This paper deals with the analysis of inflation in Kenya. We use data from the Consumer Price Index and inflation to examine, among other issues, features such as the presence of structural breaks, non-linearities and time dependence or persistence in the data. The results indicate the presence of a structural break in the inflation rate taking place in 1994, observing a level shift and a reduction in the degree of integration of the series. However, the results change depending on the assumptions made on the error term. Thus, under no autocorrelation the unit root hypothesis cannot be rejected in any of the two subsamples; however, under the more realistic assumption of autocorrelation, mean reversion takes places in the two subsample, with a slightly reduction after the break. These results are explained in terms of the policy reform and intervention during the last two decades, and some policy recommendations are also presented in the paper.