In memory of my former student and long-time friend: the late Dr. Stanley Wong
CON T E N T S
Preface xiii Acknowledgements xvii Prologue: Problems with modelling equilibrium attainment 1 P.1. A general economic equilibrium as a necessary social optimum 4 P.2. Maximization is the only behavioural assumption in neoclassical equilibrium models 8 P.3. On the road to Dynamic Stochastic General Equilibrium models 9 P.4. Outline of this book 10 PART ONE: The purpose and problems for equilibrium models 1. Equilibrium models and explanation 13 1.1. Equilibrium and explanation: Elementary considerations 14 1.1.1. Marshall’s two ‘Principles’ of explanation 15 1.1.2. Long-run vs. short-run equilibria and the role of time 16 1.1.3. Comparative statics analysis as thought experiment 17 1.2. Equilibrium implies recognition of disequilibrium dynamics 18 1.3. Equilibrium and necessary knowledge 20 1.4. Marshallian textbook explanations vs. modern economic model building 23 2. Equilibrium attainment vs. equilibrium necessities 25 2.1. Price adjustment in a formal model 26 2.2. Equilibrium attainment as an explicit process 29 2.3. Equilibrium vs. imperfect competition 32 3. Does general equilibrium attainment imply universal maximization? 35 3.1. The equilibrium actually reached by an ignorant monopolist 36 3.2. An equilibrium state as a sub-optimum 43
( vii )
( viii ) Contents
4. Time and knowledge matters for general equilibrium attainment 47 4.1. Knowledge and learning in economic models 48 4.2. Richardson on completing an equilibrium model 49 4.2.1. The informational requirements for a perfectly competitive equilibrium 49 4.2.2. The consequences of reaching a general equilibrium 50 4.2.3. Considering the disequilibrium before reaching the equilibrium 52 4.2.4. The availability of needed information in a competitive general equilibrium model 53 5. Equilibrium concepts and critiques: Two cultures 57 5.1. Two cultures 59 5.2. Equilibrium concepts involving time, dynamics and process 62 5.3. Equilibrium concerns of the formal model builders 63 5.3.1. Existence 64 5.3.2. Uniqueness 65 5.3.3. Stability 65 5.4. Concerns, beyond realism, of the critics of formal equilibrium models 66 5.4.1. Knowledge and information 66 5.4.2. Expectations 66 5.4.3. Uncertainty 67 5.4.4. Increasing returns to scale 67 5.4.5. Operational? 68 5.5. Exogenous vs. endogenous variables in equilibrium models: Cause vs. effect? 70 5.5.1. Causality among economists 70 5.5.2. Causality and economic model builders 71 5.5.3. Can economists so easily avoid causality? 74 PART TWO: The limits of equilibrium models 6. Recognizing knowledge and learning in equilibrium models 79 6.1. Modern attempts to include knowledge and learning 80 6.2. Recognizing knowledge in equilibrium models 81 6.3. Towards including realistic learning in economic equilibrium models 84 6.3.1. Does learning matter in the model? 84 6.3.2. What role do probabilities play in the model’s decision maker’s learning process? 85 6.3.3. Does the equilibrium model involve agents’ making decision errors? 86
Contents ( ix )
6.4. The problem of maintaining methodological individualism in equilibrium models 90 6.5. The problem of the compatibility of general equilibrium and psychologism 93 6.5.1. Multiple equilibria? 95 6.5.2. Psychologism 97 7. Limits of equilibrium methodology: An educational dialogue 99 7.1. A dialogue in an Economics 101 class 100 7.2. The stability in an equilibrium model must be endogenous 105 8. Equilibrium models vs. realistic understanding 107 8.1. Equilibrium attainment and knowledge sufficiency 108 8.2. Equilibrium models and the ignorant consumer 112 8.3. The market’s equilibrium price: Learning vs. knowing 115 9. Macroeconomic equilibrium model building and the stability problem 117 9.1. Rational expectations and macroeconomic equilibrium models 118 9.1.1. Rational expectations in a microeconomic context 120 9.1.2. The Rational Expectations Hypothesis in macroeconomic equilibrium models 122 9.1.3. The Rational Expectations Hypothesis and various ideas about learning in macroeconomics 123 9.2. Stochasticism and macroeconomic equilibrium models 125 9.2.1. Rational expectations vs. bounded rationality 125 9.2.2. Rational expectations and Bayesian learning 126 9.2.3. Rational expectations and econometric learning 128 9.3. Instrumentalism and the use of stochasticism in equilibrium models 129 10. Equilibrium models intended to overcome limits 131 10.1. The alleged limits of general equilibrium models 133 10.1.1. Dynamics vs. time 135 10.1.2. A list of other short-comings of the Arrow-Debreu general equilibrium 137 10.2. The current attempts to overcome the limits of general equilibrium models 140 10.2.1. DSGE models to the rescue? 141 10.2.2. Recent unrealistic efforts in formal equilibrium model building to address limitations 143 10.3. Three empirical alternatives to Walrasian general equilibrium models 144
( x ) Contents
11. Equilibrium models vs. evolutionary economic models 147 11.1. Darwin and evolutionary economics 150 11.2. Non-Darwinian theoretical evolutionary economic models 153 11.3. Alternative views of evolutionary economic models 155 11.4. Going beyond the evolutionary theory of the individual 157 12. Equilibrium models vs. complexity economics 159 12.1. Complexity economics 159 12.2. Technology, increasing returns and evolution 162 12.3. Diversity, learning, path dependency and evolution 163 12.4. Learning and ‘inductive reasoning’ in Santa Fe: A critique 163 PART THREE: Avenues for overcoming the limits of equilibrium models: Some methodological considerations 13. Building models of price dynamics 173 13.1. The analytical problem of price adjustment as presented by Arrow in 1959 175 13.2. Closure of the formal equilibrium model 176 13.3. Toward closure through posited ignorance 180 13.4. Exogenous convergence to equilibrium with forced learning 182 13.5. Endogenous convergence to equilibrium with autonomous learning 185 14. Building models of non-clearing markets 189 14.1. Unintentional disequilibria 190 14.2. Endogenously deliberate disequilibria: Keynes-Hicks generalized liquidity 193 14.2.1. A macroeconomics textbook’s simple Keynesian macroeconomics equilibrium model 195 14.2.2. Choosing not to consume in simple Keynesian macroeconomics equilibrium models 196 14.3. Deliberate disequilibria vs. methodological individualism 202 15. Building models of learning and the equilibrium process 205 15.1. Learning vs. knowledge in equilibrium models 208 15.2. Learning and methodological individualism 209 15.3. Learning without psychologism or any quantity-based theory of knowledge and learning 212 15.4. Equilibrium stability and active learning 217 15.5. Are macrofoundations needed for equilibrium microeconomics? 221
Contents ( xi )
15.6. Conjectural knowledge and endogenous expectations 222 15.7. Generalized methodological individualism 223 Epilogue: Prospects for changing equilibrium model building practice in economics 229 E.1. Behavioural and experimental economics and equilibrium models 229 E.2. Behavioural and evolutionary economics as alternatives to equilibrium models 230 E.3. Complexity economics and equilibrium models 231 E.4. Teaching with evolutionary or complexity economics 232 E.5. The issue of learning must be dealt with 233 Bibliography 237 Names Index 249 Subject Index 253
PR E FAC E
In the mid-1960s I did my graduate work in a federally financed program that was created to promote and develop what was then high-tech mathematical model building. The main textbooks I used included the large 1958 book, Linear Programming and Economic Analysis, by Robert Dorfman, Paul Samuelson and Robert Solow and the small 1957 book, Three Essays on the State of Economic Science, by Tjalling Koopmans. Except for a few elective courses and a couple of history of economic thought classes, all of the required courses involved mathematical model building or analysis. The extent of the federal financing was significant since each year of the program provided a generous tax-free three-year fellowship to three new students and it provided a salary for one professor. I say generous because when I took my first job, I had to take a pay cut. In retrospect, it seems that all of the models we were learning about were equilibrium models –usually Walrasian general equilibrium models. And as such we were learning about existence and uniqueness proofs, stability analysis, and similar issues. We were never required to actually read Léon Walras’ 1874 famous book, Éléments d’économie politique pure, ou théorie de la richesse sociale; all we were told about Walras was that he engaged in general equilibrium model building. I say ‘we were told’ to indicate also that we knew nothing of its history. Our two history of thought classes were devoted mostly to learning about economics literature of the eighteenth and nineteenth centuries with little if any mention of Walras. Based on the discussions in the theory classes, and without giving it much thought, I came away with the false impression that Walrasian general equilibrium model building was central to the study of economics from the beginning of the twentieth century. I raise these strange observations to emphasize a point which Roy Weintraub raised in a 2005 article about the concept of an equilibrium and about the criticisms of equilibrium model building. As I will discuss in Chapter 5, Roy’s point is a very important point but those of us in the graduate program in which I was involved would not have understood what Roy was talking about. The point Roy was making was that there are two very different perspectives about equilibrium and equilibrium models –which I will call two cultures in
( xiii )
( xiv ) Preface
economics. One includes those who learned about the concept of equilibrium before, let us say, 1950, and the other includes those like me and my fellow high-tech PhD students in the 1960s. For us, equilibrium was a property of a mathematical model and we had only a vague idea that it was also supposed to be something about the real world we could see out our windows. For the pre- 1950 culture –which was dominated by Marshallian economics –equilibrium was thought to be a claim about what we eventually would or should see in the real world. And the difficulty with all this is that these two cultures both talk about or criticize theoretical states of equilibrium but they are not really talking about the same thing. One of my tasks in this book is to sort these things out so that we can all benefit from each others’ criticism. In my 2014 book on economic model building, I addressed a different schism, the one between today’s model builders and those of us who learned decades ago about model building as I did when I was a graduate student. In that book I explained that models and theories were seen as two different things –specifically, we thought that the purpose for a mathematical model was to represent some given economic theory and thereby possibly provide some logical rigour to the theory. When I began working on that book I talked about model building with my colleagues, some young and some old. What I quickly learned was that the young colleagues did not see models as I did. For them the idea of a model was interchangeable with the idea of a theory. By means of a short survey I determined that roughly the year 1980 divided the younger view from my older view. My 2014 book was directed at trying to bridge these two cultures concerning what constitutes a model in economics. Interestingly, for that schism between the older and younger views of the relationship between theories and models, I was a member of the older side. But in the present book, which will be addressing the schism that Roy identified concerning the concept of an equilibrium, thanks to my training in the high-tech graduate program, I became a trained member of the younger side of Roy’s schism. Ironically, despite the best efforts of my graduate instructors, once I began teaching the ubiquitous Economics 101 class I realized how useless my graduate training was when it came to understanding the real world so that I could teach about it. Early on I deviated from my training and began teaching my students about equilibrium as something about the real world much like the older side of Roy’s schism did. In the process I discovered Joan Robinson and read many of her criticisms of the work of the newer side of Roy’s schism. When I later got to teach the fourth-year advanced microeconomics theory seminar I started looking at some interesting articles that were about how the concept of an equilibrium was problematic in economic explanations. As it turned out, these critical articles were all addressing problems with formal equilibrium models. Moreover, thanks to many of my critical students I learned a lot about economics and economic model building by later teaching an advanced micro
Preface ( xv )
theory seminar and then even more when I began teaching a graduate micro class. I think what I learned in those classes I should have learned in graduate school. As a result, I have decided that this book will be about what I learned with my students about equilibrium concepts and equilibrium models. This book will be addressing recognized problems with equilibrium models particularly from the perspective of standard economics textbooks that use equilibrium models as a basis for explaining prices or forming economic policies and especially in teaching beginning students the virtues of the competitive market. Of particular concern will be how economics textbooks almost always fail to recognize any problems with equilibrium models even though these problems fundamentally distort realistic economic explanations. So, as I go along and whenever possible, I will try to point out ideas and criticisms that are relevant today but have their origin in the ideas published by economic model builders decades ago. While my main interest is in what we teach students, eventually what will be considered here might also enable us to explain why most governments’ policy makers are failing to provide effective help dealing with real world economies. After all, most governmental economic policy makers likely were once students in an Economics 101 class. In 1986 I published a methodology book that was also about what I learned teaching both advanced and graduate microeconomics theory classes. That book proposed to offer a new methodological perspective for addressing some fundamental problems with common microeconomic models. Unfortunately, almost all of the problems I discussed there still seem to persist in microeconomic model building today, particularly with those that rely on using the analytical properties of equilibrium states. While in this book I will be dropping most of the methodological concerns of that book, I will be returning to many of the theoretical problems I discussed then, but this time by focusing instead on recognized problems involved in building equilibrium models. While methodology will play a much lesser role than it did in the 1986 book, it will be addressed briefly in Chapter 6 and a bit more in Part III, where I discuss how common methodological presumptions constrain any attempts to solve the problems I discussed in Parts I and II. I have written this book for readers interested in learning about the main tool economists use to help understand the economy. Such readers include undergraduate and graduate students, of course. But I also hope readers who may not have taken the proverbial Economics 101 –or, if they did, do not remember much from that class –will still find this book useful. For these readers I will occasionally add footnotes to help with the usual economists’ jargon that one would have learned in that class. And most important, it is this group of readers in which we will find people employed as governmental advisors and policy makers –in particular, people who should be asking economists about the assumptions that were used to reach the advice they are giving advisors and policy makers.
AC K N O W L E D G E M E N T S
I have received a considerable amount of help in the form of criticisms of early versions of chapters for this book. For this help I wish to thank my former students Senyo Adjibolosoo and David Hammes as well as colleagues Brian Krauth, John Knowles, Ken Kasa and Luba Peterson and friends Pedro Garcia Duarte and Mark Donnelly. Also, I thank Duncan Foley for helping me with Chapter 12 and Kenneth Arrow and George Richardson for answering my questions about their articles discussed in Part I. Also note that I have made use of parts of several chapters from my 1986 Methodology for a New Microeconomics: The Critical Foundations that was published by Allen and Unwin. None of those chapters are reproduced here as in each case the material I have used has been heavily revised as well as updated. Any reader interested in that 1986 book can now obtain a 2014 Routledge Revivals edition published by Taylor and Francis publisher.
( xvii )
Equilibrium Models in Economics
Prologue Problems with modelling equilibrium attainment
he idea of a state of equilibrium pervades economics research. For a definition of an equilibrium one can easily find one with a Google search and see something like this found in the 2016 edition of The American Heritage Dictionary of the English Language: ‘e·qui·lib·ri·um: A condition in which all acting influences are canceled by others, resulting in a stable, balanced, or unchanging system’. If one also looked for how a well-known economist might view the notion of an equilibrium in economics, no better example to be found would be the view of Frank Hahn: ‘Whenever economics is used or thought about, equilibrium is a central organising idea. Chancellors devise budgets to establish some desirable equilibrium and alter exchange rates to correct “fundamental disequilibria” ’ [1973, p. 1]. Critical analysis of equilibrium models is not a new topic. Of particular interest for any consideration of claimed limits to equilibrium-based explanations are three separate and different challenges presented in, ironically, the same year –namely 1959. These challenges focused on equilibrium models of the market that were very common in economics texts then and are still common today, particularly in undergraduate textbooks. One of the three 1959 articles was by Kenneth Arrow who simply pointed out that in the usual equilibrium market model (such as that illustrated in Figure P.1) it is not enough to explain a market equilibrium price (Pe) as the one at an intersection between the demand curve and the supply curve. He insisted that one must also explain in the model why that equilibrium price would be achieved. As will be explained in Chapter 2, the problem is that even to discuss a market-determined equilibrium price one needs to recognize that a stable arrangement of the market’s ( 1 )
( 2 ) Equilibrium Models in Economics
Figure P.1. Simple market equilibrium
demand and supply curves is required within one’s model. But, as Arrow also recognized, a stable arrangement is not enough because with the usual textbook models we are never told how any market participant, say the supplier, knows the market’s demand curve or at least knows when to lower the price and by how much. Similarly, how does the demander know when to bid up the price or know by how much? Textbooks just rely on some vague form of allowance of a sufficient amount of time but never say how much time this would take. As to how the price was determined within the model, Arrow suggested one possible solution for this problem of adequate explanation was to recognize that the usual textbook discussion of an imperfect competitor does involve at least a supplier setting the price1 –but, of course, this would require recognizing the supplier’s knowledge, and learning or at least identifying the available information needed to make such a decision. Ironically, in 1959 Robert Clower published an equilibrium model about a different problem but one that in effect directly addressed Arrow’s suggestion.2 Clower’s model was of an ‘ignorant’ monopolist for which allowance is made for the obvious fact that the monopolist could not possibly know the whole market demand curve it faces but instead would have to make assumptions about it. For Clower, the ignorant monopolist would at least have to 1. For those unfamiliar with economics jargon, perfect competition refers to a type of market in which no individual has a significant role or effect on the determination of the equilibrium price and imperfect competition means that individuals can affect the price. Textbooks distinguish between these two types of competition by claiming that perfect competition will exist whenever there are too many participants for any one to have an effect and competition is imperfect whenever the number or buyers or sellers is small enough that every participant can have an effect because any change in their behaviour affects either the market’s demand or the market’s supply. This distinction plays a big role in textbooks’ definition of markets and market behaviour. 2. I say ‘ironically’ because I asked Arrow (in January 2014) if he was aware of Clower’s article in 1959 and he said he was not.
P r ol o g u e ( 3 )
make assumptions about the shape and position of the market’s demand curve. Based on those assumptions, Clower’s monopolist would send a chosen supply quantity (presumed to be a profit-maximizing quantity) to the market and wait to see what market-clearing price is obtained.3 If the assumptions about the nature of the demand curve are true –such that the implied marginal4 revenue for the supplied quantity would be equal to the marginal cost for that quantity –there would be no problem.5 But there is no reason to think the monopolistic supplier has the required knowledge to assure that the assumptions made about the market’s demand curve are true. As will be explained in Chapter 3, the result is a model in which an apparent equilibrium price may be reached but it is one at which the monopolist is not actually maximizing profit even though the monopolist is erroneously thinking it is. And again ironically in 1959, George Richardson presented a model of the competitive market where it would seem that the only way to guarantee the attainment of a market’s equilibrium price is to introduce some form of collusion.6 As will be explained in Chapter 4, if Richardson is right, then this necessity would obviously fundamentally challenge what is commonly taught in ‘Economics 101’ class.7 Richardson recognized that Friedrich Hayek in a 1937 article had already raised concerns about the knowledge requirements for the achievement of a market’s equilibrium. It turns out this was preceded by a 1933 lecture in which Hayek suggested there was an even more fundamental problem concerning the information available to an investor.8 Hayek was concerned that it is too
3. For those readers who have never taken an economics class or do not remember much from of what they heard in their economics class, maximizing profit just means maximizing sales revenue net of production costs. 4. For those not remembering economics jargon, the word ‘marginal’ is just jargon for the following idea. If one is deciding about increasing the amount to produce of some commodity, marginal refers to the consequences of an increase of one unit of that commodity. In the case of revenue, it is how much more revenue is obtained if one sells one more unit of the commodity in question. Marginal cost would then similarly be about the change in total production cost if one more unit is produced. 5. Again, for those not remembering their economics jargon, this equality is just a matter of whether profit (sales revenue net of production costs) is maximized. For it to be at maximum, calculus textbooks tell us there must be an equality of marginal revenue and marginal cost. If they are not equal, then a gain in net revenue is possible by producing either more or less depending on whether the difference between them is positive or negative. 6. I also asked Arrow if he was aware of Richardson’s article when he wrote his own article and again he said he was not. So then I asked Richardson (through his son Graham) if he was aware of either of the articles by Arrow or Clower and he said he was not. 7. This is the jargon name given to the usual beginning economic principles class offered in university and college programs. 8. Hayek delivered this lecture on December 7, 1933, in the Sozialökonomisk Samfund in Copenhagen and which was first published (in German) in the Nationalökonomisk
( 4 ) Equilibrium Models in Economics
easy for governments, through a policy (perhaps by manipulating the interest rate in the money market), to actually give false information unintentionally to investors in the current capital equipment market and thereby cause disequilibria in future product markets. While Hayek may have raised the issue as a criticism of government intervention when facing the problems of the Great Depression in the 1930s, the issue he raised does recognize the limitation of models of the market that do not recognize how decision-makers in the model know what they need to know to assure the achievement of equilibria when investment decisions are involved. This is particularly the problem when a firm is placing an order for capital equipment which of necessity takes time to produce. Perhaps by the time the equipment is delivered to the firm the market does not resemble one that the investor might believe would be there as suggested by what the change in government policy promised.9
P.1. A GENERAL ECONOMIC EQUILIBRIUM AS A NECESSARY SOCIAL OPTIMUM In the Richard T. Ely Lecture to the American Economics Association’s 1994 conference, Arrow [1994, p. 4] told us about what he called the prototypical economic model: The prototypical economic model … is general competitive equilibrium. Individuals and firms take prices as given. Individuals choose consumption demands and offers of labor and other assets, subject to a condition that receipts cover expenditures. Firms choose inputs and outputs subject to the condition that the outputs be producible given the inputs. How they make these choices depends on many factors: tastes, attitudes toward risk, expectations of the future. But, it is held, these factors are individual.
As Adam Smith’s eighteenth-century view of the world would have us believe, we should never depend on authorities such as the Church or the state since the ‘best of all possible worlds’ will be achieved when everyone is independently and autonomously pursuing self-interest and nobody is inhibited in that pursuit except by the limits imposed by Nature. With this in mind, let us examine the world where everything about the economy is a matter
Tidsskrift, vol. 73, no. 3, 1935, and later (in French) in the Revue de Science Economique, Liège, October 1935. 9. Perhaps it should be noted that the going interest rate in a state of long-run or general equilibrium is sometimes seen to indicate the equilibrium growth rate of the economy –see John von Neuman [1937/45].