DP13171 Modeling of Economic and Financial Conditions for Nowcasting and Forecasting Recessions: A Unified Approach
|Author(s):||Sumru G. Altug, Cem Cakmakli, Hamza Demircan|
|Publication Date:||September 2018|
|Keyword(s):||Bayesian inference, Business cycle, Coincident economic index, Dynamic factor model, Financial conditions index, Markov switching|
|Programme Areas:||Monetary Economics and Fluctuations|
|Link to this Page:||www.cepr.org/active/publications/discussion_papers/dp.php?dpno=13171|
This paper puts forward a unified framework for the joint estimation of the indexes that can broadly capture economic and financiall conditions together with their cyclical regimes of recession and expansion. We do this by utilizing a dynamic factor model together with Markov regime switching dynamics of model parameters that specifically exploit the temporal link between the cyclical behavior of economic and financial factors. This is achieved by constructing the cycle in the financial factor using the cycle in the economic factor together with phase shifts. The resulting framework allows the financial cycle to potentially lead/lag the business cycle in a systematic manner and exploits the information in economic and financial variables for estimation of both economic and financial conditions as well as their cyclical behavior in an efficient way. We examine the potential of the model using a mixed frequency and mixed time span ragged-edge dataset for Turkey. Comparison of our framework with more conventional polar cases imposing a single common cyclical dynamics as well as independent cyclical dynamics for economic and financial conditions reveal that the proposedspecification provides precise estimates of economic and financial conditions and it delivers quite accurate probabilities of recessions that match with stylized facts. We further conduct a recursive real-time exercise of nowcasting and forecasting business cycle turning points. The results show convincing evidence of superior predictive power of our specification by signaling oncoming recessions (expansions) as early as 3.5 (3.4) months ahead of the actual realization.