CREDIT RISK ECONOMIC PERFORMANCE AND MONETARY POLICY EFFICACY CONTRASTING PREVALENCE OF AN ARDL MODEL AND A MARKOV SWITCHING REGIME FOR THE CASE OF TUNISIA.

Abstract

The historical swings between periods of excess credit risk and phase shifts of tranquil periods with relatively easier banking credit conditions stimulating investment up to a certain limit raised the debate of the prevalence of credit risk switching regimes of the type of Markov regime switching models for credit risk exerting a lagged effect on economic performance and drained by monetary policy efficacy breakpoint shifts. From another perspective, it is well documented in the literature that most applied finance models stand as ARDLs autoregressive distributed lag models for auto-regression is a feature displayed by most financial aggregates.

This instance of modeling credit risk is proven to exhibit both patterns although compromisingly contrasting apparently, but with beakpoints in unit root of monetary policy will herald obvious and fathoming key features of recent economic events driven by financial shocks in Tunisia.

The main purpose of the research is to scrutinize the impact of banking sector related effects on economic performance depending on credit to the public sector monetary policy efficacy and economic growth, in order to elucidate the relationship between financial shocks and economic performance and to forecast future short run evolution of economic situation starting from an ARDL model exhibiting the main determinants of credit risk then passing to the diagnostic of a Markov model with jump effect applied to credit risk in a time series. The first model shows a positive auto-correlation of credit risk signaling plausible self sustaining exacerbation, a positive correlation with credit to the public sector as a proportion of GDP, a lagged negative correlation with GDP growth and a negative correlation with monetary policy efficacy. Granger causality shows that credit risk granger causes GDP growth with a lag of three years. Empirical data and regression results for the case of Tunisia show prevalence of a Markov switching regime for credit risk validating the jump effect hypothesis corresponding with a lag to collapsing of economic performance and heralding a sharp decline in economic performance caused by a phase shift in monetary policy efficacy.