Ebook A Catering Theory of Earnings Management
Extant literature in earnings management proposes different theories for why firms manage earnings. Watts and Zimmerman (1978) argue that managers alter reported accounting numbers to maximize their bonus, avoid tripping debt-covenants written on accounting numbers or to reduce their firm’s political visibility. Graham, Harvey and Rajgopal (2005) survey Chief Financial Officers who indicate they manage earnings to maintain or increase the stock price of their firms. A vast literature (see Fields, Lys and Vincent 2001 for references) has validated the Watts and Zimmerman (1978) propositions. However, managerial incentives to manage earnings to address stock price concerns are relatively under-explored. In this paper, we build on the stock price motivation by proposing a hitherto unexamined incentive for why managers manage earnings.
We argue that, for psychological or institutional reasons, investor demand for stocks that report positive earnings surprises is time-varying, which, in turn, causes the relative aggregate market value of stocks with and without positive earnings surprises to fluctuate. Limits to arbitrage, presumably, fail to prevent this demand from affecting relative prices of stocks with and without positive earnings surprises. To increase, or at least maintain, their current stock price levels, managers cater to the investor demand for positive earnings surprises by increasing abnormal accruals in periods when investors place a premium on positive earnings surprises, and vice-versa.
In our empirical work, we document that the propensity of firms to increase abnormal accruals depends on the relative stock price premiums for positive and negative earnings surprises. To measure such a relative earnings surprise premium, we rely on the time-series variation in (i) the average announcement effect of firms that declare quarterly earnings with an increase (decrease) in such earnings relative to the seasonally lagged earnings announced in the previous year; and (ii) the average price-earnings ratio of all stocks in the market during a quarter. We then show that the aggregate level of abnormal accruals in the stock market is robustly associated with these proxies for the relative earnings surprise premium. This evidence suggests that managers recognize that investor demand for earnings surprises is time-varying and they cater to such demand via abnormal accruals.
Our results related to catering incentives appear robust to several alternative explanations. In particular, time-varying characteristics such as investment opportunity sets (agency cost proxy), future profitability (earnings surprise might signal future prospects), leverage and size (proxy for debt and political visibility type contracting incentives to manage arnings) do not account for our results. This gives us some confidence that the catering incentives we propose are empirically distinct from and incremental to the incentives traditionally proposed in the literature by Watts and Zimmerman (1978). Furthermore, our proxies for investor demand for earnings surprises are not a mere reflection of several macro-economic variables such as future inflation, GDP growth, growth in industrial production, growth in labor income or consumption growth.
We probe further to understand where the investor demand for time-varying premiums on earnings surprises comes from. We believe that our results are best explained by investor sentiment. Investor preference for earnings surprises will have a larger impact on prices of smaller stocks and those susceptible to more subjective valuations. Consequently, the earnings surprise premium (discount) resulting from investor sentiment will have a larger impact on abnormal accruals of these stocks. Consistent with this intuition, we find that managers are more likely to cater to investor demand for earnings surprises via abnormal accruals in stocks most likely to be sensitive to investor sentiment-driven demand (stocks of newer, smaller, highly volatile firms, firms in distress or with extreme growth potential, and firms without dividends). Moreover, we also find a positive correlation between the earnings surprise premium and several proxies for investor sentiment such as the closed-end fund discount, the share of new equity issues and the Michigan consumer sentiment index.
Posted in :
