xbrl

PART I

The eXtensible Business Reporting Language - XBRL is about to become standard also for the reporting of ESG information in Europe. When it comes to its relevance, needless to mention its almost 15-year history in the American environment for Accounting Reporting, considering only the period since the SEC determined its use. Therefore, it is the standard for digital reporting most users of financial information have available and it is likely to continue in that position for a long time.

However, despite all the alleged benefits and advantages from this tool, the possible uses for XBRL in the academic environment seem to have been ignored. The majority of empirical research in accounting is still done based on data available from the same providers that have been around for decades, although the data obtainable from XBRL filings are likely to more precise, faster to gather and provide more granularity. In fact, data can be available just a few seconds after filings are uploaded to the SEC and at a cost much lower than the usual market prices, and XBRL makes it possible to control for many elements inherent to the company's accounting reporting structure, such as segment reporting and many other multidimensional contexts. But still, very few researchers are able to find, or even use, data obtained from such filings.

The reason for that, in my opinion, is that XBRL is extremely difficult to work with. It is not some csv file that you import and create histograms from with a few clicks of a mouse. Instead, it is a nest of xml files located in different places, connected just by another xml file: the instance document filed by a company and that carries all the information available in the whole statement. It does require strong technical skills and time to be ready for any statistical analysis. However, the fact that each company can individually contribute to the complexity of something already too complicated by creating their own custom elements, although a highly interesting feature of the system, is only possible because the system complex. Because it is flexible (extensible), it has too many moving parts and academics simply do not have the time to deal with that.

Those moving parts are what make the system so interesting for academic research. For example, when working with the usual commercial data sets, the amounts recorded for "Assets" are available for the whole company, while with XBRL one can obtain total assets for product segments, countries, subsidiaries and any levels the filer company choses to report. In addition, despite the number of variables ordinarily available for research, it does not match the number available in an XBRL filing. Some examples are: "Equity component of convertible notes, net of issuance", "Deferred finance costs related to equity component", "Reclassification of equity from/to temporary", "Compensation costs capitalised in inventory balances", "Contractually obligated inventory holding period" and many more.

Because it is standardised data and contains virtually everything that is filed by a company, XBRL files can help researchers with more control variables and more interesting and impactful research ideas. To get there, we need to improve our "data department", so it includes modern sources that are also flexible enough to accommodate structual elements and new variables as they are created.