Journal of Accounting Research
Выпуск N за 2000 год
Discussion of EVA versus Earnings: Does It Matter Which Is More Highly Correlated with Stock Returns? [статья]
Опубликовано на портале: 21-06-2006Madhav V. Rajan Journal of Accounting Research. 2000. Vol. 38. P. 247-254.
A distinctive feature of the paper is its use of both theoretical and empirical approaches to understanding EVA. The authors first analyze a principal-agent model using certaine parametric assumptions. They model EVA and accounting earnings as two distinct, noisy perfomance measures of the same uderlying construct, with both measures providing information to the principal about the agent's action choices. They subsequently derive a theoretical expression for the percentage value-added contributed by using EVA as a measure for evaluating managerial perfomance. A key finding is that this expression can be restated as a function of the observed correlations of each of the metrics with stock price. In the second part of the paper, the authors estimate this value for their sample of firms using time-series data, and then correlate the estimates cross-sectionally with the decision to adopt EVA by these firms. In support of the theoretical results, the authors find a positive association, after controlling for other factors such as size, leverage, and growth oppportunities.
Опубликовано на портале: 21-06-2006Gerald T. Garvey, Todd T. Milbourn Journal of Accounting Research. 2000. Vol. 38. P. 209-245.
Dissatisfaction with traditional accounting-based performance measures has spawned a number of alternatives, of which Economic Value Added (EVA) is currently the most prominent. How can we tell which performance measures best capture managerial contributions to value? There is currently a heated debate among practitioners about whether the new performance measures have a higher correlation with stock values and their returns than do traditional accounting earnings. Academic researchers have relied instead on the variance of performance measures to gauge their relative accuracy. To formally address the above debate, we use a relatively standard principal-agent model in which contracts can be based on any two accounting-based performance measures plus the stock price. Rather than model detailed differences between EVA and traditional measures such as earnings, we focus on the problem that while the variability of each measure is observable, its exact information (signal) content is not. The model provides a formal method for researchers to ascertain the relative value of alternative accounting-based measures based on two distinct uses of the stock price. First, as is well known, prices provide a noisy measure of managerial value-added. In our model, stock prices also can reveal the signal content of alternative accounting-based performance measures. We then show how to combine stock prices, earnings, and EVA to produce an optimally weighted compensation scheme. We find that the simple correlation between EVA or earnings and stock returns is a reasonably reliable guide to its value as an incentive contracting tool. That is, a firm could reasonably gauge the merits of adding a measure like EVA by examining its correlation with the firm's stock price. This is not because stock returns are themselves an ideal performance measure, rather it is because correlation places appropriate weights on both the signal and noise components of alternative measures. We then calibrate the theoretical improvement in incentive contracts from optimally using EVA in addition to accounting earnings. Specifically, we empirically estimate the "value-added" of EVA by firm and industry. These estimates are positive and significant in predicting which firms have actually adopted EVA as an internal performance measure.