Abstract
The Balassa-Samuelson hypothesis, explaining real exchange rate volatility by the differential productivity of the tradable and nontradable sectors, was found to generally fit macro-economic developments since 1986. It turns out that the traditional measures of real exchange rate - the ratio of export/import prices to business-sector product prices -overstated the extent of real appreciation in 1993, 1997-1998, and 2001, being influenced by declining world trade prices, in comparison to the exchange rate of shekel to the US dollar adjusted by GDP deflators. The latter measure has a U-form with a turning point in 1997, suggesting robust real depreciation since then. The elasticity of this real exchange rate with regard to the appreciation of nontradable goods is estimated at 0.7-0.85, while the elasticity with regard to the terms of trade is unitary.