The article investigates role of bank-specific factors on non-performing assets (NPAs) in Indian banking system in a panel threshold framework (Hansen, 1999, Journal of Econometrics, 93(2), 345–368), using an unbalanced panel of 82 scheduled commercial banks over the period of 1995–1996 to 2010–2011. We consider capital to risk-weighted assets ratio (CRAR) and credit growth as alternative threshold variables (and regime dependent) along with relevant bank-specific variables treated as regime independent. Findings reveal that CRAR exerts negative and significant impact on NPAs once it reaches a critical threshold. Possible implication is that banks extend less risky loans in a high CRAR regime than in low CRAR regime that helps reduce NPAs. With credit growth as threshold as well as regime dependent, we observe statistically significant non-linear effect of credit growth on NPAs. Beyond threshold, credit growth exerts significant negative effect on NPAs that may imply that banks extend good quality loans. However, we cannot rule out the possibility of evidence of ‘ever-greening hypothesis’ of bad debts in Indian banking, that is, banks just roll over previous bad debts into fresh performing loans. JEL Classification: G21, G28, C13, C33

中文翻译：

---

银行特定不良资产决定因素的阈值效应：在印度银行业中的应用

本文使用面板不平衡面板调查了面板阈值框架（Hansen，1999，Journal of Econometrics，93（2），345–368）中特定于银行的因素对印度银行系统中不良资产（NPA）的作用。在1995-1996年至2010-2011年期间，有82家定期商业银行。我们将资本与风险加权资产之比（CRAR）和信贷增长视为替代阈值变量（并取决于制度），以及与银行相关的相关变量视为制度独立。研究结果表明，CRAR一旦达到临界阈值，就会对NPA产生负面和重大影响。可能的含义是，在高CRAR体制下，银行提供的低风险贷款要比在低CRAR体制下帮助减少不良资产的风险小。以信用增长为门槛并依赖于制度，我们观察到信贷增长对NPA的统计上显着的非线性影响。超过阈值，信贷增长会对不良资产产生重大负面影响，这可能意味着银行提供了优质贷款。但是，我们不能排除印度银行业坏账“永远变绿”的证据的可能性，也就是说，银行只是将以前的坏账转为新的不良贷款。JEL分类：G21，G28，C13，C33