Capital flows at risk: back-testing analysis of the left tail density in the MENA region

Purpose This study improves forecasting of extreme capital flow episodes in the Middle East North Africa (MENA) region by moving beyond mean-based models, which often underestimate downside risks, toward a distributional framework that covers the full range of outcomes and evaluates policy trade-offs. By estimating left-tail risks, this study strengthens crisis prediction in a region nearly twice as responsive to global risk sentiment as other emerging markets. Design/methodology/approach Panel quantile regression estimates the conditional distribution of future gross capital inflows (GCI) – focusing on the left tail – with a skewed-t distribution fitted to capture asymmetry and fat tails. Bayesian Model Averaging (BMA) then evaluates whether estimated tail risks predict realized crises. Findings Global factors dominate the left tail of MENA capital flows, with domestic fundamentals offering limited short-term protection. Monetary autonomy remains structurally constrained, macroprudential policy tightening stabilizes oil importers yet signals vulnerability in oil exporters, and Capital Flows at Risk (CFaR) measures prove robust early-warning predictors of capital flow crises, calling for differentiated policy responses that prioritize macroprudential instruments and precautionary reserve buffers over domestic monetary policy. Originality/value This paper provides the first comprehensive application of CFaR to the MENA region, shifting the analytical focus from backward-looking average dynamics towards forward-looking tail risks. Moreover, it models the interaction between global financial conditions and domestic policy instruments under constrained monetary regimes. It also demonstrates that predicted downside risks correspond closely with realized crisis episodes, highlighting their usefulness as an early-warning indicator for policymakers.