Demand for International Reserves: A Quantile Regression Approach

I estimate the determinants of the demand for international reserves using quantile regressions. Employing a dataset of 96 developing nations over the period of 1980-1996, I find considerable differences at different points of the conditional distribution of reserves. The ordinary least squares estimates of elasticities that were found to be insignificant in previous studies become statistically significant at various quantiles of the reserve holding distribution. In particular, I find that the coefficients of interest rate differential and volatility of export receipts are significant and have the signs predicted by the traditional reserve models, but only for those nations that hold the highest amount of reserves. In contrast, the flexibility of the exchange rate does not seem to be an important factor for the nations that are located at the tails of the distribution.