Max Mansur is a Global Market Manager in Securities Markets at SWIFT.  He works in the asset servicing domain addressing the needs of custodians, securities market infrastructures, broker/dealers, investment managers, issuers, and their agents. Mr. Mansur has 31 years of information technology experience including system design, program management, quality, and financial solutions. He can be reached by email.

This is the first part of a two-part interview.

1. In your capacity as a senior manager at SWIFT, you are working with the DT and XBRL US "to fundamentally change rporate actions announcement processing" by integrating XBRL into the system. To begin with, uld you explain what rporate actions are and what activities they entail?

A rporate action is any activity that impacts a security that has been traded, cleared, and settled on behalf of the investor. Typically initiated by the public mpany issuing an equity or offering a bond, the action needs to be described and mmunicated to all investors and holders of the security as well as the public. The announcement of the action can include dates, rates, periods, prices, options, nditions, terms, exclusions, identifiers, and other characteristics. So there’s a lot of data that is often wrapped deeply in legal verbiage to prevent or mitigate any potential litigation for the action. The more mplex the action, the longer the document will be, often into hundreds of pages.

The tradition and even regulatory mandate is all paper-oriented, hence the interest in XBRL to identify unambiguously the embedded data in acrdance with a widely used standard. There are more than 50 types of events that are nsidered a rporate action including dividends, interest payments, name changes, splits, tenders, mergers, redemptions, calls, spinoffs, warrants, exchanges, Dutch auctions, bonuses, repurchases/bids, and bankruptcy notices. Globally, about one million rporate actions are “announced” each year, 400,000 in the U.S. There are several times as many “schled” rporate actions for fixed inme securities with periodic distribution of payments, but these are very straightforward so not the focus of our initiative.

2. The data standard ISO 20022 is already in widespread use for rporate action events announcements. Why is an additional data standard like XBRL needed?

First, as an aside, let me say that ISO 20022 is not yet in widespread use for rporate actions; rather, ISO 15022 is the re securities standard that is currently being used by the SWIFT mmunity. However, ISO 20022 was developed to be easily translatable with ISO 15022 to help support a period during which both standards will exist. As such, we’re avoiding a ‘big bang’ approach to the migration of a few thousand global institutions using ISO 15022 to XML-based ISO 20022. As these institutions asynchronously obtain funding and transition their messaging technology to XML, they can expect to either translate or interface with translated messages for a period of time. SWIFT is preparing support for this transition in nsultation with our members and will address other message domains as well, such as clearing and settlement messages.

Now, to your question: First, it is important to understand that XBRL rporate reporting taxonomies are inherited from underlying GAAP or IFRS acunting standards. XBRL doesn’t attempt to create acunting standards; rather, through the design and linkbases of XBRL, it enables the use of existing standards for extensible data reporting that enhance and add value to the human readable form of business reporting. So the XBRL taxonomy for rporate actions doesn’t start from scratch; it begins with the ISO 20022 standard and, by design, will maintain through the label linkbase and style sheets the alignment with ISO 20022. The growth of TS (mmercial off-the-shelf) software tools for XBRL, facilitating a marriage between “paper” and “data” is the value that XBRL brings to the table beyond ISO 20022. 

The genesis of ISO financial messaging has been the enabling of standard business workflows between financial institutions based on a mmon business model and data dictionary. ISO 20022 is emerging for rporate actions, and will be live in late 2010, built on the very solid basis of ISO 15022, a tag-value messaging standard that was widely adopted in the early 2000s for securities, currently with several millions of messages processed globally each month. Another several million rporate actions messages per month are not ISO, but proprietary within several markets, including the U.S.’s DT messages. The advent of ISO 20022 means, with XML and a more robust business definition basis, that virtually all markets globally are mmitted to the new standard. DT will begin with rporate actions announcements in 2010. This is a large part of the opportunity to ncurrently roll out a rporate actions taxonomy aligned with ISO 20022 in 2010 and reap the rewards of clean data from the source using the most advanced standard available.

3. So ISO 20022 will be the basis for the XBRL data elements, structures, and validation in the rporate actions taxonomy?   

Yes, that’s right. Additionally, market practice guidelines help define entry points in the taxonomy, making tagging eminently more usable by the issuers and their agents. I think this mpares favorably with the way US GAAP erns the financial reporting taxonomies where XBRL only expresses data elements that are meaningful from a GAAP viewpoint while implementing entry points that are useful by the reporting industry, such as energy or airline manufacturing. Interestingly enough, we expect fewer than 500 data elements for rporate actions whereas US GAAP currently exceeds 17,000 data elements.

4. uld you describe the "scrubbing" process that rporate action announcements now undergo? How will XBRL remove the problems associated with it?

Good, now we’re at the heart of the matter. Many different institutions have the responsibility of monitoring securities for announcements of rporate actions. Upon disvering an action announced for a held security, a custodian gathers feeds from multiple sources, each of which has attempted to send the rrect data in a message. The re problem is that each source manually entered the data by reading and interpreting the source document. This is a recipe for discrepancy, thus scrubbing is necessary to create the “golden py” which is really just the best version of the truth based on all these indirect feeds.

If the issuer or their agent tagged all the key elements of data and attached an XBRL file to the filing, press release, or prospectus with which they currently announce to the market, then there uld not be any discrepancy because there would be data from the source. The additional information that a financial institution, like a custodian, adds to the message is all specific to the institution’s acunting and servicing of the holdings related to the event, not the event itself.

One important aspect to remember is that a mplex rporate action event may take several months to mplete from start to finish, with many documents filed and many updates. Each update is repeatedly going through the “scrubbing” process without re data from the source. This is the source of tremendous risk, as well as all the additional time necessary to manually check, rekey, validate, and try to prevent a stly error that uld be in the millions of dollars.

5. Slide 7 of the rporate action presentation you made in July to the XBRL technology workshop held in Santa Clara states that "Institutional investors often use multiple custodians with the potential for nflict in the custodians’ descriptions of the rporate action." How will XBRL help to overme this challenge?

This is a very deep problem and not easy to explain. The best analogy is a variant of the “telephone” or “whispers” game, except beyond the single chain of a game, it begins with the same mplex information from not one, but many origins and then, through several daisy chains, the messages are ncentrated into the investment manager. The unavoidable misinterpretation drives errors, thus the need to “scrub” to figure out which of the nflicting data is rrect, requiring significant time and introducing risk. So the proposed solution is that, instead of a “whisper,” we’re asking for a standardized “form” of the data instance to be given once from origin of the message. If each custodian and data provider and broker/dealer starts with the same data, with standard-driven precision, at the same time (via XBRL) then how much risk remains?

6. Why is XBRL particularly useful for the technology needs of rporate action announcements? What other technology solutions were nsidered and rejected? Why is XBRL superior to these alternatives?

XBRL is tailor made for “pointing out” the data of interest within a report. It has huge technical and business momentum. The data can be identified without disturbing the nditional and legal necessities wrapping the data in human readable reports. The issuers are already being enabled to use this technology, so adding a bit more seems achievable for a great deal of investor value. There are tools to help with tagging and reading/rendering. The only other options might be (a) messaging from agents using the ISO 20022 Issuer Agent rporate Actions messages that Euroclear is deploying for mmunications between registrars and their depositories, (b) some kind of paper form, or (c) re-inventing some kind of tagging interface to build into or around the ISO 20022 messaging. rporate actions announcements are business reports for public availability and investor nsumption, like rporate reporting. So XBRL is a neat solution.