He took the economic forecasts produced by the Survey of Professional Forecasters and compared them with forecasts generated by a reasonable economic model which was updated as new information emerged. The difference between what forecasters actually predicted and what the model predicted is a measure of animal spirits; professor Milani calls this the expectations shock.
And he found that this shock - changes in economic forecasts not explicable by contemporary data or the model - can account for over half of the variation in US GDP over the past 40 years. Animal spirits fell in advance of each of the last seven recessions.
Now, it could be that this is because forecasters knew something that the model and contemporary output, inflation and interest rates weren't saying - although this is unlikely given that forecasters generally do a poor job of predicting recessions. But this could also be vindication of Keynes - recessions really do happen (in part) because people become irrationally pessimistic and so cut spending, with the result that even irrational expectations become self-fulfilling.
Worryingly, professor Milani's measure of animal spirits is close to a record low.
This raises the question: if booms and slumps are due to animal spirits, what should policy-makers do?
Maybe not much that they shouldn't do anyway. For one thing, a big part of the policy response to recession should be the creation of risk-pooling mechanisms to protect innocent victims. But this job is independent of the causes of recession. As the Nobel prize-winning economist Robert Lucas wrote in Models of Business Cycles: "Policies that deal with the very real problems of society's less fortunate - wealth redistribution and social insurance - can be designed in total ignorance of the nature of business cycle dynamics."
Secondly, counter-cyclical monetary policy already operates upon expectations. One of the main effects of quantitative easing was that it put a floor under expectations, by reducing the probability of catastrophe. And Columbia University's Frederic Mishkin has argued that monetary policy can be effective in crises precisely because it improves expectations; does anyone think the US economy would be recovering as well as it is had the Federal Reserve left the funds rate at 5 per cent?
Perhaps the probability that animal spirits cause economic fluctuations is more important for ideological debates than for practical policy purposes.
