To best outline a comprehensive price risk management program, there is the Risk Management Pyramid, shown in Figure 1. As you can see, the pyramid contains three tiers.
Figure 1: Weissman's Risk Management Pyramid
Base of Pyramid
Tools at the base of the risk management pyramid are simple, robust, purely quantitative, and universally accepted throughout the speculative trading community.
The simplest and one of the most essential ingredients in the development of a robust risk management methodology is placement of stop losses. A stop loss is the cornerstone upon which all more complex risk tools are built. Why are stop losses so essential to successful risk management trading programs? Because a stop loss becomes a market order to exit once its price level has been triggered. Stops force traders to quantify risk before entry and therefore habituate us to their placement instantaneously following entry. Placing stops immediately after entry means that risk management maintains its objective, rule-based criteria as opposed to being placed after the onslaught of the greed and fear that typically characterize our emotional response to open positions in the markets. The stop order cannot rationalize or debate. It does not understand supply, demand, weather patterns, or geopolitical anomalies. It only knows that our predetermined criterion for trade exit has been triggered and therefore forces that exit despite any reason for abandonment of discipline.
Rookie traders become optimistic when studying price histories. They look at lows toward the chart's lower right-hand corner, then at highs toward the upper left-hand corner and imagine untold wealth in simply buying those lows and selling the highs. They tend to assume away all the price action in between. Unfortunately, as illustrated in Figure 2, it is not enough to have bought the 10-year U.S. Treasury note futures at 120-18 on May 25, 2010, even though they traded at 126-28 on August 25, 2010. Instead, after buying on May 25, 2010, at 120-18, we have to immediately manage the risk by placing a protective sell stop order. In other words, despite correctly assessing the market's overall bullish trend, it is quite possible that our risk management stop would have triggered a loss as the market dropped to its cycle low of 118-26 on June 3, 2010 (see Figure 2). Bottom line, it is not enough that our model makes money in general; it has to be robust enough to make money even when coupled with a stop loss order.
Figure 2: Daily Chart of September 2010 CME Group 10-Year U.S. Treasury Note Futures
Source: CQG, Inc. © 2010. All rights reserved worldwide.
Middle of Pyramid
Quantitative tools in the middle level of the risk management pyramid offer robust solutions to issues, including correlations between assets held in a portfolio as well as the volatilities of those assets.
A more robust answer regarding stop loss placement is that our stop levels should be attuned to the current volatility of the asset traded. In other words, in higher volatility environments, we will need to place our stops further from our entry price so we can avoid being needlessly stopped out of trades that would eventually result in profit, while in lower volatility markets we can place our stop levels much closer to entry without getting stopped out on false countermoves. This relationship between volatility and stop level placement is the reason we never look at stop losses in a vacuum but instead examine them in conjunction with volumetric position sizing. In other words, when the volatility of the asset is higher, we place our stop further from the entry price level but we could potentially trade fewer contracts, whereas when the volatility is lower, we place the stop closer to our entry price and could therefore potentially trade a larger number of contracts without violating rules of prudent risk management.
This relationship between stop loss placement level, volumetric position sizing, and the volatility of the asset—or assets—traded transitions us to the middle tier of our risk management pyramid and specifically to Value-at-Risk, or VaR. VaR adds two indispensable elements to our risk management models: volatility and correlations. VaR examines the historical volatility of assets held in a trading portfolio and the correlations between those assets so as to make our stop loss placement and volumetric position sizing more robust.
Apex of Pyramid
By definition, the term management discretion suggests tools that defy purely quantitative mathematical modeling. It is consequently virtually impossible to provide an exhaustive list of all the possible ways in which management discretion can supplement a quantitative risk management model. Instead, let me outline a scenario in which management experience and discretion could be used to complement such quantitative risk models. On September 11, 2001, acts of terrorism are shifting markets to heightened levels of panic. A hedge fund's risk manager checks portfolio exposures against VaR limits, even runs a stress test to determine if the fund's trading book is enduring excessive levels of risk. Despite the fact that all her quantitative models suggest exposure is within normal tolerances, she calls the fund's head trader, suggesting a reduction of portfolio exposures.
Another example of management discretion is especially instructive as it simultaneously illustrates how manager experience can be used to augment quantitative risk tools of our pyramid's lower rungs while highlighting instances in which we might ignore entry signals generated by mechanical trading models. As an example, see Figure 3.
Figure 3: Daily Chart of September 2010 CME Group Wheat Showing Extraordinary Levels of Volatility
Source: CQG, Inc. © 2010. All rights reserved worldwide.
On August 5, 2010, wheat futures closed locked limit up. The following day, August 6, 2010, it traded up almost the 60-cent daily limit, only to turn around and settle locked limit down on the day. The following trading day, August 9, 2010, saw some good follow-through selling in the market, which resulted in the triggering of a sell signal for one of my countertrend trading models. Despite the fact that a trader could have sold September wheat futures without violating volumetric position-sizing limits (or any other purely quantitative risk criteria), the trader used discretionary risk management as an overlay of those purely quantitative tools and chose to ignore the sell signal for wheat generated by the mechanical trading system.