Saturday, April 10, 2010

Cent of Fear

I was intrigued by the recent press around Renaissance Technologies, the ambitious quantitative trading fund run by the inscrutable Jim Simons. While numerous heuristic (meaning based on a series of empirical rules) quant platforms operate in the market, few have achieved the legendary status of Renaissance. Operating with an opacity characteristic of the engineers who create them, these funds buy and sell not based on performance of enterprises but on the trading behavior of the markets per se. Not surprisingly, Renaissance and some of its aspiring ilk, are well endowed with physicists, mathematicians and engineers with backgrounds in linguistics, encryption, intelligence and related unstructured data fields. Those who are interested in learning more about the man behind the legend will enjoy reading Jim’s work on topological quantum field theory.

Peter Brown and Robert Mercer, the two PhDs who have stepped in as co-CEOs of Renaissance have an interesting background. Both contributed to IBM’s natural language research effort before being spirited away to join Simons. Much of their work was on modeling the “next effect” in words – trying to rationalize what the next word in a communication sequence would be based on the pattern preceding the predicted. Simply put, their work on n-gram language models assumes that the words you’re reading are not random in their inclusion or order. Rather, they are perplexingly, predictable. They, perplexingly rather, are predictable. Predictable, they rather perplexingly, are. You may be missing my point. You see, language is an encoding process, an encryption if you will, which loads meaning not merely in words but also in word-order and context. Before Drs. Brown and Mercer could crack the real nut – how to handle metaphoric expression – they followed the sirens to a quant fund and started making money. And here’s why I find the three people’s stories interesting. Because their performance actually unveils a fundamental human limit – the “cent” of fear.

From Babylon in the 18th BCE, to 7th century CE Roman “benevolent societies” which offered to pay death benefits to “insured”, to today’s most advanced insurance enterprises an irony emerges from looking at Renaissance performance and what seems to be an improbable industry comparable. At their peak performance, they hit very similar limits of blended returns, between 35-40%. Why is it that death and automated trading both capitate their performance at 40%? I think that the answer is that we have a human constant that I would call the “Fear Premium”. Now some of you might prefer that I call it uncertainty but I don’t for a simple reason. The motivation that makes Renaissance money is the “n-gram” of the next predicted event. They have modeled a human response that they train machines to detect and trade accordingly. Insurers constantly push the margin on how much the market will pay for the perceived “control” of fundamentally uncontrollable events (like death). In both cases, the uncertainty is skewed and the skewness is around the fear of loss.

Ironically, as I’ve begun to look at this dynamic, I have come to observe the Fear Premium in numerous historical and present capital models. Over the coming weeks, I will be highlighting several of these. One of my favorite which must be mentioned is the Basel II reserve capital spread between “investment grade” credits and “junk” credits where the Fear Premium charged to banks for reserve capital is, you guessed, the same number.

At the core, what does this mean? I believe that Renaissance is not necessarily anything more than a canary in a coal mine. Equity and bond traders are letting a few people (and the number of beneficiaries appears to be shrinking as Brown and Mercer look to close the doors to more outside capital) take their money by behaving predictably. If equity and bond traders stepped away from their addiction to impulsive trading and actually began investing based on, God-forbid, actual analysis of corporations or bond issuers, the helium that is maintaining the illusion would dissipate and the game would be over. If we didn’t live in an era where dying in debt continues to justify the dizzying levels of life insurance that we buy, making sure your death doesn’t cost your loved ones would lose its cache and the Fear Premium would be exsanguinated. In short, being free from fear would put at least another 40% of money in your pocket. It would let you spend 40% more time with your family and friends. It would let you relieve 40% of your stress.

As evolved as we pretend to be, it is ironic that in our present age, we still are gazelles on the savannah living with the certainty that there must be a lion out there somewhere. As evidenced by the recent nonsense that was marketed as “health care” debate in the U.S. (a harbinger of the love fest just around the corner in the U.K. and Australia), we had conservative Christians and political activists up in arms about – God forbid – the government caring for those less fortunate. I can’t wait for the first such activist group to waive their government authorized non-profit status, operate a homeless clinic in honor of their spiritual patriarch, and then talk about the irrelevance of government. You know why they won’t do that? Guess what the taxation rate would be on an activist group that didn’t have non-profit status… You guessed it, the Fear Premium.

No comments:

Post a Comment

Thank you for your comment. I look forward to considering this in the expanding dialogue. Dave