The Federal Government is the largest producer, consumer and regulator of information in American society. Next to financial transfers and national defense, the provision of information to and from the government and its role in regulating market behavior probably ranks third in overall societal impact.
Because of government's pervasive role, the management of information technology resources is also a dominating and perennial problem to the government itself. The Paperwork Reduction Act of 1980 requires the Office of Management and Budget (OMB), together with senior officials in Federal Agencies, to assume a unified responsibility for management of information resources and to submit an annual report to Congress on the discharge of its responsibilities.
In October 1983, OMB circulated to the senior officials in the agencies a supplement to its Second Annual Report under the Act. That report, entitled"Incentives for Efficient Information Resources Management" sought: (i) agency comment and advice on improved procedures for pricing of information products and services (provided by the government), and (ii) to encourage agencies and Federal research organizations to experiment with the implementation of improved incentives for managing information technology.
In particular, OMB observed that "management incentives are important for planning the allocation of common user resources as, for example, centralized computer facilities. When many agencies utilize information resources in common, principles must be developed which will ensure efficient resource usage. In the typical government setting, the policy principles are developed by senior management on the basis of a "priority user" scheme and administered from the top down. The perspective presented in this report describes bottom-up incentive mechanisms which constrain agencies to take into account the opportunity costs that their actions and choices impose on other agencies."
Chapters 3 and 4 of the OMB Report, which follow as Section A of this note, describe, in particular, the application of demand revealing incentive procedures to the planning and management of information technology within the Federal establishment. Examples
of specific decisions to which such procedures might be applied include:
-- Decisions on whether to acquire dedicated telecommuni- cations facilities to be shared by numerous Federal Agencies or permit each agency to pursue its own strategy.
-- Decisions on the configuration of information processing support services that would serve end users within an agency; this is an intra-agency variant of the government-wide telecommunications decisions cited above.
In addition to describing these potential applications to intra- governmental decision-making (Section A.1.), OMB encouraged experimentation with the implementation of a demand revealing approach (Section A.2.) and provided a formal statement of the approach in a separate appendix accompanied by an example illustrating the incentives embodied in the rule (Appendix 1).
The second section (B.) of this note summarizes some of the more significant agency comment on the decision rule accompanied by this author's reaction to the comment, focusing in particular on some of the key distributional constraints identified by the agencies. In addition, I deal briefly with and provide some examples of potential applications of the procedure to government regulation, a topic which is treated at greater length in a companion paper, entitled "Privatization, Regulation and the Demand Revealing Governance of Enterprise", appearing in this volume.
B. Constraints on Implementation: Agency Response and Comment
About twenty agencies submitted comments to OMB on the issue of implementing the demand revealing incentives approach. Although a number of agencies thought it was an interesting approach, they pointed out a number of problems and pitfalls in implementing it. The agencies concerns are summarized below under broad topics.
Determination of Value of Benefits: One of the most commonly expressed concerns was the potential difficulty in determining
the value of the benefits for each alternative. For example, in the view of one agency:
o "Precise quantifying of price/demand relationships in information technologies has not been successfully accomplished within the information resources management community. In addition, organizational and human biases will remain an institutional dilemma thereby rendering any reported benefits suspect."
o The "bid system" used in the demand revealing approach does not appear to take into account varying information management system needs on an individual agency basis. Therefore, imposition of 'penalties' would be unfair and meaningless.
o It is not clear what the penalties are or how they would be enforced or who would monitor the program..
Variation in Agency Size and Number: A second concern to several agencies was the appropriateness of the demand revealing approach when the participating agencies are of greatly different sizes. The largest Federal agency, for example, observed that it could likely be the "dominant" agency and always determine the preferred outcome. In addition, "it would always suffer a penalty and smaller agencies could inflate the benefits of their preferred choice without concern for affecting the outcome."
Other agencies pointed out that "the demand revealing approach most likely will not accurately reveal group membes' estimates of net benefits in large groups" (because of a low probability of changing the outcome) and "one could over- estimate the magnitude of costs and benefits without fear of penalty."
Further, "agency management may have goals not consistent with the economically pure demand revealing model. Can coalitions be prevented? Multibureau agencies or other common interests provide ready mechanisms for coalitions. Can the mechanism be manipulated to allow agencies to bid out of common user systems irrespective of the best interests of the taxpayer?"
Need to Conduct Prototype Tests or Cost Benefit Evaluation; Increased Costs. Several agencies stressed the importance of establishing a prototype test for the demand revealing approach before implementing it and also foresaw considerable increases in costs in administrating and maintaining the demand revealing approach if it were implemented.
The Decision Rule; Penalties. Agencies also gave particular scrutiny to the decision rule and to the operation of a penalty system within the constraints of the normal appro- priations process. In addition, they commented specifically on the illustrative example set forth in Appendix 1 and, in one case, provided the following counter-examples of perceived pitfalls in the approach:
Author's Comment: In inviting this article on intra-governmental methods of using the demand revealing process, I was also cautioned by the editor that, in his view, many of the respondents appear to be vigorously, and sometimes not very intelligently, looking for objections to the process whenever they think it might deteriorate their current position, "... under the circumstances, repeating at length what they have to say would seem to be largely a waste of time."
The agency reactions to this proposal are, however, interesting in that the issues and concerns they raised regarding the allocation of resources to information technology a difficult planning and management problem are predominantly distributional rather than efficiency ones. Of course, they are directly impacted by such a decision-making procedure and the allocation of resources to information technology can create big differences in welfare within the Federal establishment. For further evidence on this score, see M. Schrange, "How You Can Profit from the Coming Federal Information War", Washington Post, September 29, 1985.
My comment on these issues is, however, largely directed to an important technical aspect of the demand revealing distribution which has been given a more thorough treatment in a companion piece appearing in this volume and entitled, "Privatization, Regulation, and the Demand Revealing Governance of Enterprise."
This is the role of a person or person who is given significant power to make prior allocations of costs - in this case among agency recipients of the technology. Although the basic criterion guiding these cost allocations is the long-revered benefits received principle of public finance, its application can be somewhat controversial even in the context of such a simple problem as intra-governmental cost allocations. In fact, its application becomes even more controversial in decisions affecting shareholders and workers in a non-governmental or mixed enterprise, as is amply documented in the companion paper.
To focus on this issue, consider the initial cost allocations in the OMB example in relation to the two counterexamples and the agency discussion of the perceived problems that were set forth above. 1/ Although I have replied to similar comment and zero-sum counterexamples (hypothetical example #1) elsewhere - see Frerjohn, Forsythe and Noll (1979) and my comment in the same volume -- I have several further comments that bear on important issues raised in the companion paper.
First, my view then (1979) that zero-sum games (along the line of hypothetical example #1) were to be avoided by some process of agenda control have to some extent changed in accord with Tullock's current view that the Clarke (penalty) tax is a positive
advantage in such a situation (Tullock, 1986). With respect to hypothetical example #1, the penalty tax, in fact, reduces the likelihood that such transfers (which might otherwise be carried out) will in fact be carried out or that zero sum games will in fact be played.
So too with coalitions as in counterexample #2. They (coalitions)
can clearly be a problem but only to the extent that the cost allocator has done a less than perfect job of originally assigning costs so as to achieve unanimity in the choice of the outcome (Clarke, 1980). The real problem which was to some extent muted in agency criticisms of the OMB approach is with the cost allocator. The person is given significant power to control the distribution of resources, and even if that person might do the job well (in accord with benefit-cost principles) there might be objections on equity grounds or simply that the distribution would be changed (from that which would be operative under the status quo) to such an extent that the affected parties might not accept a demand revealing regime.
This is the practical real world problem which was only given brief treatment in terms of issues for a demand revealing experiment (Section A.2.) where GSA would coordinate preferred options submitted by various agncies and estimate costs associated with these alternatives. "Each agency would then be assigned cost shares for each alternative based on estimates of the benefits to the agency."
Otherwise, following the advice of the editor, I have not dealt here at length with a variety of agency comments on the demand revealing approach that may reflect a misunderstanding of it or a view that is use, while perhaps contradicting to efficiency, might not be as beneficial relative to some other distributional outcome. However, I would note that in the context of agency discussion of counterexample #2, the observation that a party can be penalized less than it would in some other case where its benefits are lower (i.e. for getting a better deal, a lower penalty results) reflect misunderstanding of the approach and should not be mysterious, but may reflect the reaction of many that consider the implications of the decision rule independent of ex post failures in the process of allocating cost shares. That is, we seek a process of allocative costs in accordance with benefits received; when this is inherent we can have seemingly bizarre results which is almost always the result of comparing some imperfect procedure with nirvana or perfection.
Finally, by way of comment on practical application of the procedure to intra-governmental decisions, unique difficulty (in respect to intra-governmental application) with the procedure should be recognized. That is the fact that the resources avail- able to individual bureaus and agencies are not really outright entitlements and as a result, one might question whether government officials would face any real incentive in making demand revealing tradeoffs of the kind envisioned here. This problem, however, should be considered in the context of the particular applications affecting the administrative budgets of the agencies which (in contrast to the much larger program budgets) have in recent years been largely straightlined by OMB and the Congress. Therefore, to the extent that this budgetary policy continues into the future, the budget for administration, including small to medium size increases in resources for infor- mation technology, can for all practical purposes be considered a quasi-entitlement. However, for broader application, to government regulation for example, this constraint on the real incentive facing government officials must be more carefully considered. So too, we must deal with a related problem of the incentive of those who exercise budgetary oversight in Congress and the Executive Branch to actually enforce any penalties that are incurred, a topic that is considered in the following concluding section.
The DR Approach to Regulation. Some brief final comment is warranted regarding the use of the basic demand revealing decision rule for achieving efficient regulatory requirements vis a vis the private sector which has been treated at greater length (following) in the companion paper. The OMB report, while acknowledging and illustrating use of the procedure for standards and regulations within the federal establishment (e.g. information
processing standards for computers) notes obvious problems in adapting a DR process "to a joint decision involving both Federal agencies and private parties because of the pecularities of the Federal revenue collection process. Legislative authorization would clearly be required in order to permit the Federal Goverment to collect any penalties associated with a demand revealing mechanism."
Indeed, my initial interest in the DR approach was sparked by the potential emergence of new regulatory requirements (for public/private cost sharing of improvements in the Nation's telecommunications facilities) growing out of one of the landmark divestiture of AT&T during the early 1980s.
The problem concerned how the Federal establishment, including the defense establishment, was going to create a viable regulatory
interface with a highly competitive industry in comparison to a situation where most of its requirements had previously been met through bilateral relations with AT&T directly.
Appendix 2 to this note briefly describes the applicability of the demand revealing procedure for efficiently establishing national security/emergency preparedness (NS/EP) regulatory or cost-sharing requirements involving the provision of a public good (i.e. NS/EP communications requirements) as well as commer- cial advantages to private parties (i.e. common carriers in the telecommunications industry and consumers).
Appendix 2, concerning "an incentive based approach to public/ private cost sharing to meet NS/EP requirements" demonstrates the applicability of the approach in terms of a possible legislative amendment -- in that case, to the Telecommunications Act of 1980.
The amendment would have authorized the Federal Communications Commission to make cost allocations according to estimated bene- fits received along the lines elaborated in this paper. In turn, the Commission could, by rule, with the concurrence of the President or his designee, establish DR procedures for obtaining information on actual benefits before any regulatory or cost-sharing requirements were imposed in the industry. Finally, penalties could be imposed on private carriers (for purposes of buying out of an inefficient rule) or on agencies seeking to impose these requirements when private carriers reveal opportunity costs higher than those estimated by the FCC.
The basic principle outlined in this document, in my view, provides a useful means for control of the current regulatory apparatus in the United States that might also have move broad- based application to control of enterprise in other settings. With respect to regulation in the U.S., there has been much talk over the past decade, of the ideal of a "regulatory budget" that would in some way mimic the fiscal budget of the U.S. as a means of disciplining and controlling regulatory costs. See, for example, Nordhaus and Litan (1983). A basic problem, however, is the "funny money" problem in that no one can adequately measure for purposes of budgetary control the real cost (in opportunity cost) terms of the requirements that are imposed (which are inherently subjective and relate to the alternative choices available to and known only by individual parties the regulated sector). The DR approach confronts this problem directly by adopting an opportunity cost criterion and relying on the truth-telling properties of the procedure to obtain accurate information. This approach, first suggested by Portnoy and Sonstelie (1983) would, in my view, be workable at least in limited context of problems similar to those elaborated in the defense telecommunications example here, given reliable procedures for arriving at initial cost allocations, the willingness of Congress to authorize a rulemaking procedure necessary to implement a DR approach and the willingness of both Congress and the Executive Branch to actually enforce penalties against agency accounts as part of the fiscal budget process.
When I first came to think along those lines in the context of the example set forth at Appendix 2, it struck me that "charity begins at home" with potential application largely circumscribed to intra-governmental decisions and that procedural innovations building on this experience could be deferred for a time into the future. However, in the budget process as we see it today, procedural innovation is moving faster than many would have once thought and who knows when demand revealing's time will come. 2/
1/ For example, the comment ignores the initial allocation of costs in comparing the relatie level of penalties incurred in each case. If the person or organization responsible for initial cost allocations had better prior knowledge of the relative level of benefits, then parties 1 and 4 would have received higher initial allocations compared with parties 3 and 4, hopefully resulting in unanimous agreement and zero penalties in each case.
Even if this is not done perfectly, an assignment of initial costs in proportion to benefits received would change the distri- bution of net benefits and penalties incurred in each case, so as remove the possibility that parties pay higher benefits when their
benefits are lowered. The reader can easily verify this result.
2/ In a companion paper, entitled "Privatization, Regulation and the Demand Revealing Governance of Enterprise", I have treated in greater depth the application of the procedure to the governance of enterprise, both among shareholders and workers within the firm, and with respect to external constraints on the firm through regulation or other State apparatus.
The paper suggests the general applicability of the demand revealing (DR) approach to enterprise governance in a variety of socioeconomic settings. These included problems of enterprise regulation and labor relations (i) in countries such as the United States where there has been a tendency to separate the economic and political spheres, and (ii) where the value system favors their integration (corporatist socioeconomic environments in Latin America, Japan and much of Western Europe). The paper focuses specifically on the applicability of the DR approach to the governance of state or "mixed" enterprises and the process of privatizing these enterprises, both in advanced industrial societies and developing countries. It is part of a larger study of demand revealing approaches toward privatization that includes also the governance of producer cooperatives and state regulated natural monopoly.
The general approach was developed as an outgrowth of experience with the oversight of government regulation of enterprise by multiple regulatory agencies in the United States, utilizing a demand revealing approach developed by the U.S. Office of Management and Budget for intra-governmental decision-making. This approach ensures that regulations or subsidies to encourage government-induced departures from profit maximizing enterprise behavior will reflect the true opportunity costs of these induced changes and that these costs will be reflected in the state's "regulatory" or "enterprise budget" in a manner comparable to its fiscal budget.
The paper also explores the relative efficacy of demand revelation as a substitute for conventional shareholder voting (particularly when capital markets are thin or imperfect and the government is a principal shareholder) and considers how a demand revealing approach to governance could improve cooperative interaction between shareholders and employees (particularly where implicit contracts govern the distribution of enterprise resources). The paper emphasizes the role of an allocator who is given significant power to make distributional decisions in the form of wealth transfers among shareholders, between shareholders and the government, and among workers. Although the basic distributional criterion is the long-refvered benefits received principle of public finance, its particular application to the enterprise may raise new and interesting controversies for students of the new incentive or demand revealing mechanisms as well as for a range of practitioners, including those seeking to improve enterprise performance in developed and developing countries alike.
OMB's supplement to its Second Annual Report also discussed the application of the procedure to the establishment of telecommuni- cations facilities requirements vis a vis the government and the private sector. Specifically, this involves the determination of requirements that will improve the inter-connection quality of commercial communications systems that will meet certain national security/emergency preparedness requirements. The result is a joint benefit to the government in terms of provision of a public good (i.e. national security) as well as benefits to commercial carriers in the telecommunications industry. This appendix sets forth an application of the demand revealing approach to establishing such requirements with particular attention to the impact on the budgetary appropriations process of implementing a procedure that may generate payments in excess of the costs to the government and private parties.