According to standard Balassa-Samuelson effect, when the productivity increases in the tradable sector (GDP), price increase can be observed in non-tradable sector and this can lead to appreciation of real exchange rate - This is what all of other papers explain, but the paper that I follow gives different result. When Greek real GDP per capita increases 10%, the real exchange rate falls 4,8%. (depreciation). This estimation follows another paper that explains undervaluation can cause economic development and productivity. In my opinion, GDP and real exchange rate have an inverse relationship that's why the author estimates that as exchange rate undervalued, GDP level gets larger. Is this true?