Tuesday, March 29, 2011

Farmer on Farmer/Kocherlakota

Roger Farmer sent me this which addresses some of my discussion in this post:
Thanks to Stephen Williamson for publicizing my work and to Mark Thoma for providing a link and invitation to respond. Stephen: in addition to the paper you cited, I just finished an empirical paper on how to explain data without the Phillips Curve, two theoretical papers on why fiscal policy works in the short run (but shouldn’t be used) two papers on rational expectations with Markov switching and a piece on stochastic overlapping generations models.

The papers you mention in your blog, by Narayana and me, were both presented at a conference in Marseilles last week with not one but two Fed Presidents in attendance: Jim Bullard also gave a paper. Jim presented work that draws on the Benhabib-Schmitt-Grohé-Uribe paper on the perils of Taylor Rules. He sees a real danger of a Japan style deflation trap happening in the U.S.. Narayana gave a paper that combines a liquidity trap model of bubbles in an overlapping generations framework with a labor market based on the idea from my 2010 book, Expectations Employment and Prices. This book provides a new paradigm that drops the wage bargaining equation from a labor contracting model and replaces it with the assumption that employment is demand determined. This is the same assumption taken up by Narayana in the paper he presented in Marseilles.

The main idea is explained very nicely by one of the anonymous commentators on Stephen’s blog , who said,

“Think of it this way. With a centralized labor market, the real wage is pinned down by the intersection of labor demand and supply. With search, the labor market need not clear: the labor supply FOC is missing, and we need to add something else to close the model. One thing to add is an explicit bargaining model that effectively pins down the wage. An alternative is to say that output is demand-determined, and that the wage is the marginal product of labor at the demand determined level of output. Then firms are on their labor demand curve, but workers are not on their labor supply curve (but the beauty of search - unemployed workers will take a job at any positive wage).”

That’s exactly right. And once there are many possible labor market equilibria, there is room to close the model by bringing back the role of market psychology. That’s what I do in my work which has room for both involuntary unemployment and animal spirits; the two cornerstones of Keynes’ General Theory that are missing from the macroeconomics that emerged from Samuelson’s interpretation of Keynes.
Stephen professes not to understand the language of aggregate demand and supply. That’s not surprising given how many different ways it’s used. My own preferred interpretation is explained in a piece I wrote for the International Journal of Economic Theory in 2008.

The idea of aggregate demand and supply makes just as much sense as the notion of a microeconomic demand and supply curve as long as one works within a framework where the variables that shift one of the curves do not simultaneously shift the other. That is clearly not true in post-Lucas rational expectations models which is why the language went out of fashion. It is true in my work.
Thanks Roger.


  1. Maybe language isn't something to get too hung up on, but I feel like when we talk about market participants being on (or off) "demand" or "supply" curves, we're already talking about their behavior in a very particular institutional arrangment for trade: one in which, if taken literally, people and firms participate in a mechanism in which they (somehow) come to face a single linear price, and feel like they can buy or sell as much as they want to only at this price.
    But once we've decided to think about outcomes in a search model, no such thing is true, so I have no idea what being on or off a demand or supply curve means, or for that matter, why the price-taking labor supply FOC mentioned above is relevant. But perhaps I am the one confused? Thanks Stephen (and Roger)

  2. I don't really agree with Roger's final paragraph. Even in a "post-Lucas" model you have money affecting only aggregate demand and real variables affecting only supply. Also, the microeconomic distinction between supply and demand isn't as sharp as one might suppose, as Friedman pointed out. For example, speculation can affect both supply and demand.

    I suspect the reason the AD/AS terminology went out of favour is that it does not make much sense in a RBC model, since everything happens in the supply-side. Since all modern macro is based on RBC methodology, the language used by RBC economists (notably Prescott) has become standard. But whenever you have a significant role for money in determining output, the AD terminology is natural. Hence it's use in New Keyenesian models.

  3. If we are unable to use the AD-AS terminology, what should replace it? Is equilibrium outcome all we can talk about? For example, a technology, government, Fed, and/or preference shock hits the economy and the economy reaches a new general equilibrium.

  4. ^^Except that a change in the money supply has supply-side effects -- it changes the output decision of some firms. There is no "AD" or "AS" shifter, only shocks that move both.

    Give it up already.

  5. It changes the supply decision of firms because output is demand determined. In a micro model, an increase in income affects the demand side, but then induces firms to increase supply.

    Essentially, you are confusing cause and effect. An increase in money increases AD directly. The question then becomes, what happens to supply? In a classical model, supply is fixed so you just get inflation. In a Keynesian midel the increase in AD leads to an increase in AD. But its a shift in AD causing a movement along AS.