The paper sheds light on the long-run dynamics between the official and parallel market exchange rates in five South Asian countries, using the relatively new Johansen cointegration approach and a few variants of causality tests. While the evidence shows strong long-run relationship between the two rates, neither the transformed official and the parallel market rates enter the long-run equilibrium affect both the rates equally. The evidence of direct causal relationship between the two rates is very weak. However, strong feedback mechanism between the two rates is evident if the effect of error-correction term is considered in the analysis.