Measuring Financial Systemic Stress for Turkey: A Search for the Best Composite Indicator

Measuring Financial Systemic Stress for Turkey: A Search for the Best Composite Indicator

18/16

Meltem Gülenay Chadwick, Hüseyin Öztürk

English

October 2018

In this study, we aim to construct a single financial stress indicator (FSI) for Turkey adopting weekly data between April 2005 and December 2016. To do so, we compose 15 different FSIs using 14 variables that will represent five different markets, i.e. money market, bond market, foreign exchange market, equity market and banking sector. We aggregate these five different markets using variety of techniques, including principal component analysis (PCA), basic portfolio theory, variance equal weights and Bayesian dynamic factor model. We compare 15 different FSIs on the basis of their relation to and forecasting power of different variables such as the growth rate of industrial production, OECD business condition index and OECD composite leading indicator for Turkey. Our results suggest that there does not exist a simple best indicator for Turkey that will measure the financial systemic stress. Some indicators offer a good forecasting power for economic growth while others have a stronger correlation with the systemic risk. Therefore, we offer a final FSI for Turkey conducting a model averaging method via a rolling-correlation based weighting scheme to benefit from the information content of all the FSIs and observe that the final FSI successfully indicate the tension periods.

Financial stress indicators, Composite indicator of systemic stress, Principal component analysis, Bayesian dynamic factor model, Portfolio theory, Aggregation methods

C32; C43; C52