Franco Modigliani famously quipped that he did not think that unemployment during the Great Depression should be described, in an economic model, as a "sudden bout of contagious laziness". Quite. For the past thirty years we have been debating whether to use classical real business cycle models (RBC), or their close cousins, modern New Keynesian (NK) models, to describe recessions. In both of these models, the social cost of persistent unemployment is less than a half a percentage point of steady state consumption.
What does that mean? Median US consumption is roughly $30,000 a year. One half of one percent of this is roughly 50 cents a day. A person inhabiting one of our artificial model RBC or NK model worlds, would not be willing to pay more than 50 cents a day to avoid another Great Depression. That is true of real business cycle models. It is also true of New Keynesian models.
That is the background for an exchange between me and Stephen Williamson on the comment pages of this blog.
Here is Stephen:
I'm agnostic about the self-correcting nature of the economy. I do think inefficiency isn't observable (except in the behavior of my colleagues) - we need models to measure it.
and my response
It is true that in both classical and NK models, the cost of business cycles is small. In my view, it is not true in the real world and it is not true in the search models I work with where the unemployment rate is determined, in steady state, by, what I call, 'the belief function'.
I'm not even sure we should call the fluctuations that concern me 'business cycles'. They are low frequency correlations in asset prices and unemployment that are absent from data that has been HP filtered. It is possible, of course, that the persistent increase in unemployment that follows a financial crisis, is caused by the factors that a conventional search theorist would identify as fundamental. For example, increased government regulation or changes in technology that render existing skills obsolete.
I have a different explanation. The persistent increases in the unemployment rate that follow a financial crisis are caused by waves of optimism and pessimism that are driven by psychology: so called 'animal spirits'. You would probably refer to these correlated movements in asset prices and unemployment as, random shocks to the equilibrium selection device. I prefer to say that the object that drives market sentiment is a new fundamental in an economic model that would otherwise contain a continuum of multiple equilibria. Once the belief function is introduced as a new fundamental: equilibrium is unique.
Stephen responds
“Your models don’t take a stand on high frequency vs. low frequency fluctuations in inefficiency. Indeed, you can have equilibria in which the inefficiency is permanent. So that’s very different from what’s going on in standard NK/RBC exogenous shock models.”
Precisely. That's why I eschew NK and RBC models. They are both wrong. The high unemployment that follows a financial crisis is not the socially efficient response to technology shocks. And the slow recovery from a financial melt-down has nothing to do with the costs of reprinting menus that underpins the models of NK economists. It is a potentially permanent failure of private agents to coordinate on an outcome that is socially desirable. Much more on this in my new book, "Prosperity for All" available September 1st from Oxford University Press.