September 20, 2008
Complexity and Disclosure Regulation
Posted by Jeff Lipshaw
The only upside for me, I suppose, of the financial turmoil this week is that I'm teaching securities regulation so there is something exciting and topical to discuss. Reading Joe Nocera in the New York Times this morning and thinking about a timely piece of scholarship just posted on SSRN, it occurs to me there is some insight into the regulation issue here, even though I have no clue how to translate it into something that works as a practical matter.
The introductory portion of securities regulation (at least in the Choi and Pritchard casebook) deals with the appropriate form of regulation in markets where information is highly asymmetric, and the feedback loops are relatively long. (The counter-example is buying and selling salt water taffy that tastes like cod liver oil - you may not have known it would be disgusting when you bought it, but it doesn't take long to figure out that something is amiss.) What we discuss at the outset is the difference between merit regulation and disclosure regulation; in short, the 1933 Act, by and large, does not regulate the substance and merit of the securities offering, but rests on the assumption that markets will work more efficiently if material information about the issuer is disclosed.
I have long been skeptical of rational and cognitive solutions for what seem to me to be irrational and non-cognitive problems. This was my criticism of a lawyerly proposal in which venture capitalists would be required to disclose risks to entrepreneurs - since they don't even see risk in the same way, how is this going to work? Nor am I convinced that the strong or even the weak efficient capital market hypothesis holds (in terms of all or some information being efficiently worked into share prices). It also why I'm skeptical of Elizabeth Warren's proposal for a Financial Product Safety Commission on consumer credit (at least as one that has only disclosure powers, and not merit powers). But there's little doubt in my mind that at some level of widely understood kinds of information, the disclosure system works pretty well. That is, if you look at the way analysts cover companies in industrial or service industries, it doesn't take too much extraordinary brainpower to replicate the financial models and do a pretty good job of anticipating who is going to succeed and who is going to fail. That's not to say that the ordinary retail investor understands, but the institutional buyers do, and that's what drives the market.
On Wednesday, a student asked, "so what will the regulatory solution be?" (Or something like that.) I hadn't thought it through at the time, but I was pretty sure we didn't want merit regulation (e.g., the current ban on short-sales), but I wasn't sure how disclosure regulation was going to help. Of course, disclosure regulation morphs into merit regulation when it becomes apparent from the disclosure that nobody should be investing in this scam!
What if you have companies issuing securities not in pretty well understood industrial or service businesses, but in businesses that require the equivalent of understanding quantum string theory or decoding the human genome? Yes, that's great, there's disclosure, but either nobody understands it, or so few people do that truly robust and efficient markets in the securities never arises. Here, as quoted by Joe Nocera, is Henry Paulson, on the mortgage-backed securities that facilitated irresponsible borrowing and lending: "The inability to determine their worth has fostered uncertainty about mortgage assets and even about the financial condition of the institutions that own them." In short, even the experts have insufficient information.
Steven Schwarcz (Duke) has posted a number of timely and insightful pieces on this, including a Utah Law Review article on the failure of disclosure regulation in the subprime crisis. Most recently, however, he suggests in Complexity as a Catalyst of Market Failure: A Law and Engineering Inquiry that there is really a double-barreled effect here: the securities themselves are too complex to understand, and the system itself under which the securities are traded, or the contracts are signed is complex and fragile in an engineering sense.
I suggested to Steve that the thesis seems intuitively correct. Most deal lawyers know that the more complex the deal, the more difficult to get it to close. Moreover, I have experience in putting together several highly complex joint ventures between large and sophisticated multi-national companies. When you cannot reduce the structure of the business to something even the executives can understand, they will figure out something they do understand, and work to that goal, whether or not it was the one originally intended. For example, assume that for tax and public reporting purposes (i.e., what you consolidate and what you don't), you structure a joint venture reciprocally so that the European parent owns two-thirds of the equity of the joint venture entities in Europe and the U.S. parent owns two-thirds of the equity of the joint venture entities in North America. But the real opportunities are in Asia. Do the participants in the venture focus on the most efficient plan to exploit the Asian market for the overall benefit of the venture? Or do they look at the 33% spread between gaining the business in one JV entity over the other, and compete with themselves rather than for the ultimate success of the business? (I'm not making this up.)
My point is that even relatively simple business structures can promote unintended consequences because people either don't understand the dynamics, ignore them, or talk themselves into believing that they are manageable. Multiply this by an entire industry of credit instruments, and you have a market full of unintended consequences in which millions of people trade, presumably relying on efficient capital markets to protect them, when in fact there can't be an efficient capital market because too few people actually understand what's going on.
I have no idea how to draw the line I've just suggested. I have shelves full of closing books with relatively complex leveraged buyout financing or conduit municipal bond financing that most people don't understand, but which do not pose much of a threat to the financial system as we know it. The deal struck between AIG and the Fed doesn't seem all that complex to me, but that hasn't stopped much of the relatively sophisticated world from thinking that the Fed now owns 80% of AIG. I don't understand the complexities of wheeling electrical power, or how the telecommunication system allows me to be posting this note, or even how the guts of my computer do what they do. And I'm no worse off. It strikes me, however, that the problem is that those "masters of the universe" punching the computer keyboard of the financial system think they understand what's going on inside, and they don't.
TrackBack URL for this entry:
Listed below are links to weblogs that reference Complexity and Disclosure Regulation:
Jeff, I entirely agree with your ending remarks about deals being too complex for people to really understand all their ramifications, despiute their illusions to the contrary (without prejudice as to what "complexity" means in this regard; there are many definitions out there). And I also share your skepticism about the efficient markets hypothesis. But I am stumped as to what a "non-cognitive problem" is, in this context. Could you please explain?
Posted by: A.J. Sutter | Sep 22, 2008 6:28:36 AM
Heartbreak is a non-cognitive problem, it seems to me. Panic, it seems to me, is a non-cognitive problem. Likewise, over-optimism or addiction. That is, they all seem to be mind-related, but not about how we receive information or learn or intuit. When I say "cognitive" I mean that which relates to the receiving or the processing of information.
So I'm skeptical about "teaching" or "information" solutions to problems that seem to me fundamentally not to be about learning or knowledge.
Posted by: Jeff Lipshaw | Sep 23, 2008 3:48:14 AM