The profitability of currency carry trades in and of itself is 'economic' evidence against the uncovered interest parity (UIP) condition. There is a wide variety of 'statistical' evidence against UIP. Yet the relationship between these two types of evidence and their implications for time variation in risk premia is not fully understood. Furthermore most of the literature has focused on the currencies of industrialized economies. The failure of UIP in emerging market currencies and its implications for the risk premia of these currencies has received considerably less attention. In this paper I reconsider UIP the carry trade and the behavior of risk premia and draw comparisons between currencies in industrialized economies and those in emerging markets.