Philip M Halperin - Scribblings
Managing Trading Risk with Trading Signature Analysis
Article Outline Accepted for Forthcoming GARP "Risk Management Practitioner's Hand Book"
In the analysis and management of the risk and performance of traders and trading desks, there are many generally accepted methods of measuring performance. Most of the methods deployed involve the analysis of positions and are essentially static in nature. In the article proposed, the author will examine a method of returns analysis of traders through the graphic interpretation of trading P/L history results, which he devised and has employed in the management of both proprietary trading and market-making traders and desks for over ten years. This methodology, which he calls "Trading Signature" Analysis has proven a useful supplement to the widely accepted techniques of positional analysis (VAR, stress testing, etc.) because it differs from them in that it is more an analysis of trader than of his position, because it is dynamic rather than static, an analysis as much of the P/L path as of the summary performance results.
Trading "Style Variables"
While conventional measures of position and performance are useful, they by no means communicate the entire picture of what type of performance is likely to be displayed by the subject trader or desk. Indeed, our experience in the realm of proprietary trading management in derivatives, particularly commodity derivatives, has led to a quite differently based form of analysis which is complementary in quality and utility to those more widely accepted among the investments community. In the world of proprietary trading, it can be observed that there are several characteristics which define a trader or desk. These variables arise out of the psychological make-up and experiential development of the trader, as well as the objective role within the market, and can be categorised as psychological "style" variables or axes which, taken together, define a trader's natural style habitat:
- Momentum sign. Does the trader buy strength and sell weakness (positive momentum sign), or sell strength and buy weakness (negative momentum sign).
- Holding period. The holding period is another style variable, it implicitly defines the magnitude andtiming of the relative gains in individual positions, and the profitability path, including degree of dispersion of returns for a given time period.
- Risk aversion. While this is professed by many to be a stated goal, in practice the degree of risk aversion, neutrality, or even risk positivity varies not only among individual traders or desk managers, but also within the same individual as a function of portfolio size and relative diversification.
The Trading Signature
Much of the foregoing elements of psychological habitat or trading profile can be discerned from a non-parametric graphic analysis of daily returns, used in conjunction with more conventional statistical measures. The author proposes to demonstrate a proprietary empirically derived methodology in which returns, when decomposed, comprise a trader's "signature" Each individual has a written signature, characteristic of him alone, and no two signatures of a given individual are identical. Yet the signature identifies the individual. Analogously, each competent trader or desk manager has a performance signature which can typically be analysed to provide insight into the trader's psychological performance habitat, and more importantly, to shed light on likely future performance path.
Example (War Story)
The author will provide a graphic example of a trader under his management who was hired and within four months lost 1.1 million dollars in one day. During the decision whether or not to terminate the trader, the basic signature analysis methodology was developed. The author will provide graphs of actual daily P/L returns, and transforms of those returns, for that individual demonstrating
- The wrong conclusion pointed to by simple consideration of cumulative P/L and summary statistics. - The oscillator used as a basic component of the analysis. - The elements of trading style to be teased out of the analysis. (In this case, a classic negative momentum sign with positive correlation of efficiency and risk) - The projection of P/L employed. - Confirmation in the form of continued daily returns analysis covering three years for that individual. - The use of ongoing retuns analysis monitoring to improve the individual's trading efficiency, a combination of non-parametric graphics and summary statistical measures. - The epilogue ... 10 million dollars cumulative P/L with extraordinary efficiency.
The signature analysed in the forgoing real example can be shown similar to another trader in completely different market and explicit technique, yet similar in psychological make-up. Other signatures can be graphically shown:
- Long momentum sign, long holding period negative risk aversion
- Long momentum, short holding period (day trader)
etc., as prototypes
Signature Analysis in Practice
We will discuss the problems with the methodology, its efficacy in practice, and areas of application:
- Trader vetting.
- Efficiency improvement.
- Desk composition.
- VAR measure validation.
The signature modelling methodology for trading management has several things to recommend it:
- A powerful new tool to segment the definition of traders.
- A proven inductive methodology to complement existing criteria.
- An intuitively comprehensible graphic tool to inform the manager-trader relationship.