Too Good to be True? An Analysis of the Options Market's Reactions to Earnings Releases

• Author: Sugata Ray,Yan Lu
• Published date: 01 July 2016
• Date: 01 July 2016
• DOI: http://doi.org/10.1111/jbfa.12214

Journal of Business Finance & Accounting

Journal of Business Finance & Accounting, 43(7) & (8), 830–848, July/August 2016, 0306-686X

doi: 10.1111/jbfa.12214

Too Good to be True? An Analysis of the

Options Market’s Reactions to Earnings

Releases

Yan Lu and Sugata Ray∗

Abstract: Using option implied risk neutral return distributions before and after earnings an-

nouncements, we study the option market’s reaction to extreme events over earnings announce-

ments. While earnings announcements generally reduce short-term uncertainty about the stock

price, very good news does not reduce uncertainty and slightly bad news actually increases

uncertainty. We also find that left tail probabilities decrease over earnings releases while right

tail probabilities increase. We interpret these findings as evidence of maintained investor expec-

tations that very good news is generally not released during earnings announcements, combined

with skepticism in the form of lingering uncertainty at the release of such very good news.

Keywords: earnings announcements, uncertainty resolution, option implied distributions,

tail risk

1. INTRODUCTION

Earnings announcements are a standard source of information for equity analysts

performing fundamental valuation of the companies they cover. This study uses

short maturity option prices to examine uncertainty resolution during earnings

announcements, paying particular attention to tail events. It is well known that option

implied volatility increases leading up to earnings announcements and decreases

immediately after (see, e.g. Daley et al., 1988; Isakov and Prignon, 2001). By their very

nature, earnings calls reduce uncertainty as management divulges previously private

information and analysts ask questions that help them fill in cells in their valuation

spreadsheets.

However, not all earnings announcements are equally informative. Bad news may

be accompanied by management reticence. Good news may be accompanied by overly

effervescent commentary. Idiosyncratic factors, such as illness, other big news in the

∗The first author is from College of Business Administration, University of Central Florida, USA. The second

author is from Warrington College of Business Administration, University of Florida, USA. We are grateful

to David Brown, Jongsub Lee, Mahendrarajah (Nimal) Nimalendran, Jacob Sagi and seminar participants

at the University of Florida and FMA meetings for comments and thank Dominique Badoer for excellent

research assistance. (Paper received December 2014, revised revision accepted July 2016). [Acceptance date

was missing and has been added to the online version of the article].

Address for correspondence: Sugata Ray, Warrington College of Business Administration, University of

Florida, P.O.Box 117168, University of Florida, Gainesville FL 32611-7168.

email: sugata.ray@ufl.edu

C

2016 John Wiley & Sons Ltd 830

TOO GOOD TO BE TRUE? 831

market or a misstep in answering a particularly pointed question from an analyst may

lead to differing levels of uncertainty resolution, which in turn may lead to material

effects in the option and stock markets. After all, these calls are transcribed and metic-

ulously pored over by hundreds of analysts looking to profit from any fundamental mis-

valuation in stock prices. Additionally, not all uncertainty is the same: some earnings

announcements may have significant downside risk which may or may not be realized;

others may have little tail risk but a large range of outcomes nonetheless. Little is

known about the resolution of tail risk uncertainty during earnings.

We use short maturity option prices for a given company before and after the

earnings announcement to compute a risk-neutral and model-free distribution

of the company’s stock returns until the maturity of the option.1Using the risk

neutral distributions, we compute implied standard deviation and tail probabilities

for the stock before and after earnings. We use these statistics to measure and

analyse uncertainty resolution during the earnings announcement. Finally, we test

the moments of the risk-neutral distributions against those of realized distributions.

This allows us to discern potential biases in how option markets perceive uncertainty

resolution during earnings announcements.

Comparing option implied standard deviations before and after earnings, we find

that the implied standard deviation decreases significantly over the earnings release.

However, the change in implied standard deviation depends on whether good news or

bad news is released. We find large negative returns over the earnings window (‘very

bad news’) results in greater resolution of short-term uncertainty than large positive

returns (‘very good news’). A small negative return during the earning announcement

(‘slightly bad news’) leads to increased short-term uncertainty about the stock price,

while small positive returns (‘slightly good news’) lead to a significant decrease in

uncertainty about the stock price.

We also look at how tail probabilities evolve over earnings announcements. We

find that the probability of a left tail event significantly decreases and the probability

of a right tail event significantly increases.2The decrease in left tail probabilities

is intuitive, as investors are likely to believe that the risk of very bad news after an

earnings announcement, but before the maturity of the next option, is low. The

increase in right tail probability is surprising and may be related to the skepticism

with which investors greet very good news reported during earnings documented

above. Investors may have maintained expectations that very good news is generally

not released during earnings releases.

The difference between option market reactions after a very good earnings event

(not much uncertainty resolved) and realized returns after such an event (same

amount of uncertainty remaining as after very bad events) suggests that there may

be a behavioral component to the responses in the option market’s reaction. The

decrease in implied standard deviations accompanying very bad news suggests a sense

of relief and the relatively low decrease in implied standard deviation accompanying

very good news suggests misplaced skepticism. Relief at very bad news is in line with

Skinner (1994), who shows that disclosing very bad news is painful to management due

to potential legal and reputational repercussions. The skepticism at very good news is

1 An example of what these model-free distributions look like can be found in Figure 1.

2 Tail probabilities are defined as probability mass two or more implied standard deviations away from the

mean return.

C

2016 John Wiley & Sons Ltd