Michael Spence, and Joseph Stiglitz J. Barkley Rosser, Jr.
James Madison University
This paper reviews the research related to asymmetric information of George Akerlof, Michael Spence, and Joseph Stiglitz for which they jointly received the Nobel Prize in Economics for 2001. After a recounting of their overall careers, the history of the asymmetric information idea is presented and then their key papers are discussed. This is followed by an examination of various applications of the concept, including in industrial organization and microeconomic dynamics, efficiency wage theories of unemployment, credit market rationing theory, and issues of economic development and global stability. The degree to which these latter theories can be considered to be truly Keynesian is also considered.
The author wishes to acknowledge comments and useful provision of materials by George Akerlof, Michael Spence, and Joseph Stiglitz, as well as comments by Steven Pressman and several informants who have requested anonymity. The usual caveat applies.
The 2001 Bank of Sweden Prize in Economic Sciences in Memory of Alfred Nobel was awarded to George A. Akerlof, A. Michael Spence, and Joseph E. Stiglitz for their work on the economic implications of asymmetric information during the 1970s. The press release and the presentation speech by Jörgen Weibull noted specific key papers for each recipient-- Akerlof on the market for lemons (1970), Spence on signaling in labor markets through education (1973), and Stiglitz (& Rothschild, 1976) on self-screening in insurance markets.
It is a sign of how important the economics of information has become, and the key role of asymmetric information in the economics of information, that a Nobel Prize was given largely for work on asymmetric information to William Vickrey (1961) and James Mirrlees (1971) only five years earlier.1At that time, the Nobel committee signaled that the work of Vickrey and Mirrless had important consequences and cited the later work of Akerlof as evidence of this fact. By awarding a Nobel Prize in 2001 to Akerlof, Spence, and Stiglitz, whose work comprises the hard core of what is now known as the "information economics revolution," the Nobel committee has acknowledged that understanding how information is obtained and disseminated is critical for understanding how economies function.
The careers of these economists have had numerous parallels, especially those of Akerlof and Stiglitz. All three began as theoretical economists whose work was quickly recognized as innovative and of high quality. However all three were interested in real world economic problems and would come to deal with practical policy problems, although Spence has done so more through work in court cases and consulting than through holding policy-making positions. Another difference is that while Spence largely abandoned economic research after 1980, when he moved into academic administration positions where he has had a distinguished career, Akerlof and Stiglitz both continue to produce research in a wide variety of areas within economics. Stiglitz, in particular, has generated one of the most prolific output records of any economist ever.
Spence’s research has focused almost entirely on microeconomics, while both Akerlof and Stiglitz have ranged across both microeconomics and macroeconomics, with Akerlof especially motivated by macroeconomic issues even when he has been writing more specifically on microeconomics, as in his paper on the market for lemons that was cited by the Nobel Prize committee. Akerlof and Stiglitz are among the chief developers and expositors of New Keynesian macroeconomics, which seeks to use rational expectations arguments to refute the policy ineffectiveness arguments put forth by the New Classical School during the 1970s.
The next section of the paper will examine the careers of the three 2001 Nobel Prize winners. The following section will discuss the evolution of the idea of asymmetric information, focusing especially on the work of Vickrey and Mirrlees and their predecessors. The section after that presents the main arguments of Akerlof, Spence, and Stiglitz about asymmetric information. This will be followed by a section on extensions and applications with subsections dealing respectively with industrial organization and microeconomic dynamics, New Keynesian theories of unemployment, New Keynesian theories of credit market rationing, and models of economic development and global financial stability. The final section discusses the degree to which the New Keynesian approach of Akerlof and Stiglitz can be viewed as genuinely Keynesian.
George Akerlof was born in 1940 in New Haven, Connecticut, where his parents, Swedish immigrant Gustav, and German-Jewish descended Rosalie, were both graduate students in chemistry. His father, a research chemist, was involved in the Manhattan Project, as was his maternal uncle, Joseph Hershfelder, a famous physical chemist. This family emphasis on chemistry and physical science led George to feel inferior as a youth to his older brother, Carl, who would become a physicist. Akerlof was somewhat sickly when young, and admits to having been in a circle of friends “who in today’s terminology would be called nerds.” He remembers first thinking of economics at the age of 11 when he independently discovered the principle of the multiplier while contemplating the possible unemployment of his father, an early signal of his lifelong interest in the problem of unemployment.
Akerlof received his B.A. and his Ph.D. from Yale University in 1962 and 1966 respectively, following in the footsteps of his parents and his brother. From that time forward he has been located at the University of California-Berkeley where he has been Professor of Economics since 1980. While maintaining his base at Berkeley, Akerlof has enjoyed visiting positions at numerous institutions including the Indian Statistical Institute, Harvard University, the staff of the Council of Economic Advisers, the Special Studies Section of the Board of Governors of the U.S. Federal Reserve System (where he met his current wife, Janet Yellen), the London School of Economics, and the Brookings Institution. He has served as Vice President of the American Economic Association, was its Ely Lecturer in 1990, and was Director of the National Bureau of Economic Research.
He has been coeditor of Economics and Politics, and an associate editor of the American Economic Review, Quarterly Journal of Economics, and Journal of Economic Behavior and Organization, where he is now an honorary editor. He has also been a Fellow of the Econometric Society, the American Academy of Arts and Sciences, the Institute for Policy Reform, and the American Academy of Political Science, an Associate of the Canadian Institute for Advanced Research and the MacArthur Initiative on Economics, Group Values and Norms, and a Member of the Russell Sage Foundation Roundtable on Behavioral Economics.
Besides his famous work on asymmetric information, Akerlof has published many papers on macroeconomics whose major focus has been on explaining involuntary unemployment. He has been a crucial pioneer in bringing insights from psychology and sociology into economic analysis, and in recent years has focused on broader social issues of identity and social class formation. Besides great breadth and innovativeness, Akerlof's work has also long been noted for its wit. This is demonstrated in the quite amusing titles he has chosen for his articles and books-- for example, "The economics of caste and of the rat race and other woeful tales" (Akerlof 1976).
Among his many coauthors one of the most important has been his wife, Janet Yellen, who has also served in important policy positions in Washington as a member of the Board of Governors of the Federal Reserve System and as Chair of the Council of Economic Advisers during the Clinton presidency.
2.2 A. Michael Spence
Born in 1943 in Montclair, New Jersey, Michael Spence
received a B.A. in philosophy from Princeton University in 1966, a B.S.-M.A. from Oxford University following a Rhodes Scholarship, and a Ph.D. from Harvard University in 1972. His article on education and signaling, the one singled out by the Nobel Prize committee, derived from his Ph.D. dissertation, which was published by Harvard University Press as Market Signaling: Informational Transfer in Hiring and Related Processes (Spence 1974a). He followed this initial work by applying the notion of signaling to the field of industrial organization, an area in which Spence would become a leading figure (Caves, Porter & Spence, 1980; Hayes, Marks & Spence, 1983). He provides a summary of his signaling work in his Nobel Prize address (Spence, 2002). He has also published in such areas as growth theory and natural resource economics.
Spence was Associate Professor of Economics at Stanford University from 1973-75 and then was jointly Professor of Economics and Business Administration at Harvard University until 1990 when he returned to Stanford from where he retired in 2000 to Emeritus status. Named George Gund Professor in 1983, he served as Chairman of the Business Economics Program from 1981-83, as Chairman of the Economics Department in 1983-84, and as Dean of Faculty, 1984-90. He was also Dean at Stanford from 1990-99.
Spence has been a Fellow of the Econometric Society and the American Academy of Arts and Sciences as well as serving on various corporate boards of directors. He received the John Kenneth Galbraith Prize for Excellence in Teaching in 1978 and the John Bates Clark Award in 1981. In recent years he has been involved in numerous antitrust cases, many of them involving high technology industrial enterprises.
Although his work in academic administration kept him from being involved in economic research after 1980, Spence’s academic career must be noted for its exceptional quality. The Dean of Faculty at Harvard is arguably the most powerful academic officer of that premier academic institution, as is the equivalent position at Stanford. He has been described by observers as having been “incredibly smart and hard-working” in those positions and as having been likely to become President of Harvard except for having “gotten athwart of his long-serving boss, Derek Bok.”
Despite his “earnestness” in these positions he is described as having retained “a Puckish delight in extreme sports and similar activities.” One weekend during the 1980s he escaped to a family retreat in Maine where he went skinny-dipping by moonlight in a local pond. He was attacked by a group of beavers that drove him to shore and required him to receive rabies shots. But then nothing compares to the attacks deans receive from irate faculty members. Alfred Kahn, long time Dean at Cornell, once remarked that “a dean is to a faculty what a fire hydrant is to a pack of dogs.”
2.3 Joseph E. Stiglitz Joseph Stiglitz was born in 1943 in Gary, Indiana, the
hometown of Paul Samuelson whose Collected Papers he would
edit. He received his B.A. from Amherst College in 1964 and his Ph.D. from MIT in 1967. Stiglitz was a Fulbright Scholar and Tapp Junior Research Fellow at Cambridge University in 1970. Appointed Professor of Economics at Yale University in 1969, he moved to Princeton University in 1979, where he remained until 1988. Then he went to Stanford University. From 1993-97 he served on the Council of Economic Advisers, eventually becoming its Chair. During 1997-2000 he was Senior Vice President for Development Economics and Chief Economist at the World Bank. He left there after a widely publicized dispute with then Treasury Secretary Lawrence Summers about the management of global economic policy. He currently holds a joint appointment in the Economics Department, the School for International Affairs, and the Graduate School of Business at Columbia University. Since departing the World Bank he has become a very public critic of the policies of the “Washington Consensus” on international economics.
He is a Fellow of the National Academy of Sciences, the American Academy of Arts and Sciences, the Econometric Society, and the American Philosophical Society. He has also been elected to various honorary societies in Britain, Italy, France, and Germany, and has received numerous honorary doctorates. He received the John Bates Clark Award in 1979. He has served on the boards of editors of many journals and was the Founding Editor of the Journal of Economic Perspectives.
Stiglitz is one of the most prolific living economists. His list of publications extends to 30 pages in small font and includes textbooks in various fields and at different levels as well as monographs, position papers, and articles. The breadth of his interests was reported by the American Economic Association when he was awarded the John Bates Clark Award in 1979, well before a majority of his published research, as including growth and capital to the economics of discrimination, public finance to corporate finance, information to uncertainty, and competitive equilibrium with exhaustible resources to monopolistic competition and product diversity. And this is now a very incomplete list. He has summarized his own work in his Nobel Prize address (Stiglitz, 2002).
Stiglitz has long been reported to have a tendency towards absent-mindedness and mild eccentricity, characteristics associated with academic brilliance but thought not to travel well into the arena of public policy making. At Stanford, one secretary was assigned solely to him, partly because of his immense productivity, but also to keep an eye on him to make sure that he did not tie his shoelaces together or engage in other similar acts. When he became President Clinton’s Chair of the Council of Economic Advisers he attended cabinet meetings. At an early one, prior to Clinton’s arrival, Stiglitz was sitting with his tie wrapped around the outside of his collar. Treasury Secretary Lloyd Bentsen, known to be a fashionable dresser, walked over and rearranged his tie for him.
As with many academic economists who become policy makers, Stiglitz constantly confronted conflicts between the ideals of economic theory and the compromises of practical political economy, a problem he discussed in a distinguished lecture to the Society of Government Economists (Stiglitz, 1998). An anecdote not included in that lecture brings home this conflict. During an episode of high oil prices that would lead President Clinton to decide to release oil from the Strategic Petroleum Reserve into the market, Stiglitz met with Clinton and Chief of Staff Leon Panetta to discuss the matter. Stiglitz advised against releasing the oil on the grounds that doing so would have no effect on the market and that the price of oil would come down on its own fairly soon anyway. Panetta responded to this argument by declaring that Clinton should therefore go ahead and release the oil anyway because “it won’t cause any harm and we’ll get the credit for the drop in the price of oil.” Panetta’s argument carried the day.
Development of the Economics of Asymmetric Information
As noted earlier, by previously awarding a Nobel Prize for the economics of asymmetric information to Vickrey and Mirrlees, the problem of the economics of information and the special issue of asymmetries of information had been under discussion for some time prior to the crucial breakthroughs by Akerlof, Spence, and Stiglitz in the 1970s. According to Stiglitz (1987, 2000a) early economists who evinced some awareness of information issues included Adam Smith (1776), Simonde de Sismondi (1814), John Stuart Mill (1848), Alfred Marshall (1890), and Max Weber (1925). Smith observed that as interest rates rise the best borrowers drop out of the market. Marshall observed that workers are not always paid on the basis of tasks performed because of the difficulty of observing exactly what they do, and argued that information imperfections would "greatly complicate" economic analysis. However, Stiglitz argues that none of these individuals pursued the logical implications of their arguments and viewed these problems as essentially secondary issues.
Léon Walras (1874), who wrestled with the problem of achieving a general equilibrium, failed to make Stiglitz' list. Although it has not been widely noted, the problem of tâtonnement is at least partly an information processing problem. The auctioneer gathers the various responses to the proposed price vectors in order to adjust the price vector toward the general equilibrium, all prior to any market trading taking place. In the earlier editions of his work, Walras was clearer that this was an artificial mechanism and that the problem in real markets was very serious, although he tended to downplay this in the more widely read later editions (Walker, 1996). This later attitude was predominant when Arrow & Debreu (1954) formulated their now standard version of general equilibrium theory, and essentially assumed perfect information without even discussing the matter.2
Also on Stiglitz' list is Friedrich Hayek (1945), a previous Nobel Prize winner. Stiglitz argues that Hayek did not really appreciate the issue of asymmetric information, being concerned more with prices serving as efficient information signals regarding relative scarcities. But this misses an important aspect of Hayek's work. His work on dispersed and tacit information grew out of his involvement in the socialist planning controversy. The controversy concerned whether or not socialist planners could play the role of the Walrasian auctioneer and gather sufficient information in a centralized way to achieve an efficient general equilibrium. Hayek argued that socialist central planners would never be able to plan efficiently. This argument ultimately drew on an asymmetric information concept, although he did not use this terminology. For Hayek, the central planner would never be able to learn the information that is dispersed throughout the economy in a tacit way, and therefore could never play the role of the Walrasian auctioneer. Ironically, Stiglitz himself would come to emphasize such information issues when he came to discuss economic transition problems in his book, Whither Socialism? (1994).
Although Arrow and Debreu essentially ignored the possibility of imperfect information, another Nobel Laureate, Herbert Simon (1955, 1957), emphasized it forcefully when arguing for the inevitability of bounded rationality. Simon was concerned more with problems of computability, in particular the sheer scale and complexity of information, rather than with asymmetric information in specific transactions. However, Akerlof argues in his Nobel address (2002) that asymmetric information ultimately leads us to behavioral economics, and many see Simon's work as fundamental for this whole approach.
Berle & Means (1933) were among the first to specifically identify asymmetric information as a problem for firm management. What they labeled the "problem of the separation of ownership and control" we now recognize as a canonical version of the principal-agent problem, a classic problem of asymmetric information. This problem would be put into its modern formulation by Ross (1973) and Townsend (1979). It is not surprising that standard graduate texts in micro theory tend to discuss these issues either in the same chapter (Varian, 1992, Chap. 26) or in adjacent chapters (Mas-Colell et al., 1995, Chaps. 13 and 14).
In addition to the seminal work of Vickrey and Mirrlees, a number of economists dealt with problems of information during the 1960s. It was understood that because gathering information is costly, it is optimal to be less than fully informed. Stigler (1961, 1967) argued that this was not a problem and that these were just general transactions costs that are no different from any other costs and that markets can be expected to be efficient anyway. However, Radner (1968) and Arrow (1974) both noted that imperfect information can lead to incomplete contracts with resulting inefficiencies. Williamson (1979), a student of both Simon and Arrow, identified transactions costs as a source of incomplete contracts, and Hirshleifer (1971) argued that excess incentives to search for information might arise.
Although much of this literature to the present identifies equilibria in which transactions costs are accounted for, a more fundamental problem was raised by Raiffa (1968) that remains unresolved. This is how to optimize the discovery of transactions costs or the degree of imperfect information. Ultimately this involves an infinite regress of economizing on economizing on economizing. Binmore (1987), Lipman (1991), and Koppl & Rosser (2002) have studied this problem, and Conlisk (1996) identifies it as a fundamental source of bounded rationality in economic decision making.
The Core Arguments
The key paper in the economics of asymmetric information is Akerlof's (1970) study of the market for lemons, one of the most frequently cited papers in the last half of the 20th century. The “lemons” in question are used cars. Akerlof began by noting that the owner of a car knows more about it than any potential buyer. Therefore, the used car market inevitably involves asymmetric information. Akerlof (2002) claims that he first became interested in the used car market from his interest in macro fluctuations, and from the fact that the new car market exhibits large fluctuations that contribute significantly to macro fluctuations. But he soon realized that applications of his discovery went well beyond both the used car market and basic macroeconomics.
Akerlof showed that awareness of their relative ignorance would lead potential buyers to assume that any used car would have a high probability of being low quality, a lemon. This would cause them to bid down the price of used cars in general, and this would drive most high quality used cars out of the market. Indeed, in his original theoretical model Akerlof showed that in principle only lemons would be offered for sale.
However, Akerlof observed that in the real world some people would be forced by circumstance to offer a high quality car for sale, using the example of someone transferred abroad. The inefficiency arising from asymmetric information would essentially be borne by these individuals, who would be unable to receive a sufficient price for their car because it was likely to be viewed as just another lemon. Although there are markets where repeat sales and reputation may resolve the problem, Akerlof observed that in many markets they are not easily resolved, including insurance, labor, and credit. His finding of lemons driving out good cars from the market is known as adverse selection.
Just as Akerlof noted that reputation may resolve the problem, Spence (1973) pursued this particular solution in his study of signaling in labor markets. He found that it does not always work to remove inefficiency. In his example the asymmetry is between an employer and a potential employee. Because the employer is unable to accurately discern the skills of the potential employee, she relies upon signals. The most prominent signal is the costly attainment of certain educational levels by the potential employee. It is believed that only employees who are sufficiently skilled will bear the costs of getting educated and thus sending the signal. Spence notes that the signal will only signify something real if there is a negative correlation between signaling costs and productive capability. This means (for the case of education as a signal) that grades in schools (and the availability of scholarships) must be positively correlated with productive work capability.
Spence identifies the possibility of an infinite set of signaling equilibria, each associated with a cutoff level of education (or amount of the signal), which will be based on the conditional probabilities believed in by the employers. Potential employees will be sorted as being either above or below this level and, in theory, people will either invest in the signal up to that level and no further or will not invest in the signal at all. As the level of the signal rises, the higher group will be worse off due to higher signaling costs. Although the lower group is not hurt by an increase in the signal level, it is worse off than a no-signal equilibrium in which everybody gets paid their unconditional expected marginal product.
Although it is possible to construct cases where some groups gain from signaling, very often all groups would be better off with no signaling, a clear case of a Pareto inferior outcome. A crucial part of the argument is that the cost of generating the signals lowers net aggregate product, although Spence recognizes that education may have external benefits that overcome this factor. He further notes the possibility of signaling becoming entangled in various kinds of unfair discrimination in labor markets (Spence 1974a, 1974b, 1976a).
Rothschild & Stiglitz (1976) extend the analysis by introducing the concept of screening. They apply this to insurance markets which, as noted above, are rife with asymmetric information problems leading to both adverse selection and moral hazard. The screening mechanism is to offer a variety of contracts that encourage agents to reveal accurate information about their riskiness through a process of self-selection. Thus, lower risk agents will tend to select contracts that charge lower premiums but higher deductibles. However, there is an inherent conflict between the two functions of transmitting information and redistributing risk.
Stiglitz (1975) applies this argument to the education example studied by Spence and comes up with results that are partly similar and partly dissimilar. He also identifies multiple equilibria that can be Pareto ranked and notes various costs of screening. However, he argues that the Pareto optimal solution would be a full screening that properly identifies each agent's true capabilities. But this outcome is not sustainable as a market equilibrium.