Ad-hoc decision-making and trading—they go together like acne and prom night. If you've read our post outlining the damage behavioral biases can inflict on a portfolio or our post describing the success of the Turtle Traders, then you know that the best approaches to trading are systematic ones. A bad day at work, FOMO on the latest stock craze, or hand-wavy buy and sell decisions, like inopportune blackheads, can all mar an otherwise healthy portfolio profile.
The benefit of systematic, quantitative trading is to purge these interferants and standardize your decisions. By also simplifying decision-making, systematic systems leave you more time to engage in other valuable activities. Of course, many enjoy learning new financial concepts or picking stocks (we certainly do!), but the ultimate goal of trading and investing is to make money—not to be an end in itself.
So how can you construct such a strategy?
That's what we'll examine today. In line with our recent book-based articles, in this post we'll explain how to implement the systematic momentum strategy Andreas Clenow details in his book "Stocks on the Move." (Note that this article only summarizes the main strategy he presents, so if you would like access to all of Clenow's practical insights, we highly recommend that you purchase his book.)
TL;DR
We show you how to implement a complete momentum strategy for stocks based on Clenow's book "Stocks on the Move." To guide you, we highlight several rules, explaining when to buy, when to sell, and how to size your positions.
The Strategy
As with any journey, it helps to begin by looking at the end. Below is a flow chart Clenow created, illustrating the complete trading strategy.
Before we break this down step by step, it may help to review some vocabulary.
Clenow makes a point to distinguish between trend following and momentum strategies. Trend following, an approach with a proven track-record in the world of managed futures, involves going long on assets whose prices are moving up and shorting those whose prices are moving down. But what works for futures markets, which include asset classes as diverse as currencies, equities, commodities, etc., doesn't necessarily work for stocks. One primary difference is short-selling. Short-selling stocks is much less useful than it is when dealing with futures. Stocks tend to move as a group, and when the weather is fair, the seas are calm, but when it rains, the seas are volatile. That extra volatility means you can expect drops, but also sudden increases in price. In addition, the volatility may make it difficult to borrow the very stocks you're looking to short-sell. Bottom-line: short-selling stocks is not a profitable, long-term systematic trading strategy.
Subtract the short-side of a trend following strategy and you're left with a momentum strategy, a bet that stocks whose prices are rising will continue to rise—basically, trend following without the short-side. Price momentum is a well-documented phenomenon in the stock market, and there are many theories as to why it exists, many of them plausible. But regardless of the causal mechanism, it works.
Ok, enough background, let's dive into the actual strategy.
Step 1) Pick one day a week to trade.
The Rule: Always trade on Wednesdays (or whatever day you choose)
You don't want to spend unnecessary amounts of time fretting over the latest market news. The goal is passive income. And you certainly don't want your hard-earned money eaten away by trading fees. So, pick one day a week (it doesn't matter which) and stick with it.
Step 2) Rank your stocks.
You only have so much cash on hand, and it's important to get as much bang for your buck as possible by ensuring that the stocks selected demonstrate consistent upward momentum in price.
How do you do that?
There are several ways, but Clenow recommends taking the slope of an exponential regression and multiplying it by R^2, the coefficient of determination.
The Rule: Rank stocks by taking the slope of a 90-day exponential regression and multiplying it by R^2, the coefficient of determination.
There are several benefits to this approach. Like linear regression, exponential regression identifies the line of best fit for a given data set. We care about the slope of that line, or the expected change in price (y-axis) over a specified amount of time (x-axis). What distinguishes exponential from linear regression is the treatment of units. For a linear regression line, the slope would be the expected change in dollars of a stock over time. The problem with this is that a $10 upswing on a penny stock is much more dramatic than a $10 increase in the price of Tesla. We want our momentum indicator (i.e., the regression slope) to reflect that. The slope of an exponential regression, by contrast, is unit-less: it equals the expected percent change in price over a given time period. That is, the price change is scaled by the stock's starting price to reflect the speed at which the price is increasing.
While informative, the slope by itself still harbors one flaw. Between two stocks with positive momentum, we prefer to invest in the one with the least volatility, the one whose growth is consistent. Looked at another way, we don't want to invest in a stock whose price has suddenly spiked either from some external shock (like an acquisition announcement) or as the result of the stock's intrinsic volatility. The coefficient of determination, or R^2, reflects the volatility of the data around the regression line. A higher R^2 means that the data fits the regression line more closely, and thus the growth is more consistent. How can we account for both positive momentum (as embodied in a higher regression slope) and low volatility (as represented by a higher R^2 value)? One simple way is to multiply both together, and voila! We have a score with which we can rank stocks.
Step 3) Sell lagging stocks.
As part of our last step, we discussed the criteria for admitting a stock into your portfolio. But it's just as important to know when to kick stocks out. Like a broken-down car in a drive thru, a stock that's no longer performing imposes an opportunity cost. Not only does the stock drag down your returns on its own, but it also takes up space, preventing you from investing in other, more profitable stocks. On the other hand, you don't want to prematurely sell a stock and incur transaction costs just because it no longer sits at the top of the heap. You have to strike a balance.
The Rule: Sell any stock that is no longer in the top 20% of your ranking or that is trading below its 100-day moving average.
The first signal—to sell a stock if it no longer occupies the top 20% of your stocks—ensures that a lackluster ticker can't monopolize valuable space in your portfolio.
(As an aside, Clenow uses the S&P 500 for his stock universe. Therefore, the top 20% is equivalent to the top 100 stocks. If you rank a smaller or larger set of stocks, you may wish to adjust your cutoff accordingly.)
But what if the market falls? It's plausible, given the high correlation among equities, that a stock's price could plummet along with the rest of the market but still retain its top 20% ranking. In this case, you would be left holding a basket of losing stocks. To prevent this, and since you only want to hold stocks in an uptrend, we include the second sell signal: sell any stock trading below its 100-day moving average.
This protects your portfolio from both the explicit costs of a losing stock and the opportunity cost of a mediocre one.
Steps 4 and 5) Rebalance your portfolio
Now you understand what to buy and what to sell, but it's important not to neglect an often overlooked area of portfolio management: position sizing.
Many traders misunderstand position sizing as the allocation of cash. Such traders think they have equitably distributed risk by allocating an equal percentage of their cash to each stock. For instance, a trader with $3,000 looking to buy a basket of 30 stocks might think it wise to buy $100 of each stock. However, this fails to account for different volatilities among stocks. In fact, the volatility of the portfolio will be determined by the volatility of the most tempestuous stocks, resulting in an unbalanced portfolio.
(As an example, consider three stocks: two "safe" stocks that for certain will move up by 10% and one "risky" stock that has a 50-50 chance of going up 20% or going down 10%. You can see that if you allocate an equal amount to the three stocks, then the portfolio's volatility—its risk—will be entirely determined by the "risky" stock.)
Instead, we should allocate not by cash, but by risk. Although fancier tools exist, Clenow recommends the following formula to calculate the target number of shares to hold for a given stock.
The Rule: Determine target position size using the current Account Value, a Risk Factor of 0.1% , and the 20-day ATR.
The Risk Factor is an arbitrary value, representing the proportional daily impact of the stock on the total Account Value. For instance, Clenow uses 0.001 as his Risk Factor, meaning that a change in price for a given stock should only impact the total Account Value by a factor of 0.001, or 0.1%—assuming of course that the ATR stays constant (some traders use different names for this, in our Strategy Builder, we call this our risk fraction).
Thus, a lower risk factor translates into fewer shares of any stock purchased, allowing you to buy shares of a larger number of stocks. This, of course, increases diversification, although Clenow is quick to acknowledge the limited marginal returns to diversification attributable to holding more than 30 stocks.
The ATR, or average true range, is a measure of the movement up or down by a stock on a given day. The true range for a given day is the maximum of high minus low and the stock's movement since the previous day:
The ATR is then simply an average of the daily true range values over a given time period (Clenow uses 20 days). For a deep dive into ATR, see our previous article here.
Because both the ATR and Account Value are likely to change over time—the ATR may change as the stock's price movement changes and the Account Value will change due to price movements from any of the stocks in the portfolio—the number of shares needed to equitably distribute risk will also change over time. In other words, your portfolio will become unbalanced as stocks move up and down and volatilities change. To maintain your risk level, you must therefore regularly rebalance your portfolio. Specifically, Clenow recommends rebalancing bi-weekly to avoid unnecessary transaction fees, and even then, to only rebalance if your stock's position size significantly differs from your target (it doesn't have to be an exact significance level; Clenow doesn't even specify a level).
Although this isn't included in the main trading algorithm outlined above, it may be helpful to automatically rebalance if a stock makes an extraordinary move in price.
Steps 6, 7, and 8: Buy Stocks
You're nearly there, and it's finally time to buy stocks.
Well, almost...
There are a couple of additional checks to perform before purchase.
First, the index should be in a bull market, as indicated by the S&P 500's 200-day moving average. If the S&P is below that moving average, no stocks can be bought. Second, a stock's price must be above its 100-day moving average. Although this should be true for any stocks near the top of your ranking, this extra filter ensures that you don't buy a stock moving sideways (or worse, down!) because there are no other good stocks to buy. Third and finally, you’re looking for consistent momentum stocks, so disqualify any stock with a price jump larger than 15% over the last 90 days. This will prevent any particularly jumpy stocks, which may have snuck past the R^2 screener during your ranking, from infiltrating your portfolio.
As long as those conditions are met, purchase stocks according to the target position size, beginning from the top of your ranking until you run out of cash. Then rinse and repeat once a week on your designated trading day.
The Rule: Check for buy conditions. If those are met, buy stocks from the top of your ranking list according to position size until you run out of cash.
That's it!
Clenow's model is a great example of a cross-sectional momentum strategy that you can start trading today!
If you're interested in building trend following strategies yourself, check out our Strategy Builder where you can start designing and backtesting your own, custom trading algorithms without any code!