VoxEU Column Macroeconomic policy

What Keynes should have said

Originally published in February 2009, this column proposes a new paradigm to reconcile Keynesian economics with general equilibrium theory. It suggests that, just as it sets the fed funds rate to control inflation, the Fed should set a stock market index to control unemployment. This would not let every manufacturing firm and every bank fail at the same time “as a result of speculative movements in markets that serve no social purpose.”

Editor's note: New thinking on macroeconomics is shifting into to the mainstream, as evidenced by a recent exchange on Project Syndicate between Larry Summers and Joseph Stiglitz. This column, originally posted on 4 Feb 2009, is an example of this mainstreaming. The post looks at the role of stock markets in stabilisation policy.

For more than seventy years, policy makers have used Keynesian monetary and fiscal policies to control recessions (Keynes 1936). Although these policies are widely perceived to have been successful in stabilising the business cycle, academics gave up on Keynesian theory in the 1970s. The appearance of stagflation led to the adoption of the Phelps-Friedman hypothesis of a natural rate of unemployment, and it caused academics to abandon the Keynesian idea of many steady-state unemployment equilibria (Cross 1995). Mainstream economists adopted an approach in which temporary deviations from the natural rate of unemployment are caused by “sticky prices”.

In a forthcoming book (Farmer 2009), I propose a new paradigm that reconciles Keynesian economics with general equilibrium theory. Unlike existing interpretations of Keynes’ General Theory, my approach does not rely on sticky prices and does not carry the implication that the economy, if left to itself, will return to full employment. The theory implies instead that any level of unemployment can coincide with any rate of inflation.

I will make three related points. First, due to missing markets, labour market clearing may occur at many different unemployment rates, all of which are consistent with steady state equilibrium. Second, aggregate demand determines which of these equilibria will occur. Third, aggregate demand depends not on income, as asserted by simple Keynesian theories, but on wealth. My argument leads me to advocate a different policy from the trillion dollar bailout currently on the table. I argue instead for direct control of the stock market through Fed intervention in a market for indexed securities.

The labour market

Finding a job is not costless; it requires resources to match a worker with a vacancy. Search theorists (Pissarides 2000) define a search technology to be a process that takes the search time of a worker and the search time of a corporate recruiter as inputs. By embedding this technology into an otherwise standard general equilibrium model, one can arrive at a workable definition of the natural rate of unemployment – the unemployment rate that would be chosen by a social planner in this expanded general equilibrium environment.
To produce a commodity, competitive equilibrium directs firms to use the right mix of labour and capital through adjustments in the wage-rental ratio. But to produce a match between a worker and a vacancy, the corresponding price signals are missing. We do not observe markets for the search effort of unemployed workers or corporate recruiters, and there are no market prices to direct participants to use the correct mix of vacancies and unemployed workers to fill a given number of jobs.

Standard search theorists have proposed many different mechanisms that might substitute for the lack of prices and restore the uniqueness and optimality of equilibrium in a search economy. I believe that these mechanisms are not present in practice and that the multiplicity of equilibria that follows as a consequence is a real world phenomenon with the following important implication.

In high-employment equilibria, firms devote a high percentage of resources to searching for a small number of unemployed workers. In low-employment equilibria, firms devote a small percentage of resources to searching for a large number of unemployed workers. The absence of input markets allows there to be a continuum of equilibria, each one associated with a different ratio of vacancies to unemployment and all of them consistent with zero profits through adjustment of the wage.

The stock market

US households own roughly two times US GDP ($28 trillion) in the form of houses and three times GDP ($42 trillion) in the form of factories and machines. In classical theory, the value of tangible assets is equal to the present value of future rents and dividend payments, and these payments are pinned down by fundamentals – preferences, endowments, and technology.

There is an alternative theory of asset prices put forward in Keynes’ General Theory – asset values are based on confidence. According to the confidence hypothesis, stock market participants value assets based on their perceptions of how others will value them in the future. House prices and stock prices have fallen dramatically in the last fifteen months, and some observers are puzzled that private banks are not lending to each other or to corporations and firms. There is no need for puzzlement. Liquidity is frozen because market participants are concerned that the economy is moving to a low-employment equilibrium and that asset prices will fall further. If this belief is fulfilled, the US banking system will prove not just to be illiquid – it will prove to be insolvent.

According to the confidence hypothesis, there is a connection between the stock market and dividends but the direction of causation is not the one stressed by classical economics. Confidence is an independent causal factor. If confidence is low – the private sector places a low value on existing buildings and machines. Low confidence induces low wealth. Low wealth causes low aggregate demand, and low aggregate demand induces a high-unemployment equilibrium in which the lack of confidence becomes self-fulfilling.

A policy proposal

Since WWII, the US Federal Reserve has stimulated the economy by lowering the interest rate during recessions and allowing it to rise again during expansions. At the end of each recession, the economy begins a new expansion, but there is no tendency for unemployment to return to a time-invariant natural rate. Defenders of the natural rate hypothesis argue that the natural rate itself is time varying but there is, to my knowledge, no theory that can explain this variation as a function of a small number of observable variables.

In Farmer (2008a, 2009) I develop a new paradigm. I argue that the unemployment rate returns to a level that depends on wealth and, just as there are many equilibrium unemployment rates, so there are many equilibrium values for the prices of assets. This theory underlies my recent articles (2008, 2009) in the Financial Times in which I suggest that the Fed intervene directly in an index of stocks to maintain confidence in the markets. In recent years the Fed has used one instrument – the fed funds rate – to control two targets: inflation and unemployment. I argue that the Fed should add a second instrument – the rate of growth of the price of a stock market index. Here is how this would work.

First, establish a market in an index of all publicly traded common stocks. The weights of the individual securities would be set in proportion to market capitalisation and could be adjusted regularly as new firms enter and old firms exit.

Second, create a privately owned holding company that holds a basket of securities and that issues liabilities in the index. The creation of a privately owned company of this kind would allow the Fed directly to trade the index without owning shares in the underlying corporations.

Third, direct the Fed to purchase a block of shares in the index, financed by a mix of money and three month debt liabilities, guaranteed by the Treasury.
Fourth, announce a price path for the index to be reset at each meeting of the open market committee and stand ready to buy and sell the index at the announced price. One policy might be to set the fed funds rate according to an announced inflation target and to set the price path for the index to achieve an announced unemployment target. The fed funds rate and the index would be chosen independently and their values set by manipulating both the level and the composition of the Fed’s balance sheet.

Possible objections

Some will argue that the government cannot predict bubbles and crashes any better than private markets. But in practice we do have some information about bubbles and impending crashes, and this information is shared by the government and the private sector. Private individuals cannot make money by betting against the market because any one individual is too small to buck the trend. When the herd stampedes, it is best to get out of the way. The government is the only agent large enough to stop the stampede and restore optimality.

Some will argue that my plan represents a step on the slippery slope to socialism. Not so – I am arguing for no more than an extension of an already accepted role for the central bank. Just as the Fed already sets one price (the fed funds rate) to control inflation so it should set a second price (a stock market index) to control unemployment. Unlike the fiscal bailout currently under consideration, my plan involves no extension of the size of government. It puts spending power back where it belongs – in the hands of households.

Capitalism is the single greatest engine of growth ever devised and the free market allocation of capital is superior to any other system known to humanity. If the big three auto makers cannot produce cars that people want to buy – let them fail. If Lehman Brothers made bad investments – let it sink. But do not let every manufacturing firm and every bank fail at the same time as a result of speculative movements in markets that serve no social purpose.


Cross, Rod (ed) The Natural Rate of Unemployment: Reflections on 25 Years of the Hypothesis, 1995.
Farmer, Roger E. A. “The Great Depression” manuscript, UCLA, November 2008a.
Farmer, Roger E. A. (2008b) “How to Prevent the Great Depression of 2009” Financial Times Economists Forum., December 30th.
Farmer, Roger E. A. (2009) “A New Monetary Policy for the 21st Century Financial Times Economists Forum., January 11th.
Farmer, Roger E. A. Expectations Employment and Prices Forthcoming, 2009.
Keynes, John Maynard. The General Theory of Employment, Interest and Money, Macmillan Cambridge University Press, 1936.
Pissarides, Christopher. Equilibrium Unemployment Theory. MIT Press, 2000.

2,625 Reads