Where content meets code: U.S. SEC mandates, Python and financial content

Guest post by Diane Mueller
These are interesting times for Financial Content providers and consumers. The U.S. Securities and Exchange Commission (SEC) put the first stake in the ground during the Cox administration when they mandated the use of XBRL [1] and began in stages to require large public companies to submit XBRL-tagged documents to the SEC, turning corporate financial statements into more easily searchable and comparable documents.   

XBRL in a nutshell

XBRL is a freely available, open industry-XML data standard used to describe and identify each item of data in financial documents like 10K filings. This financial metadata turns corporate financial statements into more easily searchable and comparable documents and allows them to be validated and cross-checked for accuracy in a more automated fashion.

Congress tries to expand it

Back in May, Rep. Darrell Issa [2] (R-Calif.), introduced the Government Information Transparency Act (H.R. 2392 [3]) in an attempt to standardize the collection of business information throughout federal agencies. It would require agencies to use a single data standard (XBRL) and require that collected information be made readily available for public access.
This month, I found myself suddenly following legislative twitter feeds as the Members of the House of Representatives and Senate debated amendments to the Dodd-Frank Financial Regulatory Reform bill [4] that included a discussion of requirements that companies use a standard like XBRL when reporting financial data to federal agencies.

One standard to rule them all

Under one proposed amendment, credit agencies, hedge funds and investment advisers would have had to use a “single-standard data language” in the information they report to the Federal Reserve and other agencies. Hopes were high that some form of a data standard mandate would make it into the legislation. Unfortunately, the amendments were eventually dropped from the Dodd bill in one of the late night bargaining sessions on the Hill. It did show an increased awareness and understanding of the value proposition of metadata standards like XBRL at the highest levels of government. 

Another interesting approach

Another  interesting proposal came this past month in the form of another SEC proposed mandate suggesting that, in addition to filing their complex, dense and almost impenetrable prospectuses,  Asset-Backed Securities (ABS) issuers be required to file a computer program that maps out the logic to the flow of funds, or “waterfall,” provisions of the transaction itself. The SEC is proposing that this computer program be filed on EDGAR in the form of downloadable source code in Python in a similar manner as the XBRL financial content of the earlier mandate. Python is a dynamic, high-level programming language that runs on Windows, Linux/Unix, Mac OS X, and is free to use, even for commercial products, because it uses an open source license. It is an excellent choice for the SEC’s purposes here.

Where content meets code

This is where the content meets the code--XML data (aka financial content) and an open logical data model(aka a python script)combine to enable true transparency. It’s one thing to provide the open data marked up with metadata,but if the logic is so dense that even the issuers and trustees have a tough time figuring out who to pay first, as is the case with some ABS securities, it defeats the purpose. When the financial content has become so complex that the simplest way to describe it is with computer code describing the entire waterfall, then the choice to use a freely available open source technology is the right move.

Making financial filing information more accessible

The obvious goal here is that the filed source code, when downloaded and run by an investor, would provide the user with the ability to programmatically input the user’s own assumptions regarding the future performance and cash flows from the pool of assets, including assumptions about future interest rates, default rates, prepayment speeds, loss-given-default rates, and any other necessary assumptions. Along with the code, the filers would need to supply asset-level data files that would be filed at the time of the offering and on a periodic basis thereafter. Presumably this would be XML content, perhaps even XBRL that when combined with the waterfall computer program, would give investors and regulators the tools to validate the logic and perform simulations and scenarios with it. 

Final thoughts

I applaud the SEC's recognition of the need for--in addition to the metadata tagging of financial content with XBRL--an even deeper level of logical ‘tagging’ in order to truly facilitate both the investor’s and the regulator’s ability to navigate, understand, model and monitor the complexities of today’s financial content.  

The combined application of open data standards and a open source language will ensure that the content providers and tools vendors across the financial content supply have unencumbered open access to the technologies required to build the tools and infrastructure, to ensure we are all better prepared to recognize any future financial crisis, as Representative Darrell Issa and others members of Congress continue to work to bring higher levels of transparency throughout our federal agencies.

Diane Mueller is Director, Enterprise Product Management at ActiveState [5], the dynamic language experts; she serves on the XBRL International [1]’s Steering Committee, Best Practice Board, and Technical working group on Rendering and is also a principal advisor to Metamoji.com [6]. 

For more information on python and finance, visit http://www.activestate.com/activepython [5]

Related Article:
One on One with Diane Mueller, XBRL expert at Just Systems [7]