By Carlos M. Sera and Carl E. Sera, CMT
The title of this paper is Market and Investor Behavior. The reason we tie market and investor behavior together is because we don’t know where one begins and the other ends. Does the market behave in a certain way because of the way investors behave or do investors behave in a certain way because of the way the market behaves? This reflexive or feedback process is discordant with equilibrium theory but as this paper shows, markets and investors have highly predictable patterns that can be used to the knowledgeable investors’ advantage. This paper shows that the order of prices matter. While some will go so far as to say that prices influence fundamentals and that these new fundamentals change expectations, thus influencing prices, we will only deal with prices in this paper. We do not have a particular knowledge or methodology to test if prices influence fundamentals. However, nothing is more fundamental than price and we show that previous prices influence future prices in a highly predictable manner. This means that prices at time (t) can be used to predict prices at time (t+1).
In our paper Long Term Simple Moving Averages as a Risk Control Technique we showed that by using a 12 month simple moving average (SMA) or for that matter most any long term SMA, that investors could equal the rate of return of a buy and hold strategy but lower their risk by half as measured by maximum drawdown. In that paper we developed a framework that is understandable and can be easily replicated. In this paper we are expanding on that knowledge to gain insight on how markets behave when prices are above and below the 12 month SMA. More importantly, in that paper we compared a 70/30 portfolio allocation to a 12 month SMA and concluded that the difference lies in the 30% of the time that a SMA has the investor out of the stock market. This paper explains why the SMA is better.
We will take the differences between periods where markets are above their long term SMA and periods where they are below and see what we can learn about market and investor behavior. What you will conclude, as we concluded, is it makes no sense to invest in the stock market when it is trading below its 12 month SMA. Why? 98.1% of the compounded annual growth rate of the stock market from 1929-2008 transpired when prices were above the 12 month SMA and markets exhibited calm and boring behavior. By calm and boring we mean they exhibited significantly lower volatility than when prices were below the 12 month SMA. The difference is stark.
As in our other paper, we examined monthly data for the US stock market for the period from 1928 through 2008 using data from Ibbotson. We then ran a simple 12 month SMA on the data to arrive at comparable results for the 80 year period 1929 through 2008. You need one full year of monthly data to establish a 12 month SMA which is why the period tested is from 1929-2008. Once again, it is our discipline to always use 12 period simple moving averages to avoid curve fitting. Our belief is that if something does not make sense over a 12 period time frame or a consistently applied X period time frame then it does not make sense. Models must be robust. We will start with a hypothetical $10,000 in 1929 and gather statistics for a Buy and Hold approach as well as a mechanical crossover SMA system that gives us buy and sell signals for periods when we should own stocks and periods when we shouldn’t. During periods when we shouldn’t own stocks we will invest the money in US Intermediate Government Bonds based on data also provided by Ibbotson.
The following table provides a general overview of what transpired over the last 80 years in the US stock market. We can see there are 960 months of data in the 80 years. We can see that 592 of those months were winning months and 368 were losing months. When we further analyze those 960 months we see that in 682 of those months, the price at the beginning of the month was above the 12 month SMA and we see that in 278 of those months, the price at the beginning of the month was below the 12 month SMA. This is valuable information. If price behaved like a coin toss, we would observe that price would spend 50% of the time above the SMA and 50% of the time below the SMA. However, this is not the case. In this case price spends 71% of the time above the SMA and 29% of the time below the SMA. Prices are not random like a coin toss. The order of price matters.
In the next table we show what happens when you dice and slice the 960 periods. $10,000 invested in 1929 and held continuously grew at a compounded annual growth rate of 8.92% to $9,308 482 by the end of 2008. This is the Buy and Hold approach. $10,000 invested in 1929 under the 12 month SMA model grew at a compounded annual growth rate of 11.26% to $51,129,957. This is the 12 month SMA approach. Remember that when the 12 month SMA signals the investor to avoid stocks the investor has their money invested in bonds. What is most fascinating is that if you break down the 960 months into the 682 months where the price at the beginning of the month started above the 12 month SMA and only invested in those months you make 98.1% of your return. This clearly says that you don’t miss much in terms of rate of return or profitability if you are out of the stock market the 278 months when it is below its 12 month SMA. This study shows that buy and hold has had its day in the sun. In fact if you invest your money in bonds during that period you do significantly better. Investing $10,000 in bonds for the 278 months where stocks are below the 12 month SMA grew to $62,362 while investing in stocks during the same time frame grew to $11,353. We can think of no good reason to invest in stocks when they are trading below their 12 month SMA and think the Modern Portfolio Practitioner that has money invested in stocks and bonds at all times might want to rethink their strategy.
When presented with the data from the table above, most people can’t believe it. Believe it. We tested for multiple simple moving average parameters between 6 and 24 months and in every period we reach the same conclusions. It’s far, far better to be in stocks when they are above the 12 month SMA or any 6-24 month SMA and in bonds when they are below. At this point most people want to know how a 12 month SMA approach would have fared recently. They are curious because the events that started with the US stock market top in October 2007 have caused many investors to rethink their strategy. There was a significant move down in stock prices followed by a pronounced and sharp move up in stock prices. Investors that held firm to their buy and hold convictions suffered through approximately a 55% decline based on daily data followed by approximately a 65% gain as of year end 2009. The math on this still shows a substantial net loss due to the mathematics of recovery which we explain at the end of this paper. Nevertheless, the pressure on fund managers and investors has been intense. They are expected to preserve capital when markets decline and increase capital when markets rise.
From the October 2007 peak and using monthly data on the Russell 1000 and the Barclays Aggregate Bond Index, the 12 month SMA model would have had the investor out of the US stock market by February of 2008, and thus missed most of the decline that culminated in March of 2009. However, it would have missed a good portion of the recovery from March of 2009 until it would have generated a buy signal in August of 2009. The 12 month SMA investor avoided most of the pain but missed a good part of the gain. Nevertheless, missing pain is more profitable than making gain when it comes to investing and this recent period proves it once again. By the end of 2009 the 12 month SMA had exceeded the October 2007 capital peak by 6.76% while the Buy and Hold Investor was still down 23.79%. Said more plainly, an investor that had $100,000 at the end of October 2007 and used a 12 month SMA had $106,755 at the end of 2009 and only experienced a 15.85% maximum drawdown. An investor that had $100,000 at the end of October 2007 and used a Buy and Hold strategy had $76,214 at the end of 2009 and experienced a 51.13% maximum drawdown.
As we’ve said many times and will say again. The use of a 12 month or any long term SMA is not a panacea. For example, the sharp run up in stock prices from March of 2009 would have been missed until August of 2009 by using a 12 month SMA. It does not out perform in every period. What it does without dispute is reduce maximum drawdown and since what this means is reduced portfolio volatility; it has a significantly better likelihood of success than a buy and hold approach given the behavioral shortcomings of investors.
Two questions naturally flow from analyzing the recent past. The first, is it worth missing the rises in the stock market if you miss the falls as well? In other words, is the buy and hold approach still valid, is it worthwhile to always be fully invested? The second, is it a profitable endeavor to try to pick market bottoms or bottom fish?
Let’s examine these two fundamental questions based on history. We know that $10,000 invested in 1929 and held continuously grew to $9.308 million by 2008. So the answer to the first question is obviously no. Unless you have a better approach, such as a long term SMA, it is not better to miss the falls because you might miss the rises and since there are more winning months than losing months, it is a worthwhile endeavor. This is the buy and hold argument. However, what we know about the variable risk profiles of investors in conjunction with the volatility of a continuous investment approach makes it difficult to execute this approach as a long term investment strategy.
The second question is more intriguing however. Let’s reiterate it. Is it a profitable endeavor to try to pick market bottoms or bottom fish? The answer is no. The question also implies something more subtle. The mere act of trying to pick bottoms or tops implies that an investor is willing to trade the stock market and not just always be fully invested. To be able to pick market bottoms one needs capital and if you have capital it means it’s not invested in stocks, therefore you are not a buy and hold practitioner. Buy and Hold investors can’t pick bottoms because they are already invested at bottoms. This implies that a person has already decided to be a stock market trader and have periods where some or all of their capital is invested in the stock market and periods where some or all of their capital is not. So, is it worthwhile trying to pick market bottoms? In general, we think it is not and we will show you why.
This paper deals with market behavior as well as investor behavior. The following table is presented to focus on investor behavior since it explores volatility. We can see that for the 278 months when stock prices are below the 12 month SMA that they are 58.6% more volatile. Both average winning months and average losing months are more volatile. We saw this during the period from February 2008 through the period August 2009 and we are confident we will see it again the next time we have a market decline that takes us below the 12 month SMA. The reason of course is that fear is a stronger emotion than greed and people are more apt to make irrational decisions when they are losing money or have lost money. This is also why there must be more winning months in the market than losing months and by the substantial 3 to 2 margin that we have seen. Said in a different manner, prices go down much faster than they go up because fear is the dominant emotion.
Once again let’s look at volatility as it relates to periods below the 12 month SMA. The following table shows the expected and actual number of periods that we should see price moves of a certain magnitude. For example, of the 960 months tested there were 52 instances where the market either went up more than 10% or down more than 10% for the month. If you multiply the 52 observations by the proportional 29% of the time that the market spent below its 12 month SMA you would expect 15 observations where the market moved more than 10%, either up or down, and it was below the 12 month SMA. The results show that in fact there were 31 months where the market moved more than 10%, either up or down, and it was below the 12 month SMA. Looking at the table we can see that as you increase volatility you increase the probability that a period of extreme volatility will originate below the 12 month SMA. For example, a 15% move, either up or down, has a 70.58% probability of occurring when the price originates below the 12 month SMA.
We already showed that it makes no sense to invest in stocks when they are below their 12 month SMA. Nevertheless, if the largest winning months happen below the 12 month SMA isn’t it tempting to try to pick a market bottom? Sure it’s tempting, and to the victor goes the spoils, but in our opinion unless you have a disciplined bottom picking approach, which few do, we think it is better to sit it out. The reason is because the largest losing months also occur below the 12 month SMA and losses hurt more than gains. We hope the following two tables will reinforce the notion.
Let’s look at the buy and hold myth that has been perpetuated on the unsuspecting investor since the dawn of the mutual fund. We’ve all heard it. It goes like this, “Market timing does not work. You must stay fully invested because if you miss the X number of winning days in a year you miss out on all the gains for the year.” We tested the myth. We asked the question, what happens if you miss the identical number of X losing months over the last 80 years. For our study we used monthly data but it holds true for data in every time frame. This means the myth breaks down at the minute, hour, day, week, month, or quarter level. It’s a myth and you should not believe it. The results show that if you miss the 144 most profitable months over the last 80 years as well as the 144 largest losing months, a $10,000 investment does not grow. It stays at $10,000. Since a disproportionate number of these large losing and winning months occur below the 12 month SMA, we believe bottom fishing makes no sense. In this myth busting example, all the money was made in the 672 months that had the least volatility. Once again, we are advocates of the 98.1% Solution. Boring beats volatile.
The previous table dispelled the buy and hold myth. But why is it so important to miss large losing months? The reason is due to the asymmetrical nature of the stock market as well as the mathematics of recovery and the reason you can miss the 144 best and worst months. The following table is self-evident and shows the mathematics of recovery and why losses hurt more than gains.
There is much to learn from this simple paper. We can see that long term simple moving averages are a good way to gain insight on market and investor behavior. We have seen that markets spend more time going up than going down. This means that when markets go down, they must go down faster than when they go up. We have learned that markets are more volatile when they are below the 12 month SMA than above. We also learned that missing market bottoms is not costly because you miss significant losses as well.
From studies in Behavioral Finance we know investors have variable risk profiles. Their risk aversion increases during losing periods and decreases during winning periods. This means that investors are more likely to act irrationally and “Sell Everything” during periods of market losses and increased volatility than periods where they are making money and markets are relatively boring. Evidence shows institutional investors are not immune to this irrational behavior. In the recent 2007-2009 period of rapid market decline just when their asset allocation models were signaling to increase stock exposure, we can observe they failed to act. This is where a 12 month SMA is most valuable. By avoiding the volatile periods, the investor is less likely to act irrationally. A disciplined use of a 12 month SMA will reduce periods of volatility exposure and thus protect the investor from themselves. In action, this type of approach has a significantly better chance of working than the Buy and Hold approach. Lastly, it’s not a bad stand alone technique. For the 80 year period tested it significantly outperformed the buy and hold approach with a CAGR of 11.26% vs. 8.92%.