Subjective Shifting Endpoints and FX Puzzles

Joe Kupferberg, Economics PhD student

This project studies a new mechanism for exchange rate determination: time variation in market participants’ perceived long-run equilibrium exchange rates. Using daily research reports written by FX analysts at major broker-dealers, I construct a high-frequency measure of these beliefs and show that they vary substantially over time—far more than standard PPP or BEER benchmarks. I combine these text-based measures with survey data on dealers’ interim FX and macroeconomic forecasts and document that shifting subjective endpoints can account for several puzzles in the FX literature. In particular, I estimate that variation in perceived long-run anchors explains 65 to 85 percent of exchange rate fluctuations in levels, providing a new interpretation of excess volatility. I relate this mechanism to the persistence of carry premia. Finally, I study belief formation. Dealers occupy a unique position in FX and bond markets, yet survey data suggest that they do not anticipate many empirical regularities in risk premia. I examine how macroeconomic information and client flows are incorporated into shifting endpoints, shedding light on the macro exchange rate disconnect and return predictability from order flow.