Accuracy of GDP growth forecasts for transition countries: Ten years of forecasting assessed

RETRACTED ARTICLE: Estimation of the most influential science and technology factors for economic growth forecasting by soft computing technique

“…In paper Krkoska and Teksoz (2009) analyzed the performance of GDP growth and inflation forecasts for 25 transition countries between 1994 and 2007 and the empirical results were shown that there is a positive correlation between the number of forecasters covering a given country and the forecast accuracy. The GDP forecast accuracy improves with progress in transition as well as with the expansion in the information domain (Krkoska and Teksoz 2007). The relative performance of several factor models to forecast GDP growth using a large monthly dataset was analyzed in article (Dias et al 2015).…”

Analysis of forecasts of GDP growth and inflation for the Croatian economy

“…Therefore, in the last decade, research papers and reports emerged combining the above approaches as standard practice in order to assess an institution's forecasts, often written by the institution's own staff or experts working in association with them. Examples include research by Krkoska and Teksoz (2007) for EBRD's forecasts, Cabanillas and Terzi (2012) and Fioramanti et al (2016) for European Commission's forecasts and evaluation reports on their own forecasts by the IMF (2014) and the Bank of England (2015).…”

“…Regarding average MAE for the year-ahead forecasts Estonia, Latvia and Lithuania stand out with MAE values of 4.18, 4.54 and 3.18 respectively. Similarly, Krkoska and Teksoz (2007) report average MAE of 3.15 for the year-ahead forecast horizon regarding EBRD's forecasts of GDP growth rates for the south-eastern Europe transition countries. In the case of the central eastern Europe and the Baltic states the average MAE was 1.90.…”

“…They further point out that assessing the GDP forecast accuracy is not a simple problem as the official GDP data is a mobile target guessed and revised by the national statistical offices to reduce the share of approximation. Krkoska and Teksoz (2007) deal with this issue too, mentioning that: ' … even the most advanced transitioning countries continue to make substantial and frequent changes in their estimation methods … ' which can ' … lead to spurious evidence of increasing forecast accuracy over time … ' if only the most recently revised official data is used (p. 33). Regarding the state of national accounts in Croatia, it should be noted that in the 2006-2015 period it usually took the Croatian Statistical Bureau 3 years to come up with the final value for the GDP growth rate so the GDP growth rate lagged 3 years presents the final value in 8 out of 10 times (in 2009 and 2011 after 3 years the available information was still reported as an estimate).…”

“…Although the European Commission's (EC) forecasts have already been assessed twice and regardless of the fact that the EC has monitored Croatia's macroeconomic development ever since it was officially granted its European Union (EU) candidate country status in 2004, Croatia has been left out of the recent study due to insufficient data. Thus, to the best of the authors' knowledge, the research by Krkoska and Teksoz (2007) remains the only analysis of forecast accuracy to include data regarding Croatian economy. However, the research was based on EBRD's GDP growth forecasts and included 25 transition countries from central and eastern Europe and former Soviet Republics with the results only being reported at the aggregate level (for the whole sample and three sub-regions the countries were grouped into).…”