A BNORMAL RETURNS BASED ON SIC PORTFOLIOS

From the SDC financial restructuring database, we got 2145 bankruptcy events during January 1990 to December 2006, and formed 2145 SIC portfolios form each bankruptcy event during the sample period. As Lang and Stulz (1992), each bankruptcy event will form one equal-weighted and value-weighted SIC portfolio to represent the reaction of the surviving firms of the SIC industry. The bankruptcy events recorded in SDC database are of similar number as reported by Hertzel, Li, Officer, and Rodgers (2008) before 1998; however, after 1998, the events recorded is much more than Hertzel, Li, Officer, and Rodgers (2008). For the SDC database recorded each Chapter 11 bankruptcy announced by the court, and thus includes a lot of relatively small firms.

We use the SIC code provided by SDC to finds out companies representatives of each SIC industry from both of COMPUSTAT and CRSP database. Based on the return availability on CRSP, I formed value-weighted and equal-weighted portfolios of each SIC portfolio for each bankruptcy event. The average abnormal return of industry portfolios is the equally-weighted average return of all of the value-weighted and equal-weighted portfolios of certain period and of certain year. Table III and IV document the equal-weighted and Table V and VI document value-weighted abnormal and cumulative abnormal returns based on each SIC industry during January 1990 to December 2006.

【【【

【Insert Table III, IV, V, and VI here】】】】

During the sample period of January 1990 to December 2006, the equal-weighted portfolio shows significant negative cumulative abnormal return of -0.161% on the three-day window (1day before and after bankruptcy announcement) while value weighted portfolio

shows significant positive 0.3467%, 0.1017% and 0.1321% cumulative abnormal returns separately on 11-day (5 days before and after bankruptcy announcement), 3-day and 2-day window (one day before bankruptcy announcement to the event date). The contrary results of equal-weighted and value-weighted can also be seen on the abnormal returns and cumulative abnormal returns on the yearly bases. Except years of 1990, 1994, 1999 and 2002-2006, all of other years in the value-weighted portfolio during the sample period exhibit opposite results of abnormal returns while equal-weighted get the opposite trend. We think the reason might be the much larger bankruptcy events we got compared to other related literatures. Most of the literatures focus on large bankruptcy events, such as liability more than $120 million dollars, and the bankruptcy events recorded by SDC database are of greater amounts compared to the latest paper of Hertzel, Li, Officer, and Rodgers (2008). Thus, the opposite results may be affected by the relatively small bankruptcy in industry, for small bankruptcy affects smaller firms more than larger firms in the same industry. Equal-weighted portfolio is more capable of showing the effects of all firms while value-weighted portfolio is affected more by the large companies. Thus, I focus on equal-weighted portfolio to try more abnormal returns for different period. Both on the value-weighted and equal-weighted portfolios demonstrate positive abnormal returns and cumulative abnormal returns from 2002 to 2006, though not all of them are significant, there seems to exist a timely trend in the bankruptcy announcement effect. After separating equal-weighted portfolios into different periods, the 1990-2000 portfolio shows significant negative -0.561%, -0.447%, -0.479%, -0.258%

individually on 11-day, 5-day, 3-day and 2day window. But after adding years of 2001 to 2003, the significance of negative return decrease. The 11-day window cumulative abnormal returns are no longer significant, while the significance of 5-day, 3-day, and 2-day window decrease as well. As a result, I calculate the cumulative abnormal from 2001 to 2003 and find 0.402%

significant positive cumulative abnormal return on 11-day window, but insignificant results in other day windows. However, the positive significance appears after adding the year from

2004 to 2006. The 11-day and 5day window of 2003-2006 portfolio display significant positive cumulative abnormal of 0.325% and 0.260%, and 2004-2006 portfolio exhibit even more significant positive cumulative abnormal returns, 0.559%, 0.378% and 0.247%

separately, on 11-day, 5-day, and 3-day window. The empirical results shows that, contagion effect dominates the bankruptcy announcement effect during 1990 to 2000, while neither contagion nor competitive effect dominate from 2001 to 2003, and competitive effect dominate the bankruptcy announcement after 2004. That is to say, bankruptcy announcement reveals more bad news than good news to other firms in the same industry from 1990 to 2000, for the average cumulative abnormal returns display significantly negative during this period.

But after 2004, the situation changed, bankruptcy announcement becomes positive events to competing firms of the same industry, and the average cumulative abnormal returns inversely display significant positive in this period. The period of 2001 to 2003 is the transition period when cumulative abnormal returns are insignificant and neither of the effects dominates.

According to literature related to bankruptcy announcement effect, almost all of the empirical results suggest that contagion effect dominates the bankruptcy announcement effect, and only of certain circumstances that competitive effect appears, thus the empirical results from 1990 to 2000 is consistent to the former studies.

But why there is a twist in bankruptcy announcement effect of the same industry during 2001 to 2006? One of the possible reasons is that during 2000 to 2003 U.S. is in the age of turbulence. From the internet bubble in 2000 to Enron, WorldCom, and 911 attacks in 2001 to 2002, the economy in U.S. is in recession. According to Table I, the number of bankruptcy announcements increase sharply from 2000, reach the highest peak of 390 in 2001, and go back to normal level after 2004. During the age of uncertainty like 2001 to 2003, a lot of unhealthy firms announce for bankruptcy and the market enters into reorganization. The elimination mechanism in the market enhances the efficiency in the industry, and consequently results in the afterward competitive effect dominated period.