Risk allocation through securitization – Evidence from non-performing loans

Employing a unique and hand-collected sample of 648 true sale loan securitization transactions issued by 57 stock-listed banks across the EU-12 plus Switzerland over the period from 1997 to 2010, this paper empirically analyzes the relationship between true sale loan securitization and the issuing banks' non-performing loan to total assets ratios (NPLRs). We provide evidence for an NPLR-reducing effect during the boom phase of securitizations in Europe suggesting that banks in our sample may (partly) securitize NPLs as the most risky junior tranche and do not (fully) retain NPLs as a reputation and quality signal towards less informed investors in imperfect capital markets. In contrast, we find the reverse effect during the crises period in Europe indicating that issuing banks provided credit enhancement and demonstrated `skin in the game'. Our baseline result remains robust when controlling for endogeneity concerns and a potential persistence in the time series of the NPL data. Moreover, results from a variety of sensitivity analysis reveal that the NPLR-reducing effect is stronger for opaque securitization transactions, for issuing banks exhibiting higher average levels of NPLRs and for banks operating from non-PIIGS countries. In addition, a reduction of NPLRs through securitization is observed for issued collateralized debt obligations, residential mortgage-backed securities, consumer and other unspecied loans as well as for non-frequently issuing, systemically less important and worse-rated banks. Our analysis offers essential insights into the loan risk allocation process through securitization and provides important implications for the vital debate on reducing NPL exposures and the process of revitalizing and regulating the European securitization market.