Further Robustness Analysis Sample Clauses

Further Robustness Analysis. In the preceding sections we show that the data on entrusted lending facilitated by nonbank trustees serve as a powerful instrument to help identify banks’ behavior. It is therefore necessary to provide a careful analysis of the quantitative importance of this instrument. A nonbank trustee is an ideal trustee as it is the kind of trustee as intended by the law (i.e., the PBC’s 2000 guidelines on entrusted lending). Nonbank trustees face no deposit withdrawal risks and bear no credit risks as commercial banks do. Consequently, the banking regulations we have discussed thus far do not apply to these trustees. In this sense nonbank trustees act truly as middlemen. Entrusted loans facilitated by nonbank trustees capture demand from borrowers and supply from lenders. Controlling for these data with the difference-in-difference approach, therefore, enables us to identify banks’ willingness to engage in entrusted lending. A natural question is what the estimated results would indicate if we exclude from the sample the transactions facilitated by nonbank trustees. With this exclusion, the effective sample size is reduced to 650 and the triple-interaction regression represented by (2.4) is reduced to the following double-interaction regression: log Ss = α + αt + αsec + αggt−1 + γrgt−1I (Riskyi) + Controli + εs, (2.6) where an additional control variable Controli is listed in Table 2.9. The estimated results are reported in Table 2.10. As one can see from Table 2.10, all the estimates of the impact of monetary policy changes on entrusted loans channeled by banks, risky or not, have a small magnitude and are statistically insignificant. The absence of the nonbank-trustee instrument creates uninformative results. Without the nonbank-trustee instrument, one may still use the loans channeled by large banks as an instrument to identify small banks’ risk-taking behavior as in the following regression similar to the triple-interaction benchmark regression (2.5) but without the data on entrusted loans facilitated by nonbank trustees: log Ss = α + αt + αsec + αggt−1 + βsgt−1I (Smallb) + γAgt−1I (Riskyi) + γsgt−1I (Smallb) I (Riskyi) + Controlib + εs, (2.7) where additional control variables Controlib are listed in Table 2.9. The estimated results are reported in Table 2.7. One can see from Table 2.7 that the large-bank instrument helps identify the risk-taking behavior of small banks. Indeed, the impact of monetary policy changes on risky entrusted loans funneled by small ba...