Accounting measurement and beta risk measures

Prior literature provides evidence that accounting beta risk measures---based on the covariance between firm-specific and marketwide accounting rates of return---better estimate the risk implied by market prices than do stock returns beta risk estimates based on realized equity returns. Because capital markets might distort the realized equity return inputs used by traditional Capital Asset Pricing Model implementations, prior literature suggests that accounting rates of return might better capture the risk associated with firms' future earnings. I hypothesize that accounting policies can also distort accounting betas. I examine whether the price deviations for value estimates based on accounting betas are associated with firms' unconditional accounting conservatism and past investment growth. Specifically, I document that accounting betas based on return on equity are negatively associated with a firm's combined level of conservative accounting and the absolute value of the firm's past investment growth. Further, I find that the conservative accounting and past investment growth distortion in accounting betas outweighs the benefit of using accounting betas relative to stock returns betas. The evidence suggests that accounting betas best estimate risk when firms have less conservative accounting policies and smaller magnitudes of absolute past investment growth.