Ball and Brown (1968): The seed that made a difference
Australian equities research note
January 2014
1
In 1968 Raymond Ball and Phillip Brown published ‘An empirical evaluation of accounting income numbers’ in the Journal of
Accounting research. After an initial lukewarm response from the academic community it rapidly became what the American
Accounting Association now calls ‘The seed that made a difference’.
The 1968 research was the genesis of a now deep body of research on how investors utilise accounting information and the
influence of accounting information on stock price returns. Our Australian equity investment philosophy is grounded by this research
and much of the empirical and academic research that followed.
The Ball and Brown research transformed the study of accounting by showing what for us today seems obvious: there is a
relationship between stock price fluctuations and the information contained in accounting reports. At the time of this groundbreaking
paper while researchers had hypothesised on whether there was a connection between what a company said its earnings per share
would be, any changes to this forecast and its share price 12 months later, no one had empirically tested this.
It is difficult to comprehend how groundbreaking this research was at the time as it is now widely accepted that accounting
information forms the bedrock of many investment processes.
However, in the mid 1960’s the dominant accounting literature was a priori in nature with little in the way of empirical testing.
The accounting literature at the time had two central conclusions:
1. Financial statement information prepared under existing reporting rules is meaningless to investors; and
2. Radical changes in the nature of financial statement information are required.
By 1968 the above conclusions had been around for decades. There was general agreement at the time on the first conclusion that
financial statements prepared under the given accounting rules were meaningless because they involved the aggregation of data
calculated under heterogeneous rules. Consequently, the research and debate predominately revolved around the second
conclusion that the rules governing financial statements needed to be changed.
Strange as it may seem to investors today, the accounting literature then had not deemed it fruitful to investigate how financial
statement information affected or was related to share prices.
2
Ball and Brown proposed a number of reasons for this in their 2012 retrospective . One likely reason was data availability; another
was that the share market was not seen as a suitable area of accounting research. A third reason was that the accounting literature
was focused on its own accounting measurement models.
3
Even amongst this widespread mentality, a significant study had still emerged on the behaviour of share prices. Fama (1965a ,
4
1965b ) made the conceptual breakthrough of framing stock prices as a function of information flows. These days that is taken for
granted, however, in the mid-1960s this concept was an important advance in financial thinking.
In their research Ball and Brown sought to answer the simple fundamental research question; are accounting numbers useful? Their
position was summarised in Ball and Brown (1968) p 160:
“An empirical evaluation of accounting income numbers requires agreement as to what real-world outcome constitutes an
appropriate test of usefulness. Because net income is a number of particular interest to investors, the outcome we use as a
predictive criterion is the investment decision as it is reflected in security prices”
They considered both the relevance and the timeliness of income announcements to equity markets because, as Ball and Brown
saw it, “….usefulness could be impaired by deficiencies in either”.
Ball and Brown took an information perspective in their research by measuring the difference between expected change in income
and actual change in income and how this difference affected stock prices. What they were measuring was the “unexpected
component” or “forecast error” in earnings announcements (1968, 162) using a simple random walk model and a model that
controlled for market effects. This is now termed as ‘earnings surprise’.
1
Ball, R., and Brown, P., “An empirical evaluation of accounting income numbers”, Journal of Accounting Research 6 (2), 1968, pp.159-178
Ball, R., and Brown, P., “Ball and Brown (1968):A Retrospective”, American Accounting Association, 2012
3
Fama, E.F., “The behaviour of stock-market prices”, The Journal of Business 38 (1): 1965a, pp. 34-105
4
Fama, E.F.,. “Random walks in stock market prices”, Financial Analysts Journal 21 (5): 1965b, pp. 55-59
2
1
Colonial First State Global Asset Management
Because nobody had focused on earnings announcements in any detailed way before and analysts’ earnings forecasts were not
being systematically collated in any form accessible to the researchers, Ball and Brown had to determine how to measure the
newsworthy component of any earnings announcement.
To measure the unexpected income information, the difference between expected change in earnings and actual change in earnings
had to be measured. An event study was created with month zero signifying the month in which the preliminary annual results were
released. This was the time at which actual change in income was assumed to be realised by the market.
Ball and Brown ran a regression of the change in a firm’s actual earnings on the change reflected in the analysts forecast, which
gave the model for the expected change in income. Any residuals implied by the model were then assumed to be the new
(unexpected) earnings information conveyed in the income announcement.
A second regression of stock returns on market returns was performed over the same time period, and similarly, the residuals were
the abnormal stock returns. The earnings forecast errors were then matched to the abnormal stock returns over an 18 month
period, beginning a year before the annual earnings announcement date.
Ball and Brown found that when stocks had a positive income surprise, the abnormal stock price returns for the event window were
also likely to be positive. In the other case when stocks had negative income surprises, the abnormal stock price returns were likely
to be negative.
They also found that a majority of the increase in the abnormal returns was before the announcement date, which implied that
analysts have fairly accurate forecasts of whether firms will outperform or underperform. Ball and Brown postulated that this could
be through ‘leakage’ of earnings related information such as earlier announcements of dividends, earnings in interim periods,
contracts won, production figures etc.
Of the two central conclusions reached in the accounting literature of the day, the conclusion that radical changes in the nature of
financial statement information are required could not be directly tested by Ball and Brown because they could not observe the
counter-factual. They concluded that “all standard setting in accounting is to some degree an act of faith, and that applied to the
5
copious advice that the accounting literature was giving to the standard setters of the time”.
However what they could empirically and statistically test was the first proposition that financial statement information prepared
under existing reporting rules is meaningless. Their research demonstrated that this proposition failed the test.
Under the assumption that changes in share prices incorporated real changes in underlying firm value to at least some degree, Ball
and Brown were able to empirically prove that accounting information is in fact correlated with these changes. These findings thus
made it difficult to maintain that accounting information was meaningless.
While the test statistics were weak by today’s standards the evidence was sufficient to cause many of the next generation of
6
accounting scholars to abandon the prevalent way of thinking of that time. Subsequently, a new research paradigm emerged.
Ball and Brown (1969) expressed a view of information in markets that was revolutionary and contributed to a significant change in
attitudes towards investing and financial markets. By testing the connection between earnings expectations and share price
changes they were the genesis of a body of research that now underpins modern day investment processes.
Finally, it was interesting to note that at the time the paper was written it was submitted to and rejected by the US Accounting
Review.
For further information:
Chris Robertson
Senior Investment Specialist
Telephone: +61 2 9303 0002
5
6
Ball, R., and Brown, P., “Ball and Brown (1968):A Retrospective”, American Accounting Association, 2012
Kuhn, T.S., The structure of scientific revolutions, Illinois: University of Chicago Press, 1969
2
Colonial First State Global Asset Management
For further information please contact:
Business Development - Melbourne
Head of Institutional Client Relationships
Harry Moore
+61 3 8618 5532
Peter Heine
+61 3 8628 5681
Peter Weldon
+61 2 9303 6860
Institutional Relationship Management
Business Development - Sydney
Dan Bristow
+61 2 9303 6311
Liz White
+61 2 9303 2927
Rose Lor-Kershaw
+61 2 9303 2863
Hugh O’Neill
+61 2 9303 6116
Helen Squadrito
+61 2 9303 6142
Bachar Beaini
+61 2 9303 3929
Hazuki Nojiri-Kenny
+61 2 9303 2415
Edward Tighe
+61 2 9303 2416
Business Development - Auckland
Matthew Laing
+64 9 448 8440
Disclaimer
Product Disclosure Statements (PDS) and Information Memoranda (IM) for the funds issued by Colonial First State Investments Limited ABN 98 002 348 352, Commonwealth
Managed Investments Limited ABN 33 084 098 180, and CFS Managed Property Limited ABN 13 006 464 428 (collectively CFS) are available from Colonial First State Global Asset
Management. Investors should consider the relevant PDS or IM before making an investment decision. Past performance should not be taken as an indication of future performance.
No part of this material may be reproduced or transmitted in any form or by any means without the prior written consent of CFS. This material contains or is based upon information
that we believe to be accurate and reliable. While every effort has been made to ensure its accuracy we cannot offer any warranty that it contains no factual errors. We would like to be
told of any such errors in order to correct them. This material has been prepared for the general information of clients and professional associates of CFS. You should not rely on the
contents. To the fullest extent allowed by law, CFS excludes all liability (whether arising in contract, from negligence or otherwise) in respect of all and each part of the material,
including without limitation, any errors or omissions. This material is intended only to provide a summary of the subject matter covered. It does not purport to be comprehensive or to
render specific advice. It is not an offer document, and does not constitute a recommendation of any securities offered by CFS. No person should act on the basis of any matter
contained in this material without obtaining specific professional advice. Colonial First State Global Asset Management is the consolidated asset management division of
Commonwealth Bank of Australia ABN 48 123 123 124.
Copyright © (2014) Colonial First State Group Limited.
All rights reserved.
3