Scott, Financial Accounting Theory, 7th Edition Instructor’s Solutions Manual Chapter 5
153
CHAPTER 5
THE VALUE RELEVANCE OF ACCOUNTING INFORMATION
5.1 Overview
5.2.1 Reasons for Market Response
5.2.2 Finding the Market Response
5.2.3 Separating Market–Wide and Firm–Specific Factors
5.2.4 Comparing Returns and Income
5.3 The Ball and Brown Study
5.3.1 Methodology and Findings
5.3.2 Causation Versus Association
5.3.3 Outcomes of the BB Study
5.4 Earnings Response Coefficients
5.4.1 Reasons for Differential Market Response
5.4.2 Implications of ERC Research
5.4.3 Measuring Investors’ Earnings Expectations
5.4.4 Summary
5.5 A Caveat About the “Best” Accounting Policy
5.6 The Value Relevance of Other Financial Statement Information
Scott, Financial Accounting Theory, 7th Edition Instructor’s Solutions Manual Chapter 5
154
5.7 Conclusions on the Information Approach
LEARNING OBJECTIVES AND SUGGESTED TEACHING APPROACHES
1. To Appreciate the Value Relevance Approach to the Decision Usefulness of
Financial Reporting
I begin coverage of this chapter by pointing out that we are now starting to apply
decision theory and efficient securities markets theory to better understand the role of
financial reporting to investors.
The first step is to develop the concept of value relevance, as empirical testing of the
predictions of decision theory and market efficiency. I briefly review the Bill Cautious
decision theory example 3.1 and emphasize that it is Bill Cautious, not the accountant,
who has the primary responsibility, and motivation, to predict future firm performance.
The role of the accountant is to supply useful information in this regard, and not
necessarily to make direct predictions about current and future firm value. To the extent
that it facilitates investor predictions of future firm performance, historical cost-based
information can be useful even though it does not directly reveal values.
I often play “devil’s advocate” at this point and suggest to the class that since decision
usefulness implies it is not the role of the accountant/auditor to predict future firm
performance and value, is the accountant/auditor responsible if it turns out that the
financial statements did not foresee financial distress. I try to steer the resulting
discussion to a conclusion that while accountants/auditors may like this argument, it is
not clear that investors, regulators, and the courts will accept it. I also point out that
accountants are in competition with other information sources, pursuant to Beaver’s
1973 paper reviewed in Section 4.3. Repeated complaints by investors that financial
distress was not predicted can only erode the accountant’s competitive position. If time,
with a view to the measurement approach to be introduced in Chapter 6, I ask if
accountants could improve their competitive position by assuming greater responsibility
for reporting on current values.
155
2. To Introduce Empirical Securities Markets–Based Accounting Research
The body of research in this area is vast. I concentrate in this chapter on providing a
framework within which the research can be interpreted, rather than trying to cover very
much of it per se.
To establish this framework, I begin by pointing out that the empirical research
addresses some very fundamental and interesting questions — do investors use the
accountant’s product? If they don’t, of what value is financial reporting?
I then argue that the framework is provided by the decision theory model, again
referring back to Example 3.1 — if investors find financial accounting useful, then we
should see trading volume and securities prices responding in predictable ways to
accounting information.
Having said this, I then point out that actually finding a securities market response is not
easy. I discuss at an intuitive level the various research problems outlined in Section
5.2. I end up this discussion by emphasizing that the basic procedure to find a market
response is to associate some measure of market return on securities with some
measure of the information content of the financial statements.
I then review the 1968 Ball and Brown study. Since their methodology takes some
getting used to for students who have not seen it before, I stick fairly closely to the
coverage in Section 5.3, although the article itself could usefully be assigned as reading
by instructors who wish to consider BB’s procedures in greater depth. I concentrate on
explaining how BB operationalized the measurements of market return and information
content of net income. Figure 5.3 is useful in this regard. The figure also ties nicely
back to the efficient securities market theory of Chapter 4, and the discussion in
Example 3.2 of the dichotomization of factors affecting share price into market–wide and
firm–specific factors.
I go on to review the ERC research outlined in Section 5.4, as an example of an
important direction in which the Ball and Brown methodology developed. I bring out how
the ERC is a measure of earnings quality, and discuss the various measures of
earnings quality. Earnings persistence, in particular, is an important component of
Scott, Financial Accounting Theory, 7th Edition Instructor’s Solutions Manual Chapter 5
156
earnings quality. Persistence ties in nicely with full disclosure, since poor disclosure can
be used to hide low–persistence items. Discussion of one of the chapter problems can
be helpful to get across how poor disclosure can lower earnings quality by hiding low–
persistence earnings—see problems 16, 18, 20. Persistence appears later as an
important component of Ohlson’s clean surplus theory (Section 6.10.2—optional
reading).
An interesting exercise is to attempt to estimate the persistent portion of the earnings of a large
and complex company from the information in its annual report. Indeed, this could be a useful
student assignment, although one that is hard to mark. I tried such an assignment once, but the
student answers tended to be fairly superficial. In part, the problem is that earnings persistence is
hard to determine precisely. Perhaps, as mentioned above, an in-class case discussion of an actual
annual report is a better way to apply the persistence concept.
You may have noticed that most of the research outlined in this chapter is now quite
old. This is not because this research is no longer relevant to a study of accounting
theory, but rather that accounting research has moved on to other topics. I have,
however, outlined two more recent studies. Jones and Smith (2011) can be used to
introduce the concept of special items and some differences between IASB and FASB
practice. McVay’s study of classification shifting provides an interesting example of
financial statement manipulation of earnings persistence.
Instructors have considerable flexibility to augment the coverage of this chapter,
depending on their own interests and backgrounds. I have tried to design the chapter to
facilitate this. As mentioned, I provide a decision–theoretic framework within which the
empirical research can be interpreted. Also, I provide extensive references to articles
upon which the chapter material is based. I have, as well, introduced topics that could
usefully be developed further. These include the notions of narrow and wide windows,
which lead to the distinction between causation and association in Section 5.3.2, and
measuring investors’ earnings expectations in Section 5.4.3.
Section 5.4.3 also contains a brief discussion of analysts’ forecasts as a measure of
these expectations. This textbook does not give much attention to the considerable
research into analyst forecasts, other than to explain such forecasts are a way to
Scott, Financial Accounting Theory, 7th Edition Instructor’s Solutions Manual Chapter 5
157
estimate expected earnings. However, the coverage here provides an occasion for
instructors who wish to dig beeper into issues of forecast accuracy and bias to do so. In
particular, the Easton and Sommers 2007 study, and the additional reference in Note
23 provide may be of interest.
For instructors who wish to cover securities market response to non–earnings
information (hard to find), Section 5.6 considers the study of Lev and Thiagarajan
(1993) which suggests that the relationship between balance sheet information and
share price shows up via the ERC, rather than directly. Their paper is also useful as a
way of bringing out the concept of earnings quality. The paper is quite readable and
could be assigned as supplementary reading if desired.
More recently, Defranco et al (2011) studied the information content of Note
information. This study is more indicative of current capital market research, which
often considers the presence of more than one type of rational informed investor. In this
edition, I have emphasized some of the limitations of the theory in understanding how
information affects capital markets—see the discussion of the assumptions underlying
the CAPM in Section 4.5.2 and the outlines of some models that drop these
assumptions in Section 6.5 (optional sections).
3. To Appreciate the Limitations of Empirical Securities Markets Research for
Accounting Policy Recommendations
Given that empirical research has established an association between accounting
information and the market returns on firms’ shares, it may seem reasonable to suggest
that the best accounting policy is the one that produces the highest association. That is,
if net income calculated using, say, straight–line amortization is more highly associated
with changes in the market value of (i.e., the return on) the firm’s shares than is net
income calculated using declining balance amortization, then investors find straight–line
amortization policy more useful, since it is more consistent with an (efficient) securities
market’s evaluation of the firm. Such reasoning may have the potential to identify the
most useful accounting policies from the great variety of policies available under GAAP.
If so, could standard setters simply require accounting policies that are most highly
associated with changes in the market value of the firm’s shares?
158
I sometimes put this question to the class. While getting a good discussion going tends
to be like pulling teeth, some students will see the essential circularity. Adopting
accounting policies that have the greatest market response does not inform the market.
To clinch this point, ask if the best way to measure a firm’s net income is to drop
reporting of net income and, instead, simply report the change in the market value of
the firm for the period (adjusted for capital transactions). Assuming reasonably efficient
security markets, this would be an economically correct measure of income. However,
the answer is no, since nothing is added to what the market already knows. The role of
accounting information is to expand and improve the stock of information available to
the market, not to reflect it.
Another reply to the suggestion to use the association between accounting information
and share returns as the basis for accounting policy choice is to point out that the
private and public values of accounting information are not the same, leading to the
distinction between private and public goods given in Section 5.5. The distinction
between public and private goods is at the heart of the Gonedes and Dopuch (1974)
paper referenced in Section 5.5. They show that since accounting information has
characteristics of a public good, reliance on market prices to motivate firms’ information
production decisions does not result in the socially “best” amount of private information
production. For example, if a firm enjoys a stronger share price response to the GN in
net income when it uses straight line amortization than when it uses declining balance,
then the firm may prefer straight line amortization (use of straight line amortization is an
information production decision). While this is fine from the firm’s perspective, we
cannot conclude from this that straight line amortization is better from society’s
perspective. The reason, as explained in the text, is that the market does not bear all
the costs of producing the information. Consequently, society cannot use the share
price response as a signal of what information firms should produce, as it could for a
private good. If the share price response does not incorporate all the costs, we cannot
use the correlation between straight line net income and share price as an indicator of
the socially best amortization policy.
Students sometimes ask that if empirical securities markets research cannot lead to
accounting policy recommendations, what is the point of studying it? The response that
Scott, Financial Accounting Theory, 7th Edition Instructor’s Solutions Manual Chapter 5
159
it helps us to better understand how investors use financial accounting information,
while valid, does not seem to settle the issue. I usually fall back on the argument made
in Section 5.5 that, despite differences between the private and social values of
accounting information, the greater the market response to accounting information the
more useful it must be to investors (even though not necessarily to society), hence the
greater the accountant’s competitive advantage. Also, another cost of accounting
information is that managers may not like to report it (recall management’s scepticism
about RRA). Consequently, the socially best accounting policy must take
managements’ concerns into account. This issue is pursued in later chapters.
Scott, Financial Accounting Theory, 7th Edition Instructor’s Solutions Manual Chapter 5
SUGGESTED SOLUTIONS TO QUESTIONS AND PROBLEMS
1. Value relevance studies the reaction of security returns to accounting
information. It assumes average investor rationality and securities market
efficiency, under which investors base investment decisions on accounting
2. One factor that could adversely affect the accuracy of the estimate of abnormal
returns is the estimate of beta, particularly if the firm’s beta changes over time. If the
slope of the regression line in Figure 5.2 is not correct, or if beta has changed
subsequent to the period over which it was estimated, this will affect the abnormal
returns estimate.
161
be ascribed to the GN or BN in earnings rather than the other information.
A more fundamental problem is that investors may not necessarily make investment
decisions the way that the theory developed in Section 3.3 suggests. Investment
decisions may not be fully or even partially diversified, in which case beta is not the
only relevant risk measure. Or, investors may have some other method of making
possibilities are discussed in later chapters). This could dampen their reaction to the
reported earnings.
3. This anticipation up to a year ahead is consistent with the correlation argument. If the
firm in an economic sense is doing well, the efficient market learns of this from more
163
operating activities. However, they do generate realized income and the
abnormal market return will reflect this. For example, if a firm realizes a
$100,000 gain on sale of land, the assets of the firm increase by $100,000, and
the firm’s market value will increase by this amount (it would have increased
earlier if the market had anticipated the value increase). However, market value
changes. For example, suppose that a firm reports an increase in income this
year because it has changed from declining balance to straight line amortization
for its property, plant, and equipment. An efficient market would not respond to
such a change as long as it felt that there were no effects on cash flows.
Note: In both these zero–persistence cases, the market may wonder why the firm
5. It is desirable to identify the moment when the market became aware of such
information because of securities market efficiency. An efficient market will react
Scott, Financial Accounting Theory, 7th Edition Instructor’s Solutions Manual Chapter 5
164
quickly to new information. Consequently, if the researcher looks for a market
6. Yes, a negative ERC is possible. This means that the firm reports positive
unexpected earnings but the abnormal share return to these earnings is
negative, or vice versa.
166
8. a. Yes, a stock price decrease is expected, other things equal, because
unexpected earnings were negative $2 million. This conveys bad news to the
market. Security prices should react negatively to this information.
b. The share price decrease should be greater for scenario (i) because that
9. a. Trading volume may have increased in week 0 because investors revised
their probabilities of future firm performance and share returns upon receipt of
the current earnings information. These revised probabilities led to portfolio
rebalancing and resulting buy/sell decisions. Beaver’s finding that the trading
volume increase took place almost completely in week 0 is consistent with
d. No, low trading volume does not mean that share price change must be
low. The answer depends on the informativeness of the information system, that
is, on how decision useful reported earnings are.
First, assume reported earnings are highly decision useful (i.e., high main
information system diagonals). For small investors, their posterior beliefs about
However, if reported earnings are low in decision usefulness, posterior beliefs
between small and institutional investors will tend to be different, for reasons
given in the body of the question. This implies high trading volume. However,
share price change will be low since different posterior beliefs imply more equal
numbers of buyers and sellers.
Scott, Financial Accounting Theory, 7th Edition Instructor’s Solutions Manual Chapter 5
• Kim, O. and R.E. Verrecchia, “Market Liquidity and Volume Around
Earnings Announcements,” Journal of Accounting and Economics
(January, 1994), pp. 41–67.
• Demski, J. and G.A. Feltham, “Market Response to Financial Reports,”
10. a. X Ltd. would be expected to have a higher ERC. First, it uses conservative
accounting policies. Consequently, a given dollar of GN has higher implications
for future profitability and returns than for Y Ltd.
Second, Y’s reported net income may have lower reliability than X because it
uses current value accounting for its capital assets. Readily available and well–
Scott, Financial Accounting Theory, 7th Edition Instructor’s Solutions Manual Chapter 5
170
11. Theory suggests that rational investors are primarily interested in predicting
future firm performance, and will respond quickly to new, publicly available
information that is useful in updating their predictions. The empirical evidence in
text, Chapter 5 supports the theory, since it appears that security prices, hence
abnormal returns, respond to the GN and BN in current reported earnings much
in predicting amounts of future cash flows.
With respect to the timing of future cash flows, Chapter 5 does not present direct
evidence that financial statements help investors to assess cash flow timing.
However, full disclosure of unusual and infrequent items, so that investors can
determine earnings persistence, helps to assess timing. Research by Kormendi
Copyright © 2015 Pearson Canada Inc.