Abstract
While the authors has a clear preference for an event study that estimates the short-run calculations of abnormal stock returns over a short multi-day window, there are other forms that readers should be familiar with. Further, you should be aware of the challenges of conducting and interpreting these studies. Therefore, this chapter addresses long-run studies and some difficulties, criticisms, and interpretation issues of using these studies. Finally, it looks at studies that do not use abnormal stock returns but use changes in operational data. Using these alternatives can enable a project to extend its contribution by using several studies (e.g., both short- and long-run studies) in one journal article or dissertation to converge on insights into the phenomenon being studied.
TopIntroduction
This book focuses on the short-run abnormal returns estimated in an event study, based on the stock market reaction. The results show the stock market reaction and market participants’ expectation of how the event affects the companies (Sorokina et al., 2013). Other approaches will often focus on permanent or long-term effects showing or demonstrating an enduring impact. For instance, a short-run effect may be indicated based on the stock market reaction, which provides an immediate estimate of the impact of an event. However, as we know from operations management, forecasts will frequently be wrong. Therefore, a long-term study can often determine whether there is a permanent and enduring impact from the event, often over a one- to five-year time period (Bremer et al., 2011).
This chapter examines these operational and long-term event study approaches. We provide a brief overview so that the reader is aware of some of the key distinctions and differences between the study types. The chapter does not contain sufficient detail and depth for a reader to use to design and execute the studies. In this way, the reader can become familiar with the other types of event studies they might find and will be able to evaluate and assess the quality of the research they read. However, interested readers should continue to advance their knowledge of these other types of studies before conducting the rain.
It can also include a mixture of studies to extend the findings. For instance, Hendricks et al. (2014)use a short-horizon study, a long-horizon study, and an operational performance event study. They provide three reasons why the newly created positions might elicit a positive stock market reaction. They find evidence to support the third reason, that past performance is relatively normal, and the stock market participants expect that this newly created position will enable the company to lift their performance in the future. Through examining the operational data and the changes in the return on assets and sales, they find no evidence of poor performance prior to the new position being created. In the post position operational performance analysis, they find performance results consistent with stock price performance results were a relative improvement in performance. In this way, the mix of event study types provides useful supplements and enables Hendricks et al. to test alternative hypotheses. For instance, they note how they:
We also analyze the abnormal stock price performance in the period after the SCOME appointments to see whether such appointments are associated with subsequent improvement in stock price performance. We estimate CARs over a 250-day postannouncement period that starts on day 21 and ends on day 272, which spans a year that starts one month after the announcement and ends 13 months after the announcement. The reason we start one month after the announcement is to reflect the possibility that the new SCOME may join sometime after the announcement. We use day 282 to day 481 as the estimation period to compute the CAR. The reason for using a forward estimation period is that the appointment of a SCOME may cause a shift in the parameter estimates, and the forward estimation period will better reflect any such shift. The results are based on 582 firms because 99 firms did not have sufficient data during the estimation period and/or the postannouncement period to compute CARs. Panel B of Table 6 presents these results. (Hendricks et al., 2014, p. 1575)
Following the appointment, they were also interested in any change in operating performance, as measured using return on assets (ROA), sales over assets (SOA), as measured by sales/total assets, and cost of goods sold (CGS), and selling, general, and administrative expenses (SGA) summed then the total divided by sales, and CSGA divided by sales. Again, they find interesting operational performance changes, as they note how:
For the full sample as well as for all subsamples, the abnormal ROAs from year 0 to year 2 are insignificantly different from zero. Overall, the evidence suggests that new SCOME appointments are not followed by an immediate improvement in operating performance. However, there is no further decline in operating performance, suggesting that the decline in operating performance observed in the preappointment period under the existing SCOME does not appear to continue under the new SCOME. (Hendricks et al., 2014, p. 1577).
Key Terms in this Chapter
Calendar-Time Portfolio: The construction of a portfolio each calendar month of firms affected by the event the prior month to assess abnormal returns.
Model Mis-Specification: The situation where a model in the regression violates the underlying assumptions; for instance if the data distribution does not match those assumed in the model.
Buy-and-Hold Abnormal Return (BHAR): Abnormal returns are estimated by comparing the firms affected by the event (the treatment group) with a reference (a control group) that is not affected. The difference in returns, gained by holding over time, indicates abnormal returns attributable to the event.
Short-Horizon Event Study: A study that estimates an abnormal return over a short event window.
Bad Model Problem: The difficulty in assessing normal returns to determine whether the observed returns are abnormal or not. An accurate estimate of abnormal returns requires the estimation of the underlying normal returns and the abnormal returns.
Operational Performance: The measureable aspects of performance relating to outcomes of the operations department, such as the processes, quality, inventory, and reliability.
Long-Horizon Event Study: a study that estimates the abnormal return over a long period.