This paper presents tests of uncovered interest parity in Croatia, the Czech Republic, Hungary, Poland and Romania; all countries in Central and Eastern Europe with floating exchange rates. Data are monthly and the trading horizon is three months. The estimations show that the UIP hypothesis is rejected for the full sample from 1999 to 2011 for all five countries. A number of reasons for the rejection were investigated. Rolling regressions show that standard versions of the UIP essentially lose all explanatory power in 2008-10, which was a period in which the global financial crisis led to instability in currency and interest markets in Central and Eastern Europe. Two indicators of global risk aversion were also found to enter significantly in the many UIP estimations. Finally, the size of the interest rates spread also seems to be of importance, at least for Poland and Romania