The Fisher effect at the borders of the European Monetary Union: evidence from post-communist countries

Abstract

This article aims to shed some light on the Fisher effect in six non-eurozone post-communist economies (the Czech Republic, Hungary, Latvia, Lithuania, Poland and Romania). A Fisher-type interest rate equation is analysed using the PMG panel data estimator, with an early attempt to employ the Harmonised European Consumer Surveys in quantifying inflation expectations and uncertainty. The output gap is also considered as an explanatory variable. The results of the Hausman poolability test unambiguously confirm that, despite the differences in their monetary regimes and the development levels of their financial markets, all the countries observed are homogeneous in terms of the Fisher effect. It is shown that both inflation uncertainty and expectations positively and significantly feed into nominal interest rate fluctuations. The post-communist central banks seem to be the most strongly concerned about inflation uncertainty shocks, while their interest rate elasticity with regard to expected inflation is below unity. On the other hand, they do not significantly adjust their interest rates in response to demand-side pressure, assigning only a secondary role to boosting economic activity. As a robustness check, the results obtained remain rather similar when Sweden and the UK (as the remaining non-eurozone EU members) are included in the sample.

Notes

1. CS were first used in the University of Michigan in 1946. Over time, CS have become an integral part of (macro)economic modeling in numerous countries. On the European level this led to the start of integration which was formalised in 1961 through The Joint Harmonised EU Programme of Business and Consumer Surveys. This document technically regulates and unifies the methodology for conducting CS, which enables direct comparability of CS results between EU members and accession countries (European Commission 2007).

2. One possible alternative is to apply the nonlinear regression approach (Smith and McAleer 1995) which is however dismissed by Nardo (2003) as inferior to the probability method.

3. This was the case for some of the countries in this study, so the CP method had to be ruled out as a potentially suitable quantification method.

4. Bulgaria and Denmark were excluded from the sample because their CS data are not publicly available.

5. All long-run coefficients are the same: .

6. Short-run results are quite diverse among the countries observed. They may be obtained from the authors upon request.