Dani Rodrik is a brilliant progressive neo-classical economist and he has made a very constructive sideline contribution to this TPM Cafe.discussion from his blog page at Harvard University. In a post titled “If you are a progressive, you’ve got to love neo-classical economics” he expands on arguments made by Brad DeLong in his post “It’s different for Lefties and Righties”. The essence of Rodrik’s position seems to be that neo-classical economics provides a rigorous framework for studying economics, and therefore you’ve got to love it and do it.

I may be wrong, but to me that is a little bit like saying poodles are dogs, and therefore all dogs are poodles. I don’t buy that logic.

I agree economic analysis should be rigorous, and neo-classical economics is one way of doing it rigorously. However, there are also other ways of doing rigorous economics. Moreover, there are lots of deep flaws and unrecognized pitfalls in the neo-classical program that call for corrective alternative research programs. These seem to be the essential points of contention in this debate, and that leads back to the basic questions what forms of rigorous economics are allowed, what not, and why?

What follows is a long detailed post substantiating this argument and claim.

The benchmark framework for modern neo-classical (orthodox) economics is the Arrow-Debreu model of competitive general equilibrium. Under a host of extraordinarily restrictive assumptions regarding (1) the nature and structure of markets, (2) the behavior and capacities of individuals, and (3) the behavior and structure of firms, it can be shown there exists at least one market equilibrium that is “efficient” in the limited sense that one cannot make someone better off without making someone else worse off – an outcome economists term “Pareto efficient”.

The list of assumptions includes perfect competition, which means the complete “absence” of economic power by buyers or sellers (not just a balance of power), perfect information among all buyers and sellers, and costless transacting so that buyers and sellers can flit between each other to prevent exploitation. That is like saying I can switch jobs between New York and LA without any costs, be they relocation costs or costs to my reputation as a loyal worker.

Individuals must also be fully rational, unconcerned by what others do (i.e. no status issues), and excellent at math. The last requirement is because they need to be able to figure out millions of relative prices since all goods and services are available for purchase and sale in all states of nature, now and in the future (i.e. including World War III which might yet occur). The law must also be able to enforce contracts made covering these future contingencies. Moreover, this enforcement is done at no cost, or if taxes are needed to cover enforcement costs they have no impact on incentives (i.e. no supply-side effects).

Additionally, firms must have the sole goal of maximizing profits and be devoid of any corporate structure. That means there are no problems with management insiders doing things to enrich themselves (i.e. there are no Enrons or Worldcoms or stock backdating scandals).

Even assuming these outlandish conditions are met, there is still need for further restrictions regarding the shapes of demand and supply curves and restrictions on the process of transacting in markets (no sales are final until all equilibrium prices have been established) to get to the equilibrium. Absent these further restrictions, the economy may be unstable and unable to reach equilibrium even though it theoretically exists.

Now suppose we grant all of the above, in that case there does exist a competitive market equilibrium that has properties similar to those described more loosely by Adam Smith’s invisible hand. However, even then the equilibrium may be wholly undesirable from a societal standpoint in that someone could end up with almost everything and the rest with almost nothing.

Rodrik interprets orthodox “righties” (to borrow Brad DeLong’s language describing politically right neo-classical economists) as viewing this Arrow – Debreu description of the economy as a pretty good description of reality. The assumptions about markets, individuals, and firms are roughly met, as are the conditions needed to ensure that equilibrium is reached. Consequently, real world economies roughly correspond to the Arrow-Debreu model, and the model can be used to provide a good guide for economic outcomes and economic policy.

I would add that moderate “righties” also tacitly assume the distributional outcome is socially acceptable, so that the market process generates outcomes that are socially sustainable and not subject to political upheaval. Extreme “righties” don’t even care about that and believe whatever outcomes markets generate is okay. Contract is contract and the initial endowments and resources people start with are no concern of society. Ergo, the state should enforce whatever outcomes markets generate.

Rodrik then continues that orthodox “lefties” think the assumptions of the Arrow-Debreu model are not met in practice. That means the real world operates in the land of the “second best” rather than the “first best”, and in the land of the second best there is lots of room for progressive policy to correct market failures. That is also Paul Krugman’s point in “Heterodox errors”.

Additionally, orthodox lefties believe it is legitimate to have state sponsored income redistribution that corrects for initial endowment inequalities and market failures (e.g. exploitation owing to lack of perfect competition). However, such redistribution can have economically deleterious effects, giving rise to debate about how worthwhile it really is.

Heterodox economists agree with the orthodox lefty observation that the real world does not look anything like the Arrow-Debreu model. Indeed, the Arrow-Debreu “mathematical proof “of the potential theoretical existence of competitive general equilibrium is taken as real world disproof. That is because it is absolutely obvious that the necessary conditions for equilibrium do not exist in the real world and never can. Heterodox economists also largely agree that state sanctioned redistribution is legitimate – though Hayekians and Von Misians from the Austrian school of economics may differ.

Beyond these points of agreement orthodox lefty neo-classicals and heterodox economists begin to diverge. Orthodox lefties (e.g. DeLong, Krugman, and Rodrik) believe the Arrow-Debreu framework not only provides a good starting point for thinking about the economy, it is the only point. They therefore use simplified versions of it and begin peeling away particular assumptions to see what develops. Historically, the focus was on market failure – monopoly, imperfect competition, natural monopoly, public goods, and externalities.

Akerlof and Stiglitz pushed the paradigm further, investigating the implications of incomplete and asymmetric information. Akerlof has also focused on introducing richer constructions of psychology by playing with agents’ tastes and preferences (i.e. introducing new arguments such as effort, fairness, and identity into utility functions). That was the essence of his 2007 speech to the American Economic Association that was both highly praised and criticized.

This research paradigm has been productive and yielded useful insights. I am a heterodox economist and have used it myself to write papers examining what happens when people have altruistic concerns for others; preferences for wealth accumulation rather than consumption; concerns with relative consumption standings and distributional fairness; or when managers engage in strategic behavior such as following the herd or short-term thinking.

However, this fruitfulness does not justify neo-classical economics being the “only” paradigm of rigorous research. Other research paradigms can also yield fruitful insights. Moreover, there is need for awareness of how using the Arrow-Debreu model, even if modified for an imperfection here or there, can distort understanding. So much of the model is fantastically unrealistic that making one or even several pieces more realistic, may still leave biased and distorted insights.

Let’s excavate these arguments a bit. By their own admission, orthodox lefties see the Arrow - Debreu model as deeply unrealistic. I would go a lot further and say that it is drawn from Alice in Wonderland. That immediately poses the question of why start a research paradigm in Wonderland and build out from there? Isn’t it sensible to also have accompanying research programs whose starting point is closer to the real world?

The orthodox leftie neo-classical research program involves taking simplified parable versions of the Arrow – Debreu model and then making one small change after another – perhaps to the utility function, perhaps to the agents’ information sets, perhaps regarding firms’ market power or managers’ behaviors. The problem is that when you start in Wonderland, pieces of Wonderland stick with you and distort your vision even though you may try to shake off other pieces.

Nowhere is this clearer than the normative thinking that attaches to orthodoxy. Orthodoxy sells the story that, at a first approximation, free market economies generate Pareto efficient outcomes (i.e. they generate the best possible outcomes given the conditions the economy initially has and the constraints it is subject to). That is an enormously strong normative endorsement of markets.

But how plausible is this normative endorsement that so dominates public thinking? Behind it lies the triple assumption that individuals maximize their well being through freedom to choose, firms maximize profits, and markets are perfectly competitive so that there is no exploitation. Together these assumptions ensure that all agents have done the best that could be done under the circumstances. However, take away any one of them and the normative efficiency assertions about free markets evaporate.

Outside of Wonderland it is doubtful any of these assumptions hold. To quote Yogi Berra: “In theory there is no difference between theory and practice. In practice there is.” Perfect competition is a fiction. Real world transacting involves costs, and economic power and bargaining are the hallmarks of almost all economic dealings.

Firms do not maximize profits. Indeed, firms do not do anything. People do things acting within legal structures we call firms, and there is little reason to believe that this collectively organized activity maximizes profits. That leads in the direction of the economics of John Kenneth Galbraith. Moreover, it is difficult to know what it even means to maximize profits in a long-lived firm. That is why Japanese, German, and American firms differ because they have different constructions of the bottom line – though the imperialism of American neo-classical economics, channeled through business schools and MBAs, may be changing that.

Maximization of individual utility is also problematic. Personally, I have less difficulty with self-interest and rationality, which have been the historic focus of both heterodox and orthodox leftie critique. That is because it hard to know how else to systematically talk about individuals – especially if one is committed to liberty and democracy whose justifications implicitly rest on individuals being able to make rational choices. Assuming individual self-interest and rationality still leaves room for ex-post regret due to mistaken decisions or unwitting computational error. It also leaves room for inclusion of altruism and status concerns as factors that affect self-interest.

The real problem for individual choice is individuals do not live as atoms. Instead, they live in social units – families, communities, and nations. In those social units decisions are not taken in isolation, but are negotiated through socially established processes. That means cultural rules and norms governing the process matter, as do power arrangements within these social units. This is the part of feminist economics that is most subversive to the orthodox market paradigm.

It is important to recognize that none of this denies the power and creativity of the market system. That is not the issue, and heterodox economists fully acknowledge the power of the market. If you do not believe me, read Karl Marx. What is at issue is the way economists describe and study the market system, and the normative characteristics that orthodox economics attribute to market outcomes.

There is one place that even orthodox lefties dare not go. That untouchable place is marginal product theory of income distribution, which basically says that competitive markets ensure that people are paid their contribution to production. This theory provides both a justification and an explanation of income distribution.

These days it is being invoked to justify CEO pay. The argument is that CEOs are superstars whose talents are such that they multiply everyone else’s productivity, and CEOs therefore get paid part of that productivity multiplication. The sky’s the limit to this argument as corporations are getting bigger through mergers, so CEOs now have a bigger pool of workers and capital on which to work their productivity multiplier magic.

There are two critical parts to marginal productivity theory.. The first is competitive markets, which prevent exploitation – if you’re exploited you can move to work for someone who will be agreeable to paying you what you are worth. Take perfect competition away, and exploitation can occur. Orthodox lefties recognize this possibility, but most of the time (judging by journals and textbooks) that recognition is professionally suppressed.

The second part of the hypothesis is that you have to be able to objectively measure a workers contribution to production, and that is impossible in the vast majority of situations. Measurement is intrinsically subjective - one man’s art is another’s inkblot. That means it matters a lot who gets to measure and decide upon criteria for pay. Bosses and those on the boss ladder think they contribute a lot. They also have a say in pay determination, and they tend to get paid more. That said, it does not mean they can be paid any amount. After all total firm income is given and workers must get something or they will not work. However, it does leave tremendous room for range.

Social coordination also comes into play through norms. Thus, businesses and universities pay according to norms, which helps explains the loose patterns that exist across organizations. However, that is not the same as pay being set by marginal product theory.

These social dimensions and the role of institutions in pay determination are increasingly forcing themselves to the surface given the worsening of US income distribution. They are profoundly incompatible with orthodox economics and carry dramatic social and economic policy implications. However, they are fundamentally compatible with heterodox economic theory.

In the 1950s the “East Anglian” economists from Cambridge U.K. recognized the critical role of marginal product theory, and they saw the normative justification of free market society collapsed without it. For them, capitalism was enormously productive, but it was also marked by struggle. However, capitalism has none of the normative properties ascribed to it by orthodox economics. Unfortunately, the East Anglicans got bogged down in fruitless debates about aggregate production functions and the meaning of capital, when there were much easier ways of showing that the marginal productivity theory of income distribution is bogus.

At this stage there is grave danger of being pick-pocketed. That is because I suspect orthodox lefties like DeLong, Krugman, and Rodrik might agree with all of the above – and still claim “neo-classical economics rules”. That’s a version of the amoeba defense or killing with kindness.

If the above holds, it is not clear what is left of the neo-classical economics that Rodrik, DeLong and Krugman exhort us to embrace as the one truth. However, economics continues, which raise questions of what should economics look like. That is where methodological pluralism comes back into the picture.

Different economic paradigms are needed, and that can be combined with a “horses for courses” approach to practice. There will still be a place for neo-classical choice theoretic analyses set in market clearing frameworks, but with humble self-conscious caution that such analyses are parables that may bear only slight resemblance to the real world and are easily misinterpreted.

There is also a place for earlier aggregate macroeconomic modeling (old Keynesian style) in which market demand curves are taken as the primitive building block. That approach is pragmatic, sensible, theoretically inclusive and susceptible to empirical testing. Market demand curves are observable, whereas what goes on inside individual’s heads is not. Moreover, the orthodox “individual utility maximization” approach provides almost no insight into markets since it can theoretically generate almost any and every shaped market demand curve. That makes it entirely non-constructive.

There is also the question of how to represent market processes. Historically, this has been in terms of equilibrium, but viewing the market in evolutionary or historical terms produces different representations that heterodox economists have long emphasized (especially those pesky East Anglian Cambridge U.K. economists).

When it comes to equilibrium, there are also different choices. Orthodox neo-classical economists assert that output approximates the economy’s full employment supply, and demand matches this. Keynesians and other heterodox economist say that output is equal to demand or expected demand, but both may be below full employment supply. Neither theory nor empiricism allows us to choose between these representations, which suggest they should be allowed to co-exist (which is not the opinion of today’s orthodoxy, left and right).

The appropriate use of mathematics is also an issue. Mathematics is a useful representation device that can provide insights and provide logical consistency checks. But it proves nothing about reality. Non-constructive applications of mathematics that translate the obvious into the complex and inaccessible is not good economics – though it may awe students and the public who misinterpret it as science. That suggests need for good judgment in the use of mathematics, and heterodox and orthodox economists differ regarding that judgment.

Methodology matters. Today, orthodox economics is prone to math fetishism, and that fetishism distorts economics discourse, keeping out ideas not suitable to mathematical modeling and distorting communication. It also changes the community of economists, changes how economists engage their discipline, and changes how and who they communicate with. Together, that changes economics. Among heterodox economists the joke is that orthodox economist prefer to be precisely wrong rather than roughly right.

There are two last charges that need to be dismissed. One charge is that heterodox economics allows an “anything goes” approach. That is nonsense. For instance, heterodox macroeconomic models embed demand curves that describe reasonable economic behavior; supply rules that describe reasonable production and pricing behavior; production relations that link inputs (including labor) and outputs; balance sheet constraints that properly account for investment; saving, and wealth; and national income accounting constraints that fully account for income and production. These are rigorous models based on logically coherent theories of market economies. It is just they do it differently (and I would say more realistically) than orthodox counterparts.

A second charge is that neo-classical economics is “the only game in town”. That is something I remember Robert Barro, a poster-child of the orthodox mafia, once saying at an American Economics Association meeting. That too is nonsense. What he should really have said is that neo-classical economics is “the only game allowed in town”.