荷爾蒙與交易

周末大部分時間都交給了冬奧，但還是看了篇很有趣的文章，原來是否適合當高頻交易員與雄性激素水平有關，但當雄性激素處於高水平時,就得停止交易,否則會出現高風險交易，同時，雄性激素的含量還可以預示的是交易員承受風險的大小，而非技能高低。非常之有趣。 原文如下： Male traders, like animals in the wild, take more risk when their testosterone levels rise. Research by myself and my colleagues found that moderately elevated levels of this hormone increased the profits of high-frequency traders – although at higher levels it can cause overconfidence and risky behaviour, morphing traders into Masters of the Universe.What we could not say, however, was whether testosterone was having its beneficial effects by increasing the trader's skill or merely by increasing his appetite for risk.In a study published today in PLoS ONE we found that testosterone had little to do with trading skill. Traders with higher testosterone did indeed do better at this type of trading, because they took more risk. But there was no link between the hormone and their trading skills, as measured by the Sharpe ratio (of which more later). Testosterone alone was not enough.Showing this requires an understanding of how to judge whether a trader's profits are due to skill or luck – a vital question for banks and hedge funds when allocating capital and paying bonuses.Trading managers commonly equate skill with a trader's profit and loss (P&L). This measure is simple, but also misleading and dangerous. Knowing that a trader made $100m says nothing about the skill involved unless we know how much risk they took. What if that trader could just as easily have lost $500m?A better way of sorting the skilful from the brave is to look at the Sharpe ratio, a measure of the steadiness of returns calculated as the P&L divided by the variability of P&L. The higher the variability (or risk), the lower the Sharpe. A trader making $100m but whose P&L regularly swings back and forth by $500m will have a low Sharpe ratio, around 0.20. A trader making $100m but whose P&L varies less, swinging by $100m, will have a higher ratio, around 1.0.I and my colleague Lionel Page looked at the Sharpe ratios of a group of male high-frequency traders between 2005 and 2007. The experienced traders among them achieved an average Sharpe of 1.02, significantly higher than their benchmark index, Germany's Dax, which averaged 0.53. Furthermore, these traders continued to make money in 2008, a year when many banks and hedge funds gave back five years of returns.We thought testosterone levels might predict the traders' Sharpe ratios. But they did not. Testosterone levels predicted the amount of risk taken, but not skill.How were these traders harnessing a high-testosterone/high-risk style of trading into high Sharpe ratios? We found that a trader's Sharpe ratios increased markedly with the number of years they had traded, with first years having a Sharpe around zero, second years matching the Dax, and the most experienced, those in their 12th year, achieving a ratio close to 2.00. Were the traders getting better or were employers culling low Sharpe traders? We found traders increased their Sharpe ratios significantly during the two years of the study – indicating they were learning to make more money per unit of risk.Learning was encouraged by the compensation scheme at the trading company where they worked. Banks, with their yearly bonuses, may attract traders with an appetite for risk rather than prudence; but the traders in our study received only profit sharing, so if they lose money for the firm they lose it for themselves. These traders have, therefore, a strong incentive to damp the swings in their earnings.As an aside, learning, like outperformance, is incompatible with the efficient markets hypothesis, according to which the markets follow a random walk and you can no more learn to trade them than you can improve at flipping coins. Our data therefore suggest the markets are not in fact random.More usefully, the data also suggest that banks could use an improving Sharpe ratio over time as a measure that indicates a trader has developed a skill worth paying for. A high tolerance for risk – predicted by pre-natal testosterone exposure – is needed, but, like height or speed in sports, may count for little without proper training and management. In trading as in sports, biology needs the guiding hand of experience and the right incentives.