Real exchange rate behavior is characterized by both medium-run variability and long-term trend movements. The implication of such movements for production, consumption, and the external balance depends upon the underlying causes of the real exchange rate movement. This paper aims to assess the relative importance of such alternative factors over the short and long horizons. Conceptually, it distinguishes between demand- and supply-side causes of real exchange rate adjustment.
On the supply side, a long-standing view asserts that countries enjoying higher productivity growth in the tradable goods sector will experience a trend appreciation, reflecting higher wage growth. An alternative view focuses on the demand side, arguing that a spending shift toward nontradables, reflecting either changing consumer tastes or an increasing share of government expenditure, tilts demand toward nontradables and thereby generates the real exchange rate appreciation. The two approaches differ both in their assumed causal structure and in their implication for sectoral composition.
The paper assesses the empirical importance of demand and supply factors using a disaggregated data set for fourteen OECD economies between 1970 and 1985. Sectors are classified as tradable or nontradable based on the fraction of output that is actually traded. For each group, the paper constructs measures of total factor productivity growth. Demand-side measures include the share of government expenditure and the level of income.
The paper presents some stylized facts. The data reveal a trend increase in the relative price of nontradables for all countries studied except Canada. For 8 of the 14 economies in the sample, the increase in the relative price is accompanied by an increase in the production share of nontradables, with the notable exceptions of Japan and Norway, suggesting that demand-side factors were of some importance in determining relative price movements.
Turning to common currency inflation rates, the paper detects a fairly small overall correlation across countries for both tradable and nontradable inflation rates, which rejects PPP as a useful characterization for the data. However, when the sample is split, it is found that within the subset of core EMS economies, the correlations of sectoral inflation are close to unity. While the correlations of demand- and supply-side variables are also somewhat higher between the EMS economies, a potential role for an “EMS effect” remains.
The results of a regression analysis reveal that both demand- and supply-side effects played a significant role in determining relative price movements. Separating by duration, demand factors are found to explain more than half of the short-run variation while supply factors explain virtually all of the longer-term movements.