Event study

from Wikipedia, the free encyclopedia

Event studies are financial statistical methods that can be used to check whether and how certain events, such as B. share buyback programs, affect the valuation of companies and their security prices (e.g. share or bond prices ). In order to determine the influence of an event in the context of an event study, the actually realized returns on the security prices are compared with the expected (normal) returns, e.g. B. can be estimated using prices from a historical period, compared to the time of the event. The expected return represents the so-called "counter factual".

The theory of efficient capital markets forms the theoretical basis for the methodology of the event study. This theory is due to the Nobel Prize winner Eugene Fama. It states that in an efficient market, the prices of securities reflect new and value-relevant information immediately. A price change as a result of a certain event (e.g. the announcement of a share buyback program), assuming an efficient capital market and in the absence of competing value-relevant events (English confounding events), can have an effect of this event on the company's security price and consequently its revaluation below Taking this information into account.

methodology

The difference between the realized and the expected (normal) return on a security price is known as the abnormal return (also known as excess return, or AR for short). Various methods are available for estimating the expected return, i.e. the return that could have been expected without the event examined. The market model, which is based on a regression analysis of historical returns on a security and a corresponding security index within an estimation window, is a frequently used method (see e.g. MacKinlay (1997)) ). Alternative calculation models include: a. the “constant mean return model”, the capital asset pricing model (CAPM) and multi-factor models are used.

An event window defines the period in which the potential impact of an event is expected or over which it is to be calculated. If the impact of an event on the value of a company or its security prices can extend over several trading days, then the overall influence of the event is determined using the cumulative abnormal return (CAR) for the entire event window of several trading days .

Event studies typically determine the average impact of certain events on security prices for a sample of companies that are affected by such events. The average impact on the security prices of all companies examined at a certain point in time (such as the day the event is announced) is determined using the average abnormal return (AAR for short). The AAR is obtained by cumulating the abnormal returns for all companies in the sample at a given point in time and then dividing them by the number of companies in the sample. The average overall influence of the observed event on all companies within the sample for the entire event window is determined using the cumulative average abnormal return (CAAR for short). The CAAR is the sum of the AARs for the entire length of the event window. Alternatively, the CAAR can also be determined as the sum of the individual CARs, which is divided by the number of companies. If prices are available for the entire event window, both approaches lead to the same result.

During the CAR approach is more suitable for short-term event studies used, are often used for long-term analysis, the so-called. "Buy-and-hold abnormal returns" (Engl. Buy-and-hold abnormal returns or short BHARs) determined (see eg. B Ritter (1991), Barber and Lyon (1997)). A “buy-and-hold return” is based on the idea that investors usually hold their investments for a certain period of time and do not realize returns every day. As a result, a buy-and-hold rate of return when acquiring a security at the beginning of the event window and then selling it at the end of the event window is calculated as the product of [1 plus return on each day in the event window] minus 1. The BHAR is determined as the difference between the realized and expected “buy-and-hold return”. A matching approach is often used to estimate the expected return, in which the expected return for a given day corresponds to the contemporaneous return of a reference index, a reference portfolio or a comparable company (see Lyon, Barber and Tsai, (1999)). In the context of long-term event studies, the calendar-time portfolios method, also known as Jensen's Alpha approach, is used. With this method, a portfolio of companies that are affected by a certain event is first formed. It is then examined whether this portfolio has an abnormal return that is not captured by a risk factor model (such as CAPM or multi-factor models) (see, inter alia, Mitchell and Stafford (2000)).

Finally, as part of an event study, it is important to check whether the calculated returns are statistically significant . Numerous parametric as well as non-parametric test methods can be used for this purpose . These test methods include, for example, the T-test (Brown and Warner (1980 and 1985)), the Standardized Residual Test (Patell (1976)) and the Generalized Sign Test (Cowan (1992)). As part of these tests, the validity of the null hypothesis that the abnormal returns (AARs, CAARs) have a value of zero is checked. The p-value that can be determined for these tests is compared with the predefined error probability for the false rejection of the null hypothesis and allows a quick assessment of the statistical significance of the results.

In summary, the typical course of an event study can be presented as follows:

  1. Definition of the event
  2. Definition of the sample and the sources of information
  3. Definition of the exact time of the event
  4. Elimination of events that became known along with other information relevant to the assessment
  5. Obtaining the necessary price data
  6. Definition of the model for determining the expected returns
  7. Determination of the estimation and event window
  8. Calculating Abnormal Returns
  9. Examination of statistical significance

A detailed description of the methodology of event studies can be a. Brown and Warner (1985), Campbell, Lo and MacKinlay (1997), MacKinlay (1997), McWilliams and Siegel (1997), Kothari and Warner (2008).

Types of price-relevant events

Basically, a distinction can be made between different types of price-relevant events, which can be traced back to different value-relevant information. Company events , such as announcements of share buyback programs, company takeovers ( M & As ) or changes in dividends , represent the primary type of value-relevant information. Regulatory events, e.g. B. include legal requirements for companies, also represent price-relevant information. Furthermore, macroeconomic events, such as announcements of monetary policy measures or other messages with regard to essential production factors, are to be mentioned as a type of value-relevant information. Events such as natural disasters or soccer world championships, which initially have no economic nature, can also represent price-relevant information because such events u. a. can have economic consequences for certain companies.

For example, an event study examined the reaction to airline share prices to what extent the four coordinated aircraft hijackings followed by suicide bombings in the terrorist attacks against the United States of September 11, 2001, had an impact (Carter & Simkins, 2004). One can also examine how news (announcement) about important free trade agreements ultimately impacted national financial markets (Moser & Rose, 2014).

Applying the methodology of event studies to other economic measures

Event studies are primarily used to determine stock returns. However, the methodology is also used to calculate bond yields (Bessembinder et al. (2009)) and credit derivative yields (Andres, Betzer and Doumet (2013)) as well as to determine abnormal trading volumes (Campbell and Wasley (1996)) and other measures of liquidity (Corwin and Schultz (2012)) and for determining voting rights premiums (Kalay, Karakaş and Pant (2014)).

Software for conducting event studies

To carry out short-term and long-term event studies, which can include a large number of events, there are special software solutions that facilitate the calculation of the abnormal returns and numerous corresponding test statistics. Event studies can also be programmed by the user himself in the common statistical software packages such as Matlab or Stata . There are also packages for the free programming language R that can be called up online with additional functions for carrying out event studies. An event study can also be carried out with Microsoft Excel.

Event studies in litigation

In civil, criminal and administrative law suits, incident studies can serve as evidence. In this way, for example, the damage can be substantiated or the causality between an act (or omission) and the damage that has occurred can be proven (see e.g. Müller (2015) with further information and examples).

Individual evidence

  1. ^ EF Fama: Efficient Capital Markets: A Review of Theory and Empirical Work. In: Journal of Finance. 25, 1970, pp. 383-417.
  2. ^ A b C. A. MacKinlay: Event Studies in Economics and Finance. In: Journal of Economic Literature. 35, 1997, pp. 13-39.
  3. ^ J. Ritter: The Long-Run Performance of Initial Public Offerings. In: Journal of Finance. 46, 1991, pp. 3-27.
  4. ↑ BM Barber, JD Lyon: Detecting Long-Run Abnormal Stock Returns: Empirical Power and Specification of Test-Statistics. In: Journal of Financial Economics. 43, 1997, pp. 341-372.
  5. ^ JD Lyon, BM Barber, C. Tsai: Improved Methods for Tests of Long-Run Abnormal Stock Returns. In: Journal of Finance. 54, 1999, pp. 165-201.
  6. ^ ML Mitchell, E. Stafford: Managerial Decisions and Long-Term Stock Price Performance. In: Journal of Business. 73, 2000, pp. 287-329.
  7. ^ S. Brown, J. Warner: Measuring Security Price Performance. In: Journal of Financial Economics. 8, 1980, pp. 205-258.
  8. ^ A b S. Brown, J. Warner: Using Daily Stock Returns - The Case of Event Studies. In: Journal of Financial Economics. 14, 1985, pp. 3-31.
  9. ↑ J. Patell: Corporate Forecasts of Earnings per Share and Stock Price Behavior: Empirical tests. In: Journal of Accounting Research. 14, 1976, pp. 246-276.
  10. ^ AR Cowan: Nonparametric Event Study Tests. In: Review of Quantitative Finance and Accounting. 2, 1992, pp. 343-358.
  11. ↑ eventstudymetrics.com
  12. ^ JY Campbell, AW Lo, AC MacKinlay: The Econometrics of Financial Markets. Princeton University Press, Princeton / New Jersey 1997, ISBN 0-691-04301-9 .
  13. ^ A. McWilliams, D. Siegel: Event Studies in Management Research: Theoretical and Empirical Issues. In: The Academy of Management Journal. 40, 1997, pp. 626-657.
  14. ^ SP Kothari, JB Warner: Econometrics of Event Studies. In: BE Eckbo (Ed.): Handbook of Corporate Finance: Empirical Corporate Finance. Elsevier / North-Holland 2008, ISBN 978-0-444-53265-7 , pp. 3-36.
  15. ↑ David Carter, Betty Simkins: The market's reaction to unexpected, catastrophic events: the case of airline stock returns and the September 11th attacks . In: The Quarterly Review of Economics and Finance . tape 44 , no. 4 , 2004, ISSN  1062-9769 , p. 539–558 ( repec.org [accessed June 21, 2019]).
  16. ↑ Christoph Moser, Andrew K. Rose: Who benefits from regional trade agreements? The view from the stock market . In: European Economic Review . tape 68 , May 1, 2014, ISSN  0014-2921 , p. 31–47 , doi : 10.1016 / j.euroecorev.2014.01.012 ( sciencedirect.com [accessed June 21, 2019]).
  17. ↑ H. Bessembinder, KM Kahle, WF Maxwell, D. Xu: Measuring Abnormal Bond Performance. In: Review of Financial Studies. 22, 2009, pp. 4219-4258.
  18. ^ C. Andres, A. Betzer, M. Doumet: Measuring Abnormal Credit Default Swap Spreads. 2013.
  19. ^ CJ Campbell, CE Wasley: Measuring abnormal daily trading volume for samples of NYSE / ASE and NASDAQ securities using parametric and nonparametric test statistics. In: Review of Quantitative Finance and Accounting. 6, 1996, pp. 309-326.
  20. ^ S. Corwin, P. Schultz: A Simple Way to Estimate Bid-Ask Spreads from Daily High and Low Prices. ( Memento of the original from December 26, 2015 in the Internet Archive ) Info: The archive link was inserted automatically and has not yet been checked. Please check the original and archive link according to the instructions and then remove this notice. In: Journal of Finance. 67, 2012, pp. 719-759. @1@ 2Template: Webachiv / IABot / www.afajof.org
  21. ^ A. Kalay, O. Karakaş, S. Pant: The Market Value of Corporate Votes: Theory and Evidence from Option Prices. ( Memento of the original from July 16, 2014 in the Internet Archive ) Info: The archive link was inserted automatically and has not yet been checked. Please check the original and archive link according to the instructions and then remove this notice. In: Journal of Finance. 69, 2014, pp. 1235-1271. @1@ 2Template: Webachiv / IABot / www.afajof.org
  22. ^ Event Studies with Stata. Retrieved June 21, 2019 .
  23. ↑ Dr Simon Mueller: EventStudy: Event Study Analysis. March 14, 2019, accessed June 21, 2019 .
  24. ↑ L. Müller: The evidence of causality by means of 'Event Study' reports in the context of capital market law. In: Current Legal Practice. 24, 2015, pp. 251-268.