452 Demonstration of Final Research Paper (May 2000)

1. The hypothesis is that the dollar exchange rate and the U.S. unemployment rate are positively related. This is based on the theory that as the dollar rises in value, net exports become more expensive to foreigners and decline. This reduces the IS curve as well as AD and hence reduces GDP and raises unemployment. This of course assumes that the Mashall-Lerner condition is above one and consumers in both the US and our trading partners are responsive to price changes as signaled by the changing exchange rates.

2. The variables for the test are the following:

(1) the dependent variable, the US unemployment rate, is taken from Robert E. Gordon, Macroeconomics. The independent variable, the number of pounds per dollar, is taken from the same source. Data were compiled on a quarterly basis from 1974.1 through 1992.2. You will note I began after 1973 when the world went to flexible exchange rates.

3. The results of the test were the following:

U = -1.2 + 0.345 $

The sign was appropriate for the hypothesis. The test t-value of 4.53 for 72 degrees of freedom was statistically significant at the .005 level seeming to confirm the hypothesis. The R2 was 22% indicating that over a fifth of the variation in the unemployment rate was associated with variation in the exchange rate. However, the Durbin-Watson statistic was 0.96 indicating the presence of auto-correlation and suggesting that the relationship as measured was unreliable, in fact high biased.

To overcome auto-correlation, the data were transformed into first-differences, and the regression re-run. The results were: U = -1.0 + 0.026 $. Although the sign was appropriate, the test t-value was 1.70, statistically significant at the .05 level for 70 degrees of freedom. The R2 fell to 10%, but the Durbin-Watson rose to 1.76 indicating no serious auto-correlation.

In summary, the hypothesis was confirmed but nine-tenths of the variation in unemployment was left to the operation of other factors outside the model tested. These include all the factors affecting GDP growth like fiscal and monetary policy, the climate of investment, factors affecting consumption and saving as well as international events and trade partner income levels affecting net exports. Part of the weak explanatory power of the model may also be due to the operation of low demand elasticities, the so-called Marshall-Lerner condition, which might reduce the responsiveness of spending and unemployment rates in the short run to exchange rate changes.

In conclusion, this paper successfully confirmed the positive relation between the exchange rate and the unemployment rate for the US since 1974. Decision makers must be made aware that any policy favoring a stronger dollar will have a negative effect on employment. On the other hand, the employment effects are weak. Further research in the future using lagged unemployment responses to changes in the value of the dollar may reveal a stronger impact.