On Business Cycle Asymmetries in G7 Countries

Time Varying Cyclical Analysis for Economies in Transition

“…In fact Chauvet and Potter (2001) find that the US business cycle was in line with the G7 from the mid 70s, but then diverged thereafter. Kiani and Bidarkota (2003), Watson (2002, 2003) likewise find divergence caused by structural breaks. All these results suggest a time-varying approach is going to be necessary if we are to analyse the state of convergence among economies that are in transition from planned to market economies.…”

Fluctuations in Economic and Activity and Stabilization Policies in the CIS

“…In this context Auerbach (1982), Gordon (1986), Kling (1987), Koch and Rasch (1988), Diebold and Rudebusch (1989), Hamilton (1989) and Estrella and Mishkin (1998a,b) have all provided guidance but there is still ample need for additional non-linear models in business cycle research. A number of studies that include Neftci (1984), Brunner (1992Brunner ( , 1997, Beaudry and Koop (1993), Potter (1995), Ramsey and Rothman (1996), Bidarkota (1999Bidarkota ( , 2000, Anderson and Vahid (1998), Anderson and Ramsey (2002) and Kiani and Bidarkota (2004) demonstrated significant evidence of asymmetries in business cycle fluctuations in macroeconomic time series data. The results from these studies show most macroeconomic time series are nonlinear; therefore, linear models should not be employed for anticipating fluctuations in economic activity.…”

“…Andreano and Savio (2002) used Markov switching models for testing asymmetries in business cycle fluctuations in the Group of Seven (G7) industrialized countries' output but failed to detect the possible existence of business cycle asymmetries in France, Germany, and the UK. Likewise, Kiani and Bidarkota (2004) employed alternate regime switching and other highly sophisticated time series models encompassing long memory to account for persistence of the process, conditional heteroskedasticity to account for time varying volatility, and stable distribution to account for outliers that might have been present in the series for testing business cycle asymmetries in G7 countries' real GDP growth rates. Kiani and Bidarkota (2004) made some progress over Andreano and Savio in testing possible existence of asymmetries, but contrary to their initial hypothesis that "business cycles in all the countries were alike," Kiani and Bidarkota were not able to detect asymmetries in the France and UK series.…”

“…Likewise, Kiani and Bidarkota (2004) employed alternate regime switching and other highly sophisticated time series models encompassing long memory to account for persistence of the process, conditional heteroskedasticity to account for time varying volatility, and stable distribution to account for outliers that might have been present in the series for testing business cycle asymmetries in G7 countries' real GDP growth rates. Kiani and Bidarkota (2004) made some progress over Andreano and Savio in testing possible existence of asymmetries, but contrary to their initial hypothesis that "business cycles in all the countries were alike," Kiani and Bidarkota were not able to detect asymmetries in the France and UK series. This would pose a challenge to macro-theorists for building new business cycle theories.…”

“…Artificial neural networks, approximated with genetic algorithms (GAs), have demonstrated superiority over other sophisticated nonlinear and various switching time series models in testing business cycle asymmetries in G7 countries (see Andreano andSavio 2002 andBidarkota 2004) and have exhibited superior performance for predicting recessions (see Kastens 2006 andKiani 2007) when compared to recession forecasts obtained from probability models, such as those employed by Estrella and Mishkin (1998a,b). Therefore, the present study seeks to predict fluctuations in economic activity in the selected CIS economies using ANNs.…”