OPEN-ECONOMY MARKET: Model Calibration 2

To examine the properties of this model, we have calculated impulse response functions for the main endogenous variables in response to the system’s five shocks. For this exercise, we have used the following policy rule, which features
Impulse response functions, depicting the reaction to unit shocks to zRti (1 – ph)~]vh ey*h eKl, and еш are plotted in Figures 1-5, respectively.
The main interest, probably, resides in Figure 1, which gives responses to a unit shock in the monetary policy rule (7.1). In the upper left panel of Figure 1 we see that output drops by 0.4 units in response to a one unit unexpected increase in Rt other.

The largest effect is in the first period after the shock and there is considerable persistence, so that about half of the effect still remains 10 quarters after the shock. This response pattern for output is fairly similar to that depicted by Rotemberg and Woodford (1997, p. 306) — and used as one of the three impulse response functions that their estimation procedure seeks to match. The other two functions considered by Rotemberg and Woodford (1997, pp. 321-323) are the responses of inflation and Rt to an Rt policy shock.

For the Rt response, our pattern matches the Rotemberg-Woodford VAR pattern rather nicely, although theirs returns to approximately zero after two periods while some effect remains in ours (see the bottom right panel). As for the inflation variable, the maximum response in ours (bottom left panel of Figure 1) is much larger than Rotemberg and Woodford’s, but only for a very few periods. There is, evidently, less inflation persistence in our model than is the case in reality, but there is some present in our Figure 1 plot, nevertheless.

The remaining panels in Figure 1 show responses of the price level, the nominal exchange rate, and the real exchange rate. The price level response begins only after one quarter, of course, and bottoms out after four quarters, which is perhaps a bit sooner than in reality. But the contrast with st, the (log of the) nominal exchange rate, is qualitatively as one would expect. Thus the exchange rate response (an appreciation) occurs in the first period and is almost four times as large as the maximum price level response. Then st moves back, as in “overshooting” models, to reflect a much less strong long-term effect. The real exchange rate qt moves with st in the first quarter, but returns more quickly to its original value, which reflects long-run monetary neutrality.

Tags: , ,