Friday, October 17, 2008

supercrunching into oblivion

OK…. Just to get it out of my system it may be time to give my thoughts on the financial crises as we know them. Best take this all with a grain of salt at the very least and is really just my diversion on a 10 hour plane flight thanks to a new laptop with a battery that no longer goes on siesta after less than an hour.

Any comments I have on the whole financial markets miasma is at the very least above my paygrade. Certainly not anything I have looked at in depth in years although I did graduate courses in international finance and macroeconomics at one point. That only helps a bit in understanding what has been happening. What is scary is how little I really can explain the news despite the fact that I did once work at Lehman Brothers which looks to be one of the epicenters of all this. Note I don’t really believe Lehman had a unique role in all of this, but it was like having a ringside seat at the circus. Little did I know at the time that I was watching the seeds of destruction for capitalism as we know it.

What went wrong? It really is hard hearing all the varying explanations of what happened. Most things being identified as the causes are at best symptoms, most things are more likely collateral damage, and in some cases just other bad events that are all getting thrown into the one big pot whether that makes sense or not . When at Lehman in the early 1990’s I actually worked on the Derivatives Products trading desk when it was a rapidly expanding operation at the time. It was a remarkable time because of the changes taking place. What Lehman and others were working on at the time was once a vast expansion in the universe of financial products being traded in entirely new markets. Nothing wrong with innovation and finance needs to evolve as much as everything else. The question is how far and how fast can financial markets change for there to be some hope for efficient market theory to come into play.

The problem was that new financial products were being developed far faster than they could ever be understood by the markets that would trade them. Without being understood they would inevitably be mispriced by the markets. These new products were in many forms. The media seems to be focused on the exotic world of credit market swaps. I think this is the media story in play only because it borders on explainable in a finite number of column inches. The scary problem is that is it is just one new market that came into being, or at the very least expanded beyond all recognition in recent years. What I remember was mostly the vast world of interest rate derivatives which must dwarf the global market for credit default swaps, although I really have not seen any data on the total size of those markets… I bet nobody knows for sure how big many of these markets are. Without writing a tome, some of which would be beyond my explanatory ability anyway, the basic description is much the same as in credit default swaps. Where one product was essentially taking bets against the future credit worthiness of firms or other specific debt, interest rate derivatives are bets on how interest rates will change in the future. Seems simple enough, except there are virtually infinite number of interest rates out there and they all interact in ways that makes a complete description of the dynamics inconceivable. The various contracts that were being created to 'bet' on the possible changes in interest rates around the world and across time was no less creative than the odd bets you can make on the Superbowl.

So my different view on all of this. (what? You want to hear what CNN has said 1,000 times at this point? ) At the core of the problem is a really fundamental misunderstanding of what role sheer information processing plays in just about everything. There is something of a debate that we have now reached an End of Science because of the power of sheer information technology these days. Basically the idea is there no need for the the traditional scientific method any more. If we can just set the computers to chomping though all possible answers to everything, we will get the correct answers without all that messy theory requiring things like hypotheses and testing. (yes, I really simplify, but it’s a huge debate unto itself). This is exactly what was happening at Lehman when I was there. The folks I worked with included whole banks of PhD physicists and other frighteningly smart folks who some thought were the only people able to work out the hedging that is essential for these products to be bought and sold responsibly. I actually believe their math was probably impeccable and a vast improvement over whatever was done before they worked on these problems. I am sure they created probability tables for things folks never thought of quantifying ever in the past. It just was not the right problems for them to be working on, or at the very least not the only problem to be working on. Not their fault either, the folks in charge were not asking the right questions of them if even they would have been the right folks to ask.

In social sciences there is a deep aversion to ‘data mining’ which is one form of this blunt force application of reverse engineering to what others consider scientific problems. It is certainly true that sheer quantitative power can be applied to work out innumerable problems that before require a lot of analytical work to figure out. The question is whether this is merely a tool to be used to push the knowledge boundary out farther or knowledge unto itself. One problem is that the most interesting questions remain well beyond the fastest computers. I know that may sound like heresy to some, and I point out I have a degree in electrical engineering so I am not a Luddite in any way. Consider how much time, effort and concentrated computer processing power has just barely gotten to the point where machine will beat the best chess players in the world. Evidence of progress for sure, but think how simple chess is as a game. A fixed number of little squares, a smaller number of pieces and a few rules to govern play. The world is exponentially more complex than that. If IBM could even predict IBM’s stock price next week they would close up all other parts of their business and just play the market. Put much more simply, Hari Seldon has not been born yet.

Doubt this IT based explanation? Consider a few things that have happened in the last couple of decades, the period when technology has exploded into the world of finance. The cases of not one but several ‘rogue traders’ that brought immense losses and in a couple cases brought down whole institutions that had previously survived wars, plagues, famines and the other vicissitudes of history. In each of these cases a single trader supposedly brought huge losses upon their employers. Yet, we really only hear about the really large cases that can’t be hidden. How many less well known cases there have been will never be known. The first question you have to ask is just how such things happened and why there were not controls. Such individual traders really could never have leveraged as many transactions as they did without immense IT infrastructure. But in each case you have to wonder why that same information technology was not able to provide greater and more comprehensive oversight. Nick Leeson in Singapore actually brought down the multi-century extant Barrings Bank. Nobody above him noticed that his positions were so badly hedged?? A failure in process? Or something much more systemic.

The more recent French case of Jérôme Kerviel is even more puzzling. In his case he was more a computer person and no matter what a very junior operator more credentialed just to be doing the paperwork for the trades he was executing. That he was able to (supposedly I still wonder) engineer $billions in losses is really a mystery. A case that I still ponder happened again when I was working on Wall Street. Joe Jett who was a 36 year old trader at Kidder Peabody whose trading incurred only a $350 million dollar (back when that was real money) loss. Unlike Leeson who was put in jail for fraud in Singapore, or even Kerviel in France who is at least still under investigation, Jett never suffered any criminal penalty, suffering only a minor SEC hit for bad accounting that he probably shouldn't’t have been responsible for anyway. He claimed then and to this day not only that he was innocent of any indiscretions , that he himself thought he was making the profits that didn't exist. I am not quite sure I disbelive him either Even if the profits were mythical, the bills he created were not and Kidder Peabody would disappear due to the losses he incurred. You have to go consider again why there was not enough information technology in place to make sure more financial controls were in place in each of these situations. But in each case there was no inkling until it was too late.

In these cases of rogue traders you see evidence that information technology actually created complexity exponentially faster that it could manage. All of this would play out in a more systemic failure when the Minotaur like creation known as Long Term Capital Management would come about a few years. LTCM was but a single hedge fund that exemplified the use of this technological hubris to think they were able to make money exploiting minute like inefficiencies across financial markets that supposedly were never exploited in the past. Turns out there was a reason for that.

The problem was that the inefficiencies LTCM thought it was exploiting did not really exist, or at least paled in comparison to the inefficiencies in their own understanding of the markets they were trading in. They engineered such a highly leveraged and risky position that when it all came crumbling down it struck fear in the heart of the ever-implacable Alan Greenspan. The Fed would take an unprecedented and direct part in unwinding the positions of LTCM because the dollar amounts were so large and the risks to the entire banking system were great. What is playing out now is just the LTCM story multiplied out many many times across the world. We can place blame in lots of policies, individuals and institutions for creating the system that put us in this situation. Clearly lots of bad choices were made, but bad decisions are made all the time. In fact bad decisions are part of the learning curve that makes efficient markets just so. Why there was no attempt to fix the problems that caused LTCM is the question. Never has been there such a clear canary warning of the failure that was to come as there was with LTCM. And never has that canary been so completely ignored. LTCM made clear how bad the situation was and why there was not a wave of regulatory reform after it folded in 2000 may be the single reason we are in the situation we are in now. LTCM was once called the greatest private decimation of capital in Wall Street history at the time. If that was remotely true, how would you describe the last month?

What does it mean and where do we go from here? That goes way beyond my paygrade. I am sure there will come a realization that the world has not ended and most of what we call the economy is continuing to function as it must. I’ve spoken with some smart folks of late who are convinced the new depression is here. It will not be a sudden realization, and some uncertainty will be a drag for a long time, but compared to any similar crisis in the past the shortness of pure denial is remarkable. Remember that from the Wall Street crash of 1929 it would be years before the New Deal was really implemented. What is happening is not all that different from where Japan found itself in the early 1990’s and they chose mostly to pretend there was no problem. Thus they didn’t have the cataclysmic spasms we are experiencing now, but they suffered a decade of recession. The level of fear prompting such quick action is itself a defining action that could not have been factored into all of those probabilistic models that got us into this situation. The human factor is yet to be quantified.

Lest anyone think I am being terribly optimistic, I don’t think we are through the worst yet. Most of what I see as being fixed by all the world’s financial policymakers have been addressing symptoms and not causes. There are a number of huge markets that have to unwind because the sheer amount of counterparty risk is just not known at the moment. I hear there are whole markets (markets may be a generous term in a lot of cases) where price discovery is nonexistant at the moment. They say credit default swaps are a $60+ trillion market, but I really don’t think that the biggest problem. Interest rate swaps collectively must dwarf that number let alone the highly leveraged loan markets overall are real concerns that have not been addressed by the $trillions deployed by the world’s centrals banks as yet. There are innumerable other markets out there for everything from arbitrage on the future of the Icelandic Krona-Swiss Franc exchange rate out to even uranium futures that have huge unquantified losses for many counterparties. Many of the mostly unregulated markets are virtually chaotic at the moment. Until they unwind enough to be priced by the market they will act as huge anchors on the world’s collective money supply. If there is no believable price discovery going on then people don't know what their assets are worth. Until everyone gets a grip on their losses it remains a dangerous situation for all. All the policymakers have been doing is basically trying to buy time for these markets to come to some equilibrium. So we may not observe the final phase of the crisis as these markets deleverage, but when that finally happens we will be at a point to start over again.

If you really read most of that… I owe you a beer.


Blogger Unknown said...

Great article.
I'll assume that the models accounted for a housing downturn. I'm sure the "brains" warned of this risk. The "folks in charge," as you call them, who were not asking the right questions have committed criminal negligence.
Also, I'm very unimpressed (as is the head of the FDIC) with the fact that (essentially) nothing is being done to help homeowners who are underwater or unable to make payments on their exploding arm loans. (Let's not be so naive as to think they understood what they were signing at the closing table.) This was the source of the problem and until the government addresses this source, the "hits" will just keep coming.
Thanks, Lawton Stokes

Friday, October 17, 2008 8:13:00 AM  
Blogger The Urbanophile said...

There was a recent book that made a similar argument called "A Demon of Our Own Design". Part of the problem is that these strategies are not proprietary, and the models do not take account of the fact that a large array of market participants will mimic the same strategy. The best case in point is probably the portfolio insurance wave of the 80's that triggered the Crash of '87.

Friday, October 17, 2008 9:04:00 AM  

Post a Comment

<< Home