Industry classification misfits: identification, consequences and guidance
利用两种行业分类方案的差异，识别出行业分类错配的企业，发现它们有更大的绝对异常应计利润，且错配比例高的行业整体异常应计更大；对行业核心企业，异常应计与未来重述相关，但对错配企业则不然，这源于应计模型估计中的测量误差。最后提出用固定同行组来缓解问题。
Abstract We exploit differences in two industry classification schemes to distinguish between industry classification misfits and industry core firms. We posit that misfits differ from their industry peers, and we document consequences of this heterogeneity. Misfits have larger absolute abnormal accruals, firms in industries with a greater proportion of misfits have larger absolute abnormal accruals, and contemporaneous abnormal accruals are associated with future restatements for industry core firms but not for misfits. We attribute these results to measurement error generated by the inclusion of misfits in the estimation of accrual models. We then provide guidance to alleviate this issue. For both misfits and industry core firms, using fixed peer groups based on the largest firms in a given industry significantly outperforms other peer selection methods in detecting abnormal accruals. In additional analyses, we highlight other economic consequences of industry classification misfits such as higher information processing costs.