Money demand and disinflation in selected CEECs during the accession to the EU

Abstract

A panel data set for six countries (Czech Republic, Hungary, Poland, Romania, Slovakia and Slovenia) is used to estimate money demand with panel cointegration methods over the recent disinflation period. The basic money demand model is able to convincingly explain the long-run dynamics of M2 in the selected countries. However, money demand is found to have been significantly determined by the euro area interest rates and the exchange rate against the euro, which indicates possible instability of money demand functions in the Central and Eastern European countries. Therefore, direct inflation targeting is an appropriate monetary regime before the eventual adoption of the euro.

Notes

¹ We concentrate in this contribution on the Czech Republic, Hungary, Poland, Slovakia and Slovenia, which joined the EU in May 2004, and on Romania, which followed in 2007. Slovenia introduced the euro in 2007.

² Given the objective of these countries to fulfil the inflation Maastricht criterion (that is, to reduce the inflation differential to the three best performing EU countries below 1.5% points), the Czech Republic, Hungary, Poland, Romania and Slovakia (in a combination with the ERM II participation) have recently introduced official inflation targets (Jonas and Mishkin, 2003).

³ The changes in the monetary regime took place in the Czech Republic in 1997, in Slovakia in 1998, in Poland in 2000 and finally in Hungary in 2001.

⁴ Estimations with the longer, unbalanced sample were used in order to check the robustness of the parameter estimates to the inclusion of earlier transition periods (available upon request from author). In general, the parameters remain in the range of those presented for the balanced sample.

⁵ The results of the ADF test and of the test according to Kwiatkowski et al . (1992) for all variables are available from the authors on request. For the interest rate in the euro area, which is used in the subsequent analysis, the ADF test with two lags is −1.236 for the levels and −4.889 for the first differences (critical values are −2.889 at 5% and −3.493 at the 1% significance level).

⁶ Banarjee (1999) provides detailed surveys of panel unit root tests.