Managers’ Private Information, Investor Underreaction and Long‐Run SEO Performance

• DOI: http://doi.org/10.1111/j.1468-036X.2011.00616.x
• Date: 01 November 2013
• Published date: 01 November 2013
• Author: Pawel Bilinski,Norman Strong

European Financial Management, Vol. 19, No. 5, 2013, 956–990

doi: 10.1111/j.1468-036X.2011.00616.x

© 2011 John Wiley & Sons Ltd

Managers’ Private Information,

Investor Underreaction and Long-Run

SEO Performance

Pawel Bilinski

Lancaster University Management School, Lancaster University,Lancaster, LA1 4YX, UK

Email: p.bilinski@lancaster.ac.uk

Norman Strong

Manchester Business School, University of Manchester, Booth Street East, Manchester,

M156PB, UK

Email: norman.strong@mbs.ac.uk

Abstract

Fora sample of 2,879 SEOs by US stocks from 1970 to 2004, this paper decomposes

an average three-year post-issue buy-and-hold abnormal return of −25.9%

(relative to size- and B/M-matched non-issuing stocks) into two components. One

component, representing41% of the total, is due to lower risk exposure. The second

component, representing the remaining 59%, is abnormal performance related to

the surprise element of the issue decision, which the paper attributes to managers’

private information that the market does not incorporate into the announcement

return. This second component results in abnormal returns during the 16 months

after the offering.

Keywords: managerial private information,investor underreaction,seasoned eq-

uity issues,long-run performance

JEL classification: G1, G2, G3

1. Introduction

Seasoned equity issuers underperform benchmark stocks over the long run.1We test the

behavioural explanation for this finding, namely the underreaction hypothesis, which

We would like to thank the EFM editor, John A. Doukas, and, in particular, Jay Ritter, the

referee, whose comments led to significant improvements in the paper.We also thank Michael

Brennan, Ning Gao, Ian Garrett, Kai Li, Weimin Liu, Roberto Mura, and participants at the

2009 Spanish Finance Association meeting, the 2010 Midwest Finance Association meeting,

and the 2010 Eastern Finance Association meeting for helpful comments. Correspondence:

Pawel Bilinski.

1The discussion of Table 1 below cites relevant evidence.

E-mail: p.bilinski@lancaster.ac.uk

E-mail: norman.strong@mbs.ac.uk

© 2011 John Wiley & Sons Ltd

Long-run SEO Performance 957

assumes that managers act rationally on their private information about stock overvalu-

ation when announcing an equity issue. Investors underreact to the SEO announcement,

stock mispricing persists at the issue date, and subsequent underperformance occurs as

manager and investor valuations gradually converge. This is consistent with Loughran

and Ritter’s (1995, p. 48) statement that ‘ ...our numbers imply that if the market

fully reacted to the information implied by an equity issue announcement, the average

announcement effect would be −33 percent, not −3 percent.’

Our proxy for manager’s private information about stock overvaluation is the inverse

Mills ratio (lambda) from a probit model that predicts the issue decision based

on publicly available infor mation. Less anticipated SEO announcements have higher

surprise components, implying greater stock overvaluation and giving higher lambdas.

The underreaction hypothesis, therefore, states that higher lambdas predict lower post-

announcement returns as the announcement effect fails to fully incorporate managers’

private information.

For a sample of 2,879 equity issues in the USA over 1970–2004, we decompose an

average three-year post-SEO buy-and-hold abnormal return of −25.9%, relative to size

and book-to-market (size–B/M) non-issuing stocks, into two components. The first is

a−15.43% abnormal return related to the surprise element of the equity issue decision,

which we attribute to investor underreaction to managers’ private infor mation revealed

by the issue announcement. The second component of −10.44% is due to lower post-

SEO risk exposure. We show that SEOs are larger, more liquid, with higher investment

rates and B/M ratios, and lower gearing and profitability than benchmark stocks after

the offering. A detailed analysis of the underreaction to the issue shows that it corrects,

on average, within 16 months. Thereafter, SEO returns are consistent with returns on

similar assets. This gives a much shorter period of market correction than the five-years

in Loughran and Ritter (1995) and Spiess and Affleck-Graves (1995).

The relation between post-issue returns and managers’ private information is robust

to controlling for pre-issue (abnormal) return performance, SEO attrition before the

end of the holding period, delisting returns, and hot issue periods. The results are also

robust to possible misspecification of the empirical model that could cause estimates

of managerial private information to include public information affecting returns over

a cross-section of stocks, since we find no relation between our estimate of managerial

private information and returns for size–B/M benchmarks or randomly drawn samples

of non-issuing CRSP stocks.

This study decomposes SEO post-issue performance relative to size–B/M benchmarks

into a short-term underreaction effect and a long-run discount rate effect. We propose an

unbiased and consistent test of the underreaction story compared to previous studies that

investigate whether investors rationally interpret publicly available information around

the equity offering. A relation between public information and post-issue returns could

reflect a relation that affects all stocks, rather than an effect specific to the equity issue.2

Consequently, past studies offer weak support for the underreaction hypothesis. Our

tests control for the discount rate hypothesis, which says that lower risk explains low

2For example,Rangan (1998) attrib utes lowSEO post-issue returns to investor underreaction

to pre-issue earnings management. Xie (2001), however, f inds that investors overprice

discretionary accruals in a cross-section of stocks and high discretionary accruals predict

low returns over the following two years. Xie (2001, p. 359) argues that ‘[discretionary

accruals] mispricing is not limited to settings that give managers opportunistic incentives to

manipulate earnings, such as before IPOs or seasoned equity offerings’.

© 2011 John Wiley & Sons Ltd

Pawel Bilinski and Norman Strong

958

SEO returns. We show that SEO post-issue expected return estimates are biased and

inconsistent in the absence of a formal model of managers’ private information and their

decision to make an SEO. In contrast to some previous findings, however, we find that

a discount rate effect alone cannot explain SEO underperformance.

Evidence of less-than-rational investor behaviour around various corporate events

suggests that adjusting for private information should accompany any cross-sectional

regression of post-announcement returns.3Our study has implications for related

international research, since the literature documents SEO underperformance in non-US

markets with different regulatory and institutional regimes and other equity flotation

methods.

The paper proceeds as follows. The following section reviews previous literature that

examines the underreaction and discount rate explanations for post-SEO returns and

presents our empirical hypothesis. Section 3 develops an econometric model to test

this hypothesis. Section 4 presents the data and we confirm previous evidence of SEO

underperformance in Section 5. We test our hypothesis and present empirical results in

Section 6. Section 7 examines the length of the delayed market reaction and Section 8

concludes.

2. Previous Literature and Development of the Underreaction Hypothesis

Several studies document abnormally low SEO returns up to five-years after equity

issues. Table1 summarises US and international evidence on long-run SEO perfor mance

using buy-and-hold abnormal returns. Underperformance ranges from −9.0% over three

years for German SEOs from 1960–1992 (Stehle et al., 2000) to −53.3% over three years

for Canadian SEOs from 1993–2004 (Carpentier et al., 2010) and to −59.4% over five

years for US SEOs from 1970–1990 (Loughran and Ritter, 1995).

Direct evidence on whether SEO underperformance is due to irrational investor

behaviour around equity offerings is difficult to obtain. Previous studies investigate

whether investors rationally interpret manager and analyst actions around equity

offerings. Rangan (1998), Teoh et al. (1998), and DuChar me et al. (2004) show that

increases in pre-issue discretionary accruals predict post-issue returns, consistent with

investors incorrectly extrapolatingpre-issue ear nings performance.4Shivakumar(2000),

however, f inds that the market rationally anticipates pre-issue earnings management

and undoes its effect at the issue announcement. He argues that test misspecification

explains Rangan’s (1998) and Teoh et al.’s (1998) results. Jegadeesh (2000) suggests that

investors are over-optimistic about issuers’ future earnings at the issue date and adjust

their expectations in response to disappointing earnings results after the issue. Jegadeesh

reports that SEOs underperform by twice as much around post-issue quarterly earnings

announcements as outside these periods. Brous et al. (2001), however, f ind no evidence

of abnormal returns around quarterly post-issue earnings announcements. As most of

these studies assume investor underreaction to publicly available information, failing to

model investor reaction to managerial private information may explain their different

results.

3See Subrahmanyam (2007) for a recent review of the behavioural finance literature.

4Iqbal et al. (2009) provide corresponding evidence on earnings management before UK

open offers.