From discrete to continuous-time transition matrices in intra-distribution dynamics analysis: an application to per capita wealth in europe

Abstract

Previous studies focusing on the intra-distribution dynamics analysis have usually computed, in a Markov chain framework, discrete-time transition matrices. Such an approach, however, can involve some limitations, especially when using stock variables. In order to illustrate the importance of the time-scale issue when estimating transition matrices, this paper applies both discrete and continuoustime approaches to a set of cross-national European data on per capita wealth for the period 2000–10. The results reveal, on the one hand, that the continuous-time estimation provides a most accurate estimation of transition probabilities and, on the other, that the differences between both approaches are especially remarkable in the long-term equilibrium distribution.