Thursday, 22 July 2010

Rational expectations - is it real?

One of the big divisions in macroeconomics is the idea of rational expectations theory. This is the proposition that people behave as if they have a perfect prediction of the economy's future path, and therefore they collectively fulfil that prediction. This idea is used by some to claim that government borrowing cannot boost the economy because people will reduce spending to pay the future taxes they expect to incur; and by others to propose that price or NGDP level targeting must work, because people will act as if the target will be met and therefore - in a self-fulfilling equilibrium - the target will be met!

While there must be some truth in the idea, there are three main points where it may fail:
  1. Psychologically, people do not act consistently with their rational prediction of the future. People are often over or underconfident even when they make an accurate prediction of future outcomes.
  2. Often there is no single stable future path to the economy. Economic growth is dependent on other people's actions - because investments and purchases have externalities. Therefore, it may not be possible to determine a stable rational expectation for the economy without coordinating your actions with other people.
  3. Different people have different prediction horizons. Especially, the lower one's income the more liquidity constraint one suffers; and the less one can afford to take into account the future path of taxes, inflation or national debt. Therefore a full rational expectations model will break down because the short-term behaviour of the economy is likely to be stronger or weaker than its long term.
In addition, we can surely expect economic actors to learn over time. One reason that Keynesian deficit spending may have appeared to work from the 1930s to 1950s but fail by the 1970s is that people learned of the existence of inflation, and started to demand pay rises to match it. Prior to the 20th century inflation barely existed. Even until the 1950s - outside of Weimar Germany and a few other special cases - significant inflation was not considered an important risk. Now it's a basic part of all macroeconomic calculations.

Rational expectations, therefore, are only likely to account for the phenomena that we have learned so far. What new phenomena might lurk in our future?

I will explore the psychological basis of, and challenges to, rational expectations in a more detailed post soon.

No comments: