Abstract
This chapter introduces some challenges and opportunities of working with data in the event study method. There are some similarities and differences in collecting, managing, and analyzing the data. The author also looks at some of the best practices in some issues that researchers may need to be aware of if they are not used to working with the types of data used in the event study method. First, he briefly looks at issues relating to identifying the company at a point in time as required to access the historical financial data. Next, the management of data from several sources is addressed. Finally, the chapter examines the data management lifecycle issues, focusing on the archiving processes at the end of a study. This chapter is not intended to be a comprehensive examination of data management issues, but to provide researchers with enough insight to understand when to get additional help and what sort of support they may need.
TopMain Focus Of The Chapter
It is not the intent or the purpose of this book to advocate for particular databases or software packages. For example, it is possible to undertake an event study with yahoo finance and a spreadsheet (e.g., open spreadsheet) and could, therefore, be conducted for free. However, such approaches have several limitations. First, the quality of data is low, and there is a high likelihood of errors. Second, is the inability to adapt calculations quickly (e.g., running a different estimation model) and tests to determine whether a calculated CAR is statistically significant. A gold standard may be using well-cleansed sets of data, such as provided by the CRSP for U.S. stocks, and software such as Eventus (which has been used for years with the calculations carefully examined and checked). In between, there is a continuum of options. If your institution does not have access to these packages, several options are available. It may be possible to collaborate with a colleague who has an institutional subscription to the services/software required. Alternatively, affordable options can be explored.
Many researchers may be more familiar with spreadsheets for managing data. We cover some fundamentals of spreadsheets to use the type of data we will use in event studies. In terms of calculations, I have found it easiest to run most of my analysis and models in “R” (R Core Team, 2012) because:
- 1.
It is designed for statistical analysis and has a heavy academic/scholarly following. This ensures that researchers constantly update packages and incorporate the latest techniques and tools into packages for wide use.
- 2.
It is free, so I am not locked in and do not need to worry about leaving my employer and finding funding for a new license.
- 3.
It is command-line driven. I admit this is a double-edged sword; there is a steep learning curve, but the model’s history and tests are clear, and adjustments can be rapidly made.
- a.
A key drawback is the steep learning curve, and it is often challenging to understand what is wrong with a model or a command if it is not working.
- b.
On the other hand, you can rapidly develop and change models when you know what you are doing. The model includes the list of inputs, and you can cut-paste these as a record of the model development and analysis. This speed of model development and adjustment can be valuable later as you record the analysis decisions for writing up a manuscript or dissertation.
- 4.
R can generate useful images and figures that may be helpful during the analysis or communication of the study results.
Key Terms in this Chapter
Data Manipulation: The process of modifying the data structure and data types to make sure we can use it effectively in the event study analysis.
Firm Identifiers: An identifier for a firm. This identifier should be unique (i.e., not shared with another firm) and can be used to extract appropriate information from databases or data providers such as CRSP. Examples include the ticker symbols and the CUSIPs.
Archive: The process of taking the data used in the study, cleansing it, and creating a copy in another destination. The process should create files that other researchers will quickly understand if they need to access the archived files.
Data Management: The set of approaches used to process, store, and organize the data used in the study.
Data Accuracy: Ensuring that the data is accurate and is free from errors. Given the large amount of data we may use in an event study, data accuracy is best assured by using reliable sources, like CRSP.
Backup: The process of creating a copy regularly that can be used if the working file is damaged or the data is lost.
CRSP: CRSP is the Center for Research in Security Prices, an affiliate of the University of Chicago Booth School of Business. The team at CRSP curates the data and ensures data accuracy.