Markov Chain approach to Purchasing Power Convergence in the 15 European Union

In the present paper we study the degree of convergence in the European Union from the Purchasing Power Parity (PPP) point of view. The price of the shopping basket can be the cause of disparities in a global market in construction that, like the European Union, is formed by different countries with different consumption habits. In addition, in this construction process twelve out of fifteen countries of the EU have left its national currency to adopt the Euro like common currency. Therefore, it is necessary for the stability of the Union process in the long run that, among others, purchasing power of the different state members tends towards a same common value. Moreover, the question is whether that process of convergence within the European Union is taking place or not. In order to solve this question, the series of the Absolute Purchasing Power Parity (APPP) are estimated through the suggestion of Rodriguez et al (2004). These authors use the Harmonized Consumer Price Index in the European Union and the nominal exchange rates of the different currencies with euro. Monthly estimates of the APPP series for the 1995-2002 period are obtained for each of the fifteen countries. These figures show, for each country, their relative position to the average value of the European Union. Using these series we applied the Markov Chain methodology to study the time evolution of the distribution of APPP in the European Union. This methodology has been very used by its facility of calculation and interpretation of the results. Nevertheless, with the purpose of obtaining good estimations it is necessary to solve the discretization problem of a continuous variable. This is, to use a finite set, and relatively small number of states, for a variable with infinite values. In the present work different approaches are used to solve the problem. We test for structural change on the estimated probabilities using adapted test to Markov Chains. This allows us to study if an effect exists on the Purchasing Power Parity with the entrance of the Euro. Markov Chains are estimated by Maximum likelihood, and allow us to do different analyses. In the first place, we can study the mobility of the distribution, measured through the probabilities of permanence or not in the same state, and in the degree of diagonal structure of the resulting matrix. This objective can obtained by direct observation, calculating Mobility Index, or using expected time of first passage. Secondly, we can obtain the ergodic or long term distribution. This one shows the temporary evolution in the long run of the distribution, under the hypothesis of maintenance of the present conditions. This distribution would show the possible convergence or not of the whole distribution. We also estimate elasticities of ergodic probabilities, to analyze the effect of each probability in the Markov chain in the long run distribution. Results show differences with the Euro Entry, mobility towards convergence within the distribution is slow, with high elasticities of the ergodic distribution to changes in the transition probabilities.