Abstract:
In March 2008 the Bank of Israel started buying foreign currency, one of the aims of which was to affect the exchange rate. This study examines the effect on the NIS/US$ nominal exchange rate of the Bank's intervention in the foreign currency market. It does so by comparing the dynamic and static forecasts of the exchange rate, using an unrestricted VAR model estimated using the Bayesian method, with the actual exchange rate in the intervention period. Before performing the comparison the quality of the model's forecasts was examined, and it was found that the model produced good forecasts (given the actual values of the model's exogenous variables). In particular it was found that the forecasts were better than those obtained from the random walk model. This study found that the changes in the Bank of Israel's policy of intervention in the foreign currency market in March 2008, July 2008 and August 2009 resulted in shekel depreciation. The strongest estimated effect was recorded in the period after July 2008 when the scale of the purchases was increased, and the deviation of the actual exchange rate from its forecast level exceeded 10 percent. It was also found that the effect started to wane at the end of 2008, and that in the first half of 2009 the gap between the actual exchange rate and its expected level in the absence of Bank of Israel intervention narrowed, until it widened again when the Bank changed its intervention policy from one of fixed daily purchases to ad hoc purchases.