Event studies enable researcher to study and evaluate the effects of particular anticipated or unanticipated event on securities prices.
It should be noted that in this part reviewed papers had employed the standard classic event study scenario, which could be defined as follows: firstly the events of interest are identified then normal performance of selected variable is obtained, thirdly the abnormal component is calculated and lastly statistical significance of analyzed events is determined. The alternative way of conducting event study was proposed by Ryan and Taffler (2002). It was proposed to conduct the event study backwards: starting with identification of statistically significant variations in stock prices volatility and trading volume (having normal and abnormal performances identified) and just then move to the identification of events that generated statistically significant share price or trading volume movements. Such backward event study conduction has a benefit of capturing all value-relevant information releases are captured, as opposed to the classical procedure where not all relevant events could be selected for consideration. For conduction of proposed research both methods will be used to verify the value-relevance of quarterly reports announcements. (For more detailed explanation of event study methodology see XX section).
Event studies let to measure “the magnitude of abnormal performance” (p. 4) Kothari at al. (2006) generated by some events (corporate information announcements, stock splits, ..). The ability to identify nonzero abnormal returns after some information releases let to evaluate stock market efficiency and makes event study methodology the major tool in investigating efficiency level of the market. The superiority of event study was also highlighted by Fama (1991), ho indicated “cleanest evidence on market-efficiency comes from event studies” p. 1607. If systematic abnormal returns are identified it is concluded, that the market is inefficient, but as it was noted, the test of stock market efficiency is treated as a joint test of market efficiency and correct asset pricing. Meaning that inefficiency could be caused either by inefficiency of the market or by incorrect definition of intrinsic asset value. As a major toll to increase the credibility of stock market efficiency evaluation event studies are used to reduce possibility of deceptive conclusions. In regard to the same issue Fama (1991) argues for event studies as the best tool for testing market efficiency: “they (event studies) come closest to allowing a break between market efficiency and equilibrium pricing issues, event studies give the most direct evidence on efficiency” Fama (1991) (p. 1577).
The relation of event studies to tests of market efficiency receives considerable attention in Fama (1991), and in recent summaries of long-horizon tests in Kothari and Warner (1997) and Fama (1998). Smith (1986) presents reviews of event studies of financing decisions. Recently, Kothari (2001) reviews event studies in the accounting literature. Kothari and Warner (2004)
The first event study (analysis of nominal price changes at the time stock split) was conducted by Dolley (1933) . But the first event study having the same methodological structure as it is used in nowadays was conducted by Fama, Fisher, Jensen and Roll (1969). It should be noted, that Ball and Brown (1968) in fact were the first to publish it. Fama et al. (1969) paper was also concerned with stock splits as Dolley (1993), but the difference is …. . Fama et al. (1969) in particular were analyzing the stock prices reaction to stock splits and research to what extend stock splits and other variables influences changes in returns. It should be noted, that monthly stock prices data was used. As indicates Fama et al (1969) it was assumed by scholars that stock splits should be treated as good news for investors, as dramatic increase in expected dividend payment arises before the announcement of the stock split. But the results showed otherwise - the abnormal returns at the time of stock split were almost zero. It should be noted, that abnormal returns were captured few before the actual stock split, but this could be explained by the fact that the stocks splits were performed at the periods of “boom” Fama eta al. (1969) periods (p. 11). After the information on the split and dividend changes were observed and properly considered by the market at the stock split by it self do not initiate any value-relevant stock price movements. This result of prices aggregating stock split and further dividend performance information quickly was with accordance with efficient market hypothesis.
Further more analysis of short time horizon effects surrounding corporate information release is very valuable for understanding market participants' behavior and expectations related with particular company. Also, evaluation of reaction magnitude to particular news is helpful to managers for understanding corporate decisions role.
Event studies in accounting field were started by Ball et al. (1968) analysis of Earnings announcements content and timing. For the analysis only net annual earnings were taken, event study was employed two models of expected earnings forecasting (regression and naïve models) and then mismatching real and forecasted earnings expectations generated reactions were analyzed. Scholars find out that “accounting income numbers capture about half of the net effect of all information available thought 12 months preceding their release” (p. 176). And only around 20 per cent of annual earning announcement month abnormal stock price performance could by explained annual earning announcement. This could be explained by the fact that final annual earning number contains previously released quarterly earnings announcements and analysts provided earnings expectations. The valuable insight of the research is not all information is valuable, and not all new information is value-relevant - only new and different then forecasted or known annual earnings information is valuable. As the weak part of this work narrow scope of analysis, monthly data employed and other assumptions used.
Empirical studies in accounting field were also enriched by Barber and Lyon (1996) by analysis of accounting-based methods used for abnormal operational performance determination employed in event. It should be noted that they already used daily stock return as proposed by Brown and Warner (1985). The major contribution to event studies literature was that scholars indicated that while constructing sample of companies to be researched “unusually well or poorly” performing companies should not be included in the sample as such companies lead to miss specified tests' statistics.
Already mentioned Brown et al. (1985) researched what effect has daily stock returns on the firm specific events studies' methodologies compared to monthly returns as it was commonly used by scholars. The major focus was on type I error (falls rejection of null hypothesis of zero abnormal performance) and the power of different tests. This paper had also made valuable input to empirical event studies by detailed analysis and suggestions how to overcome problems related with daily returns. For estimation stock return performance on event window 3 procedures were used: mean adjusted returns, Market adjusted returns, OLS. And the major conclusions derived were that „little difference in the power of alternative procedures“ (p.12) were identified between the methods, also tests power is larger using daily data. It could be noted, that OLS and market adjusted returns methods are superior than mean adjusted returns model in cases of event-date clustering. Sensitivity analysis also showed that no significant difference were identified and specifications of tests are not altered significantly (p.14). Scholars also showed, that longer than (-5, +5) event widows lead to decrease.
Other group of studies employing event studies was concerned with stock returns variances and trading volume volatility generated by some events. First one to show the statistically significant fluctuations of returns and trading volume during event widows by analyzing quarterly earnings announcements was Beaver (1968). Later his results were verified by Patell (1976).
Also variety of event studies history, methods valuation and explanations on usage of event studies could be found. MacKinlay (1997) paper provides explicit but concise explanation of event studies methodology procedure, overviews models used for normal stock performance on event windows, provides, guidelines of usage of statistical test by providing an example of event study on quarterly earnings announcements.