Economic Empiricism, circa 1875
Robert Heilbroner’s brilliant book The Worldly Philosophers devotes a single page to William Stanley Jevons, the 19th century economist whose claim to fame lies in his “sunspot theory” of the business cycle.
Jevons had analysed huge piles of economic and meterological data and discovered the periodicity of the business cycle (from boom to boom) was 10.46 years, while the periodicity of sunspots was 10.45 years. This was “too close to be accidental”, and as such Jevons had “explained” the business cycle.
Jevons argued that there was a connection between the timing of commercial crises and the solar cycle. The basic chain of events was that variations in sunspots affect the power of the sun’s rays, influencing the bountifulness of harvests and thus the price of corn which, in turn, affected business confidence and gave rise to commercial crises. (Source)
This anecdote would be funny if the economics world had learnt from this failed approach. But it hasn’t. The same failed methodology, pure empiricism still rears its ugly head. The prime example of this is the Phillips’ Curve.
Economist A. W. Phillips plotted the relationship between the rate of price inflation and unemployment in the United Kingdom between 1913 and 1948 and came up with this:
As we clearly see, there is an inverse relationship between the rate of inflation and unemployment. This discovery was a major rationale for central banks need to manage the money supply in order to keep this relationship in check.
But something happened in the 1970s. The relationship between unemplyment and price inflation severed. It became just a random cloud of dots!
But never fear, the economists managed to manipulate the data and “mutate” the curve to fit the data. Now the New Phillips Curve represents the relationship between expected inflation and unemplyment. Who needs a theory when you can make the evidence conform to whatever relationship you want?
So there it is. The folly of economic empiricism revealed once again. You cannot effectively examine any economic data without building a theory beforehand. Any attempt is bound to fail.