<![CDATA[This study aims to investigate the dynamic and asymmetric responses of Islamic banking financing risk (Non-Performing Financing/NPF) in Indonesia to macroeconomic variable shocks, specifically Economic Growth (IPI), Inflation, and Exchange Rate. This research is motivated by the urgency of risk mitigation amidst the post-pandemic "pseudo-stability" phenomenon and the inconsistency of prior literature findings dominated by linear approaches. Such conventional approaches are considered less accurate in capturing the behavior of fluctuating macroeconomic variables. The unit of analysis is the national Islamic banking industry using monthly time-series data from January 2015 to September 2025. The data analysis technique employs the Non-Linear Autoregressive Distributed Lag (NARDL) estimation, which decomposes independent variables into positive and negative shock components, complemented by the Wald Test for statistical asymmetry validation and the Error Correction Term (ECT) to measure adjustment speed. The results reveal variable-specific patterns: (1) Economic Growth (IPI) exhibits a dominant negative effect but is symmetric, indicating that the profit-loss sharing mechanism effectively distributes risk proportionally according to the business cycle; (2) Inflation has no significant long-term effect, proving the resilience of Islamic banking customers supported by cash flow certainty in Murabahah contracts; and (3) The Exchange Rate is proven to have a significant asymmetric effect. Rupiah depreciation dynamics have a far more destructive impact on asset quality compared to the improvements during appreciation, confirming the validity of the non-linear balance sheet channel. Systemically, Islamic banking demonstrates robust stability with an estimated self-recovery time of approximately 4.4 months. This study recommends the implementation of asymmetric hedging strategies for banking management and the calibration of dual-scenario stress testing policies for regulators to maintain financial system stability.]]>
