Chapter 3: Methodology
3.4 Measurement of a Firm’s Performance
3.4.1 Accounting Performances, Tobin’s Q, and Stock Return
Barberis, Shleifer, and Vishny (1998) develop a behaviorally-based explanation for short-term underreaction and long-term overreaction anomalies, in which investors misreact to a string of news such as earnings announcements. However, they do not specify what kind of earnings leads to market misreaction. Accordingly, I choose the three most salient and available accounting performances in the income statement, Tobin’s Q, and stock return as proxies for a firm’s performances.
I employ the three definition of earnings commonly found in empirical studies.13 The first is revenue, which represents actual or expected cash inflows generated by the sales of goods and services from a firm’s major or central operations. Thus, the recurring aspect of revenue allows it to be considered as a good firm’s core business performance indicator. Lakonishok, Shleifer, and Vishny (1994) find that future returns are negatively related to past sales growth, suggesting that investors tend to overreact to this information. In contrast, Daniel and Titman (2006) find no evidence of any link between sales growth and future returns. The quarterly revenue growth is defined as:
12 The computation of global and local representativeness scores is motivated by the argument of Tversky and Kahneman (1974): “People expect that the essential characteristics of the process will be represented, not only globally in the entire sequence, but also locally in each of its parts.”
13 A branch of the earnings management literature examines if investors see through firms managing earnings about accruals (e.g., Sloan, 1996; Xie, 2001), and if managers take advantage of investors’
limited attention with bloated balance sheets (Hirshleifer, Hou, Teoh, and Zhang, 2004). The issue examined in my dissertation focuses on the string of past high- (low-) growth performance that leads to the systematic behavioral bias in the stock market, and firms’ past performance may imply some degree of, but do not require, earnings management. Thus, if managers undertake earnings management to take advantage of investors’ representativeness heuristic, then it becomes a promising issue for future research.
1
Second, operating income is related to a firm’s normal business activities, including recurring items and management discretionary expenditures. For example, Sloan (1996) employs operating income as the definition of earnings, because it excludes non-recurring items such as extraordinary items, discontinued operations, special items, and non-operating income. The quarterly operating income growth is defined as:
Third, net income represents the bottom line of a firm’s recurring and non-recurring items over a time period, and it is the most salient and available earnings component of a company. Dhaliwal, Subramanyam, and Trezevant (1999) do not support that comprehensive income is a better performance indicator than net income,14 and they also find that net income has better predictive power in explaining the market value of a firm. The quarterly net income growth is defined as:
1
14 SFAS 130, “Reporting Comprehensive Income” effective in 1997, requires several valuation adjustments such as foreign currency translation, market values of investments, and minimum pension liability.
15 The computation of growth rates in operating income and net income is deflated by total assets in
Unlike accounting performances that are backward-looking measures,16 Tobin’s Q is a forward-looking measure which offers a different time perspective. By and large, Tobin’s Q is used as an indicator of the expected profitability of future investment (Denis, Denis, and Sarin, 1994), an indicator of investors’ acumen, optimism, or pessimism (Demsetz and Villalonga, 2001), and also an indicator of investors’ change toward future expectation (McGahan, 1999). Lang and Stulz (1994) argue that Tobin’s Q is a measure of the contribution of the firm’s intangible assets to its market value. Since a firm’s intangible assets include its organizational capital, reputational capital, monopolistic rents, investment opportunities, and so on, any management action directly affects the value of the intangible assets.
The definition for Tobin’s Q is the market value of the firm divided by the estimated replacement cost of the firm’s tangible assets. Although recent studies do not attempt to measure the replacement cost in the denominator of Tobin’s Q, they do use the depreciated book value of tangible capital instead of the replacement cost.
Hence, the measure of Tobin’s Q used in this study is similar to that found in Demsetz and Villalonga (2001) as:
t t t
t
MV PS D
TA
+ +
, (3.13)
where
MV
t is the market value,PS
t is the book value of preferred stock,D
t is the book value of debt, andTA
t is the book value of total assets. All variables are measured in quarter t.
quarter t-1, and by doing so it avoids any possible negative number in the denominator.
16 Researchers generally agree that accounting profit reflects a firm’s historical advantages, indicating an estimate of what management has accomplished (e.g., Demsetz and Villalonga, 2001; Kao, Chiou, and Chen, 2004). Theoretically, high accounting profitability arises, because the firm’s management has obtained assets at some previous date at less than their full value in the planned use. The fact that assets were acquired and booked at low cost reflects the competence of management at some point in the past (McGahan, 1999).
Finally, prior research has shown different horizons of momentum in returns (DeBondt and Thaler, 1985, 1987; Jegadeesh and Titman, 1993, 2001), which also offer a different aspect of momentum in comparison with financial accounting performances. Thus, the quarterly stock return growth is defined as:
1 1
t t
t
P P
P
−
−
− , (3.14)
where
P
t is the closing price in quarter t.3.4.2 Additional Proxies of Performance
Barberis, Shleifer, and Vishny (1998) construct a model of investor sentiment aimed at reconciling the empirical findings of overreaction and underreaction. Their model hypothesizes that investors do not realize that earnings follow a random walk, in which they shift back and forth between two different regimes. Specifically, in the trending regime, investors overweight the strength of past performance trends and hence overreact, exhibiting the representativeness bias, whereas in the mean-reverting regime investors are slow to change their beliefs in the face of a recent performance announcement and hence underreact, showing the conservatism bias.
Behavioral theory does not precisely indicate what kind of information is strong and salient but low in weight, leading investors to overreact, and what kind of information is weak but high in weight, leading investors to underreact. The theory also does not tell us how sensible investors are in the face of a positive (negative) streak or reversals in a firm’s performance. Thus, I examine three different measures of performance. One natural measure is the return-on-equity (ROE) － a purely accounting rate of return that is of great interest to the shareholders (or potential shareholders) of a firm and defined as earnings divided by book common equity.
Since ROE, like other accounting performances, only reflects a firm’s past advantage, I also use book-to-market (B/M) and cash flow-to-price (C/P) as additional proxies of performance.