The Persistence of Special Items

Abstract

Prior research shows that negative special items are less persistent than core earnings. However, the allocation of special items to periodic earnings entails significant measurement error which may induce the observed lower persistence. In this paper, we control for this measurement error by aggregating core earnings and special items over periods of one to five years, and testing whether the differential persistence is robust to longer aggregation periods.

We find lower persistence of special items only for firms with low core (pre-special items) profitability. For these firms, negative special items have no association with future bottom line earnings for all aggregation periods. In contrast, for high core profitability firms, negative special items exhibit significant persistence, and the persistence increases with the aggregation period. In particular, when earnings are aggregated over five years, we find no difference in the persistence of negative special items and core earnings.

Keywords: special items, measurement error, persistence

JEL Classification: M41, M44, M45