Jon’s Blog

The 1972 article by Stephen J. Turnovsky and Michael L. Wachter used direct observations of wage and price expectations to test the accuracy of the “expectations hypothesis”. The expectations hypothesis was one of a few ideas developed to better explain the theory behind the Phillips curve. The hypothesis says that the current rate of change in wages is dependent on the rate of change in prices, wages, and unemployment rate in the future. It also implies there is a vertical Phillips curve passing through the “natural” rate of unemployment in the long run.

The data on price and wage expectations was used to accomplish to things. First, itwas used as explanatory variables to test the “expectations hypothesis”. Second, the data was used to explain the actual expectations in wage equations. Below is the basic wage equation used to determine expectations:
- Equation: wt= ao + a1 Ut -1 +a2 p*
Where,
- wt = average percentage change in the money wage rate in period t
- Ut = average rate of unemployment in period t
- p* = average expectations of the future rate of inflation, held during period t
However, this is only the most basic equation used, as several other values can be added to it. A “profit rate” can be used to represent informational input. Also, there could be a “catch-up” term or lagged variables to reflect adjustment effects.

The data used in the paper was collected from past issues of the Philadelphia Bulletin. The editor of the paper surveyed between 40 and 60 business economists in different professions, asking them their estimation of various economic indicators. The indicators included the Consumer Price Index (CPI) and average weekly wages in manufacturing. The survey was conducted every two years from 1949 until 1969, and asked for expectations both six and twelve months in the future. The data is not perfect, as the economists may not reflect the expectations of decision makers. However, it is the best available data on expectations.

Numerous regressions were run using ordinary least squares but only the “best fitting” results were discussed in the paper. Two major findings came from the results. The first was that price and wage expectation series are significant in explaining actual wage changes. Second, the coefficients for price and wage expectational variables are consistent for the different equations. While the results were positive, there was a possibility of error in variables or equation bias. As for the actual wage expectations, the sum of the coefficients were near zero. This suggests that in the short run money wage expectations are likely to be influenced by current conditions in the labor market.

“A Test of the “Expectations Hypothesis” Using Directly Observed Wage and Price Expectations”
Stephen J. Turnovsky; Michael L. Wachter. The Review of Economics and Statistics, Vol. 54, No. 1. (Feb., 1972), pp. 47-54.