Event-Study Analysis

Economists are frequently asked to measure the effect of an economic event on the value of a firm. On the surface this seems like a difficult task, but a measure can be constructed easily using financial market data in an event study. The usefulness of such a study comes from the fact that, given rationality in the marketplace, the effect of an event will be reflected immediately in asset prices. Thus the events economic impact can be measured using asset prices observed over a relatively short time period. In contrast, direct measures may require many months or even years of observation.

The general applicability of the event-study methodology has led to its wide use. In the academic accounting and finance field, event-study methodology has been applied to a variety of firm-specific and economy-wide events. Some examples include mergers and acquisitions, earnings announcements, issues of new debt or equity, and announcements of macroe-conomic variables such as the trade deficit.1 However, applications in other fields are also abundant. For example, event studies are used in the field of law and economics to measure the impact on the value of a firm of a change in the regulatory environment,2 and in legal-liability cases event studies are used to assess damages.* In most applications, the focus is the effect jof an event on the price of a particular class of securities of the firm, most bften common equity. In this chapter the methodology will be discussed in jerms of common stock applications. However, the methodology can be applied to debt securities with little modification.

Event studies have a long history. Perhaps the first published sturdy is Dolley (1933). Dolley examined the price effects of stock splits, studying nominal price changes at the time of the split. Using a sample of 95 splits

Wc will further discuss the first three examples later in the chapter. McQueen and (1993) provide an illustration using macroecouomic news announcements. -SecSchwert (19H1). See Mitchell and Netter (1994).

Roley

f. Event-Study Analysis

j.

from 1921 lo 1931, lie found dial die price increased in 57 of the cases and the price declined in only 20 instances. There was no effect in the other 12 cases. Over the decades from the early 1930s until the late 1900s the level of sophistication of event studies increased. Myers and Bakay (1948), Barker (1950, 1957, 1958), and Ashley (1902) are examples of studies during this time period. The improvements include removing general stock market price movements and separating out confounding events. In the late 1900s seminal studies by Ball ami Brown (1908) and Fama, Fisher, Jensen, and Roll (1909) introduced the methodology that is essentially still in use today. Rail and Brown considered the information content of earnings, and Fama, Hishcr, Jensen, and Roll studied the effects ofstock splits after removing the effects of simultaneous dividend increases.

j In the years since these pioneering studies, several modifications of the basic methodology have been suggested. These modifications handle complications arising from violations of the statistical assumptions used in the erly work, and they can accommodate more specific hypotheses. Brown aijd Warner (1980, 1985) are useful papers that discuss the practical iin-pwrtancc of many of these modifications. The 1980 paper considers implementation issues for data sampled at a monthly interval and the 1985 paper deals with issues for daily data.

j This chapter explains the econometric methodology of event studies. Scttion 4.1 briefly outlines the procedure for conducting an event study. Section 4.2 sets up an illustrative example of an event study. Central to any event study is the measurement of the abnormal return. Section 4.3 details the first step-measuring the normal performance-and Section 4.4 follows with the necessary tools for calculating the abnormal return, making! statistical inferences about these returns, and aggregating over many evehl observations. In Sections 4.3 and 4.4 the discussion maintains the null hypothesis that the event has no impact on the distribution of returns. Section 4.5 discusses modifying the null hypothesis to focus only on the mean of the return distribution. Section 4.6 analyzes of the power of an event study. Section 4.7 presents a nonparamctric approach to event studies which eliminates the need for parametric structure. In some cases theory provides hypotheses concerning the relation between the magnitude of the event abnormal return and firm characteristics. In Section 4.8 we consider cross-sectional regression models which are useful to investigate such hypotheses. Section 4.9 considers some further issues in event-study design and Section 4.10 concludes.

4.1 Outline of an Event Study

At the outset it is useful to give a brier outline of the structure ol an event study. Wlnlc there is no unique structure, the analysis can be viewed

4. I. Outline of an Event Study

as having seven steps:

1. Event definition. The initial task of conducting an event study is to define the event of interest and identify the period over which the security prices of the firms involved in this event will be examined-the event window. For example, if one is looking at the information content of an earnings announcement with daily data, Ihe event will be the earnings announcement and the event window might be the one day of the announcement. In practice, the event window is often expanded to two days, the day of the announcement and the day after the announcement. This is done to capture the price effects of announcements which occur after the stock market closes on the announcement day. The period prior to or after the event may also be of interest and included separately in the analysis. For example, in the earnings-announcement case, the market may acquire information about die earnings prior lo the actual announcement and one can investigate this possibility by examining prc-cvcnl returns. 2. Selection criteria. After identifying the event of interest, il is necessary to determine the selection criteria for the inclusion of a given firm in the study. The criteria may involve restrictions imposed by data availability such as listing on the NYSF. or AMFX or may involve restrictions such as membership in a specific industry. Al this stage it is useful to summarize some characteristics of the data sample (e.g., firm market capitalization, industry representation, distribution of events through time) and note any potential biases which may have been introduced through the sample selection. 3. Normal and abnormal returns. To appraise the events impact we require a measure of the abnormal return. The abnormal return is the actual ex post return of the security over the event window minus the normal return of the firm over the event window. The normal return is defined as the return that would be expected if the event did not take place. For each firm i and event date r we have

e* = Л1(-К[ I X,), (4.1.1)

where e*, /?,- and F.(/t ) arc the abnormal, actual, and normal returns, respectively, for time period Л X, is the conditioning information for the normal performance model. There arc two common choices for modeling the norma! return-the constant-mean-rcturn model where X, is a constant, and the market model where X, is the market return. The constant-mcan-rcttirn model, as the name implies, assumes that the mean return of a given security is constant through lime. The market model assumes a stable linear relation between the markei return and the security return.

4. Evenl-Slntly Analysis

I. Estimation procedure. Once a normal performance model lias been selected, die parameters ol die model must be estimated using a subset olthe data known as the estimation wintlmn. The most common choice, when feasible, is lo use the period prior to the event window for the estimation window, for example, in an event study using daily data and the market model, the market-model parameters could be estimated over the 120 davs prior lo ihe event. Generally the event period itself is not included in the estimation period to prevent the event from influencing the normal performance model parameter estimates. 5. Testing procedure. Willi the parameter estimates for the normal performance model, the abnormal returns can be calculated. Next, we need lo design the testing framework for the abnormal returns. Important considerations are defining the null hypothesis and determining the let hniques for aggregating the abnormal returns of individual firms, ti. Empirical results. The presentation of the empirical results follows the .formulation ol the econometric design. In addition to presenting the basit empirical results, the presentation of diagnostics can be fruitful. ()<< asionallv, especially in studies with a limited number of event observations, the empirical results can be heavily influenced by one or two firms. Knowledge of ibis is important for gauging die importance of the results.

7. Interpretation and conclusions. Ideally the empirical results will lead to insights about the mechanisms by which the event affects security prices. Additional analysis mav be included lo distinguish between competing explanations.

4.2 An Example of an Event Study

lhe Financial Accounting Standards Hoard (FASH) and the Securities Exchange Commission strive lo set reporting regulations so that financial statements and related information releases are informative about the value of the firm. In selling standards, the information content of the financial disclosures is of interest. Event studies provide an ideal tool for examining the information content of the disclosures.

In this set lion we describe an example selected to illustrate the event-study methodology. One particular type of disclosure-quarterly earnings announcements-is considered. We investigate the information content of quarterly earnings announcements for ihe thirty firms in the How Jones Industrial Index over the live-year period from January 1989 to December 1993. These announcements correspond to ihe quarterly earnings for the last quarter of 1988 through the third quarter of 1993. The live years of data lor thirty firms provide a total sample of 000 announcements. For

43. Models for Measuring Normal Performance

each firm and quarter, three pieces of information are compiled: the date of the announcement, the actual announced earnings, and a measure of the expected earnings. The source of the date of the announcement is Datastream, and the source of the actual earnings is Compustat. !

If earnings announcements convey information to investors, one wiuld expect the announcement impact on the markets valuation of the firms equity to depend on the magnitude of the unexpected component of the announcement. Thus a measure of the deviation of the actual announced earnings from the markets prior expectation is required. We use the mean quarterly earnings forecast from the Institutional Brokers Estimate System (I/B/E/S) to proxy for the markets expectation of earnings. I/B/E/Scom-piles forecasts from analysts for a large number of companies and reports summary statistics each month. The mean forecast is taken from the last month of the quarter. For example, the mean third-quarter forecast from September 1990 is used as the measure of expected earnings for the third quarter of 1990.

In order to examine the impact of the earnings announcement on the value of the firms equity, we assign each announcement to one of three categories: good news, no news, or bad news. We categorize each announcement using the deviation of the actual earnings from the expected earnings. If the actual exceeds expected by more than 2.5% the announcement is designated as good news, and if the actual is more than 2.5% less than expected the announcement is designated as bad news. Those announcements where the actual earnings is in the 5% range centered about the expected earnings are designated as no news. Of the 600 announcements, 189 are good news, 173 are no news, and the remaining 238 are bad news.

With the announcements categorized, the next step is to specify the sampling interval, event window, and estimation window that will be used to analyze the behavior of firms equity returns. For this example we set the sampling interval to one day; thus daily stock returns are used. We choose a 41-day event window, comprised of 20 pre-event days, the event day, and 20 post-event days. For each announcement we use the 250-trading-day period prior to the event window as the estimation window. After we present the methodology of an event study, we use this example as an illustration.

4.3 Models for Measuring Normal Performance

A number of approaches are available to calculate the normal return of a given security. The approaches can be loosely grouped into two categories- statistical and economic. Models in the first category follow from statistical assumptions concerning the behavior of asset returns and do not depend on