Tuesday, May 12, 2015

Multiple Equilibria and Financial Crises

Models of sunspots and multiple equilibria were developed in the 1980s as an alternative to the dominant Real Business Cycle agenda. For the last couple of decades, these models have taken a backstage role as explanations of the macroeconomy. Now they are back with a vengeance. 

On Thursday and Friday of this week, Jess Benhabib and I are running a conference at the San Francisco Fed that showcases new research on multiple equilibria and financial crises. The papers at this conference trace their roots to an agenda on sunspots, developed at the University of Pennsylvania in the 1980s.

The sunspot agenda began with the seminal paper by David Cass and Karl Shell, Do Sunspots Matter?, the pathbreaking paper on Self-Fulfilling Prophecies by Costas Azariadis and a paper by myself and Michael Woodford in which we developed techniques that form the basis for dynamic models of indeterminacy that are now widely used to understand monetary policy regimes

Another important landmark was the 1994 conference at NYU, published in the Journal of Economic Theory as a symposium on Growth Fluctuations and Sunspots.  Here is a link to a survey paper that explains the history of the sunspot agenda and its connection to endogenous business cycles.

I'm looking forward to the dinner talk on Thursday night by Karl Shell and I'm also looking forward to seeing the tremendous range of papers that are moving this agenda forwards in new and exciting directions. Here is a link to the conference papers. 


  1. Multiple equilibria are a great as a way to endogenise in some sense what are otherwise purely exogenous shocks in state of the art New Keynesian DSGE or in business cycle accounting models. But I still don't understand why you're so focused on deriving them from bargaining inefficiencies in labour markets? The question always remains in your models: wouldn't workers in the end agree to a real wage cut to avoid being fired? i.e instead of taking the wage as given as a function of some aggregate belief function, wouldn't there in the end be some rebargaining towards the core or the competitive search equilibrium, as e.g in
    Why not focus the multiple equilibria in financial markets? The large number of players in financial markets and the complexity of their balance sheets makes it much more likely that inefficient bargaining resolved via "animal spirits" is a factor there. For sure, it would make your models look much more like models of a "financial" crisis, instead of just a labour market crisis masquerading as a financial crisis.

  2. Daniels
    Competitive search equilibrium of the kind, for example, described by Espen Moen, requires that there be competition between 'market makers'. Those markets are, in my view, missing.

    Why not focus on the asset markets? I agree that asset market inefficiencies are also critical: but they are not enough. Persistent inefficient movements in asset prices cannot explain the permanent component of unemployment rate movements. The price earnings ratio is persistent but mean reverting. It is the ratio of a stock to a flow. The flow (earnings) is tied to unemployment which has a permanent (non mean reverting) component.

    A shock to the asset markets will eventually dissipate leaving the economy at its 'natural rate of interest'. But that natural rate of interest may be associated with many different unemployment rates. In my view, asset market inefficiencies (asset price bubbles) and labor market inefficiencies (involuntary unemployment) are both necessary to understand the data.

  3. Yes, but I get the impression that credit market shocks can be long lasting, more so than what you describe for stock market P/E. Private sector debt/gdp ratios can be very persistent (e.g 1980's to 2000's household debt levels in the US were much higher than in 1950's-1970's, with the financial crisis bringing a partial reversal which looks like it could be again a permanent shock). I'm thinking about models with a steady state bubble like this
    (with potential switching between steady states like in your models).
    Kocherlakota also had a paper on bursting of quasi-permanent financial market bubbles back in 2009, though I'm not sure if the manuscript is still around.
    For labour markets, I agree that core/competitive search equilibria are idealisations that are never reached in the real world. But taking prices as given in a search equilibrium goes too far for me. Once you have search frictions, there are incentives not to take prices as given. I think recent research suggests much more wage flexibility at the margin for new hires, as well as flexibility in terms of back loading of wage cuts (i.e not cutting wages now but reducing wage increases in the future). And once you have a model of persistent credit market booms/busts, you can link it to labour markets through financing constraints on labour demand or through movements in the employer stochastic discount factor used to evaluate hiring related investment (as in recent papers by Bob Hall). Maybe your or some student of yours should work on integrating the multiple credit market/multiple labour market equilibria models.

  4. Roger, what is your view of the importance of sunspots in explaining medium-term cycles versus, for example, Comin and Gertler (2006) technology-based explanation? Also, are multiple equilibria models and models with slow reversion observationally equivalent (or can be made to be)? If so, can we ever hope to figure out "who is right"?

    1. Sunspots that shift the economy across equilibria and slowly moving technology shocks that shift the natural rate are observationally equivalent in some models but not in others. It depends on how much detail one is prepared to place on what generates technology shocks and how that feeds into the "natural rate". Anyone who is committed to the NRH interpretation of the labor market will be able to find an interpretation of data that fits their mental model. It depends on whether the NRH is in your hard core as in Lakatos's view of the philosophy of science.

    2. Makes sense. Thanks so much for the response!


Note: Only a member of this blog may post a comment.