Preliminary Draft Not for Quotation

Prior studies have found that black market rates are good predictors of the official exchange rates. Typically, there is a black market premium for each country. In the case of recently-debated the Central American Free Trade Agreement (CAFTA) nations, the black market rate possesses long-run informational efficiency. This study examines the long-run informational efficiency and short-run predictability of the black and official exchange markets in the DR-CAFTA partners to shed fresh light on the two markets way of processing information because CAFTA wants to follow suit of NAFTA to increase the trade and financial capital flow. Applying a battery of non-parametric as well as time series models, this study finds that the black market information exhibits a non-random behavior. Furthermore, there exists a long-run cointegration relations between the black and official pair for the four of the six countries. Finally, the predictability runs from the black to official rates confirmed by the vector error correction model and the weak exogeneity (and Granger causality) tests, The results can have implications for financial managers and to manage exchange rate risks and for policy-makers to use black market exchange rates as a guide for setting the official rates