臺大管理論叢 NTU Management Review VOL.30 NO.2

183 NTU Management Review Vol. 30 No. 2 Aug. 2020 “net effect” of reporting non-zero ICWs depends on which effect is dominating. When a bank reports that it has zero ICW, the first force is zero, but the second force is negative because investors do not believe that a bank would actually have zero ICW, and would interpret such contention as that the bank making little effort in identifying and reporting ICWs. Therefore, the net effect of reporting zero ICW is highly likely to be negative. In contrast, when a bank reports non-zero ICWs, although investors may perceive the bank is having some deficiencies in operations, and hence value the bank negatively to some degree, investors may also perceive that the bank takes the internal control process seriously, and such effort would benefit the bank in the long run. Therefore, reporting non- zero ICWs may have some positive effects on firm value. The net effect of reporting non- zero ICWs depends on which force is dominating. In sum, it is reasonable to expect that when empirical comparisons are made, banks that report non-zero ICWs are positively valued by investors. The discussions above lead us to hypothesize that disclosing non-zero ICWs is positively valued by the market. 5 3. Research Design This study investigates whether ICW disclosures are valued by the market using the Ohlson model. However, the endogenous variable, disclosures of ICW, may cause bias in ordinary least squares (OLS) parameter estimates (Wooldridge, 2002). Hence, Heckman’s two-stage method is used to mitigate the potential selection bias. In the first stage, the study regresses ICW disclosures on a set of determinants for reporting ICWs suggested by prior studies. Specifically, the following Probit model is used for estimation: Probit( ICW it ) = α 0 + α 1 SIZE it + α 2 ln(AGE) it + α 3 AGGLOSS it + α 4 ROA it + α 5 sd(ROA) it + α 6 DIV_ASSET it + α 7 DIV_REV it + α 8 FOREIGN it + α 9 EXGROWTH it + α 10 BDSIZE it + α 11 BDMGER it + α 12 BDHD it + α 13 MGERHD it + α 14 MODEV it + α 15 BDPLEDGE it + α 16 BLOCKHD it + α 17 LEGALHD it + α 18 GOVHD it + ε it (1) 5 In contrast to Li et al. (2016), who also document some positive valuation implications, we do not impose the restriction that ICWs have a positive valuation effect only when ICWs are remedied. In other words, they view ICW disclosures as negative events that the market reacts positively to when they are remedied. Nevertheless, we propose that ICW disclosures may simultaneously contain both negative and positive information. While the effect of negative information should be included in the assets (or liabilities) related to the specific ICW, no prior study has examined the effect of positive information.