I am trying to read John Cochrane’s comments on Paul Krugman’s article on why economists got it so wrong. I tend to get upset while reading. I have managed to get through the first paragraph in which Cochrane compares Krugman to someone who denies that HIV causes AIDS and compares developments in economics to progress in the natural sciences. Now in the natural sciences, there are counter intuitive models which are consistent with all available data and have withstood many tests. Also there are once mysterious facts which are explained by theories to the satisfaction of all people familiar with the theories.
So I wondered what puzzles have ceased to be puzzling due to work by macroeconomists in my lifetime (November 9 1960 – present). I especially wonder if any mysteries were explained by macroeconomists not named Milton Friedman. After the jump and some more throat clearing, I will get to examples.
I’m counting growth theory as another field.
I have often written about the fact that economists (especially but not only fresh water economists) are unperturbed by regection of their theories by data. They identify “theory” with “a set of models” and note that models are false by definition. This is not at all true of theories in the natural sciences. It is very likely that all existing theories are false, but they are not false by definition and the discovery that they are inconsistent with data shocks scientists when it occurs. It also causes them to, first try to modify the theory to fit the new data and second (and very often) to abandon the theory. This is not the approach taken by economists. Models of say labor markets or financial markets which are rejected by the data are, nonetheless, used in macro models, based on the argument that if it is false therefore it is a useful approximation.
But what about mysteries resolved. There was a mystery — why were so many gay men in New York dying following a collapse of their immune system. Almost all scientists familiar with the data agree that an explanation has been found. AIDS remains horrible, but it is not mysterious.
So what macroeconomic mysteries were resolved ? And were any resolved by anyone but Milton Friedman ?
1) in a cross section consumption increases less than proportionally with income. As a result, many brilliant and famous economists (yes one was named Keynes) expected that as average income rises, the ratio of consumption to GNP would fall. This didn’t happen. The two facts are no longer mysterious at all. They are both implications of the permanent income hypothesis (PIH) or of a model in which part of consumption is according the the PIH and part is by liquidity constrained consumers who live paycheck to paycheck. The PIH explained many strange facts in the cross section of consumption. Just for example African Americans consume less than White Americans with the same income. Why ? Seems an odd fact, but it follows since average African American income is lower so if you have Black and White with the same income it is probabally an unusually good year for the Black family and/or a bad year for the White family.
2) why the hell is there a Phillips curve which related unemployment and inflation in the USA and UK ? How could this happen ? Well it would happen if expected inflation were about constant because there had not been persistent shifts in inflation. So Friedman predicted that the relationship would disappear if there were persistent inflation. And so it did. Now this wasn’t exactly a new insight. Everyone knew there is no Phillips curve during hyperinflations.
As a matter of fact Keynes specifically warned that economists shouldn’t hope to find a simple stable relationship between increased nominal demand and nominal wages which would imply a relationship between inflation and unemployment. Discussing the hope of finding, well something like a robust valid Phillips curve he wrote “they do not readily lend themselves to theoretical generalisations.” (General Theory Chapter 21 section VII paragraph 2) and, after deriving a Phillips curve from implausible assumtions
I do not myself attach much value to manipulations of this kind; and I would repeat the warning, which I have given above, that they involve just as much tacit assumption as to what variables are taken as independent (partial differentials being ignored throughout) as does ordinary discourse, whilst I doubt if they carry us any further than ordinary discourse can. Perhaps the best purpose served by writing them down is to exhibit the extreme complexity of the relationship between prices and the quantity of money,
Oh in chapter 21, Keynes also noted that the Phillips curve is vertical during hyperinflations but I won’t look up the paragraph.
chapter 21 section VI third paragraph from last. Keynesian economists were warned and ignored the warning.
Friedman and Phelps didn’t really go fundamentally beyond Keynes on that one.
3) The Peso problem. Why are mexican peso interest rates higher than US dollar interest rates when the exchange rate is fixed ? Answer, people consider a peso devaluation more probable than a dollar devaluation. One only concludes that one can make useful predictions from a finite data set if one is willing to make assumptions about the distribution of random variables. To assume that all variables are normally distributed is to make a gross mistake. Some, like the Peso dollar exchange rate have distributions so that asymptotic results only tell us what will happen when we are all dead.
OK so 2.5 mysteries resolved before I graduated from junior high school (1973). And since then ? hmmm.
Maybe Hall on how Solow’s residual is positive when there is an increase in demand if price is greater than marginal cost (competition is imperfect).
Over near the great lakes there was, for a while, great interest in working out the implications of
“(1) For a time at least, rising prices may delude entrepreneurs into increasing employment beyond the level which maximises their individual profits measured in terms of the product.”
The General Theory … Chapter 20 section III paragraph 4. For some reason this is called the Lucas supply function.
Also they worked out at great length the implications of assuming perfect competition and perfect labor and capital markets and technological progress which is often negative and which is not persistent. Last I checked (which was long ago) the idea was that with great effort one could fit some summary statistics. I don’t think anything mysterious was explained. Actually, I think basically all noticible patterns other than those that fit the PIH (or thePIH with some liquidity constrained households) are anomalies which they will explain asymptotically.
Back near Oceans, there was work on putting some nominal rigidity in dynamic stochastic general equilibrium models. The aim was to show that the assumptions of rational expectations and perfect capital markets didn’t imply that the market outcome was optimal or that policy was irrelevant. When I was a graduate student in the very dead sea of saltiness, no one explained to me why this was the aim. It does seem to be a strikingly irrelevant project at the moment doesn’t it ?
The line isn’t “if you accept these assumptions, then this mystery is resolved” but rather “you must make these assumptions so this fact which seems totally unsurprising to you is a mystery.”
update: Thanks to a suggestion by Rob in comments, another possible case of actual explanation
An idea which explains things might be the dynamic inconsistency of optimal monetary policy (Kydland and Prescott). The strange facts that it explains include the moralistic, almost religious, tone of support for tight monetary policy. I’m not really sure how new the idea was, but someone I know says he personally knows that K and P were surprised by the mathematical result. Then a related idea is that it can be rational to delegate monetary policy making to an anti inflation zealot. This helps explain why apparently sane people appoint apparently insane people as central bankers (I think the argument is due to Runkle).