Are Our FEERs Justified?
Paper number: 00/02
Paper date: 2000
Paper Category: Working Paper
Giacomo Barisone, Rebecca L. Driver and Simon Wren-Lewis
Estimates of ‘equilibrium’ values of real exchange rates often play an important role in macroeconomic policy debates, both under fixed and flexible exchange rate regimes. Although PPP based estimates are frequently used in this context, there remain serious doubts about their theoretical and empirical validity, particularly over a medium term time horizon. A number of alternatives to PPP have been proposed, the most widely used of which are concepts related to Williamson’s Fundamental Equilibrium Exchange Rate (hereafter FEER).1 Although there exists an extensive literature on testing PPP, there has been relatively little research on testing alternatives.
FEERs can be calculated in a number of ways, but the most widely used approach is based on a partial equilibrium analysis, which inputs exogenous estimates of trend output and structural capital flows into a model of aggregate trade. (For an extensive review of this and other approaches, see MacDonald and Stein, 1999). Estimates using this method have been influential in a number of policy debates (e.g. Williamson, 1991). However, FEER estimates usually relate to a specific point in time, so the approach has never been systematically tested in a time series context.
In this paper we present a test of FEER estimates for six major industrialised countries based on cointegration analysis. If PPP does not hold over the medium term, then real exchange rates over the last 20/30 years rate might well be non-stationary, and this is the case for the data we use. A natural question to ask of FEERs is whether they can account for this non-stationary behaviour i.e. whether FEER estimates and real exchange rates cointegrate. We examine this for each country both individually, and collectively using panel unit root tests.