Working Paper: NBER ID: w23341
Authors: Woo Jin Choi; Alan M. Taylor
Abstract: We document a new international stylized fact describing the relationship between real exchange rates and external asset holdings. Economists have long argued that the real exchange rate is associated with the net international investment position, appreciating as external wealth increases. This mechanism has been seen as central for international payments equilibrium and relative price adjustments. However, we argue that the effect of external assets held by the public sector—reserve accumulation—on real exchange rates may be quite different from that of privately held external assets, and that capital controls are a critical factor behind this difference. For 1975–2007, controlling for GDP per capita and the terms of trade, we find that a one percentage point increase in external assets relative to GDP (net of reserves) is related to an 0.24 percent real exchange rate appreciation. On the contrary, a one percentage point increase in reserve accumulation relative to GDP has virtually no effect on the real exchange rate in financially open countries (low capital controls), and is related to a 1.65 percent real exchange rate depreciation in financially closed countries (high capital controls). Results are stronger in developing countries and in more recent periods. Gross rather than net positions matter and we present a new theoretical model to account for the stylized fact. The framework encompasses so-called precautionary and mercantilist motives for reserve accumulation, and also explains how the optimal capital account policy—the mix of reserve accumulation and capital controls—is determined. Further empirical support arises from evidence that reserve accumulation is associated with a trade surplus, along with higher GDP and TFP growth in countries with high capital controls, findings that are consistent with the mechanisms of our model.
Keywords: Reserve Accumulation; Capital Controls; Real Exchange Rate; External Assets
JEL Codes: F31; F38; F41; F43; O24
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| External assets relative to GDP net of reserves (F34) | Real exchange rate appreciation (F31) |
| Reserve accumulation relative to GDP (H60) | Real exchange rate depreciation (in financially closed countries) (F31) |
| Reserve accumulation relative to GDP (H60) | Real exchange rate (virtually no effect in financially open countries) (F31) |
| Reserve accumulation (F32) | Real exchange rate (F31) |
| Reserve accumulation (F32) | Real exchange rate (stronger effects in developing countries) (F31) |
| Reserve accumulation (F32) | Real exchange rate (stronger effects in more recent periods) (F31) |