Determinants of Growth in European Transition Economies – a Bayesian Averaging of Classical Estimates (BACE) Panel Data approach
Can Erbil (),
Zuzana Kristkova and
Bram Smeets
No 3275, EcoMod2011 from EcoMod
Abstract:
This paper applies the Bayesian Averaging of Classical Estimates (BACE) approach to empirically study the determinants of economic growth in a panel data set that focuses on Central and Eastern European transition economies in the period 1995-2008. Although economic growth is used as a key criterion to assess the performance of countries and policy-makers, insights on what factors drive this growth are mixed and often contradictory. By today, a rich literature has evolved on the topic, studying the relevance of a large set of potential candidates as growth drivers. One of the interesting findings in this literature is that, even though the theoretical growth models lie at the core of the macro-economic practice, in empirical studies there are many factors that correlate significantly with economic development. The relevance of development factors in specific models depends on details such as the region and time period under consideration, but often also on just the model specification. In an extreme illustration, Sala-i-Martin () found over 60 variables to be significantly correlated with growth, by running a substantial number of regression analyses and by varying the model specification in each of the replications. In a study that addresses the model uncertainty around the specification of empirical growth models, Doppelhoffer et al. (2003) use Bayesian Model Averaging to assess the robustness of candidate drivers across many regression analyses. The resulting BACE approach is an elegant application of Bayesian techniques, that use simulation methods to more conventional OLS regression results. In a cross-sectional analysis based on 88 countries, Doppelhoffer et al. (2003) are able to isolate 18 variables as robust determinants of economic growth. This paper extends the BACE methodology to a panel data framework. It focuses on 12 countries in Central and Eastern Europe during the period 1995-2008, most of which have been candidate member states of the European Union or joined the Union during this period. Working with a panel data set brings several additional insights compared to the cross-sectional examples mentioned above. First, by adding the time dimension, drivers of fluctuations in growth over time can be analyzed, which can provide more relevant insights for individual countries. In the cross-sectional analysis, time-variant variables (such as economic growth or educational enrollment rates) need to be averaged over time, which fails to capture structural change in the economy of individual countries. Particularly for the European transition countries, it is interesting to study the constituents of the strong economic development that most of them have experienced. Second, adding the time dimension increases the number of data points, allowing for a reduction of the number of countries, which will enhance the relevance of the insights for the specific set that is under consideration. Initial results suggest that out of 15 candidate factors, 5 are robustly correlated with economic growth. Growth in investments and the education enrollment role are found to show the strongest correlation.
Keywords: 12 countries in Central and Eastern Europe during the period 1995-2008; most of which have been candidate member states of the European Union or joined the Union during this period.; Growth; Macroeconometric modeling (search for similar items in EconPapers)
Date: 2011-07-06
References: Add references at CitEc
Citations: Track citations by RSS feed
Downloads: (external link)
http://ecomod.net/system/files/
Our link check indicates that this URL is bad, the error code is: 403 Forbidden (http://ecomod.net/system/files/ [301 Moved Permanently]--> https://ecomod.net/system/files/)
Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.
Export reference: BibTeX
RIS (EndNote, ProCite, RefMan)
HTML/Text
Persistent link: https://EconPapers.repec.org/RePEc:ekd:002625:3275
Access Statistics for this paper
More papers in EcoMod2011 from EcoMod Contact information at EDIRC.
Bibliographic data for series maintained by Theresa Leary ().