The short-term event study method, grounded in the Efficient Market Hypothesis, is one of the most widely used tools for quantifying the impact of a specific event on a firm's shareholder value. As the short-term event study method has been increasingly employed by researchers to investigate various operations and supply chain management (OSCM) events, it is timely to conduct a systematic review of the method to examine how it has been implemented in the OSCM literature and what could be improved to deploy it for future OSCM research. Analyzing 29 short-term event studies published in renowned OSCM journals between 1995 and 2017, we find that OSCM researchers generally follow the standard procedures in conducting event studies, but pay less attention to some methodological issues ranging from addressing the confounding events to expanding the event windows. Based on our analysis, we provide several recommendations for future event studies in OSCM, such as the opportunity for studying external events in the non-U.S. context, the caution of expanding the event windows, and the need to deal with the self-selection bias.
Over the past few decades, there is growing recognition of the strategic importance of operations and supply chain management (OSCM) in creating shareholder value. OSCM plays a vital role in generating shareholder value through the mechanisms of revenue growth, operating cost reduction, and efficient use of fixed and working capital (Martin and Lynette, 1999). Following this theoretical logic, researchers have conducted various empirical studies to analyze the connection between OSCM and shareholder value, among which the event study method represents one of the most popular methodologies adopted in the literature. Grounded in the Efficient Market Hypothesis (Malkiel and Fama, 1970), the short-term event study method relies on the premise that the value of market information will be reflected almost completely in the equity prices in financial markets. By detecting the abnormal equity price changes in response to new market information available in the financial market, the short-term event study method enables researchers to quantify the impact of a specific event on a firm's shareholder value (MacKinlay, 1997).
With its growing popularity in the OSCM literature, the short-term event study method has been employed by researchers to investigate various OSCM topics such as supply chain disruptions (Hendricks and Singhal, 1997; Zhao et al., 2013), environmental management (Jacobs, 2014; Klassen and McLaughlin, 1996), and quality management (Lin and Su, 2013; McGuire and Dilts, 2008). In addition, short-term event studies in OSCM are evolving as a result of advances in asset pricing models and statistical analysis. The method has been modified to address potential statistical issues specific to different research settings (Fama and French, 2015; Kothari and Warner, 2007). In view of its increased popularity and recent methodological improvements, it is timely to conduct a systematic review of the method to examine how it has been implemented in the OSCM literature and what could be improved to deploy it for future OSCM research.
Reviewing 29 short-term event studies published in renowned OSCM journals between 1995 and 2017, we have the following observations: (1) The majority of the short-term event studies in OSCM focus on internal corporate events in the U.S. context. (2) While most studies set standard event windows including at most three days around the event, theoretical justifications are not commonly provided for short-term event studies with longer event windows. (3) Researchers often rely on multiple data sources to identify the events under study, but pay less attention to the issue of confounding events. (4) The market model is the most popular estimation model in the OSCM literature, but some researchers also employ multiple estimation models to increase the robustness of the analysis. (5) Researchers are wary of possible violations of the assumptions for the significance test, so adopting various modifications of the traditional t-test according to different research contexts. (6) Researchers often conduct subsequent cross-sectional regression and ANOVA to probe into the operational determinants of variations in abnormal returns.
Based on our analysis, we provide several recommendations for future event studies in OSCM. First, we urge OSCM researchers to take advantage of events external to the firms concerned and occurring outside the U.S. context, advancing our understanding of the financial impacts of these under-studied events. Second, researchers should be careful about expanding the event windows, and provide theoretical explanations to justify the window lengths. Third, removing confounding effect is a critical step in conducting short-term event studies. Fourth, the possible self-selection bias should not be ignored, especially when the events under study are initiated by firms voluntarily. Fifth, employing alternative models to estimate the expected returns could enhance the robustness of the analysis. Sixth, modifications of the traditional t-test might become necessary in some research settings such as external events and industry-specific studies. Finally, independence is a vital assumption in testing the significance of cumulative abnormal returns. It thus is important to address the issues arising from time and industry clustering.
Our research is important in several ways. First, it serves as a practical guide for OSCM researchers interested in employing the short-term event study method in their research. We document the detailed steps of conducting a short-term event study and discuss some common issues encountered in each step, thus enabling OSCM researchers to have a better understanding of how a short-term event study should be conducted. Moreover, to the best of our knowledge, this is the first comprehensive review of event studies in the OSCM literature. Given the increased prevalence of event studies in OSCM, it is imperative to provide an overview of the current state of knowledge and best practices adopted in the OSCM literature. Finally, our research identifies several important research design issues that are often ignored by researchers of past shortterm event studies in OSCM, as well as some emerging opportunities specific to the OSCM context, so helping advance the adoption of the event study method for OSCM research.
2. Literature review
The first event study reported in the literature was perhaps conducted by James Dolley in 1933. Based on a sample of 95 stock splits from 1921 to 1931, Dolley (1933) investigated the nominal stock price changes at the time of the stock splits. Modern event studies were initiated in the two seminal works of Ball and Brown (1968) and Fama et al. (1969). Modern event studies are developed into different categories in terms of the event window length and performance measurement. Long-term event studies detect abnormal stock returns over a period normally ranging from one to eight years with calendar-time portfolio abnormal return (CTAR) or buy-and-hold abnormal return (BHAR) (Barber and Lyon, 1997; Lyon et al., 1999), while short-term event studies examine abnormal stock returns over a maximum window length of 40 days (Brown and Warner, 1985; MacKinlay, 1997). A broader definition of event study goes beyond the scope of stock market reaction as it also measures other firm-level outcomes such as operating performance (Barber and Lyon, 1996). In parallel with advances in asset pricing models and statistical analysis, the event study method is still evolving to account for possible deviations from the fundamental assumptions. However, the gist of modern event studies remains the same, which is measuring the significance of sample securities' mean and cumulative abnormal returns around an event period (Kothari and Warner, 2007).
Originally applied in accounting and finance, the event study method has expanded its application to virtually all the business disciplines including management, information systems, marketing, operations and supply chain management (MacKinlay, 1997; McWilliams and Siegel, 1997). For example, in the marketing literature, researchers adopt the event study method to examine the financial impact of such marketing events as new product release, CMO appointment, brand acquisition and disposal, and Internet channel addition (Sorescu et al., 2017), while events attracting information systems researchers' attention include IT outsourcing, IT investment, IT excellence award, software vulnerability, and security breaches (Konchitchki and O'Leary, 2011).
Table 1 summaries previous literature reviews of event studies in different business disciplines. It indicates that the literature reviews in accounting and finance emphasize the econometric and statistical fundamentals and provide guidelines for applications in other fields. For instance, MacKinlay (1997) and Binder (1998) reviewed the use of event studies in finance, outlined the standard procedures for conducting event studies, and discussed the power of analysis and the subsequent regression analysis. Corrado (2011) reviewed variations in the basic short-term event study method to adjust for non-normality, event-induced volatility, and cross-sectional weighting. Kothari and Warner (2007) conducted a comprehensive survey of over 500 studies published in five of the top finance and accounting journals from 1974 to 2005. They found that the properties of the event studies reviewed were different depending on the time period and sample firm characteristics. They also indicated that, compared with short-term event studies, long-term event studies suffer from several important limitations.
As the event study method evolves over time, its statistical properties become well-defined and its applications are widely acknowledged. Literature reviews in other business disciplines place a greater emphasis on the research design issues and economic interpretations of the study results. McWilliams and Siegel (1997) conducted a survey of 29 event studies in three of the top management journals from 1986 to 1995. They discussed several concerns about the validity of the assumptions and research design issues. By replicating three studies in management with alternative research designs, they called for adequate attention towards the aforementioned concerns. They also indicated that the abnormal returns only reflect the effect on the shareholder wealth, rather than the welfare of all the stakeholders. Konchitchki and O'Leary (2011) examined the use of the event study method in over 50 information systems studies. They focused on the research design issues without investigating the actual results and conclusions in specific studies. Sorescu et al. (2017) identified over 40 event studies published in the marketing journals included in the list of Financial Times' 50 top business journals. In addition to research designs, their review examines interpretations of event studies as well. They provided economic inferences from the event studies by summarizing the main findings and common determinants of abnormal returns in the marketing literature.