Merrill Dimensions Newsletter – July 2012
An Interview with Suzanne Morsfield, CEASA
The Center for Excellence in Accounting and Security Analysis (CEASA) at Columbia Business School recently completed a detailed study on how investors, analysts, and other data-seekers in the market are using XBRL-tagged financial data. In the following interview for Dimensions, Director of Research Suzanne Morsfield shares some of the findings and recommendations from that study. Part 2, which will appear in next month's issue, details the discussions with analysts and investors.
What is the Center for Excellence in Accounting and Security Analysis (CEASA)?
CEASA is a research center within Columbia Business School located in New York City, which seeks to gather the best thinking from academia and practice to improve both accounting and security analysis. Note that the term “practice” includes more than practicing CPAs: we do include accountants and auditors, but we also include investors, analysts, technologists, and regulators; and the “best thinking from academia” implies those writings and studies from scholars that can inform real-world practices and approaches, not only those works that are publishable in leading academic journals. As a result, the Center's output is not directed just to other academics, though we would hope they would be interested in our work. Our intent is to provide an independent and unbiased influence over the real practice and regulation of financial reporting, as well as the real practice of security analysis.
How did XBRL became a topic of research interest for the Center?
CEASA's interest in studying XBRL derives from the fact that XBRL, which is a method for communicating a firm's financial results, has the potential to enhance security analysis. In the United States, XBRL-tagged financial information is now required by the SEC for most of the significant portions of 10-K and 10-Q filings. The main target audience, according to the SEC mandate, is investors and other security analysts.
Beyond this theoretical motivation to examine its usefulness to investors and analysts, two CEASA team members have prior experience with developing and implementing an XBRL-like taxonomy and analytical framework at Morgan Stanley. That particular taxonomy and framework, as well as the resulting datasets, were designed specifically for use by precisely the audience envisioned by the SEC when it decided to require XBRL-tagged 10-K and 10-Q information.
From this previous experience, we knew that XBRL-tagged information would have great potential. We were also aware that how the mandate was implemented would have significant ramifications on the quality of the underlying XBRL data, which in turn could have serious repercussions on the end-user community's confidence in and desire to use the data. So, the combination of the Center's mission, our personal experiences, and the apparent need for an unbiased view about how useful the XBRL data are currently for investors and analysts converged quite easily in this project.
How did this study come about?
Our study originated with the seemingly straightforward goal of understanding the usefulness of the SEC's XBRL-tagged data to investors and analysts. The process began with a closed-door session that included representatives from all of the key stakeholder groups affected by or influencing the XBRL mandate in the United States-ranging from filers and data providers to sell-side analysts and portfolio managers. The goal of this session was to provide a means for the stakeholders to hear from each other firsthand about their respective views on the mandate itself, as well as its implementation. A related portion of the discussion was devoted to what kinds of financial statement data or other information are important to each type of stakeholder when assessing the true economic performance of an SEC-reporting firm.
Without disclosing anything specific from the ensuing discussion, we can note that the type, amount, and timing of required financial information varied by stakeholder in many instances, while in other cases, the responses were the same or similar. A main outcome of the session was that each type of stakeholder learned, on a direct basis, about the varying information requirements of the other stakeholders, as well as about their varied views on the pros and cons of the XBRL mandate thus far. Many stated that they hoped more such sessions would follow. From this session, the need for an independent study on the usefulness of the XBRL data to investors and analysts became apparent to CEASA. We formulated a plan to study this issue, which resulted in six months of interviews, discussions, and surveys of additional representatives from the key XBRL stakeholder groups, culminating in a relatively more formalized study focused solely on investors and analysts.
What insights did you gain in the course of your conversations with stakeholders?
In the course of our study, we observed several things related to the SEC's mandate and to pre-implementation expectations, which likely added confusion around the usefulness and usability of the XBRL data file to investors and analysts. We provide a more comprehensive discussion of these items in our forthcoming full report, but two that I find particularly interesting are an appearance gap and an expectations gap:
1. XBRL tagging follows traditional paper filings too closely
The SEC mandate does not specifically require that the XBRL data have the same appearance as that of the related traditional paper filings. However, the tagging appears to us to follow the traditional filings enough that the usability of the underlying data is often obscured by the technical tagging and formatting requirements which enable a filer to meet the presentation-related portions of the mandate-i.e., those requirements described in the relevant sections of the initial educational seminars presented by the SEC, and in the current EDGAR Filer Manual.
For example, all line item names in a filer's XBRL data file must match exactly those in the related HTML filing. And while this does not appear, at first glance, to be that onerous or confusing of a requirement, it has led to the use of extensions where none are needed and has made searching for data across a large sample of reporting firms more difficult than necessary. Portions of the regulation compel filers to use the XBRL technology to reconstruct financial statement headings, table headings, and column and row headings of the related HTML filing. The emphasis, even if by accident, shifts from the data items themselves to their presentation without much difficulty and without much added value. In my opinion, there are other ways to ensure that the XBRL data files contain nothing more and nothing less than the related official filing.
2. There was an expectations gap
A hard-to-avoid issue arises when a long-awaited, potentially much-hyped product or technological advancement is finally released-i.e., a gap between the expectations raised or developed along the way and the actual experience of the first users of the new product or technology. A related expectations gap can also occur between creators and end users who foresaw something other than how the product turned out or how it was used upon release.
The gap between certain investors' expectations and their perceptions about the usability of the currently available XBRL data-as well as between filers' expectations about how much the data would be used upon release and the realization of those expectations to date-is significant. This gap can be best described as follows: Filers and regulators appeared to have believed that investors would use the newly available XBRL-tagged data in whatever form it was in, while investors who knew about XBRL were under the impression that the XBRL-tagged information would be relatively more accurate, timely, inclusive, and usable upon its release than other financial information available to them.
Did discovery of this gap lead to further questions?
It is easy to understand how and why these gaps may have occurred, but it is little discussed in the conversations about why “no one is using our data” that understandably come from filers. Whether or not the data are usable to investors is for investors to decide. And whether or not investors are using the data is a question that requires a methodical and meaningful approach to measurement. Both of these issues presented CEASA with a further opportunity to conduct an independent study on both the usability and the use of XBRL data by investors and analysts.
To study these questions, we chose the simplest and most reliable method in this case, which was to ask the questions directly of as many analysts and investors as would be willing to share their thoughts with us.
A more detailed description of your discussions with analysts and investors will appear in the next issue of Dimensions, but please share some highlights now.
Our questions for analysts and investors were quite straightforward. We wanted to understand two separate but related topics:
- Their use of certain financial and other data, as well as their data sources and their timeliness requirements for these data items
We limited our focus to items in the footnotes, the MD&A, earnings releases, non-SEC filings data, or nonfinancial data. We did this on purpose because we know that investors and analysts have ample sources of data from the body of the financial statements, but what about the other types of data that they might need to use to make their investment decisions and conduct their own research?
Their knowledge of, use of, or willingness to use XBRL-tagged SEC data
- We knew anecdotally that some investors and analysts had heard a lot about XBRL, while others may not have known anything about it. Here are some of our findings from the investors and analysts we surveyed about their actual use of items from the footnotes, the MD&A, earnings releases, non-SEC filings data, or nonfinancial data.
- Although exact use varies by investor/analyst type, virtually all of those we spoke with use the following: detailed footnote and MD&A numerical data; numerical data items that are not comparable with similar or peer companies; and various nonfinancial detailed measures found in SEC filings.
- Virtually all of these investors and analysts manually collect these data points either because they are not available from the large data providers or because they do not trust the accuracy of the large data providers.
- Knowledge of XBRL varies by size of institution. Larger investor or equity research institutions generally have heard of or used XBRL data. However, none of these end users went to corporate websites to view or access the data; instead, they relied on other access points from the SEC or XBRL US.
- Smaller institutions generally have never heard of XBRL.
- Once they know about XBRL, many investors and analysts say they are eager to use XBRL detail-tagged footnote data (in particular) in significant and noticeable amounts and ways. Most investors and analysts would like more detailed tagging (e.g., tagged MD&As and earnings releases), not less.
- Most of the demand is pent-up demand, however, because they are waiting for the data to meet their needs, which means, in part, that they would like to see present and past tagging errors cleaned up, as well as those situations where the tagged data do not match the EDGAR HTML filings data (even document dates are wrong in some cases).
- They would like to see the elimination of unnecessary use of extensions, again for both the present and the past (detailed-tagged items are key), or a solution to the difficulty in accessing and using the data tagged with extensions.
- Some investors would like to see the XBRL data file audited for tagging correctness, as well as numerical accuracy.
Do you think changes are needed in the way XBRL data is prepared and presented?
In the full report, we describe what, in our opinion, are some of the most important steps to enhance the usability and usefulness of the XBRL data, at least for investors and analysts. If the target audience is only filers and regulators and the XBRL technical community, then a different set of recommendations or, perhaps, no recommendations at all may be appropriate. But, if XBRL data are to reach the target audience for which the mandate was written, then changes are needed, and needed yesterday from regulators, filers, consumption tool designers, and the XBRL technical community. At a minimum, existing data quality problems and nonproductive tagging requirements must be looked at in earnest by filers, regulators, and yes, even by auditors.
We learned recently that the XBRL technical community in the United States is attempting to move forward on the technical aspects. In a statement on the XBRL.US website,the technical community acknowledges that formatting of paper financial statements and XBRL-specific technical requirements have contributed, in part, to at least some of the usability and analysis limitations. To their credit, the technical stakeholders are beginning to address the aspects of those issues that are in their domain.
Suzanne Morsfield is Director of Research at Columbia Business School's Center for Excellence in Accounting and Security Analysis (CEASA). Before joining the faculty of Columbia Business School, Dr. Morsfield held roles in major investment banks and was an audit professional at Ernst & Young. She has taught at the University of Arizona, the University of Michigan, NYU's Stern School of Business, and Baruch College.
NOTE: The views expressed here are entirely Dr. Morsfield's and do not necessarily reflect those of CEASA or any other organization.