A New OBSA Data Collection Database

Hello again,

As promised, I’ve returned to fill you in on how the second phase of my data collection is progressing.

Overall, I am confident that the new database that we have constructed in excel is formatted most efficiently for this part of my project. There were a lot of concerns up front about the structure of this database given We decided to change the format from that of the previous database so that I have now started to create multiple rows within the spreadsheet for a single company year if that company has multiple types of Off Balance Sheet Arrangements. In the original database, we limited each company year to a single line in the database. This change poses a potential problem should we wish to combine the data from each database. Nevertheless, the current structure is the only one that allowed us to categorize the kinds of data I am collecting in a clear, concise manner.

To arrive at these categories, we returned to the SEC study that this project hopes to expand upon. Using the study and some of the subsequent final ruling, we devised some specific quantitative disclosures that should be collected for each of the four types of Off Balance Sheet Arrangements focused on by the SEC (guarantees, retained interests, derivatives, and variable interest entities).  While these references were not specific enough to say exactly what needed to be collected, they did provide general guidelines, such as disclosing the amounts of revenues, expenses, and cash flows of the Off Balance Sheet Arrangement. We compared these guidelines to some of the categories that I had already created in my first round of collection to arrive at our final list of categories (ie. the columns of our spreadsheet).

We also found several other transparency measures within the first study that we have also integrated into this database. Along with the other data that I have been collecting, I am now also collecting data concerning the changes in disclosure text between years, the business purpose of the arrangements, the quality of any cross references within the text, as well as a few other measures described in the Final Rule.

Having made the changes to the database, the task of collecting the quantitative amounts has been fairly easy and straightforward. The real time consuming portion of this project has been comparing the filings from year to year. It has become perhaps the most frustrating part of this project since I don’t feel myself getting as far as I would like to at the end of each work day.

Nevertheless, I am really happy with the overall progress that I am making and look forward to giving you one last update before summarizing what I have learned at the end of the summer.

–Joe