ABSTRACT
Exchange rate determination has been the “holy grail” of international finance and macroeconomics ever since the collapse of the Bretton Woods regime in 1971 and the ensuing period of high exchange rate volatility. The work by Obstfeld and Stockman (1985) is an excellent survey of models put forth in the 1970s and early 1980s. Of these, perhaps themost successfulwasDornbusch’s (1976) overshooting model, centered on the assumptions of uncovered interest parity (UIP) and sticky prices. Dornbusch clarified how exchange rate volatility was indeed consistent with rational behavior in the presence of sticky prices, which would cause the short-run response of the exchange rate to shocks to overshoot the new long-run equilibrium level. Sadly for a generation of promising theoretical work, Meese and Rogoff
(1983) documented evidence that the assumption that the exchange rate is simply described by a randomwalk processwould performbetter than the theoretical competitors at predicting the path of the exchange rate at business cycle frequencies. Since then, Meese and Rogoff ’s (1983) result has been among the major hurdles that theoretical work in search of the “exchange rate grail” has had to overcome. Anothermajor stumbling bloc has been the evidence in favor of delayed overshooting in Clarida and Gali (1994) and Eichenbaum and Evans (1995). Dornbusch’s overshooting model predicts that the exchange rate should overshoot its new longrun position on impact in response to a monetary shock. But empirical evidence suggested that overshooting actually takes place several periods after shocks, a finding that was interpreted as evidence against the importance of UIP in exchange rate determination. Theoretical research on exchange rates developed renewed momentum with
the publication of Obstfeld and Rogoff ’s (1995) seminal article, “Exchange Rate Dynamics Redux.” There, Obstfeld and Rogoff put forth a fully microfounded, general equilibrium model of international interdependence and exchange rate determination with an explicit role for current account imbalances.1 Nevertheless, the non-stationarity of the Redux model led most of the subsequent literature in the so-called “new open economy macroeconomics” to develop in different directions and “forget” the insights of the model on the dynamic relation between
the exchange rate and net foreign asset accumulation by de-emphasizing the role of the latter.2 (The assumption of purchasing power parity (PPP) was admittedly another weakness of the Obstfeld and Rogoff (1995) model on empirical grounds, addressed by several subsequent contributions. Yet, it was not PPP that motivated most scholars to de-emphasize the role of net foreign asset dynamics.) US data show a growing and persistent current account deficit over the 1990s,
that is, capital inflow and accumulation of a large foreign debt. During the same period, the dollar has appreciated steadily. It is a commonly held view that the advent of the “new economy” has been the most significant exogenous shock to affect the position of the US economy relative to the rest of the world in recent years. We can interpret this shock as a (persistent) favorable relative productivity shock. A story that one could tell about the behavior of the dollar andUSnet foreign assets in the 1990s is that the shock caused the United States to borrow from the rest of the world and the capital inflow generated exchange rate appreciation. This story could be reconciled with models of exchange rate determination developed in the 1970s and early 1980s.3 If the shock is taken as permanent, the story can also be reconciled with Obstfeld and Rogoff’s model. Nevertheless, the argument cannot be reconciled with the overwhelming majority of new generation models that followed. We returned to the original intent of Obstfeld and Rogoff’s work in Cavallo and
Ghironi (2002). In that article, we developed a two-countrymodel of exchange rate determination in which stationary net foreign asset dynamics play an explicit role. We dealt with indeterminacy of the steady state and non-stationarity of the original incomplete markets setup by adopting the overlapping generations framework illustrated in Ghironi (2000). If exogenous shocks are stationary, the departure from Ricardian equivalence generated by the birth of new households with no assets in all periods is sufficient to ensure existence of a determinate steadystate distribution of assets between countries and stationarity of real variables. Unexpected temporary shocks cause countries to run current account imbalances, which are re-absorbed over time as the world economy returns to the original steady state.4
