금융산업 구조조정과 금융기관 위기대응력 연구

This paper analyzes which is the more appropriate for Korean financialinstitutions between the learning hypothesis and the risk culture hypothesis. Kimand Yang(2013) supported the learning hypothesis and concluded that Koreanfinancial institutions raised the confrontational ability to crisis after crisisexperience. But this study has the limitation which is the ignorance of effectivecontrolling for survivorship bias. Thus there is possibility of contamination ordistortion from it. Analyses in this paper are directly related to those presented inKim and Yang(2013) and indirectly to those in Fahlenbrach et al.(2012). First, this paper shows the results after controlling for the survivorship bias. Isort the financial institutions who has the possibility to cause survivorship bias bylooking into the merger and acquisition events of each financial institution. Theregression results without these financial institutions is supportive of the learninghypothesis. The regression coefficient on the preceding crisis (1997 crisis) returnsto the posterior crisis (2008 crisis) returns is statistically significantly negative. The absolute value of regression coefficients are increased as the strength ofcontrol for the survivorship bias is increased. Also, the coefficient on determinationis increased with the stronger control for the survivorship bias. Second, I find additional evidence to the learning hypothesis. The negativeassociation between the two crisis returns is more strongly found for financialinstitutions who relatively performed bad during the preceding crisis. More detailed search for learning contents and the differences by type ofinstitutions is left as the future assignments.