You can’t really get a handle on what to do about the current financial crisis and how bad it might be without at least beginning to figure out what has actually happened and why. I’ve spent a good deal of time studying this and thinking about it, and I’ll offer some observations about the what and why of how things got to the state they’re in.
I think one of the main root causes of what has happened is the impact of advances of information processing technology on financial systems and institutions. Twenty years ago, the kinds of financial instruments that are at the core of the current crisis and the institutions that developed around those instruments simply wouldn’t have been possible. The kind of large-scale creation and trading of what I call the “layering and repackaging” of mortgages requires a huge amount of information processing, information processing that would have been inconceivable just a couple of decades ago.
The basic LEGAL process involved in creating mortgage-backed securities isn’t new. The tools involved in loan syndication are very old, almost as old as modern commerce itself – say, close to 300 years. What’s new is the quantitative complexity of the syndication involved in creating the sort of mortgage-backed securities that have ended up causing the current crisis. The amount of financially-relevant information incorporated into each step of the “layering and repackaging” involved in creating the securities that make up the bad sections of the balance sheets that are the targets of the proposed bail-out is staggering.
One of the things that is said in the news reports on the subject from time to time, but that I think doesn’t really get enough attention, is the fact that, at this point, it really isn’t possible to VALUE these securities. One of the main reasons for this is that there are so many layers of combination and recombination of financial information in the “stack” of syndications that ultimately leads to the problematic final securities.
So, what happened? As I see it, one of – if not THE – key causes of the problem is that risk assessment did not keep pace with the “complexification” of mortgage-backed securities. The traditional one-on-one risk assessment involved in secured lending is relatively simple and is even older than modern commerce – I’m sure there was some kind of pawn shop in Sumer or Ur, and it wasn’t long before the pawn shop owner hired a big guy, and made his business more attractive by allowing the borrower to hang on to the collateral until the loan was repaid, or until the big guy had to go grab the collateral. I know that documentary indentures of loan security predate the modern age – for modest sums you can buy nicely printed medieval collateral indentures to frame for your law office.
With residential real estate, lending risk assessment – “mortgage underwriting,” as it’s called – is also a well-understood exercise: The lender engages in an analysis of the value of the real estate and the probability that the borrower will be able to and will in fact repay the loan, builds in some margin of required equity to cover that risk and, if the numbers match – bingo, you have a real estate mortgage.
Now, herein enters a second factor – “the bubble.” With rising real estate prices and decreasing interest rates, the “greater fool” factor began to operate in the period after about 2000. Of course, speculative bubbles are nothing new in capitalist systems; as someone above pointed out, they’re as old as the tulip bubble.
The problem is that it became far, far easier to “lay off” mortgage loans with the advent of “hyper-syndication” made possible with the amplified information processing capabilities of the last 20 years or so. Before then, there were simple practical limits to how many individual mortgages could be “packaged” to back a security imposed by such mundane things as how fast a person could type or add columns of figures or collate information.
I remember when I first started practicing law 21 years ago that “residential mortgage-backed securities” were a NEW THING. I remember the “aha” expressions on the banking and real estate lawyers’ faces when they “got” what was possible. In those days, only the really biggest players could even consider experimenting with this new financial tool, because they were the only ones that had access to the information processing capacity and institutionalized manpower necessary to marshal the information required to put together the bond indentures that were the legal mechanism upon which these new-fangled securities were based.
But little by little, the technology got more and more powerful and then easier and easier to use. The institution of the “mortgage broker” began to change – it went from a little-known, behind-the-scenes, labor-intensive specialty business, to one that was open to small entrepreneurs. Smart real estate brokers and small-time bankers began to at first dabble in mortgage loan syndication, then jumped in with both feet: It was a business that required little or no capital, and, as time went on, became wildly profitable, especially as the natural post-2000 real estate bubble began to inflate. There were hundreds of these kinds of operations going full-blast in the US, creating the raw material that was feeding the input side of the mortgage-backed securities market.
The problem is that, at each step of the process of syndication, the basic mortgage underwriting judgments became more and more diluted. And with the ease of selling individual mortgages into the syndication market, loan originators had less and less of an incentive to engage in sound underwriting in the first place: Why bother, when you can ship the risk off of your balance sheet in days, at most, and probably hours or even minutes by the time things began to come to a grinding halt a year or so ago?
So the basic discipline of the consequences of bad underwriting was decoupled from the process of making loans. The highly “granulated” nature of the underlying collateral made any kind of regulation nearly impossible: The ratio of individual loans to the size of the “tranches” of mortgages backing specific bonds was huge; so auditing the process once it was underway was functionally impossible. This is what is being talked about when you see reporters saying that valuing the currently problematic securities is impossible – you’d have to unwind the syndication process and get back to all the individual underwriting. This is theoretically possible, but would involve staggering amounts of tedious effort to do. On a cost basis, it’s more efficient to just wait and see how the underlying residential home loans actually perform. But runs on the financial institutions that are holding the highly syndicated securities as assets take place on a time-scale of hours and days, while the home loans perform on a time scale of months and years.
I’m out of time to write this morning, but I’ve been meaning to set these ideas down in writing for my own purposes, and I thought I’d share them here. I hope to return to discuss how the above relates to what ought to be done to address the problems later this coming weekend.
