Perspective

Grand prognostications are always problematic, but the November issue of Scientific American had an article by David H Freedman, A Formula for Economic Calamity, which, inadvertently, highlighted some critical problems with basic analysis. To summarize, the article contends that the 2008 crash was largely brought about by inadequate mathematical models which failed to correctly assess market risks and protect participants.  Specifically, current models are both insufficiently sophisticated and missing relevant data. Considering the source of the article, Scientific American, it's acceptable that there's a “science can fix it" bent to the argument, but the most important lesson to take away from the article is that Freedman, and probably the vast majority of readers, clearly view the financial markets as a natural ecosystem and expect science to one day find the correct model. 


Financial markets are always being compared to “nature”, and the attempt to transpose robust scientific models to the field of financial analysis is a temptation that neither Goldman Sachs or regulators are able to ignore. Fundamental to this quest is an implicit acceptance that the markets are “natural”.   As is pointed out in the article, it may be that the data available to us is insufficient and current models aren't up to the task; this leaves us with the assumption that one day we'll have all the data and a model capable of handling the task.

Critics of financial modeling, black box trading algorithms, and even the dismal science of economics, tend to unite in announcing that the quantity and complexity of the data will always surpass our ability to both gather the relevant data and arrive at robust models capable of accurate predictions. With economies and markets booming and busting, hedge funds imploding, bank bailouts, and a slew of other unpredicted events occurring on a daily basis, the anecdotal evidence obviously seems to support the view of the critics. And, even the most ardent believer, has to accept that we currently live in a world without access to the right formula.

It would be easy to join the critics and conclude that the only defensible perspective is that markets are unpredictable and nothing can be modeled, yet this is simply an excuse to stop thinking and analyzing data.  Also, active participants recognize that there are models which work dependent on “market conditions”.  Concluding that we're having a bull run in a secular bear market turns high P/E ratios into red flags for fundamental analysts. If market conditions change, and the most widely held view is that we're in a bull market, then the same analysts will largely ignore P/E and look at other indicators – perhaps revenue growth and market share become the new favorites. Technical analysts are happy to accept that a chart formation in one type of market is a sell signal, while the same signal in a different market, or perhaps just a different security, can be a buy signal. To market participants it's clear: the nature of markets and securities dictates which models are useful/correct, and it's futile to apply one grand unifying model and attempt to reconcile these differences. As we have witnessed: it's no good to say that book values of 1:1 makes for safe investing; it's no good to say that when volatility increases the markets are about to correct (up or down); and “broken” correlations shout out that markets and asset classes don't have the cozy causative relationships we wish they had.

So, back to the Scientific American article, the belief that the markets are a natural system and that it is worthwhile pursuing a “correct” model is the main problem. Clearly, markets trade based on the actions of its participants, and market participants (whether they are central banks or day traders) act in accordance with their chosen models – so let's adopt a fresh perspective. Let's forget about attempting to develop a master theory that allows us to model all financial markets.  Let's turn our analysis up-side-down and assess which models are in the ascendancy and predict where these models will lead the markets. Let's maintain that the fluctuations of financial markets are not the behavior of natural ecosystems, but, merely, natural outcomes of the dictates of fit (from an evolutionary stand point – remember, we're trying to “borrow” from hard science here) financial models (which exist in an ecosystem of ideas attempting to gain supremacy and followers).  Models rule – markets follow.

An easy example: in a world where generally accepted models tell us Japanese government debt is a safe haven investment, this prediction leads to high demand and the models are “correct”, but in a world where our models tell us that Japan is hopelessly indebted and in risk of default – capital flees, and Japanese debt becomes junk.