Accrual persistence and accrual anomaly

Abstract

The first essay, "Inter-temporal accrual persistence and accrual anomaly" investigates whether accrual persistence and accrual anomaly vary with the state of economy. Prior accounting research argues that diminishing marginal returns on new investments drive lower persistence of accruals relative to cash flows. Macroeconomic research documents that marginal profitability is counter-cyclical, which implies that diminishing marginal returns on new investments are more pronounced during periods of expansions than recessions. Linking the cyclicality of diminishing returns on investments with the argument that diminishing returns to investments contribute to lower persistence of accruals relative to cash flows, this paper predicts that the differential persistence of accruals is greater during expansionary periods than recessionary periods. Using a U.S. sample from 1972 to 2003, I find that the differential persistence of accruals is greater during economic expansions than recessions. When I focus on the components of accruals, I find that depreciation, change in accounts receivable, change in raw materials,and change in finished goods are the main drivers of cyclical differential accrual persistence. These findings are robust to alternative conditioning sets, estimation procedures, and measures of the business cycle. I also find that investors are unable to assess the cyclical differential persistence of accruals, leading to higher returns (both raw and abnormal returns) from an accrual-based trading strategy during expansionary periods. The second essay "Can cyclical property of accrual persistence explain the accrual anomaly?" examines whether cyclical accrual persistence documented in the first essay can provide an explanation to accrual anomaly based on consumption based assets pricing theory. Specifically, I posit that accruals decrease in consumption risk because of cyclical property of accrual persistence (i.e., accruals are less persistent during economic expansions than during recessions). The implication is that the observed abnormal returns from accrual-trading strategy represent compensation for the underlying consumption risk. Using a U.S. sample from 1972 to 2003, I find that consumption risk decreases in the level of accruals. I also show that after controlling for other known risk factors, pricing kernel (a proxy for the state of economy) can explain about 11 percent of abnormal returns from accrual-based trading strategy. These findings are robust to alternative conditioning set and estimation procedures.