Chapter 5
The Information Approach to
Decision Usefulness
Copyright © 2009 by Pearson Education Canada
5-1
Chapter 5
The Information Approach to Decision Usefulness
Copyright © 2009 by Pearson Education Canada
5-2
The Information Approach
• Assumes securities market efficiency
• Investors responsible for predicting future firm
performance
– Role of financial reporting to provide useful information for
this purpose
• Usefulness of financial information evaluated by
magnitude of security price response to that
information
Copyright © 2009 by Pearson
5-3
5.2.1 Reasons for Security Price
Response
• An application of decision theory model
–
–
–
–
–
–
Investors have prior probabilities of future firm performance
Investors obtain useful information from financial statements
Investors revise their probabilities
Leads to buy/sell decisions
Security price changes
Return on share changes
Copyright © 2009 by Pearson
5-4
Abnormal Share Return
• Total share return = return due to market-wide
factors ± abnormal share return due to firmspecific factors
– Only abnormal share return can be attributed to financial
accounting information
– If good news in financial statements leads to positive
abnormal share return (and vice versa), conclude financial
statement information is useful.
– To reach such a conclusion, need to separate market-wide
and firm-specific returns
Copyright © 2009 by Pearson
5-5
5.2.3 Separating Market-Wide and
Firm-Specific Factors
• Firm releases financial information
– Most studies look at release of earnings
• Use market model to estimate market-wide return
on that day (or narrow window)
– Assumes market efficiency
• Abnormal share return during narrow window =
total return – market-wide return
• See Figure 5.2 for details
» Continued
Copyright © 2009 by Pearson
5-6
5.2.3 Separating Market-Wide and
Firm-Specific Factors (continued)
Copyright © 2009 by Pearson
5-7
Unexpected Earnings
• Investors have expectations of current earnings
• Investors’ expectations are built into share price
prior to release of current earnings
– Assumes market efficiency
• Investors will react only to unexpected earnings
• Investors’ earnings expectations unobservable
– How to separate expected and unexpected earnings?
Copyright © 2009 by Pearson
5-8
Estimation of Investors’ Earnings
Expectations
• Time series approach
– Based on earnings in prior years
• Analysts’ forecasts
– Available for most large firms
– Now the most common approach
Copyright © 2009 by Pearson
5-9
5.3 The Ball and Brown Study
• The First Study to Document Statistically a Share
Price Response to Reported Net Income (1968)
• Methodology Still in Use Today
Copyright © 2009 by Pearson
5-
B&B Methodology
• For Each Sample Firm:
– Estimate investors’ earnings expectations (proxied by last
year’s actual)
– Classify each firm as GN (actual earnings > expected
earnings) or BN (vice versa)
– Estimate abnormal share return for month of release of
earnings (month 0), using procedure of Figure 5.1
» Continued
Copyright © 2009 by Pearson
5 - 11
B&B Methodology (continued)
• Calculate Average Abnormal Share Return for GN
Firms for Month 0
• Ditto for BN Firms
• Repeat for Months -1, -2,…,-11, and Months +1,
+2,…,+6
• Plot Results
– See Fig. 5.3, next slide
Copyright © 2009 by Pearson
5-
B&B Results
Copyright © 2009 by Pearson
5-
B&B Conclusion
• Stock Market Reacts to Accounting Information,
but Begins to Anticipate the GN or BN in Earnings
12 Months Prior to Month of Earnings
Announcement
–
Consistent with securities market efficiency and
underlying rational decision theory
Copyright © 2009 by Pearson
5-
5.3.2 Causation v. Association
• Narrow Window Studies
– Evidence that financial statement information causes
security price change
• Wide Window Studies
– Evidence that financial statement information is associated
with security price change
• Narrow window studies more consistent with
decision usefulness
Copyright © 2009 by Pearson
5-
5.3.3 Research in Years Following
Ball & Brown
• Does Amount of Abnormal Share Price Change
Correlate With Amount of GN/BN? Yes
• With Quarterly Earnings Reports? Yes
• On Other Stock Markets? Yes
• Response to Balance Sheet Information? Hard to
Find
Copyright © 2009 by Pearson
5-
5.4 A Different Question
• Earnings Response Coefficients (ERC)
– Do characteristics of unexpected earnings affect magnitude
of abnormal share return? Yes
Copyright © 2009 by Pearson
5-
5.4.1 Factors Affecting ERC
• Risk (ß): higher ß → lower ERC
• Capital structure: higher D/E → lower ERC
• Earnings quality: higher quality → higher ERC
– Earnings persistence: higher persistence → higher ERC
» Continued
Copyright © 2009 by Pearson
5-
5.4.1 Factors Affecting ERC (continued)
• Growth opportunities: higher opportunities,
higher ERC
• Similarity of investor expectations: more similar,
higher ERC
• Informativeness of price: more informative, lower
ERC?
Copyright © 2009 by Pearson
5-
More On Earnings Quality
• How to Measure?
– Conceptual: main diag. probs. of info. system
– Earnings persistence: higher persistence → higher quality
• Line-by-line evaluation (Ramakrishnan & Thomas (1991))
– Accruals quality (DeChow & Dichev (2002)): higher accruals
quality → higher earnings quality
Copyright © 2009 by Pearson
5-
5.5 Unusual, Non-Recurring, and
Extraordinary Items
• Hierarchy of income numbers
– Net income before unusual and non-recurring items,
also called core earnings
xx
– Unusual and non-recurring items
xx
– Income from continuing operations,
also called operating income
xx
– Extraordinary items
– Net income
xx
xx
» Continued
Copyright © 2009 by Pearson
5-
5.5 Unusual, Non-Recurring, and
Extraordinary Items (continued)
• Definition of extraordinary item
– Infrequent
– Not typical
– Do not depend primarily on decisions of managers or
owners
• If item is not extraordinary, it is part of operating
income
Copyright © 2009 by Pearson
5-
A Financial Reporting Problem
• Manager motivation to put core earnings “in the
bank” by overstating unusual, non-recurring, and
extraordinary writeoffs
• Overstating writeoffs overstates future core
earnings
– Effect is to overstate earnings persistence, thereby
misleading investors
– See Theory in Practice 11.1 re: Nortel
Copyright © 2009 by Pearson
5-
5.6 Accounting Information as a
Public Good
• A public good is a good such that use by one
person does not destroy it for use by another
person
• Accounting information has public good
characteristics
– Use by one person does not prevent its reuse by others
• Thus firm cannot charge users for accounting
information
» Continued
Copyright © 2009 by Pearson
5-
5.6 Accounting Information as a
Public Good (continued)
• Investors who do not pay for accounting
information will demand more of it than socially
desirable
• Implication is that standard setters cannot be
sure that an accounting policy that has a higher
ERC than another is socially better.
• Complicates standard setting
• Still true, though, that an accounting policy with
higher ERC is more useful to investors
Copyright © 2009 by Pearson
5-