A Brief Outline for FEER Framework

When I was at Osaka three weeks, I was asked by couple of fund managers and academic about my evaluation of the equilibrium value of the various Asian currencies.

Here I wish to share the brief outline of the framework behind my Fundamental Equilibrium Real Exchange Rate (FEER), defined as the rate consistent with the economy operating at potential output while at the same time registering a sustainable current account balance. In essence, the underlying current account balance is arrived when the determinant variables are calibrated to values congruent with internal balance. My framework is very much based on the direct estimation method of Clark and McDonald (1998) but calibrated in the conventional FEER manner as in Williamson (1985).

In short, the final reduced form equation is as follow:

FEER = f (yd/yf, TOT, TNT, RIR, NFAR)

where:

yd/yf : relative domestic to foreign output ratio
TOT : terms of trade
TNT : relative ratio of tradable to non-tradable prices
RIR : real interest rate differential between asian-USD assets
NFAR : net foreign asset to nominal GDP

I include a real interest rate differential in order to assess the effects of a shorter-term fundamental determinant to exchange rates and a dummy variable for the effects of exchange controls imposed. Co-integration methods were used to estimate the final model with inclusion for the Phillips-Hansen Fully Modified estimation that allows estimates and standard errors to be derived free from the distortion due to nuisance parameters through a semi-parametric adjustment for serial correlation of innovations and endogeneity of the variables, which is suitable for the relatively small sample size.

Table 1 shows the final model estimation results with relevant diagnostic and it is interesting to note the significant findings on the relative productivity differential and terms of trade variables, which suggest the information contribution of these variables above what is contained in relative output levels. It is shown that exchange control dummy variable was significant and having an opposite sign to relative real interest rates indicating the effects of the capital restrictions in muting the response of capital flows to interest rate differentials after the imposition of the controls.

Table 1: Summary of Results
Variable Coefficient Standard-Error
Constant 2.908 0.892
yd/yf -0.301 0.066
TOT 0.614 0.003
TNT -0.201 0.004
RIR 0.158 0.003
NFAR -0.153 0.005
Capital Control
Dummy -0.201 0.054
Overall F-Test 290.8 9c.v. = 2.5@99% significance level