Neorational Keynesonomics

This article by Brian Kantor is probably best understood as a defense of the rational expectations hypothesis by demonstration 1) of its (apparently) common-sense consequences and 2) the empirics that bear it out.  It is divided into a history of the hypothesis’s development and relevant precedents, an account of its movement from hypothesis into a utile theory, and a discussion of its various actual and potential applications.  Kantor’s position is clearly stated in the conclusion: Rational expectations is here to stay and has changed macroeconomics for the better.

The impetus for a rational expectations theory appears to derive, as it seems do most new developments in the economics of the 60s and 70s, from a recognition of the failure of policies informed by the long-run Phillips curve.  Rampant inflation in the 60’s followed by stagflation in the 70s demanded a “new” approach, which, the likes of Kantor thought, demanded a partial return to classical ideas and a critical appraisal of Keynes and the economic policies derived from his General Theory.  In particular, it was suspected that policy changes were broadly anticipated by market participants and that, therefore, prices mostly - if not entirely - reflected these anticipations.  If I am reading Kantor correctly, he seems to be suggesting that a strong version of the idea that economic actors might actually be able to develop expectations of the future based on current trends and events and past experiences was nigh revolutionary.  In other words, it was not broadly suspected that such information was widely accepted and integrated into economic actors’ supply and demand schedules.  It was as if economists considered it implausible that their ilk might be employed by the private sector and use standing theory to make forecasts about the future.  Besides, if economists held the monopoly on knowledge of future market movements, then wouldn’t it stand to reason that all economists would invest and become quite wealthy?

The general outline of the rational expectations hypothesis is that there are three factors which market participants observe and for which they adjust.  There is the information from the previous period about market movements and responses (especially in relation to policy changes and reactions), anticipation of the market movements of the current and future periods (which are functions of that history as well as current real and nominal factors), and a stochastic disturbance term with a mean at zero.  After the time period is up, actors observe outcomes and adjust accordingly.  The stochastic terms are assumed to not be serially correlated; that is, as Kantor observes by way of the Austrian economist Lachmann, consistent systematic error on the part of particular actors is weeded out by competition applying the available information. 

Throughout the seventies (this paper was published in December 1979), the rational expectations hypothesis was born out by empirical study.  Kantor particularly focuses on studies of the equities markets, since they produced ample if not definitive evidence against Keynes’s notion that changes in investment were, it seemed, entirely explicable by the idea that it was driven by speculation and “animal spirits.”  The efficient markets hypothesis - the idea that prices reflect all the available information about past, present, and future market conditions - garnered a great deal of support from these studies, and would definitely seem to go against Keynes’s implicit preference for a strong role of a more ‘rational’ central authority.  However, critics have suggested that this is all well and good when discussing investment, but other markets - such as those for labor - are less apt to be explicable from a rational expectations perspective.  The answer here for Kantor and others is that both sides of the labor market are liable to make decisions based on the information available, but they may no doubt have some difficult distinguishing whether shocks to the economy are temporary or permanent, positive or negative.  Evidence suggests that this uncertainty leads to a regular deviation about some trend (in terms of employment or economic activity).  The lags in bringing new information into decisions, or delays in the acquisition of relevant information, may be as Robert Lucas suggests, an important source of economic fluctuations which we know as the business cycle.

What does this say for stabilization policy?  Well for one it suggests that it was entirely misguided insofar as it had been conducted up to that point. Unless policy makers can somehow find a way to surprise the market every time they attempt to guide it, it would be very difficult to ’shock’ it in the way that they might hope.  It may be difficult to keep one step ahead of the market at every turn. 

Instead, Kantor suggests that efforts and resources expended in smoothing the business cycle may best be spent in improving the function of existing institutions and in bringing new, helpful institutions into existence.  On this, at least, Kantor remains substantially vague, and I think this is to his credit.  There are, it seems to me, two roads here.  On the one hand, it may provide support for those who believe government has as a tendency to get in the way of and disrupt otherwise well-functioning markets.  To wit, one might suggest that deregulation, lower taxes, etc. etc. are the key.  This will give added incentive for entrepreneurs to find means to improve flows of productive factors and information between markets.  On the other hand, for those more inclined for government-supported solutions, there is the road of policy taking an active role in creating institutions that serve this same purpose.  Nonetheless, we must grant the other side that, even in this, the private sector just might be able to do it better and more efficiently.  The debate is unavoidable, and it continues. 

Good article, fairly intelligible.  I have avoided, despite the voices of my better angels, discussion of some of the philosophical implications here. Suffice it to say that while there is ample evidence and good old-fashioned intuition that rational expectations is the ‘right’ way to go, we may want to question (as I did in my previous and equally long-winded entry) just how far we want to take it without wondering what sort of queer and possibly inefficient or disastrous paths our ‘rational’ expectations, taken in the aggregate over a long period of time, might lead us down.

Like many other people in the class, I chose to read Friedman’s (apparently) influential 1967 presidential address before the Eightieth Annual Meeting of the American Economic Association.  I have also perused a couple of the other articles from the reader, so if some of their content seeps into this review, that is why.

 Friedman’s goal is as the title states; he desires to outline what the goals of monetary policy should and should not be, and much of this is based on his work in resurrecting the quantity theory of money over the previous decade.  After an introduction that relates a brief history of the valuation of monetary policy in the decades preceding and following Keyne’s General Theory, Friedman divides his talk into three sections:  What monetary policy cannot do, what monetary policy can do, and how monetary policy should be conducted.

As for what monetary policy can not do, Friedman claims that it can peg neither a specific interest rate nor a particular unemployment rate for the long run.  Here Friedman is explicitly referring to his natural rate hypothesis.  When interest rates are raised or lowered above the natural level (as through open market operations), individuals will begin to see an unexpectedly larger quantity of money in their pockets.  They will feel their income has increased, and as that money moves through the economy, the price level may rise.  Additionally, greater income means a greater demand for loans and the like, bidding up interest rates.  The only way to prevent this is to increase the rate of money growth at an increasing rate, and this is even more true as expectations adapt, requiring even greater acceleration.  Even when the rate of money growth is slowed to previous levels, interest rates will probably continue to rise and overshoot the natural rate, below which the money infusions were meant to keep interest rates in the first place.  Thus, monetary policy can not peg interest rates in the long run.  

It is not hard to see how this same reasoning may be applied to the unemployment rate, and here we see Friedman’s proposition of the vertical long-run Phillips Curve.  More money means greater income means greater demand for goods and services.  This leads to an increase in the value of the marginal product of labor, increasing labor demand.  But here aggregate demand/supply analysis tells us that the price level will increase, increasing real wages and reducing labor demand.  The only way to keep unemployment below its natural level for more than a short period is, as with the interest rate, to increase the acceleration of money growth, and this leads to inflation and its various and sundry dire effects.

Friedman’s ideas of what monetary policy can do is very much reflective of Friedman’s position on the role the government should play in the economy (not much).  First, monetary authorities shouldn’t screw up.  Monetary policy can have powerful effects on the economy whether it is done well or not, as the Great Depression attests.  Second, monetary policy should be conducted conservatively as often as possible so that, when its powers truly are needed, it can achieve maximum effect.  Essentially, people should not expect the Fed to act except in dire circumstances so that it remains a potent tool.  Finally, monetary policy can indeed act to great effect in times of serious economic disturbance, but should not engage itself in fine-tuning the economy.  This is not so much a third point as it is a necessary conclusion from the first two points.

Finally, Friedman concludes with a note on the conduct of monetary policy, in which he prefers that a rate of monetary growth be the goal rather than an interest rate.  Here, at least, history has not born out Friedman’s thought.  With deregulation of the financial markets and the volatility of the 1970’s and 1980’s, sudden and drastic changes in the money supply became the norm, and it became inadvisable to trust any one measure of money (M1, M2, etc.).  Now, as I understand it, the Fed controls money demand through interest rates, letting the chips of money supply fall where they may.  Edit: The previous sentence is incorrect. The Fed lets money demand fluctuate and targets interest rates by adjusting money supply.

If this article truly is one of the most influential of all works in economics, I think I can see why.  It is clear and concise and represents a powerful rejection of the standing thinking of the time.  One thing bugs me, however, and my guess is that more current theory would bear this out.

Specifically, I simply don’t buy the position that changes in the money supply are neutral in their effect on the economy in the long run.  It strikes me, and Mayer discusses this only in analytical detail in his article “The structure of monetarism,” that the aggregate effects of monetary policy changes depend a great deal on 1) transmission mechanisms and 2) the state of individual sectors of the economy at the time of the change.  I would argue (against a very strict rational expectations assumption) that nominal changes in money supply are registered in the near-term as real changes.  If investment and consumption decisions are based on perceived real changes, this can have long-lasting effects on the structure of the economy.  Investment may be undertaken that has no basis in “real”-ity so that excess money is filtered into the hot sectors of the day.  But if such investments are long-term in nature, such as in real-estate or expensive equipment, a nominal change in the money supply may nonetheless bring about long-run real changes in the economy.  Empirically, I suspect this hypothesis is incredibly difficult to prove.  As Mayer also discusses, monetarists and the schools into which it evolved rely on reduced form equations that perceive the process by which interest rates are changed as black box.  On the other hand, large structural models preferred by Keynsians and those of a similar vein are bound to be criticized, and rightly so, for failing to account for certain vital factors or for violating ceteris paribus constraints. 

Yet my instinct remains - monetary policy can give impetus to actions that bring about economic restructuring, possibly by reinforcing network effects and leading to path dependence.  A neutral monetary policy, it seems to me, would require policy-makers to judge the amount of money needed in each sector (a seemingly impossible task bound to be erroneous in its conclusions) and find ways to bring money into each of those sectors, accordingly. 

A complete theory of monetary policy effects would have to account for this possibility. Anyway, that’s my two cents.  Thoughts?