The Balassa-Samuelson effect
Various factors contribute to the change in the real exchange rate of the koruna over time. One of these factors is differential trends in labour productivity in different sectors of the domestic and external economy. The contribution of this factor to the change in the real exchange rate is generally referred to as the Balassa-Samuelson (B-S) effect. Analyses of the B-S effect conducted by the CNB have focused on quantifying this effect on the CZK/EUR real exchange rate. Emphasis has been placed on the comparison with Germany, the dominant trading partner of the Czech Republic. The analyses assume that all items (i.e. goods and services) consumed in the Czech Republic and in Germany can be broadly classed into tradables (for which the costs of international arbitrage are low) and nontradables (with high costs of international arbitrage). Let π T and π N stand for the growth rates of prices in these two sectors, w for the share of tradable output in the consumer basket, and aT and aN for the growth rates of labour productivity; and let an asterisk denote values for Germany.
When quantifying the B-S effect, we assume for simplicity that (a) the nominal CZK/EUR exchange rate changes in accordance with changes in the relative Czech and German prices of tradable items, i.e. that the nominal appreciation of the koruna exchange rate is approximately equal to π T* - π T, (b) that owing to "wage contagion", wages in the tradables and nontradables sectors of the economy rise at a similar rate, and (c) that the share of labour in value added in the two sectors of the economy is approximately the same. Using the standard microeconomic model of a two-sector economy, we can deduce from these assumptions that higher labour productivity growth in the tradables sector than in the nontradables sector will give rise to correspondingly higher growth in output prices in the nontradables sector than in the tradables sector, that is,
π N - π T = a T - a N a také π N* - π T* = a T* - a N*
Therefore, the essence of the B-S effect is that differential trends in labour productivity in the tradables and nontradables sectors of the Czech and German economies affect the differential between the inflation difference in the two sectors in the Czech Republic and the same difference in Germany. It is this differential which determines the change in the real CZK/DEM exchange rate, because we obtain the real appreciation of the koruna's exchange rate when we deflate the nominal exchange rate by the ratio between inflation in the Czech Republic and that in Germany, i.e.: real appreciation of koruna = (π T* - π T) + [ w . π T + (1 - w) . π N] - [ w* . π T* + (1 - w*) . π N*]
To obtain a rough quantification of the B-S effect, it is necessary to make assumptions about the rate of productivity growth in each sector of the Czech Republic and Germany. These assumptions can be based inter alia on past developments in the Czech Republic and on experience in other countries. If, for example, we assume that labour productivity in Germany will rise in future at a rate of 3% in the tradables sector and 2% in the nontradables sector, while labour productivity growth in the Czech Republic will be 5% in the tradables sector and 2% in the nontradables sector, and if we further assume that the share of the tradables sector in the consumer basket will be around 50% in both countries, then the aforementioned methodology implies a B-S effect (i.e. a contribution to the real CZK/EUR exchange rate trend from labour productivity growth) consisting of a real appreciation of the koruna of around 1% a year. This estimate differs little from the results of various studies for the economies of the other Central and Eastern European countries that are candidates for EU and EMU membership.
If the CNB fulfils its inflation targets, and assuming German inflation of 2%, consumer price inflation in the Czech Republic could be 1-1.5 percentage points higher than in Germany on average over the next few years. This inflation differential, coupled with the real appreciation of around 1% derived above, would imply approximate stability of the nominal exchange rate. Consequently, if the aforementioned assumptions hold true, there is little support to be found in the B-S effect alone for the frequently mentioned expectations of a medium-run nominal appreciation of the koruna against the euro.