February 28, 2023 Reading Time: 6 minutes

In the (1955) edition of his top-selling Economics text, Paul Samuelson claimed economists had finally determined how economies worked. As he put it:

In recent years 90 percent of American economists have stopped being “Keynesian economists” or “anti-Keynesian economists.” Instead they have worked towards a synthesis of whatever is valuable in older economics and in modern theories of income determination. The result might be called neo-classical synthesis and is accepted in its broad outlines by all but about 5 percent of extreme left wing and right wing writers.

By means of appropriately reinforcing monetary and fiscal policies, our mixed-enterprise system can avoid the excesses of boom and slump and can look forward to healthy progressive growth. This fundamental being understood, the paradoxes that robbed the older classical principles dealing with small-scale “microeconomics” of much of their relevance and validity – these paradoxes will now lose their sting. … Perhaps for the first time – the economist is justified in saying that the broad cleavage between microeconomics and macroeconomics has been closed.

The 1955-1972 period might be labeled the “all is known by us” period of macroeconomics. There was one authoritative model, all-seeing and all-explaining. By twirling the right fiscal knob-settings, and pulling the right monetary levers, the priesthood of planners could fine tune a previously unpredictable world economy.

But models were at best retrospective approximations, quickly rendered obsolete by attempts to exploit measured relations for policy intervention. For example, the attempt to exploit the “Phillips Curve” trade-off between unemployment and inflation ended up producing more unemployment and more inflation, because markets are about people, not billiard balls. While the government generally cannot make things better, aggressive state intervention can certainly make things worse.

This is the heart of the Austrian economics response to the idea of fixed models, based on equilibrium. The problem for the Austrian perspective is that the politically perceived need for a model to represent the aggregate economy and capture the effects of policy interventions is overwhelming. Austrians hold that the very idea of the synthesis “macro model” is incoherent. For one thing, the level of aggregation required to define the variables—GDP, employment, price level—rules out any identifiable vectors of cause and effect, because production, capital, and jobs are not homogeneous things. Aggregating the “capital stock” in particular is a problem.

Suppose I did a careful inventory of the “equipment” of a hospital. On my ledger I can report that the hospital has 31,718 “pieces of equipment;” is that enough?  As I discussed in my own recounting of the experience of seeing highly specialized, capital-intensive medical operations, having “equipment” is not enough.  It is the capital structure, organized in a particular way in a particular point in time, extending into an uncertain future, that makes capital valuable or useless. Claiming that a bedpan, a heart monitor, and a set of surgery tools is “three equipments” is to write down nonsense and pretend you know something.  

But even if aggregates had meaning, the empirical relationships among these variables are never “in equilibrium.” The causal relationships are not even stable at any level, but are changing like a kaleidoscope in complex ways. A snapshot of the data, or the aggregation of data over a month in order to write down a model, is already hopelessly out-of-date by the time it is used to analyze a simulated “shock,” because the kaleidic changes in economic relationships are by definition unpredictable. This is a general problem in the social sciences, but the problem of even short-term parameter stability is acute in macro models.

Why is the Austrian perspective so little heeded, and in fact almost ignored, in the halls of power? Bryan Caplan has argued that the insistence on the essential difference of the Austrian approach has been counterproductive, and I think he makes some sensible points in that regard. But I think there is a simpler explanation, one based on the needs of political discourse. The difficulty faced by the Austrian critique is the oft-cited requirement that it takes a model to beat a model.

This view was hardly limited to the Keynesians; here are two pro-market neoclassical economists who hold the same view:

The answer is that it takes a theory to beat a theory; if there is a theory that is right 51 percent of the time, it should be used until one comes along that is better. (Theories that are right only 50 percent of the time may be less economical than coin-flipping.) — George Stigler, 1987; emphasis added.

I play by the rule that it takes a model to beat a model. — Thomas Sargent, 2011; emphasis added.

Saying that economics is a competition among models is already missing the point. The Austrian critique of central planning is founded on the claim that government officials, even if motivated by the best of purposes and goals, cannot get accurate, timely, local information. This an empirical claim about the world, in the sense that it makes a (falsifiable, in principle) claim that it is impossible — not difficult, impossible — to “calculate” the values of resources and outputs necessary to “run” an economy.

To the advocates of planning, this seems like a rejection of models because our measurements “are not good enough, yet.” In fact, that was precisely the response of Oskar Lange, who famously went so far as to congratulate (sarcastically) the contributions of Ludwig von Mises in highlighting the calculation problem:

Socialists have certainly good reason to be grateful to Professor Mises, the great advocatus diabol of their cause. For it was his powerful challenge that forced the socialists to recognize the importance of an adequate system of economic accounting to guide the allocation of resources in a socialist economy. Even more, it was chiefly due to Professor Mises’ challenge that many socialists became aware of the very existence of such a problem. And although Professor Mises was not the first to raise it, and although not all socialists were as completely unaware of the problem as is frequently held, it is true, nevertheless, that, particularly on the European Continent (outside of Italy), the merit of having caused the socialists to approach this problem systematically belongs entirely to Professor Mises. Both as an expression of recognition for the great service rendered by him and as a memento of the prime importance of sound economic accounting, a statue of Professor Mises ought to occupy an honorable place in the great hall of the Ministry of Socialization or of the Central Planning Board of the socialist state. — Lange, 1936.

Obviously, planners think the calculation problem can be solved by the same kind of trial and error, invoked by advocates for market processes. Traders in a particular market — say, for tin — buy and sell to arbitrage away price differences, and “the price” then reflects the opportunity cost of the resource at that moment, without anyone fully understanding the sources or the uses of that tin. As Hayek pointed out in 1945, it is very important that no one needs to know why the price is at the level; all anyone needs to know is the value of the price at this moment to be able to decide whether to buy, hold, or sell tin.

It seems that central planners are thinking along the same lines, but the analogy is misleading. Instead of using trial and error in one market, planners want to substitute centralized trial and error along thousands of dimensions simultaneously. There is a world of difference between groping for price, and guessing at “the price level,” in the aggregate.

Frustratingly, we are left with two sharply contradictory imperatives, both of which are actually true.

  • First, attempts to exploit observed statistical “relationships” measured on aggregate data in the past are at best unrelated, and at worst actively harmful, to the ability to manage aggregate economic activity, employment, inflation, or really anything that we care about.
  • Second, a recommendation to do something, and the claim that “we have a plan,” is always going to be a competitive advantage in politics and elections; the claim that we don’t know anything, and state intervention is likely to do more harm than good, is simply untenable as a policy recommendation.

Ultimately, some policy recommendations are better than others, and the political imperatives to “do something” are simply irresistible. To the extent that the Austrian perspective puts itself on the bench by refusing to play, the field is occupied by teams of economists who make things much worse. Because bad models, advanced with smooth confidence and the authority of rafts of dogmatic-but-refereed journal articles, are much worse than no model at all.

Still, that suggests a defense of economics that we often forget. Most economists, if you have a sober conversation with them in private, readily admit that models are at best temporary approximations, and can be misleading. But anyone who wants to work for the government, or to be listened to by important officials for advice, must pretend to believe that models are specifically informative, and in particular, that their personal pet models are the best. In short, the problem is not so much our reliance on economics as our reliance on politics. Politics insists that we must pretend to believe in the fiction of central direction of markets as effective policy.

Michael Munger

Michael Munger

Michael Munger is a Professor of Political Science, Economics, and Public Policy at Duke University and Senior Fellow of the American Institute for Economic Research.

His degrees are from Davidson College, Washingon University in St. Louis, and Washington University.

Munger’s research interests include regulation, political institutions, and political economy.

Books by Michael Munger

Get notified of new articles from Michael Munger and AIER.