Empirical evaluation of accounting income

With the rise of multinational companies and the growth of transnational investments, there has been an increasing demand for an empirical evaluation of accounting income numbers. The Ball and Brown study in 1968 broadens this area of accounting research. With the rapid development of the literature, there are more than 1000 published papers in academic accounting and finance journals in the last thirty years (Kothari, 2001). However, several accounting experts assert that there are some weaknesses in the research of Ball and Brown. Meanwhile, critics argue that their empirical evaluation contributed to the accounting research since late 1970s. The purpose of this essay is to introduce the strengths of Ball and Brown’s paper and identify its influence on the development of accounting. This essay will argue that the limitation of the paper is not serious, but its advantages are very significant. First, arguments for and against the empirical research are discussed. Then, an evaluation is explored .Lastly, a conclusion is presented.

2. The weaknesses of the paper

The major motivation of the Ball and Brown study is to use empirical evidence to prove whether the accounting income numbers convey information of firms’ financial performance. They pioneered capital markets research in accounting. However, their study still has some weaknesses.

2.1 Unreliability

The reliability of their results depends on the sources of information, which may be a potential limitation of the study. Although we are not disputing the reliability of the said sources, they should be investigated. Research should be conducted to document various types of institutions. According to Zhang (2007), the selection of the sample did not include companies meet the following four conditions. The first one is the failed company. The second one is a company whose financial year does not end on the 31st of December. The third one is the company that is not recorded in the CRSP database of stock price research centers. It also includes the young firm that is not described by the Wall Street Journal. Given those situations, the generality and reliability of their results may be reduced.

2.2 Limitations

The research firmly supposes that some information of stock prices can be reflected by the earnings. However, the Ball and Brown study did not use a statistical test to ascertain whether there is any difference between earning information and the information reflect in the stock prices.

Rayburn (1986) asserts that stock returns have been statistically compared with cash flows, earnings, and accruals during the 1980s’. Aside from providing a formal test, sketchy information of cash flow was used in previous research. The succeeding studies used more comprehensiveness expectation models to separate the unanticipated components of earnings, accruals, and cash flows, given that returns in an efficient market only reflect the unexpected components (Livnat et al., 1990).ListenRead phonetically

3. The strengths of the paper

Although the weakness of this paper is obvious, the design of variables and the results of the test have made outstanding contributions to research of accounting income.

3.1 Ingenious design of variables

After identifying the research themes, Ball and Brown selected the most representative accounting income numbers to measure the information content. Considering that shareholders are more concerned about EPS and net income, Ball and Brown used them as variables. To determine the information content of the accounting income, they used market efficiency hypothesis and the capital asset pricing model as references.

According to the efficient market hypothesis, the authors maintained that observable stock price is fluctuant and linked with information. This can mean that accounting income numbers are useful. The key to determining the relationship between accounting income and stock price is to distinguish the security prices of a particular company as well as all the other companies. Therefore, the authors built two models of market expected return to examine how the market reacts to accounting income numbers.

Ball and Brown separated factors that affect earning into two different kinds: specific factors and system factors. System factors affect all companies, thus, the surplus of one company and other companies can be linked with each other. If the link is suitable, then stability can be shown using a fixed function, enabling us to achieve the income conditional expectation of one company based on other companies’ data. Thus, changes in the unexpected surplus of earning can be estimated by calculating by the changes in both actual income and conditional expectation. The authors defined the difference as the gain of current information. At the same time, they assumed that the changes in policy and corporate have been enacted before the first estimation. Thus, the influence of macroeconomic and policy changes can be estimated jointly.

The authors initially used the Ordinary Least Squares (OLS) to estimate the linear regression coefficient and intercept of samples’ annual income changes and average income changes. Secondly, the changes in the market income average were used as independent variables and included in the regression model to calculate the expected value of income changes. Finally, the changes in the value of unexpected earnings (prediction residual) referred to the actual value of the earnings change minus the change in earnings expectations. Therefore, market effect was removed from the estimation of expected surplus, which means the authors did not take system factors into account.

Similarly, the factors affecting the impact of no stock or stock returns can still be separated into two factors: system factors and specific factors. First, the authors used capital asset pricing model to separate the system factors and from non-system factors, then calculated the deviation of actual rate return and expected stock return. Secondly, Ball and Brown calculated the company’s stock return residuals (abnormal returns). Given that markets are efficient, stock prices will adjust quickly and effectively about the new information, so residuals can show the impact on stock prices of new information. To test the validity of statistics, Ball and Brown used an alternative model- naive model. In this model, an alternative variable is the expected surplus for last year’s actual surplus. The naive model does not eliminate the market effects, but verifies the earnings per share targets.

3.2 Remarkable results

This paper distinguished expected changes and unexpected changes in accounting incomes to estimate the abnormal return and changes in unexpected accounting incomes. In addition, it theorized that when the surplus prediction error is negative, it is both advantageous and disadvantageous. This paper presented the hypothesis that if the accounting income numbers could be linked with stock prices, then the announcement of accounting income numbers could result in changes in the stock returns. In the empirical test, the authors defined the month of annual report as 0, API representing the abnormal performance index of month M. In the process of calculating the API, the changes in unexpected earning are first separated into two groups (positive and negative), and then all samples are calculated together. Ball and Brown thought that if accounting income is related to stock returns, it can be assumed that if the changes in unexpected surplus are positive, then API is larger than 1; if they are negative, then the API is less than 1. In the combined sample, API is close to 1.

Ball and Brown thought that annual income report can provide new information; however, it cannot be transmitted in time, because most of its content comes from various sources (interim report and non-accounting information, annual accounting report is only one of them) .The authors also found that after the announcement of the annual report, API has a tendency to drift on, that is, revenue projections’ residual error signal and the relationship between stock returns in the annual report may be continued for two months. After analysis, which may be caused by the transaction costs, and excluding the impact of transaction costs, the market reaction to the numbers tends to be unbiased.

4. Evaluation

At first glance, one weakness of this empirical research is the limitation of the sample which may decrease the generality and reliability of the results. However, the selection of the sample was similar to the tests used in related literature, which could lead to consistent results if used the same way on other samples.

The weaknesses of the paper also include the limitation of the statistical tests. Nonetheless, it had a significant impact on later research. In 1968, the Ball and Brown study shows that accounting earning announcements contain information content. In the meantime, they link the sign of monthly abnormal stock return with the sign of the previous year’s earnings change. Starting with Ball and Brown, many studies began to compare alternative accounting performance measures, such as historical cost earnings, current cost earnings, residual earnings, and operating cash flows with stock returns. Under the similar circumstances, the Ball and Brown study also facilitated Watts and Zimmerman’s positive accounting theory in the late 1970s (Watts and Zimmerman, 1979). As Watts and Zimmerman (1990) point out, most accounting research since Ball and Brown has been positive, and the role of accounting theory is no longer normative.

According to Kothari (2001), Ball and Brown heralded the positive-economics-based empirical capital markets research in the late 1960s. The theoretical and methodological innovation of the early accounting research was founded by the concurrent developments in economics and finance. Without this historical exploring of Ball and Brown in early capital market research, even seasoned researchers may have to detour or lose continuity.

In addition, this paper analyzes the insufficiency of theoretical studies using empirical testing to find out whether the accounting income numbers are useful. It initially provides reliable evidence that stock markets can influence annual reports. Then researchers began to do a lot in reflect of stock market. Furthermore, the method used is also applicable to a large number of accounting and financial issues, including dividend announcements, earnings announcements, mergers and acquisitions, and investment spending.

5. Conclusion

This essay has discussed the weaknesses and strengths of the empirical evaluation of accounting income numbers by Ball and Brown. Although this research has some limitations, the merits far outweigh the disadvantages. It preceded the positive-economics-based empirical capital markets research of the late 1960s. The theoretical and methodological innovation of the early accounting research was founded by the concurrent developments in economics and finance. Therefore, it is suggested that this research plays a great role in the development of accounting .In this essay, only a few strengths are mentioned; the ways of putting these benefits into practice need further exploration.