisfied or neutral with respect to the quality of services provided. The questions related to quality were asked of both contract officers and management, with equal weight given to customers of the service (for instance, residents of military housing).

The CNA reports an important problem in competitions. A rationale for contracting is that it provides productivity incentives. The customer can reward or punish with incentive payments and penalties, based on results. CNA says that “as a ru
le, the PWSs [i.e., the specification of work to be done] we reviewed were not performance-based or results-oriented” (p. 45). The CNA further elaborates that the PWS typically prescribes what is to be done, not how, while contracts often stipulate how but not what (p. 47-48).

II. Personnel Savings in Competitively Sourced DoD Activities: Are They Real? Will They Last? Susan M. Gates and Albert A. Robbert, RAND, 2000.

This study covers six competitions. The choices were extremely limited because few competitions had been held that met the requirements of the study. Candidates for study were provided by the services.

RAND (a research organization) found savings rates to be high, resulting from lower personnel costs, as opposed to the downgrading of positions or substitution of capital for labor. Here again there is no pretense by the authors that any kind of representative sample has been constructed.

The purpose of this analysis is to develop a better understanding of the sources of efficiency improvement generated through the A-76 process, how those improvements are achieved, and whether they could be achieved outside of an A-76 competition (RAND 2000, p. 7).

Note that the above statement refers to the old A-76 regime and nearly takes the likelihood of efficiency improvements as a given.

A basic issue for the RAND study is that it does not consider capital or material costs. Savings under analysis pertain to labor and nothing else. The authors note, “[N]onpersonnel costs (material and equipment costs and the costs of reimbursables) are not treated in a consistent way in these cost comparison studies” (p. 26). By the same token, however, observed savings rates could be biased by unobserved capital and other expenditures.

In the simplest arithmetical sense, limiting savings determinations to one category of cost is likely to inflate overall savings. If nonlabor costs between compared projects are identical, personnel savings of one relative to the other are necessarily in excess of overall relative savings rates.

One finding by RAND is that no regulations prevent government from achieving personnel savings as described in case studies. This is interesting in light of the common notion that personnel regulations are too inflexible for efficient work organization. RAND instead finds that government managers do not face sufficient incentives to achieve personnel savings. RAND claims that only the threat of A-76 competitions motivates government managers to achieve savings from “lean bids,” either through the reorganization of public employees or the use of a contractor (p. 58).

The truth of the matter is that from 1990 to 2002, total federal employment fell from 2,250,000 to 1,818,000 (OMB 2003). The notion that “you can’t fire anyone” in the federal government is one of the most spectacularly mistaken axioms of public policy debate.

This suggests the possibility of gains from reform of incentives facing managers, rather than the expensive and time-consuming A-76 process. An example is so-called “gain-sharing,” where some part of savings a manager is able to achieve becomes available to be plowed back into the agency in question, for the sake of augmenting resources available to expand the service provided.

The data for RAND’s analysis were from site visits and interviews. Their site visit protocol stipulated meetings with management and vendors for six hours and with civil service or union organizations for 30 minutes.

As for the CNA study, the interesting findings speak to the cases studied, not necessarily to any wider universe of possible candidates for competitions. In particular, two sources of personnel savings are detailed on the following pages, along with an analysis of each.

The contractor reduced labor costs by downgrading positions so that the average grade level is lowered. This downgrade changed the grade mix, reducing the average skill level and wage of the workforce (RAND 2000, p. 91) .

The failure to incorporate productivity into “savings rates” has already been noted. The CNA finding that PWSs and contracts fail to systematically stipulate contract expectations has also been detailed. Taken together, these shortcomings underline the danger of output degradation in the competition process. Less qualified workers would be expected to be less productive.

The contractor can easily adjust the workforce to respond to changing work volume. The contractor can hire in a few weeks and can fire with minimum due process. Because the contractor does not have to staff for exceptional workloads, staffing can be more streamlined (RAND 2000, p. 91) .

Here the question of cost raises the problem of what economists call an externality. The public sector is limited, on this account, by labor standards that protect job tenure and enforce regular work hours. However, job insecurity and nonstandard work arrangements impose costs on the larger economy. An example is the failure to provide health insurance benefits, a potential factor in lower personnel costs for private sector contractors. Workers not provided with health insurance are obliged to rely on the insurance of their spouses, on public sector programs such as Medicaid, or on uncompensated care (e.g., visits to the emergency room). In this vein, savings to the DoD may be borne in part by the Department of Health and Human Services, taxpayers, or workers whose health suffers as a result. Other factors that could bias savings estimates upwards are listed below.

• Benefits aside, RAND claims that contractor pay is not lower. RAND had no data on actual contractor wages and benefits. The authors admit they cannot be certain that contractors adhere to Department of Labor standards on prevailing wages. These standards include imputed fringe benefits.

• RAND notes that personnel savings are not realized if the military or other personnel from the service are shifted elsewhere (p. 104). As stated earlier, in this respect, competitions are vehicles for expanding the DoD.

• RAND believes that overhead costs imputed to keeping work in-house (done by federal employees) are “substantially overstated” (p. 106). On this account, outsourcing savings rates would be overstated as well.

III. Moving Toward Market-Based Government: The Changing Role of Government as the Provider, Jacques S. Gansler, IBM Endowment for the Business of Government, June 2003.

This study cites two sets of what it describes as “case studies” in the first two sections of the report. The report provides no original empirical research on the outcomes of A-76 competitions.

The first set is grouped under the generic heading of “outsourcing.” No mention is made here of the use of A-76 procedures. It turns out that all four studies included in this set were winners of awards in a competition sponsored by Government Executive magazine. (This is not mentioned in the report.)

No external studies on three of these four projects have been conducted, despite the report heading of “case studies” (p. 13). Accounts in the report are no more than recapitulation of the news account in Government Executive (Laurent 2000) .

The one exception mentioned above was the Navy contract with EDS for the Navy Marine Corps Intranet (NMCI). The study in this case was done by Booz Allen Hamilton (BAH). BAH did not find cost savings, but it did claim that the contract led to added capabilities (Dorobek 2002). This is in keeping with the observation above that contracting is a vehicle for expanding services. The “savings” found by BAH are projected, since their study was cond
ucted before fulfillment of contract.

A second section lists additional “case studies” under the heading “competitive sourcing.” As above, there is no pretense that the examples selected are representative. Taken at face value, they show that the use of competition can be successful. Of course, a similar report could be prepared showing that in-house provision can be successful too.

The first case in point pertains to the city of Indianapolis, Indiana, and has nothing to do with A-76. Savings in the realm of road repair are attributed to a “threat of competition.” This case is discussed in detail in Sclar (1997, 2000). Sclar reports that savings derived from the mayor’s willingness to eliminate redundant supervisory positions, a commendable deed for which he suffered politically. This model could prove useful in the federal sector, but it would work differently than an A-76 competition.

Several other examples were cited as evidence to support the claim made in the study that competition can be successful:

• A second case concerns a Navy shipyard. There was no study, nor was it conducted under A-76 rules.

• A third case pertained to Fort Rucker. The source is a CNA study.

• The fourth case refers to the RAND study discussed in the previous section. It is inaccurately summarized as follows: “RAND concluded that the labor savings came from three major factors…[one of which is] capital-labor substitution.” In fact, as noted above, RAND explicitly discounts the importance of capital-labor substitution in its analysis of savings.

• The fifth case is the CNA study discussed above.

• The final case pertains to the Defense Finance and Accounting Service (DFAS). The source for reported savings is the DFAS.

• A table is provided (p. 22), but it reports expected savings, not observed savings. A second table illustrates realized savings, based on the CNA study.

There is no original case study research showing cost savings in this report. All of its evidence, such as it is, is gleaned from other reports.¹

One point of interest in the report is a citation to the effect that competition results in few layoffs—8% according to a study by the General Accounting Office. This emphasizes the earlier point that competition is a vehicle for expanded government. Personnel are redeployed, and if they take “early outs,” that is additional expense not related to service provision. Agency savings do not translate to budget savings.

Conclusion

The research cited by the Bush Administration provides scant evidence that competitions under new or old A-76 rules will save money. In the most basic sense, the studies cited by Angela Styles, Administrator of Federal Procurement Policy in the OMB, are not designed to provide any such evidence. The extent of data, the manner of data collection, and the depth of information provided cannot lead to useful generalizations about whether or not existing or additional competitions will save money.

The extent of data is minimal. The most expansive study encompassed 16 cases. One study provided no new data at all. All three of the studies were prepared by contractors. Their data were obtained from management groups that employ contractors, contracting officers, and the contractors themselves. The savings data fail to reflect quality differences, and some of the data do not cover overall contract costs.

The studies do provide many hints of the pitfalls of competitions, though as with the savings estimates, it is not possible to generalize from these findings to the universe of potential competitions. That said, one hint of potential problem areas is that when savings are attained, they are not returned to the taxpayer. Another is that costs often exceed agreed-upon bids, even taking into account some expansion of the work to be performed. A third is that savings under competitions could be achieved at lower cost with workplace restructuring under current law and regulations. Finally, contractors that fail to replace health insurance enjoyed by public employees could be adding to taxpayer costs.

Absent better evidence, the federal government would be well advised to defer wholesale changes in A-76 rules and campaigns for expanded outsourcing. Its first priority should be to develop adequate management information systems that would usefully inform procurement decisions. Data and analysis are logically prior to the “make-or-buy” decision, much less to sweeping institutional reform.

Endnote

1. Citations in the report by the IBM endowment reflect some lack of care. Footnote number 37 just says “RAND,” with no further explanation (p. 57). One citation described in the text as a “study” (footnote number 30) is actually an unpublished presentation entitled “A-76: The View from DoD.” The footnotes cite the same studies multiple times, while the text has language such as “a study,” failing to reflect that it is referring again to
a study already cited. The overall effect is to imply more research than has actually been referenced.

References

Commercial Activities Panel. 2002. “Improving the Sourcing Decisions of the Federal Government: Final Report.” U.S. General Accounting Office.

Dorobek, Christopher J. 2002. Study: NMCI worth the cost. Federal Computer Week. July 15.

Executive Office of the President, Office of Management and Budget. 2003. Budget of the U.S. Government, Historical Statistics, FY2004.

Executive Office of the President, Office of Management and Budget. 2003. Competitive Sourcing: Conducting Public-Private Competition in a Reasoned and Responsible Manner.

Gansler, Jacques S. 2003. Moving Toward Market-Based Government: The Changing Role of Government as the Provider . IBM Endowment for the Business of Government.

Gates, Susan M. and Albert A. Robbert. 2000. Personnel Savings in Competitively Sourced DoD Activities: Are They Real? Will They Last? RAND.

Kettl, Donald F. 1998. Reinventing Government: A Fifth-Year Report Card. Washington, DC: The Brookings Institution.

Laurent, Anne. 2001. Award-winning acquisition. Government Executive. August 22.

Rosenblum, Cheryl, Murrel Coast, and Elaina Smallwood. 2001. Long-Run Costs and Performance Effects of Competitive Sourcing, Frances Clark . Center for Naval Analysis (CNA).

Sclar, Elliott D. 1997. The Privatization of Public Services: Lessons from Case Studies . Washington, D.C.: Economic Policy Institute.

Sclar, Elliott D. 2000. You Don’t Always Get What You Pay For: The Economics of Privatization. Ithaca, NY: Cornell University Press.

Styles, Angela B. 2003. Statement Before the Subcommittee on Oversight of Government Management, the Federal Workforce, and the District of Columbia, Committee on Governmental Affairs. U.S. Senate, July 24.

Walker, David M. “Competitive sourcing: Implementation will be challenging for federal agencies.” Testimony Before the Subcommittee on Oversight of Government Management, the Federal Workforce, and the District of Columbia, Committee on Governmental Affairs. U.S. Senate. U.S. General Accounting Office, GAO-03-1022T, July 24.

Walker, David M. Testimony Before the Subcommittee on Oversight of Government Management, the Federal Workforce, and the District of Columbia, Committee on Governmental Affairs. U.S. Senate. Transcript in Congressional Quarterly, July 24.