24 January 2010

It is NOT different this time, and XBRL will not avoid the coming crisis


There has been a lot of discussion suggesting that if there had been an XBRL taxonomy for MBSs and other CDOs (etc) and a taxonomy for government stimulus spending, that somehow the next crisis can either be averted or reduced - that the lack of transparency was a cause of the last crisis, and will contribute to the next.

I disagree with this view, and while I advocate the expanded use of XBRL for an increasingly wide range of business reporting applications, I do not think that a taxonomy or MBSs and CDOs (etc) would in any way either have avoided the crisis, or reduce its impact. Yes it would have increased transparency, but no, it would not have avoided the last crisis or the next one.

More data is a good thing, but data is not the same as political or business courage. Reasoned thinking has been disengaged, and "Its different this time" is the prevailing view. The last crisis was caused by a failure of trust, while the next one will be caused by a lack of political courage.

Trust, not Detail

The crisis was caused by a common group-think that said that house prices would continue to go up, that there was no bubble, and that when an agency gave a CDO a AAA rating, that the rating could be believed. Ratings serve as a form of trusted assessment of detailed information, and as such it was (and is) the responsibility of the rating agency to assess the risk of default on a total CDO package, and on the individual tranches of the package. Equally, no matter how complex the merging and re slicing of such tranches, it remained (and remains) the responsibility of the rating agency (or other rater) to understand what is in each tranche and to assign an appropriate rating.

The facts are that investors trusted rating agencies, and added their own prejudices, allowing that to drown out market indicators. "But this time it is different" is the common refrain during all bubbles, and this one was no exception.

But some were watching, and some were speaking about the potential risks. Others were watching, and were making their own investment decisions based on what their models told them.

I remember Chris Whalen in his weekly article (www.instituationalriskanalytics.com) warning of CDOs and subprime resets as far back as 2005 and 2006. Most recently, Dennis Santiago, CEO of Institutional Risk Analytics, told me "bottom line is that as long as people are allowed to create opacity on purpose in finance these problems will resurface from time to time."

Likewise, there have been a number of stories about the major players, and Goldman Sachs in particular. From Vanity Fair, January 2010 article on Goldman Sachs:

In the aftermath of the crisis, criticism erupted that Goldman had continued to sell mortgage-backed securities to its clients while betting against those very securities for its own account. Clearly, in the simplest terms possible, this is true: while Goldman was never the biggest underwriter of C.D.O.’s (collateralized debt obligations—Wall Street’s vehicle of choice for mortgage-backed securities), the firm did remain in the top five until the summer of 2007, when the market crashed to a halt.
Goldman argues that the buyers of their C.D.O.’s were themselves sophisticated investors who were capable of making their own decisions. In other words, they were counterparties. “You don’t shut your franchise down just because you have a view of a market,” says Cohn. “When we do an I.P.O., people don’t ask us our view of the stock market.” But a less generous interpretation was given in a recent McClatchy Newspapers series, which quotes an analyst report that describes Goldman as being “solely interested in pushing its dirty inventory onto unsuspecting and obviously gullible investors.” (A Goldman spokesperson says, “The statement is not true. The McClatchy series was characterized by unsubstantiated claims, innuendo, and outright falsehoods.” McClatchy, however, stands by its work.) And so, if the old Goldman was defined by its refusal to do hostile takeovers, the new Goldman is defined by its skill at protecting its own interests.

Clearly this was a case of a seller being happy to sell a product that the client wants to buy, regardless of how good or bad it is for the client. Sort of like a non-smoker selling cigarettes. All the additional XBRL tagged information would not have changed investment decisions, certainly not on a scale that would have influenced the outcome.

To use another analogy, paper based maps have improved in their level of detail and quality of content. In addition, technology has allowed the development of the GPS (Global Positioning System) and Google Maps applications, to mention just two. Yet the data is still there at the detailed level for anyone to look at. It is called the Yellow Pages and the local Printed Map.

Yet to day we get in cars are program in a destination, and blindly follow the instructions given to us, regardless of the outcome. We TRUST the GPS system. Hell, I bet if the GPS system suddenly went out, half the moms in America would not be able the find the soccer fields that they've been taking their kids to for the past years. They would be lost without the trusted system.

We TRUSTED the rating agencies and sellers of securitized debt instruments.

More data would not have changed the outcome, because investors would not have looked at the data, they TRUSTED the analysts and rating agencies. They also listened to the voices that said what they wanted to hear - "Its different this time".

The Next crisis

Just like the are listening to the same voices saying "Its different this time" about the 10% of GDP that the United States is borrowing last year, this year and for years to come. "It will be different this time, we are not Argentina". But it is not different this time.

And yet the data is there for everyone to see. I.O.U.S.A. was frightening before the crisis. It is just plain terrifying now, yet nothing is being done about it. The data is all there. Yes, a taxonomy for public spending to push that information where it belongs, into the public domain, is important and must happen. We will then be able to actually, as individuals, dissect government spending at the lowest level. The give winners will be the opposing political parties who will use the data to prove that one part or the other is spending money wastefully (usually in amounts that appear huge - $10s or even $100s of millions).  These exposures will certainly bring some "discipline" to spending, but will do nothing, absolutely nothing, to avert the coming crisis.

Tagging data at the lowest level is about efficiency and effectiveness, and must happen.And XBRL is the most effective standard for the tagging of business (that's what this is after all) information.

But all the stimulus money, line by line, tagged in XBRL at the lowest level will not change the simple fact that the next crisis is not about the data, it is about the decision making that says "Its different this time". So lets make the case for XBRL on the basis of what it will do, improve efficiency and effectiveness, not avert or limit the impact of the next crisis. That is not a data issue, and is a political issue.

Detailed data does not make political courage.


  1. I consider it an honor to post a comment, ahead of my friend and colleague, Paul Wilkinson. I suspect he is writing while he flies across country from CA to DC for our meetings together here and in Belgium later in the week....

    Your post keeps alive an important discussion on the balance between transparency and opacity. You are quite correct in your assessment that data, simply being available and transparent, is only part of the answer to "it being different next time." Data, that remains simply data, indeed has no ability to change markets or behavior. As Charlie Hoffman and Liz Watson make clear in their book, "XBRL for Dummies," data is at the bottom of the knowledge continuum. Information, knowledge and ultimately wisdom are the vital components of the continuum, and this is where focusing on who interacts with transparent data becomes critical.

    CLOUD, Inc. (www.cloudinc.org) takes the view that the next phase of the evolution of the Internet requires standards for personal information that moves us beyond the silos of web pages and links data to its ultimate connection point, people. One-sided transparency, although vital, is indeed limited, as you point out. The magic of XBRL only happens when transparency is two-sided, and that is the goal of CLOUD, to create wisdom out of data, from business information, wrapped in XBRL, to health information, wrapped in HL7, to education content, wrapped in SCORM or Common Cartridge. The wisdom is in the crowds.

    The wisdom of crowds was seen in the election of 2004, when the blogosphere did a better job 'rating" the selective service evidence uncovered by the "rating agency" of CBS and Dan Rather. The various examples you highlight in the financial sector, from MBSs and CDOs, does an excellent job of pointing out how trusting institutions, rather than markets can lead us astray and away from wisdom. In the case of former President Bush's selective service records, the market indicators (i.e. blogosphere) beat the rating agency (i.e. CBS's editorial board) to wisdom and accuracy.

    Of course, the WHO matters as much as the WHAT. Simply having transparent access to the various detail behind the tranches is one step in the process, ensuring that the market can serve as its own rating agency. Like eBay, with its buyer and seller ratings, the transparency and control advocated by CLOUD will bring what XBRL has brought to business information to people. It will allow investors to know immediately how to compare our current market aggregators and brokers (i.e. Goldman Sachs).

    Equally vital to this discussion is the point you tuck away in the middle about Google Maps and the printed Yellow pages. In both of these cases, the content (data) is exactly the same. The only difference is that in one case the data is delivered on paper, and in the other, it is delivered via silicon. The data does not change based on delivery vehicle. However, keeping information trapped in the physical plane of paper does limits its ultimate richness, just like trapping data in institutions limits its usefulness as the wisdom of the market interacts with it. Ultimately, you can only trust data, if you trust people. Until there is a language for people on the Internet, as robust as XBRL for business information, then trust will indeed be misplaced.

  2. I'd merely point out three things:

    First, in his new IRA newsletter, http://us1.institutionalriskanalytics.com/pub/IRAMain.asp, Chris Whalen says the following:

    "The basic issue that still has not been addressed by Congress and most federal regulators (other than the FDIC with its proposed rule on bank securitizations) is how to fix the markets for OTC derivatives and structured finance vehicles that caused losses to AIG and other investors.

    "Neither prop[rietary] trading nor the size of the largest banks are the causes of the financial crisis. Instead, opaque OTC markets, deliberately deceptive structured financial instruments and a general lack of disclosure are the real problems. Bring the closed, bilateral world of OTC markets into the sunlight of multilateral, public price discovery and require SEC registration for all securitizations, and you start down the path to a practical solution. But don't hold your breath waiting for President Obama or the Congress or former Fed chairmen to start that conversation."

    SEC registration like Chris wisely suggests typically requires disclosure of all material information. So far, the only effective way to do that has been with the XBRL taxonomy described in last year's Wired magazine article -- the taxonomy that hedge funds used to justify short sales related to MBS. If there's a better way to disclose all material information, that's fine. Obviously, the ASCII and HTML posted to EDGAR under Reg. AB didn't work -- or took until Bear Stearns collapsed in March 2008 to work. Next time, exposing work like Wired described to market sunlight might at least help the truth be discovered before things get quite so far out of control.

    Second, there's a detailed 13-page plan to use XBRL to restart securitization. See http://paulwilkinson.com/2009/12/10/xbrl-data-tagging-standards-advance-on-two-fronts/#econtrans . Again, if there's a better idea to provide for genuine market pricing of MBS and other ABS, let's hear it.


  3. [continued]
    Third, the statement that "[d]etailed data does not make political courage" appears sadly accurate. Having spent more than two decades in the political arena, I agree. That's also why I believe that the most effective way to improve markets is to improve the information available to buyers and sellers so that they need not trust. Instead, the market should be able to verify, a process that has generally worked well with public companies that don't seek to avoid scrutiny by hiding their activity from the good old structured disclosure language called U.S. GAAP. For all the pain caused by their inaccessibility, the facts that would need to be disclosed about basic ABS are remarkably less complicated than the facts that are disclosed every day under U.S. GAAP. And once ABS were made transparent, the opportunity to create complex instruments on top of them would diminish as buyers came to understand the futility of building a house of cards on top of future revenue streams. GAAP has helped bring order to public company securities markets and XBRL is working well in those markets. If there's a better way to make ABS markets work well, let's hear it. It is not important HOW what Mr. Whalen suggests happens -- it is important THAT it happen.

    The coming crisis can be best mitigated by economic growth far in excess of the current baseline. The only hope for that may be a much larger scale repeat of the mid- to late-90s, when computer technology helped boost productivity and therefore federal revenue to significantly accelerate the accomplishment of a balanced budget. Even that failed to solve the problem of unfunded liabilities, particularly Social Security and Medicare, but in those areas too, technology remains, in Joel Mokyr's words, the lever of riches. Gary Thompson, whose comment appears above, and I had the pleasure of studying economics under Professor Mokyr at Northwestern, who wrote in 2007:

    "Growth was possible through capital accumulation, increasing trade, better internal allocations, freer markets, and improved institutions. But all of those processes would eventually run into diminishing returns. It is technology that remains at the foundation of modern economic growth."

    Professor Mokyr wrote in the context of the Enlightenment and the Industrial Revolution; XBRL is a modern tool that combines the knowledge and standardization that were the basis of the Enlightenment and the Industrial Revolution. Hope it works.

  4. Gary and Paul,

    Thank you both for your very thoughtful comments. There clearly is a need for greater transparency, and XBRL is the best way to accomplish that. There is also the need for political courage.