Working Paper: NBER ID: w1406
Authors: John H. Makin; Raymond D. Sauer
Abstract: This paper presents results of estimating an exchange rate equation in light of theoretical considerations regarding changes in sterilization and intervention policy and tax policy which imply that the coefficients in the equation will not behave as fixed parameters in a given sample period,as standard econometric practice assumes. We compare the results of ordinary least squares and a random coefficients model of the Japanese Yen-- U.S. dollar exchange rate during the floating period of July 1973 through June 1982.When systematic end of year policy changes affecting Japanese reserves are explicitly modeled, both OLS and the random coefficients model show increased explanatory power. The random coefficients model appears to be superior to OLS however; by allowing the coefficients to vary over time as required by the economic theory discussed above, estimates of the mean response coefficients for the floating period all have the hypothesized sign, and explanatory power is sharply increased.
Keywords: exchange rate; policy regime changes; random coefficients model; intervention; sterilization
JEL Codes: F31; E52
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| systematic and unsystematic policy regime changes (E65) | coefficients in the exchange rate equation (F31) |
| intervention policies (J68) | exchange rate (F31) |
| systematic end-of-year policy changes in Japanese reserves (E63) | explanatory power (C20) |
| relative excess money supplies (E51) | exchange rates (F31) |
| failure to account for systematic changes in intervention policy or tax rates (E65) | biased estimates (C51) |