Real exchange rates: are they dominated by fundamental factors?

ABSTRACT

Using Bayesian model averaging, we determine which fundamental pair-wise differences suggested by the literature on optimum currency areas give the best explanation of medium-term variability of bilateral real exchange rates. The intercept in the best specification is statistically insignificant, implying that for a hypothetical pair of economies for which the differences were zero, the bilateral real exchange rate would not move. Thus, the ‘non-fundamental’ element of the medium-term real exchange rate variability is, in our sample at least, negligible on average. In other words, floating exchange rate does not in itself imply, on average, more real exchange rate variability in the medium term than an exchange rate peg.

KEYWORDS:

• Real exchange rate
• asymmetric shocks
• optimum currency areas
• Bayesian model averaging

JEL CLASSIFICATION:

• F31
• F45
• C11

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplemental data

Supplemental data for this article can be accessed here.

Notes

¹ RER = NER*.(P^F/P), where RER is the real exchange rate, NER is the nominal exchange rate in the sense of the price of one unit of foreign currency, P^F is the foreign price index and P is the domestic price index.

² Later on, when we report the results of our regressions, MAC(R^τ_IJ) will be referred to as MACRERτ, e.g. MACRER3 for τ = 3.

³ For mutual trade links and for policy variables such as fiscal surpluses and monetary policy autonomy and transparency we take, where possible, their values lagged by three years to tackle potential endogeneity.

⁴ As is clear, our approach does not allow us to assess the degree to which the real exchange rate changes due to the inflation differential versus due to changes in the nominal exchange rate.

⁵ The basis here is the IMF de facto classification of exchange rate regimes (see the appendices of IMF Annual Reports during 1999 - 2008). Up to 2004, the Latvian currency was pegged to the SDR, where the weight of the euro was about 40%; the currency was pegged to the euro only as from 2005 (i.e. for the last four years of our sample period). Nevertheless, for simplicity, in the regressions, we take the currency as a peg to the euro.

Additional information

Funding

This work was supported by the Grant Agency of the Czech Republic under Project No. 16-22540S. The views expressed in this paper are those of the authors and do not necessarily represent those of the Czech National Bank. All errors and omissions remain entirely the fault of the authors.