There are many unemployment rates for any r*

Eric Lonergan has a new blog post in which he draws attention to the importance of the asset markets in business cycle dynamics. While there is much to like in that post, I find parts of Eric’s argument misleading. Here is Eric:


“This leads us to the how [stet] the equity market determines the ‘neutral policy rate’. I have argued in the past that we need to get away from the idea of some magic level of the interest rate that somehow brings inflation to 2% and the economy to full employment.”

Eric is even clearer in his tweets in which he asserts that "stock prices determine r*”. I disagree for the reason I outlined on page 195 of Prosperity for All. The stock price is a stock. r* is the ratio of a stock to a flow. If r* is 5% (for example) the (real) stock price could be 100 bushels of corn and earnings could be 5 bushels of corn or the stock price could be 200 bushels of corn and earnings could be 10 bushels of corn. Both cases would lead to a value for r* of 5%.

If there are 100 people in the economy and it takes one person to harvest a bushel of corn, the first case would correspond to an unemployment rate of 5% and the second case would correspond to an unemployment rate of 10%.

In a second passage, Eric points out that recessions are typically preceded by stock market falls.

“Stock markets are a measure of recession risk. They are not predictors of recessions, as such. Paul Samuelson famously quipped that the stock market has predicted seven of the last three recessions – or words to that effect. But the joke is only partly insightful. Yes, there is excess volatility in stock prices for a host of reasons, but the probability of recession fluctuates, so stock prices should fluctuate with these probabilities. In that sense, the stock market should ‘predict’ more recessions than actually occur.”

Here Eric and I agree. But I would go further. Persistent drops in the stock market are not simply indicators of a future recession: they are causes of a future recession. This is demonstrably true in the sense of Granger Causality, as a I show here. And the word ‘persistent’ is important in this statement. A one-day crash that comes back a week later has no impact on the unemployment rate. A one day crash in the stock market that persists for three months does have an impact.

To get beyond Granger Causality, one needs a theory of how stock market crashes, triggered by a loss of confidence, so called animal spirits, can generate a causal chain that increases unemployment. I have developed such a theory here. This theory explains why ANY unemployment rate can be consistent with an equilibrium in which no market participant has an incentive to change his or her behavior. And it explains why there is no correspondence between r*, the so called natural rate of interest, and u*, the socially optimal rate of unemployment.