An Empirical Evaluation of the Adjustment of Stock Prices To New Quarterly Earnings Information, Journal of Financial and Quantitative Analysis by Ronald, J. Jordan (1973)

No comments

The study focuses on earnings as a key variable in evaluating the firm's share price. If these theories are correct (namely, that earnings are important in determining the value of a firm), then it is a logical extension of this train of thought to hypothesize that at some point in time this earnings information should be reflected in stock price movements. The tests conducted on this Journal Article work were meant to evaluate the speed of price adjustments to various kinds of information, also one could easily argue that these forms of measurement are too broad for this rather delicate operation. However the objective of the study is to explore the adjustment process of stock prices to new quarterly earnings information.
That is the effect of earnings reports on changes in expectations as reflected in share price movements.  The primary conclusions of this study are: That the market evaluates third quarter and annual earnings differentially from first and second quarter earnings, and That share prices of growth companies adjust differentially from the share prices of non-growth companies, a result anticipated from the simple evaluation model E /(r-g).
This sample consists of 45 firms chosen randomly from Forbes' 21st Annual Report on American industry. The statistic of interest in this report is the five-year annual increase in per-share earnings in percent.1 From this population three stratified samples of 15 firms each according to its five-year annual compounded growth rates were selected. These subsamples comprise firms with increasing earnings (Group I), decreasing earnings (Group II), and stable earnings (Group III). The period covered by this study is April 1, 1963 - December 31, 1968. It’s selected because it was the first complete year for which daily price data were available on the ISL tapes, and the latter date chosen because the American Institute of Certified Public Accountants changed reporting standards regarding earnings per share figures effective for fiscal years beginning after December 31, 1968. However, King has found that a substantial influence on stock prices is brought about by market-wide factors, it becomes necessary to abstract from general market conditions before attempting to examine the adjustment of stock prices to new quarterly earnings information. The model for achieving this goal was first suggested by Markowitz and later further developed by Sharpe, and Lintner. Inasmuch as all "omitted" variables have by construction been impounded in the error term "u," it should be apparent that a quarterly earnings announcement will have an impact on a security's rate of return in the days surrounding that announcement and, consequently, will have an impact on the behavior of that security's regression residuals during the affected period. Thus, a major part of the paper will be concerned with an analysis of this abnormal residual behavior and its various implications. This residual behavior will be examined through both the mean residuals and the absolute mean of residuals surrounding the quarterly earnings announcements. The mean of the residuals captures all the impact on a firm's share price on a given day relative to earnings announcements which cannot be "explained" by the market index relative. Since the absolute mean of the residuals disregards signs and looks only at the numerical value of the residual, it measures uncertainty. Essentially, the absolute mean captures the same information as does the mean with the absence at least in part of one item random noise. After initial testing and analysis of the data, the 50-day period surrounding earnings announcements (25 days preceding the announcements to 24 days following the announcements) was selected as the appropriate time period for the two-way Analysis of Variance Tests in this study.

No comments :

Post a Comment