Ambiguity Aversion, Company Size and the Pricing of Earnings Forecasts

Date01 June 2014
DOIhttp://doi.org/10.1111/j.1468-036X.2012.00651.x
Published date01 June 2014
European Financial Management, Vol. 20, No. 3, 2014, 633–651
doi: 10.1111/j.1468-036X.2012.00651.x
Ambiguity Aversion, Company Size
and the Pricing of Earnings Forecasts
Constantinos Antoniou
Xfi Centre for Finance and Investment, Universityof Exeter, Exeter, EX4 4ST, UK
Email: c.antoniou@exeter.ac.uk
Emilios C. Galariotis
Audencia PRES-LUNAM, Centre for Financial and Risk Management, 8 route de la Joneli`
ere,
Nantes, 44312, France
Email: egalariotis@audencia.com
Daniel Read
Warwick Business School, University of Warwick, Coventry CV4 7AL, UK
Email: daniel.read@wbs.ac.uk
Abstract
Several authors have reported an unconditional size effect in returns around
earnings announcements. In this study we show how this finding can be understood
as resulting from ambiguity aversion. We hypothesise that analyst forecasts
for smaller companies are relatively more ambiguous; hence they are priced
pessimistically by ambiguity-averse investors. As the quarter comes to a close and
ambiguity gradually subsides, the stock prices of smaller companies rise to correct
this pessimism, creating the size effect. Our results support these hypotheses.
Keywords: ambiguity aversion,size premium,analyst earnings forecasts
JEL classification: D03, D81, D84, G11, G14
1. Introduction
Neoclassical valuation theory postulates that investors calculate a projection of earnings,
determine its present value, and trade accordingly. An assumption of this framework
is that the distribution of expected earnings can be fully calculated from available
information, and so decisions are made in conditions of risk. In some cases, however, the
available information maybe of such low quality that investors are unable to confidently
We wish to thank two anonymous referees, John Doukas (the Editor), Nicholas Barberis,
Donald Brown,Richard Har ris, DavidKelsey, Tigran Melkonyan,and par ticipants at the 2008
European Financial Management Conference and the 2009 Risk and Ambiguity Seminar at
Yale University for helpful comments and suggestions. Correspondence: C. Antoniou.
C
2012 Blackwell Publishing Ltd
© 2012 John Wiley & Sons Ltd
Warwick Business School, University of Warwick, Gibbet Hill Road, Coventry, CV4 7AL, UK
Email: constantinos.antoniou@wbs.ac.uk
Constantinos Antoniou, Emilios C. Galariotis and Daniel Read
634
estimate the distribution of expected earnings. In these cases decisions are made in
conditions of ambiguity. Epstein and Schneider (2008) show theoretically that in such
conditions pricing will be initially pessimistic due to ambiguity aversion,1with this
pessimism being corrected with an upward price movement as ambiguity gradually
subsides.
A pattern observed in stock market data is a ‘size effect’ around earnings announce-
ments (Ball and Kothari, 1991; Atiase 1985; Foster et al., 1984; Berkman et al., 2009).
Ball and Kothari (1991) find that, compared to large firms, small f irms have higher
abnormal returns around earnings announcements irrespective of the sign or magnitude
of the earnings surprise. This size effect is puzzling because it is economically significant
and does not appear to reflect compensation for risk. Foster et al. (1984) extend
this result and show that the unconditional upward movement in the prices of small
companies begins months before the earnings announcements. In this paper we provide
an ambiguity based explanation for this finding. Our hypothesis is that the analyst
forecasts for smaller companies used by investors to form earnings expectations2are
relativelymore ambiguous because there is less ‘hard’ information with which to estimate
forecast accuracy (something which wedemonstrate quantitatively).3These forecasts are
consequently priced pessimistically due to ambiguity aversion.As the quarter comes to a
close and more information about upcoming earnings is revealed, ambiguity is gradually
resolved and so the stock prices of smaller companies drift upwards to correct this
pessimism, generating the documented size effect.
Our first test examines whether smaller firms are likely to be perceived as more
ambiguous by investors. It is motivated by Daniel Ellsberg’s (1961) proposal that when
there is little reliable information about the relative likelihood of events,decision makers
will feel they face ambiguity.4In our framework the event of interest is a forecast, and
the likelihood of interest is its accuracy. When this accuracy can be predicted from
1Ambiguity aversion is one of the most well-established results in behavioural economics.
A large literature, starting with Knight (1921) and Keynes (1948), and continuing through
Ellsberg (1961) and up to the present day (Ahn et al., 2009), showsthat situations that involve
ambiguity are treated differently from those that involve risk. Hsu et al. (2005) present
evidence that ambiguous situations produce a unique neurological fingerprint, suggesting
that ambiguity aversion is rooted in the fundamentals of human cognition. See Camerer and
Weber (1992) and Keren and Gerritsen (1999) for reviews on the evidence on ambiguity
aversion. As noted by Hirshleifer (2008) ambiguity aversion has a broader economic impact
as it also affects financial regulation.
2The importance of analyst forecasts to investors is evident from the robust evidence that
they exert a powerfulinfluence on asset prices (Givoly and Lakonishok, 1979; Stickel, 1992;
Gleason and Lee, 2003; Dische 2002). The impact of analyst forecasts is so large that it is
common in the finance literature to treat analysts’ forecasts as a proxy for market expectations
(Livnat and Mendenhall, 2006).
3Along these lines various authors have suggested that the information environment of
smaller companies is poorer (Atiase, 1985; Zhang, 2006; Hong et al., 2000).
4Frisch and Baron (1988) proposed that ‘ambiguity is uncertainty about probability, created
by missing information that is relevant and could be known’ (P. 1988). Einhorn and Hogarth
(1985) suggest that ambiguous situations arise when the available information is vague, and
does not allow one to confidently rule out alternative possibilities, while G¨
ardenfors and
Sahlin (1982, 1983) argue that feelings of ambiguity are produced when the relevance of the
available information is low. For all these authors the underlying theme is that ambiguity is
C
2012 Blackwell Publishing Ltd
© 2012 John Wiley & Sons Ltd

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT