Heterogeneity in the relationship between NPLs and real economy: Evidence from the Mongolian banking system

The paper analyses the relationship between the banks' credit risk and macroeconomic conditions by addressing the following questions; (i) How are macroeconomic shocks transmitted to lending risk depending on the ban-specific features? (ii) Are the effects of macroeconomic shocks different across the loan portfolios in various economic sectors? Unlike the common assumption in the literature, the empirical analysis considers banks' heterogeneity and diversification across borrowers. It employs heterogeneous panel SVARs and standard SVAR models on a dataset from 2002. Q1 to 2019.Q1. The results suggest that the deterioration in credit quality is affected by both macroeconomic and bank-specific factors, with substantial heterogeneity in the magnitudes and timing in terms of the type of loans in various business sectors and bank characteristics. In particular, we find strong evidence of cyclical sensitivity of loan quality, and about 1/4 of banks' NPLs increases stronger in response to the shocks to growth, exchange rate, interest rate, and profitability. The highly profitable banks tend to less engage in excessive risk-taking, resulting in lower NPLs, whereas the relation of asset size to NPLs is not significant for the sample. A growth shock plays a prominent role in explaining the variation of NPLs for the trade and mining sectors. Similarly, the loan supply shock is the main determinant for the construction sector's NPLs, while the exchange rate shock is the most responsible for the manufacturing sector. The interest rate shock and exchange rate shock are the most effective factors on NPLs of consumer loans. Finally, the feedback effect of NPLs shows that deterioration of credit quality slows down economic growth.