Download files
Abstract
This thesis investigates the pricing (or mispricing) of earnings quality in Australia. It investigates whether information in earnings quality is used by investors in valuing firms, evidenced by an association between earnings quality and the cost of equity. In the alternate form, the question may be posed as whether earnings quality is mispriced by investors such that there may be opportunities to earn abnormal profits from trading strategies based on earnings quality. Ten earnings quality constructs are studied: total accruals, unexpected accruals, cash-to-profit, accrual quality, persistence, predictability, smoothness, relevance, conservatism and timeliness.
In the cost of equity pricing tests, when earnings quality is proxied using one construct (accrual quality), it is found to be associated with the cost of equity. However, when the additional nine constructs are included in the regression models, accrual quality loses statistical significance. Various other constructs are found to be associated with the cost of equity depending on the choice of the cost of equity proxy.
In the trading strategy tests, there is some initial evidence of trading strategy opportunities for firms with high quality earnings. However, after deleting outlier observations with annual buy-and-hold returns of greater than 200% the potential for earning abnormal returns from a hedge portfolio strategy disappears.
The existence of Australian evidence on the accruals anomaly provides a convenient basis to validate the results of the earnings quality trading strategy tests. Although no clear evidence on the accruals anomaly is found, results are obtained which appear to be consistent with prior Australian evidence of the accruals anomaly, depending on the research design choices made.
Overall, the evidence on whether earnings quality is priced or mispriced in Australia is best viewed as inconclusive. It highlights the importance of conducting thorough robustness tests and suggests a need for caution by researchers in making inferences from a narrow set of earnings quality constructs and research design specifications.