You've heard it before –
“You can't time the market.”
“Time IN the market matters more than timing the market.”
“Nobody has a working crystal ball.”
“The only value of stock forecasters is to make fortune-tellers look good.”
It's part of Investing 101, and it's very important that you understand why market timing is generally a futile exercise. We are naturally emotional creatures who constantly flip back and forth between greed (aka Fear of Missing Out) and fear, which causes us to constantly buy high and sell low, incurring transaction costs and taxes all the while. Following a written, systematic, buy and hold approach is a great way to beat our own emotions and achieve investing success.
However, in teaching those extremely important principles to investing beginners, we fudge the truth a little bit. You see, the truth is that market timing has worked in the past and that it did NOT require a crystal ball to be successful. Today, I'm going to discuss how the market can be successfully timed. I'll describe not only what I mean by “successful” (because it's probably not what you think) and also tell you exactly how to do it.
Trend-Following
So here's how you do it. You use a Simple Moving Average (SMA). When the price of stocks (preferably using a broadly-diversified, low-cost index fund) goes above the SMA, you buy. When it goes below the SMA, you sell. What is the SMA exactly? It is the average of the price of that fund over the last time period. Common time periods are 10 days, 50 days, 100 days, 150 days, and 200 days. The longer the time period, the more of a downturn you capture before selling and the less of an upturn you capture before buying, but the less frequently you are “whipsawed” (i.e. buying just before it goes back down and selling just before it goes back up) and the less frequently you pay for the transaction (in commissions, bid-ask spreads, taxes, and the value of your time). The nice thing about this technique is that it doesn't require you to know anything about the future, just the past. Nor is there anything complicated or emotional about it. It can be done very rationally and automatically.
So here's what the chart might look like. You can see where the technique is successful and where it is not. The idea is to be in while the market is rising and out while it is falling. You can see in 2000 there were a bunch of times that getting out when the trend fell below the average was the wrong move (since it then rapidly went up.) But you can also see that it had you out of the market for a good chunk of the downturn in 2001-2002. That had to be beneficial to returns, no? And you got back in before it went up too much in 2003. Then you can see a bunch of whipsawing in 2004-2006, which isn't good, but then the system had you out for most of the 2008-2009 crash. Again, you got whipsawed a few times in 2010-2012, but mostly you were in for the bull market.
So the question is how does following the Simple Moving Average stack up against buying and holding? Well, the honest truth is it wins. And it loses. What? Let me explain.
Returns vs Risk-Adjusted Returns
While the data is all period-specific, in general, you will find that a trend-following technique like this has lower returns than a simple buy and hold portfolio. It just doesn't work well enough to give you higher returns, despite avoiding a good chunk of the major crashes. Take a look:
This data comes from ETFreplay.com, but it looks like all the other data you get when you look at this stuff. This follows an ETF that tracks the Dow Jones Industrial Average (DIA) and applies the 200-day moving average from 2002 to 2018 and compares it to just buying and holding DIA. As you can see, buying and holding had a higher return, a cumulative 302% vs 271%. However, it also suffered a much larger maximum drawdown- 52% vs 18%. You can also see some other interesting stuff, like six of the eleven trades boosted returns and five did not.
You can do it with other asset classes too, such as foreign developed market stocks using an ETF like EFA. The data looks similar:
Lower returns, but lower risk as well. There are other sites that can give you free back-tested data as well, such as etfscreen.com (although only up to 10 years). If you plug in the Vanguard Total Stock Market Index ETF (VTI) you get this:
Same story — lower returns, but significantly lower risk as manifested by maximum drawdown, standard deviation, Sharpe ratio, and other measurements of risk.
How Does it Compare to Bonds?
So how does this compare to another well-known method of reducing risk at the cost of reduced returns, adding bonds to the portfolio? Vanguard has a nice article about asset allocation and historical returns and risk. Here's the meat of it:
So if you're willing to give up 1.5% of your annualized return by using a 200-day moving average, you could simply use a 60/40 portfolio and get a similar return. So which one is less risky? Well, by one measure of risk — the maximum drawdown, the risk was about the same (note the Vanguard chart is looking at 1931 and the etfscreen.com chart is looking at 2008-2009). So historically, you got about the same thing by market-timing using a 200-day moving average and using bonds. Thus my conclusion that market-timing works, and it doesn't. Want more opinions than just mine? Here you go.
What Do Experts Think About Trend-Following?
So why does this work? Larry Swedroe explains that it is due to the momentum risk factor.
As an investment style, trend-following has existed for a long time. The data from the research provides strong out-of-sample evidence beyond the substantial evidence that already existed in the literature. It also provides consistent, long-term evidence that trends have been pervasive features of global stock, bond, commodity and currency markets.
Addressing the issue of whether investors should expect trends to continue, the AQR researchers concluded: “The most likely candidates to explain why markets have tended to trend more often than not include investors’ behavioral biases, market frictions, hedging demands, and market interventions by central banks and governments. Such market interventions and hedging programs are still prevalent, and investors are likely to continue to suffer from the same behavioral biases that have influenced price behavior over the past century, setting the stage for trend-following investing going forward.”
The bottom line is that, given the diversification benefit and the downside (tail-risk) hedging properties, a moderate portfolio allocation to trend-following strategies merits consideration.
In the other corner is Mark Hulbert, who suggests that although trend-following used to work, perhaps it no longer does.
Over the past decade…there have been at least a half-dozen times since the bull market began in March 2009 in which the index dropped below the [200 day moving average] — without triggering a bear market. On many of those occasions, in fact, the stock market rebounded almost immediately after the S&P 500 fell to below its moving average. Far from marking the beginning of a bear market, in other words, breaking below the 200-day moving average often signaled a reason to buy, not sell….Still, the 200-day moving average has had some successes. It protected followers in the financial crisis, getting them out in December 2007 and back in June 2009. The market had started rebounding in March. But its missteps during the subsequent bull market frittered away its bear-market gains.
Nor is this recent experience a fluke. Over the past three decades…an investor would have lagged a buy-and-hold strategy if he got into stocks whenever the S&P 500 was trading above its 200-day moving average and moved his portfolio to a money market fund whenever the S&P 500 was below its average.
To be sure, the 200-day moving average’s track record was better than this over the six decades through the 1980s. But its track record over those earlier decades comes with a huge footnote: It is calculated assuming no transaction costs. And that’s unrealistic. Prior to the advent of no-load index funds, and particularly exchange traded funds (ETFs), it was quite cumbersome and expensive for an investor to move out of stocks into cash, or vice versa.
It seems ironic that the 200-day moving average stopped working in the early 1990s at the very point when ETFs and discount-brokerage commissions became widely available. But Blake LeBaron…suspects there is a connection between the two. After all, it’s a hallmark of market efficiency that previously successful strategies stop working when too many investors try to follow them. In this regard, LeBaron notes that moving-average systems also stopped working in the foreign-currency market around the same time that they did for equities. It increases our confidence that the markets have undergone a fundamental shift that these strategies stopped working more or less simultaneously in two entirely different markets.
Final Thoughts
I don't market time. I use bonds to reduce risk instead. If you're not interested in using bonds because you don't want to reduce your potential long-term returns, then you probably shouldn't be interested in market timing either. But if you do choose to do market timing using a trend-following strategy, here are some tips to boost your risk-adjusted returns as much as possible:
- Use good investments. Most of the time you're buying and holding the investment anyway since the market trends upward most of the time, so you might as well use investments you would use to buy and hold. That means low-cost, broadly diversified index funds. This trend-following data isn't some excuse to introduce uncompensated risk to your portfolio.
- Do it in a tax-protected account. One of the reasons buy and hold works so well is that it reduces your investing costs, particularly transaction costs — the largest of which is capital gains taxes.
- Do it in an account with free trades. Again, reduce your transaction costs. So long as the trades are free, mutual funds have even lower transaction costs than ETFs as there is no bid:ask spread.
- Consider the value of your time. One of the best parts of buy and hold is that I can pretty much ignore my investments for months at a time. While a 200 day moving average doesn't require very frequent trades, you do have to keep an eye on the markets.
- Write down your plan and stick to it. The biggest problem in investing is the investor, not the investment. Once you have a reasonable plan, it is critical that you stick with it.
- Do it only with liquid, transparent, inexpensively transacted asset classes. Don't try to do this with your income properties.
- Remember that this is all based on back-tested data and will only work insofar as the future resembles the past. Don't bet the farm.
What do you think? Do you engage in market-timing using a trend-following strategy? Why or why not? Comment below!
Terrific article Jim. I am going to share it with our newsletter subscribers. I have been using market timing on all or a part of my portfolio since 1983. When I was an investment advisor (1983-2012) I worked with hundreds of investors who, like myself, have half of their portfolio in buy and hold and half using market timing. I know from decades of experience that most people want timing but very few are willing to put up with the regrets that trend following systems create. The regrets include often selling at a loss, sometimes having many losses in a row, selling at a loss and then getting back in the market at a higher price than they last got out, under performing during bull markets and not making money in a bear market. Of course market timers expect to do better than a buy and hold in a declining market but whipsaws can lead to under performance even in the worst markets. The other emotional trap is investors are more likely to be attracted to timing after a very lucky trade. My biggest day as a timer was after getting out of the market a month before Black Monday, October 19, 1987 . Mark Hulbert tracked our newsletter performance in 1987 and he could confirm we had some accounts up over 50% in 1987. The trade made the press sufficiently to be invited as the the weekly guest on Wall Street Week. Plus, I enjoyed a handful of appearances on Nightly Business Report with Paul Kangas. When given credit for having “called” the ’87 crash I was always quick to explain that there is a big difference between predicting a crash (which I didn’t) and, by chance, being out of the market when it crashes (which I was). But there is more to the story. Another timer (Dick Fabian), with a far bigger following than I had, used a longer term trend following system. It triggered a sell signal on October 16, 1987 and that meant his followers got out of the market the day of the crash! I was lucky, Dick was not. When the market was crashing my clients were all in cash. Many of them fired me after thanking me for having protected their investments and their feeling of loss they would have had without the timing. The reason they fired me was they asked the right question. The question was simple and one every timer should be willing to answer. “When the next crash comes are you confident that your systems will be able to protect against similar losses.” In fact, Dick Fabian cared just as much about his clients as I did ours. I was very clear that any system has much greater risk than most investors, and even many professionals, expect. No, we couldn’t suggest our systems would protect them from a market pattern that didn’t fit our system. Of course that is true of buy and hold as well. What made our timing better than most timers was not about the return, as that is an unknown, regardless how good a system has worked in the past. To the best of my knowledge we were the first timing manager who used the research that came out of Fama and French for buy and holders. Our portfolios were built with many different equity asset classes, as well as many different fixed income asset classes. Our equity portfolios included timing large, small, value, growth, U.S. and international asset classes. Plus, very conservative clients could have bonds in their portfolio as a second level of defense. Also, in 1995 we started a hedge fund (Leverage Global Opportunity Fund) that combined broad asset allocation, timing and leverage. The hedge fund has recorded the real time monthly performance of more than 25 years. As would be expected the results could either much worse or much better than buy and hold. Call it luck or skill the real time results have been terrific. Risk wise it has about the same downside standard deviation as the S&P 500. In other words, it is not low risk. For those interested in reading the thoughts of the best market timer I know, I hope you will check out the blog of Dennis Tilley. (http://www.assetclasstrading.com/author/dennis/) He knows far more about timing than I will ever know. My final thought, after being around market timing since 1966, is no more than one out fo 100 investors are built for market timing. It asks the investor to accept to much emotional pain. I would never do it on my own, which is why I have it professionally done. At 77 years of age I don’t have the interest or the time to devote the few years of my life to worrying about the market. I want my portfolio to be on automatic. I don’t want a manager to be involved in making what I consider to be emotional decisions. So, my buy and hold portfolio uses 10 equity asset class funds that are passively managed along with timing similar kinds of funds (mostly ETFs) using mostly trend following with a percentage using momentum. I have absolutely no interest in predicting the future or being involved in the day to day noise of the financial markets. I have no idea how my portfolio will perform. If I live 10 more years maybe there will be a huge bear market that will “prove” timing was a good thing to do. On the other hand, the timing may be a wasted effort. The same is true about the buy and hold portion of my portfolio. Maybe the small and value over weightings will pay off and maybe they won’t. Maybe I should have my son give a final performance report at my funeral as that is the only point that I can judge how the strategies performed. The two ways I judge the financial outcome of a lifetime of saving and investing include: 1. How much did we take out of our portfolio to live on and give away and 2. How much did we leave to others. The rest is noise.
Love that last couple of sentences.
Paul first off I appreciate your effort for financial literacy and I am so happy reading Jim’s work had led me to you as well. As a teacher of buy and hold I was always fascinated that you still have half your portfolio in market timing, but from what you just said we just don’t know which will come out ahead. Also, though you are removed from the market timing portion of your portfolio, does your advisor who is managing the market timing portion of your portfolio have nerves of steel? Isn’t here or she bringing their own behavioral biases to the table when trying to market time? Or are they truly objective and robotic, given it’s your money, not theirs?
also in regards to the last couple of sentences, I would add a 3rd point that is pertinent, especially to docs. It is minimizing the sacrifice away from friends and family to build that wealth. The reason I hate having been duped by a NWM financial advisor in the past is not that I had to pay the stupid tax, but that money represented blood, sweat, and worse, time away from my wife and kids. In the end, no matter how extravagantly I spend and give away and leave to others, I don’t want the cost to be losing my relationships and never realizing potentially precious times with friends and family.
I just buy and hold (UBH)…. my market timing approach is : “If I have spare cash to invest… it is Time to put it in the Market”
The main problem with trend following is that it is overly marketed as the holy grail because there are so many people selling “trend following systems” for $2997. That is why it gets so much attention. They focus on the fact that you get stock like returns with bond like drawdowns, but leave all the other negative parts of it out. The fact that markets can get more volatile, it might not work, it is extremely emotionally draining especially in a long bull market where you underperform for 10 years or so with no guarantee it will pay off.
The people selling these systems take advantage because this strategy has worked in the past. Big deal, as Jim once said, “If you torture the backtested data long and hard enough you will get it to admit anything.” Its no different than creating some magic bitcoin course to get rich because it happened to get great returns for several years.
Every investment approach has its cycles and seasons where it will underperform or outperform.
As Paul Merriman said, trend following is good for 1/100 people. People say they can handle it but they really can’t and will give up years in showing nothing but a loss since need to stick with this for decades. Reminds me of those people in college that were all going to be doctors. They were positive. They were all premed. Then magically after one chemistry course they all dropped it and found another major. Trend following is like that.
Love the pre-med chemistry analogy.
usually its orgo that’s the killer, not freshman chem
Excellent article that does a great job explaining the basics of market-timing, using a trend-following strategy.
I invest mostly from column A and a little from column B. I am primarily a buy and hold investor, with a generationally focused timeline. My default allocation is 60% stock, 20% bond, 10% real estate and 10% broad commodities/hard assets, using predominately low-cost index ETFs. Cash is managed separately, not as a percentage, but as a target amount, since my cash needs don’t fluctuate with the market. One third of the stocks and one fifth of the bonds are international. I use a trend following, market timing strategy to slightly reduce overall portfolio risk. During periods of extremely high valuations (three in modern market history, two in my investing lifetime, i.e., 1929, 2000 and recently), I reduce my stock allocation to 50% and increase my bond allocation to 30%. The trigger is my broad US stock market ETF dropping below the 10-month rolling average. This occurred in late-2018. I will return to 60% once valuations are no longer extreme and the index rises above the 10-month rolling average. The difference between the two position states is minor, but it does ratchet down risk a bit. I have no expectation that this approach will increase my ROI, but it “may” improve my risk-adjusted return.
FWIW, since I don’t automatically reinvest dividends, I have the flexibility to choose where they are invested. I reinvest into the ETFs that are below their target allocation, which keeps me closer to my preferred asset allocation. Rebalancing is only done when an investment moves more than 15% out from its target allocation. This tends to happen very infrequently, requiring some fairly large market moves.
I hope this info has some value to your readers.
Another momentum perspective, which I follow, are the two portfolios (from Brian Livingston) based on research by Meb Faber and Steve LeCompte; MuscularPortfolios.com. Note: All the info is free.
Thanks for the nice article. I’d never thought of trend following it in comparison to bonds as a way to reduce risk until you put it that way and showed the data. I agree, bonds seem much simpler!
Really nice insight on the bond equivalence.
One thing I’ve never understood about trend following – the buy and sell signals only look obvious in hindsight. If you truly bought or sold at every wiggle where the index touched the SMA, it looks like there would be dozens of transactions in the time period displayed, not just 11. I suspect if you looked intraday when the index and SMA are about the same, it would be even worse. So to use it in the real world, you would have to have some kind of “confirmation” parameter. Which would make it less likely to give false signals, but also less likely to protect you.
I’ll say one surprising impact that “dry powder” had for me.
I kept to my retirement investing plan with maxing all available deferred options but have been putting away extra into taxable every month too.
I stopped doing that at the pandemic outset, wanting to have cash on hand for any opportunity or investment but really because if I lost my job, I’d have a bigger emergency fund.
Well, I didn’t lose my job and didn’t buy the big dip in March.
But now I have enough cash on hand that I feel really comfortable leaving corporate healthcare employment and starting my own practice.
Could I have invested that money and then sold it when time came to start my own practice? Yes, of course, but the question is, since I’ve been holding and holding and holding my index investments and never selling, would I have sold it? I’m not so sure about that.
Thanks for your great articles as usual.
It’s funny, I was thinking about Paul Merriman as I read the article, and then he shows up as the first comment. I learned how to frame market timing from his podcast, as far as looking to reduce maximum drawdown not necessarily improve long term returns. I think that his approach of automating it, or at least taking emotion out of the equation would be key. Without that, it would honestly be too stressful, even setting aside the likelihood that emotional based market timing will underperfom over time.
So nice article and comments section!