Join our community of smart investors

Smarter than the average?

We talk a lot about selling on the 10-month moving average but how can investors use this rule in practice without making unnecessary trades and are there better ways to manage your positions?
May 12, 2017

When to sell? It’s one of the toughest conundrums for investors and inevitably gives rise to some of the psychological and behavioural biases that human beings are prone to. Does the answer lie in disciplined rules-based investing? One method that we have flirted with in some of our model portfolios could be to look at a security’s 10-month moving average price and sell when the current price falls below this level. The time to buy back in is when the price is back above the 10-month moving average.

Potential advantages of this approach are that investors have a built-in mechanism to ensure they ride momentum – which when in full flow has been the most significant outperformer in terms of factor returns that the stock market has known. On the other side of the coin, investors also give themselves an automatic stop-loss to get out of positions with negative momentum and avoid value traps. The main drawback is that movements in some volatile asset prices can precipitate trades that, in hindsight, seem unnecessary and result in irksome charges.

 

Does the rule work in practice?

There are some practical difficulties and questions to answer before attempting to implement the moving average rule. Firstly, what data is best to use? In the strictest sense, the 10-month moving average refers to the 200-day moving average of a security price. In the 10-asset portfolio that we constructed to achieve a balanced portfolio allocation using exchange traded funds (ETFs), however, we simply used the average end-of-month figure for total return (TR) indices over the 10-month period.

This greatly simplified method was back-tested to early 2005, which was the furthest back we could get data for the full set of indices tracked by the exchange traded funds (ETFs) used in our portfolio. As well as taking less effort to monitor than a daily moving figure, following total returns (rather than price only) indices helps reduce trading frequencies, as the dividend component can mitigate volatility caused by price falls.

Focusing on the FTSE 100 in isolation it is possible to delve back further and examine the returns over time had you invested using the 200-day moving average or the end-of-month moving average (MA) rule. Following the FTSE 100 price index 200-day moving average since 15 July 1988 and using it as a buy and sell signal, the price returns of the index would have delivered a compound annual growth rate (CAGR) of just 2.03 per cent, which is less than the 4.75 per cent overall growth rate of the FTSE 100 price index over that time. Using the price index as the basis for trading signals but tracking the total returns index for profitability, the 200-day moving average rule would have made 4.79 per cent a year, on average. Just following the total return index, however, would have been worth 8.74 per cent a year.

There are two reasons it has been more worthwhile to stay invested in the market for blue-chip shares. Firstly, some of the best price returns have occurred in the early stages of rebounds from bear markets. Secondly, as counting total returns demonstrates, the significance of dividends is huge and being out of the market means missing out on this income. In our calculations, we have assumed that you would have made the trade at the end of the last day used in the average calculation. This is to give a realistic simulation where you would have been in the market up until the sell signal had been computed and out of the market up until the buy signal was clear. This has a significant impact on the way returns are calculated.

For the sake of comparison, what is the difference if we look at the moving average of the price index at the end of the month, over 10 months? From the end of June 1988, the annualised price returns would have been 3.82 per cent and the total returns 6.29 per cent. This is slightly better than the 200-day rule because we miss some of the signals and just adds to the weight of evidence that the less you trade, the better.

These figures also ignore the returns from cash when you would have been out of the market. The reason for this is that money market funds often have minimum investment periods, so you would in all likelihood have had to park cash in low-interest dealing accounts. Of course, over longer bear markets, interest could have accrued in these accounts, but this certainly wouldn’t have been the case for many of the shorter spells, especially with the 200-day iteration where trades could have occurred within a few days of one another. Given the high interest rates on offer in the 1980s and 1990s, assuming zero returns from cash has a major impact on the system. It is possible for private investors to move seamlessly between tracker and money market funds within Sipp wrappers, but the analysis here has been based on using a dealing account. Also, given the current low rate environment, it felt more appropriate to focus on the effect using the moving average has on equity returns.

 

 

 

 

 

Psychological benefits limiting peak-to-trough losses outdone by overtrading

Trading the FTSE 100 this way would have helped you to avoid some of the heaviest peak-to-trough drawdowns experienced in the past 29 years. For example, following the 200-day moving average rule on the price index, the worst fall in the value of capital invested would have been 34 per cent in the 2000-03 bear market. This compares favourably with the FTSE 100 total return index, which lost 48 per cent, and the price index more than halved from its December 1999 peak. Trading on the end-of-month moving average would have been better, with the worst losses being 22 per cent.

These are still significant drops, however, and any psychological satisfaction gained from missing some of the worst bear markets still needs to be set against the frustration felt due to making too many wasteful trades. Using the 200-day moving average rule would have involved being out of the market a total of 2,039 out of 7,270 trading days since 15 July 1988. The strategy would also have entailed buying in and out of the market on 270 occasions. That’s under 10 trades a year, but would still eat into your profits. It would also be emotionally challenging on the occasions when the rule would encourage buying back into the index several times in a short space of time. Using the end-of-month method lessens this negative effect, but being out of the market 118 out of 334 months would still have necessitated more than 200 trades.

In Investors Chronicle’s 10-asset ETF portfolio, we only tracked the FTSE 100 total return index. As mentioned, this lessens some of the impact of price movements, so some of the smaller price falls didn’t result in trades. Our 10-asset portfolio just used the end-of-month moving average method and, applying this to the total returns FTSE 100 index back to mid-1988, we would have only been out of the market for 87 months. The CAGR achieved using this method would have been 7.39 per cent. Total returns would have been marginally less than buying and holding the TR index, but we begin to see a better risk/reward trade-off as the worst peak-to-trough drawdown here is 19 per cent as opposed to 44 per cent for buy-and-hold. Over a near 29-year period, the extra one-and-a-half per cent each year from just holding the total returns index would have a significant compounding effect, but at least there is some benefit to the investor in terms of having gotten out of very bad downturns. But, especially considering trading costs too, buy and hold still looks a more attractive option overall.

 

 

 

Heading down the market cap scale

Going down the market cap scale, how does the end-of-month moving average rule work? Looking at the total returns version of the FTSE 250 index of UK mid-cap stocks, the average annual return since the beginning of June in 1988 has been over 11 per cent, with a worst peak-to-trough drawdown of 47 per cent in the 2007-09 financial crisis. Our end-of-month moving average rule would have limited the worst losses experienced to 36 per cent but at the cost of lower annualised total returns of 8.49 per cent.

The main benefit of the moving average system, or indeed any form of stop-loss, is in limiting downside in the worst periods for the stock market. If you are investing for the long term and can afford to ride out paper losses, then the impact of falls will be mainly psychological. Therefore it is worth asking yourself, if the losses aren’t going to be crystallised, how much worse would you feel seeing the value of investments fall by half rather than by a third? If you are investing towards a long-term goal (for example a person in their 20s or 30s saving for retirement) then chances are the opportunity cost of losing out on recovery moves and dividends by being out of the market is going to be far more damaging – so the price paid for smaller paper losses in distressing periods may not be worth it.

For investors with more immediate needs, limiting short-term falls is of a higher priority. However, if this is the case, it is arguably more sensible to choose a more cautious strategic asset allocation, than trying to time investments in equities using moving averages.

 

What about for individual stocks?

At the index level it looks as though, in the absence of simple access to good cash returns when not in the market, following a simple buy-and-hold strategy is the best way to invest for the longer term, but is this also true for individual companies? Stocks have different levels of beta – the measure of how correlated their returns are with the market benchmark – so we might reasonably expect there to be some stocks where the price moves are of a different magnitude or direction to the overall index, and the moving average rule can prove useful as a buy or sell signal.

Firstly, focusing on some random FTSE 100 companies, if we take Royal Dutch Shell (RDSB) (market cap £166.4bn), British Telecom (BT.A) (market cap £30.8bn) and BAE Systems (BA.) (market cap £19.7bn), we can see whether buying in and out according to moving averages would have enhanced profits. The results are revealing, but not unexpected. Working according to the end-of-month moving average of each stock’s total returns (with a 100 base starting at 31 December 1997), we note that Shell, which has consistently been among the very biggest companies on the FTSE 100 index, does not benefit from employing the rule. The FTSE 100 is a very concentrated index, with companies like Shell, HSBC and BP making up a large part of the overall. The fortunes of these companies most influence the index and as the moving average rule hasn’t worked for the FTSE 100 index, it is unsurprising that it doesn’t work for Shell either. Just buying and holding Shell would have delivered a 6.94 per cent annualised total return up to 31 March 2017, but buying in and out on the end of month, 10-month moving average would have seen a much lower CAGR of 4.92 per cent.

For the other two large companies, following the 10-month average would have had a more positive impact. In the case of BT, buying and holding would have delivered 3.45 per cent annually, but the figure was transformed to 10.45 per cent by buying in and out on the 10-month moving average. A similar effect can be seen in returns from BAE Systems – the defence and aerospace manufacturer making 5.97 per cent annual returns when bought and held but a much more impressive 13.56 per cent when bought and sold on the moving average rule.

Looking at three more arbitrarily chosen companies, this time from the FTSE 250, there seems to be little benefit from following an end-of-month moving average strategy. From a start date of 31 December 1997, Berkeley Group (BKG) would have made an average annualised return of 15.19 per cent, but this falls to 12.12 per cent buying in and out on the end-of-month 10-month moving average. Investors in Amec Foster Wheeler (AMFW) would have been better off with buy and hold (10.71 per cent versus 5.86 per cent), as would people who owned Synthomer (SYNT) (8.28 per cent versus 7.59 per cent). Using the 10-month rule for pub chain JD Wetherspoon (JDW) would have achieved marginally better returns (9.22 per cent using the rule versus 8.29 per cent without), but, although an extra 0.93 per cent a year, on average, compounded over 19-and-a-half years is probably worth the dealing costs, in practice simply holding the stock would undoubtedly have been easier.

 

 

 

Not a rule for trading mid-caps and be wary of survivorship bias

Of course, picking a few companies at random doesn’t tell us very much, other than that applying this rule in isolation to your individual mid-cap holdings is likely to be hit and miss! This is even more likely to be the case with other companies on the index that haven’t been listed for as long as those we have looked at. There is bound to be a bias in our random selection because they have survived as listed companies for the past 20 years. Their performance is not going to be reflective of 250 companies that have not been around for as long. More importantly, it is not representative of the fact that many companies will have failed or been taken over in the past 20 years.

In spite of these drawbacks, however, it is too early to discard the moving average just yet. The next stage of our investigation will be to explore the connection between stocks’ betas and the moving average of the market value. The difference in returns for some FTSE 100 companies using the moving average rule warrants further study and it would also be interesting to test some rules for FTSE 250 companies without retaining a bias to companies that have been long-term index constituents.

This will be the focus of a follow-up article, but the immediate conclusion to be drawn from our analysis thus far is that following the 10-month moving average rule slavishly is not necessarily an optimal way to invest. There have been periods, such as the 2000s, when a combination of decent cash interest rates before the financial crisis and worse returns from equities would have made the system advantageous for Sipp investors. The 10-month moving average rule still protects against some of the very worst losses in bear markets, but this has to be set against dealing costs and the dissatisfaction of making too many poor trades. Nowadays, however, with rates low and investors looking to equities to provide capital gains and income, it may be a better policy for long-term investors to accept the downside risk of stocks they hold and get comfortable enough to sit tight through any period of falling markets.

There is no substitute for good solid investing principles and, before applying a rule like the 10-month moving average, it is important to consider what a given stock is in our portfolio for. If we are holding a growth or a value stock, then it is appropriate to use the rule as a way to sell out or buy back into opportunities that may not pan out as smoothly as we had hoped. In the case of growth stocks, it can be a disciplined way to get out of a story that is just not finding its legs and, in the case of value stocks, it can help us to get out before we get into a ‘value trap’. On the other hand, if we are holding a stock for income, then other factors like its ability to cover its dividend and how cash generative it is will be more important than 10-month average price momentum. Of course, the two are not mutually exclusive. Overall, there may be instances where the rule can help make a selling decision easier, but unfortunately, from our analysis of stock market indices, it doesn’t appear to be a silver bullet to improve risk-adjusted returns.