The Reaction of Reduced-Form Coefficients to Regime Changes: The Case of Interest Rates
NBER Working Paper No. 1379
This study investigates whether the apparent intertemporal instability of a particular reduced-form equation (that for interest rates) can be explained by changing government policy parameters, or regimes, and otherwise stable structural parameters. We hypothesize that major fiscal, monetary, and regulatory policy parameter shifts have been important sources of that instability. Direct tests imply that reduced-form coefficients move by statistically significant and economically meaningful amounts in response to policy parameter change. Allowing for this systematic parameter variation produces greater stability in the remaining parameters. Furthermore, in-sample and out-of-sample forecasts from the proposed model out perform those from the non-responsive parameter specification.