ED - Trading Strategy
TRADING APPROACH (SUMMARY)
ED is basically a market neutral, equity long/short strategy. It trades a 100 stock portfolio of the most highly capitalized US companies. Between 50 and 70 out of 100 stocks are bought, with various component weight coefficients, while on the short side there is only one position: the S&P 500 index tracking ETF (SPY - "Spiders"). The strategy is exactly market neutral (long positions value is equal to short positions value), although a long/short ratio above unity can be chosen to follow the prevailing upward market direction. In order to provide zero correlation with the overall market, the long/short ratio is currently set to unity. The portfolio is adjusted daily, although there are versions of the strategy which trade weekly. The strategy can trade any number of stocks (minimum 50, and maximum is not limited although back testing shows there is no benefit from trading more than 100 stocks).
Unlike most trading systems which trade a single stock or a long/short pair, this strategy trades a stock portfolio, which makes it unique and ahead of the competition. Another important feature is that the program trades stocks on the long side only and constantly holds a short a position in SPY ("Spiders"), thus avoiding liquidity issues associated with short selling (some stocks are difficult to borrow).
The strategy uses statistical methodology (time series analysis and statistical forecasting) and not classical fundamental or technical analysis. Basically, it looks for over/under valued stocks, relative to the overall market and attempts to exploit these deviations. There are no classical indicators which attempt to forecast market direction, no pattern recognition techniques, no fundamental analyses and no trading rules based on trader's experience; It trades by numbers and does not forecast anything - it is a "numbers crunching machine" or a quantitative, "Black Box" trading system. Since this strategy is not correlated to swing and trend following strategies, it can be seen as excellent complement to them.
The portfolio composition (position size management) is calculated by a proprietary computerized program. Unlike other pair trading strategies, this program is more sophisticated and seeks for arbitrage opportunities between all of the 100 portfolio stocks. It gives specific position size adjustments for the 100 stocks comprising the portfolio. The resulting positions are then implemented in the market. The trading system is always in the market, but portfolio weights are gradually adjusted on daily basis. Complete portfolio turnover is completed in 7.7 days.
The leverage is set to 1.0, both on the long and short side of the portfolio. Depending on the preferred risk tolerance level, the leverage can be further reduced (or increased), i.e. smaller (or larger) position sizes can be traded then those calculated for the nominal account size. Due to the large number of stocks, trading in small adjustments and commission costs, the recommended minimum account size for the strategy is $2,500,000.
Fig. 1. Stock price chaos: ED portfolio stock prices, adjusted for dividends and splits
PORTFOLIO APPROACH VS. SINGLE STOCK TRADING
Instead of betting on/against any single stock, this trading system analyses a 100 stock portfolio - the Dow Jones Industrial Index universe plus other 70 very liquid NYSE stocks. Stocks of well-established companies ("blue chip stocks") were selected for trading for their relative safety, reliability and excellent liquidity. By trading 100 stocks we achieve the following objectives:
- We diversify trading. By trading 100 stocks simultaneously and in small quantities (maximum single stock exposure is 3%) we cannot make a single major mistake by entering long (or short) at the wrong moment and with the wrong stock. We do not attempt to time or predict the market. Anyone who claims otherwise is deeply wrong.
- We minimize the exposure to the systemic market risk. Since all stocks tend to move following the overall market trend, by trading long/short we minimize the exposure to this aggregate risk.
The ED strategy extracts profits from the price chaos. Apart from the systemic risk, there is also a "idiosyncratic risk", which is specific to individual stocks (Fig. 1). By adjusting portfolio weights (position sizes of individual stocks), we try to keep our 100 stock portfolio relatively insensitive to the systemic risk, while extracting profit from the relative "statistical miss-pricing" of individual stocks. This doesn't imply, however, that this type of trading is riskless. Based on back testing results (Fig. 2) it is remarkable that the ED long/short strategy was profitable in a rising market (1991-1998), declining market (2001-2003 and 2008) and sideways moving market (1999-2001 and 2004-2006).
Fig. 2. Value Added Monthly Index (VAMI). Theoretically, initial $1,000 with profit reinvestment would grow 15,409% to $155,091 in 22 years.
By reinvesting trading profits and compounding monthly, a $1,000 investment would hypothetically grow into $155,091 in 22 years (note the VAMI is shown on a logarithmic scale). That's 15,409% profit on the initial investment, or average compound annualized rate of return of 25.88% (note these figures are before performance fees, but net of commissions). The same strategy without profit reinvesting (by extracting profit and keeping portfolio size at $1,000) would only yield 517.5% during the same period.
Fig. 3. Monthly rates of return (ROR)
PROFITABILITY VS. RISK
Back tested profitability was pretty stable, with average compound annual rate of return of around 26%, net of commissions and bid/ask spreads (0.06% of the trade value). The monthly rate of return (ROR) plot (Fig. 3) clearly shows that most months finished in the positive territory. The profit factor, defined as a sum of positive monthly returns divided by the sum of negative returns is 10.34, i.e. $10.34 were gained per each $1 lost on monthly basis.
The profitability looks to be persistent in long run. Despite the long term profitability, there are occasional draw downs (Fig. 4), which is the ultimate source of risk in this type of trading. To limit the downside risk, the leverage is carefully determined, i.e. the position size relative to Net Asset Value (NAV). The leverage is constant and fixed to 1.0 both on the long and short side.
The Calmar Ratio (average annual ROR to maximum intramonth drawdown) was 1.88 in back testing. The worst drawdown, on month-end basis was (6.21%), while the worst intramonth, peak-to-valley drawdown was (13.77%). Interestingly, the October 1987 market crash did not produce a remarkable drawdown in ED.
Fig. 4. Strategy Drawdowns (on end-of-month basis)
More performance analytics can be found on the ED Back Testing page.