Using information markets to improve public decision making.

• Author: Hahn, Robert W.
• Position: Twenty-Fourth Federalist Society Student Symposium, Law and Freedom

1. INTRODUCTION II. AN INTRODUCTION TO INFORMATION MARKETS III. GETTING BETTER INFORMATION FOR MAKING POLICY CHOICES A. A New Approach B. Extending the Framework C. The Potential for Improving Fairness D. Potential Problems with This Approach 1. Versatility of the Approach 2. Project Governance 3. Measurement Issues 4. Market Design Issues IV. A BENEFIT-COST ANALYSIS OF INFORMATION MARKETS A. A More Informed Assessment of Policy Proposals B. Greater Transparency and Accountability in Decision Making C. Greater Availability of Assets for Financing Projects and Spreading Risks D. Cost of Information Markets E. Comparison of Direct and Indirect Approaches for Information Markets V. SIGNIFICANCE FOR POLICY DESIGN AND EVALUATION A. Information Markets in the Policy Process B. Solving Some Difficult Government Oversight Problems C. The Potential Role for Government and Researchers VI. CONCLUSION VII. APPENDIX I. INTRODUCTION

   Many legal scholars have studied how to improve public decision making. Justice Breyer, for example, argues that technical problems could benefit from greater scientific expertise, and suggests using such analysis to help prioritize among competing social needs. (1) Cass Sunstein argues for the judicious use of cost-benefit analysis in a variety of areas, but also points out its limitations. (2) Sunstein's proposal, and the proposals of other scholars, would rely heavily, albeit not exclusively, on cost-benefit analysis to evaluate public policy decisions. (3)

   Cost-benefit analysis is a tool used by decision makers to help inform the policy process. Cost-benefit analysis examines how different policies affect the overall level of net benefits to society, or benefits minus costs. A cost-benefit analysis may also be used to explore equity issues, examining how the distribution of net benefits varies across key groups, such as minorities or small businesses. (4)

   A fundamental problem with cost-benefit analysis of new policies is that the analysis is conducted before such policies are implemented. When conducting ex ante analyses, it is difficult to predict the future values of key variables that could be affected by a policy. (5) For example, an analyst might predict that a worldwide carbon tax of $100 per ton would reduce world GDP by 1% in 2010. (6) How confident should we be in such a prediction?

   In this paper, we present a new framework for addressing such uncertainty; this framework has the potential to substantially improve public decision making. We argue that decision makers can be more confident in analytical results if these results are based more directly on market data. Our framework introduces "information markets" that allow people to profit from superior knowledge about the future. (7) For example, if an information market suggested that expected GDP would fall by 1% with a carbon tax, (8) this estimate would theoretically incorporate all publicly available information about that policy's effects. We also argue that if these information markets are designed well, information from the prices in these markets is likely to be much more accurate than other forecasts.

   An information market allows individuals to purchase contracts, using real money, that yield returns to their owners contingent upon the uncertain outcome of a future event. (9) With the advent of the Internet, information markets are becoming more common. They are used in a number of contexts, ranging from assessing the likelihood that the Federal Reserve will raise interest rates to assessing the odds that a particular presidential candidate will be elected.

   As an example, consider the online exchange at TradeSports.com. This exchange allowed its members to trade contracts that yielded $10 to their owners if President Bush was reelected in November 2004. The contracts yielded $0 if Bush was not reelected. The prevailing price of these contracts on July 6, 2004 ($5.42) (10) revealed the price at which the supply for Bush contracts equaled the demand for these contracts. Assuming that the market was efficient, this price suggested the probability of Bush's reelection was 54.2% on that date. (11)

   The idea of making greater use of information markets to promote social objectives is not new. Indeed, several scholars have suggested using information markets in a number of different contexts. Robin Hanson suggests that governments use information markets to identify whether particular policies will improve national welfare; he proposes relying on such markets exclusively, for example, when information markets predict a policy will increase GDP. (12) Michael Abramowicz also notes the potential of such markets and advocates using them to predict the results of a cost-benefit analysis that would be done in the future by a designated expert. (13) Ronnie Horesh offers a novel proposal for creating incentives for private investors to aid the government in achieving social policy objectives: government-issued "Social Policy Bonds" that pay a fixed amount after a certain performance objective is met, such as achieving a particular unemployment rate. (14) His proposal has some similarities to ours, but does not address how to obtain information on the costs and benefits of a policy prior to implementation. (15)

   To this point, no one has explicitly linked the improved information obtained through these markets to the subsequent implementation of policies that can maximize net benefits. Our central contribution is to propose an efficient way to implement well informed policy decisions. (16) Our hope is that this general approach will induce the government to make more efficient decisions--and, in particular, curb the appetites of politicians for introducing policies that have costs far exceeding their benefits. (17)

   Our new approach builds upon the literature on information markets as well as the literature on the design of efficient policy mechanisms. Our claim is that the prices in information markets can inform the mechanism design process, thereby making previously infeasible mechanisms feasible for the policy maker. Specifically, information markets substantially broaden the scope for offering pay-for-performance contracts that yield useful information on the net benefits of a policy.

   First, we show how it is generally possible to design markets whose prices will convey useful information on the costs and benefits of a number of policy choices, ranging from regulation to public works projects. Second, we describe one way of providing incentives for self-interested agents to implement policies that maximize net social benefits. Third, we show how information markets can be used to provide a stronger foundation for implementing a variety of government oversight mechanisms that, up to this point, have been stymied because of difficulties in estimating costs and benefits. We also show how legislators can use traditional budgetary controls in conjunction with information markets to exercise more effective oversight.

   Finally, we identify and analyze the strengths and limitations of using information markets to help improve policy. While we focus on public-sector decision making, our analysis also holds for cost-benefit analyses used to inform private-sector and not-for-profit sector decision making.

   This Article proceeds as follows: Part II provides a brief introduction to information markets. Part III provides a template for designing contracts that can supply better information on costs and benefits. Part IV discusses key costs and benefits of using information markets. Part V addresses the significance of these contracts for policy design and evaluation, and suggests a new approach to legislative and regulatory oversight. Part VI presents our conclusions.

2. AN INTRODUCTION TO INFORMATION MARKETS

   All markets can be thought of as providing some kind of information. We use the term "information market" to denote a market for contracts that yield payments based on the outcome of an uncertain future event. (18) In this paper, we specifically consider contracts that yield financial payoffs to their owners contingent on the status of key policy variables. The relevant policy decisions could be private or public, but this paper focuses on public decisions. Information markets could provide information related to costs, benefits, net benefits, or the likelihood that a certain event will occur, such as the probability that a President is reelected. (19) They can also be used to address potential market failures in the provision of information on public policy matters. (20)

   Information markets differ from traditional equity markets in that they are not typically tied to a claim of an ownership stake in a firm. Instead, the assets are claims that will pay off an amount that depends upon the state of the world, such as the monetary value of actual policy benefits. (21) Although information markets for claims on benefits do not correspond directly to traditional equity markets for claims on corporate profits, there is a clear analogy. By monetizing policy benefits, information contracts allow organizations implementing policies to transfer risks from these policies, just as traditional equity contracts allow corporations to transfer risks from their projects.

   Information markets have already been used in a variety of contexts. The most well known information markets are for small-stakes political contracts; researchers at the University of Iowa, for example, conduct an electronic market for political futures contracts. In the corporate world, Hewlett Packard has experimented with information markets to forecast sales, while Eli Lilly has used these markets to help predict the success of pharmaceuticals. TradeSports.com offers information contracts in a number of areas including sports, politics, finance, law, entertainment, and even the weather. Goldman Sachs supports an exchange called economicderivatives.com, which hosts call auctions...