Author: mungofitch

Subject: Getting away from the bear

Date: 9/27/2008

Recommendations: 93

There has been a lot of discussion about market timing in the past, but frankly most systems are too complicated to use for the average person. Plus, though all timing systems work in backtest, many fail in real life. So what`s a person to do? Herewith, a really dumb, simple timing system that gives only one signal per year on average, and reliably improves risk adjusted returns. It`s aimed at the broad market buy-and-hold crowd, with the goal of just a little bit better sleep during bear markets. But, I find it`s a very good simple bear market detector for more sophisticated uses too. From a free investment newsletter I write. Jim ======================================================================= The Informed Hunch, your free investment rag "Words are cheap, especially ours!" Executive summary: beat the market with 2/3 the risk and only one simply determined trade per year on average. Now you can decide if you want to keep reading! You may have noticed that it has been a fairly exciting month, quarter, and year in the stock markets. But, in order to resist confusing the trees for the forest, let`s start with the broad view. In short, if you hold stock and aren`t buying much more right away, it`s annoying when the stock you own goes down in price for a long time. This is a quick discussion about bear markets, aimed at those folks who don`t think it`s very worthwhile trying to beat the market, and yet---and yet---discover that bear markets are no fun. If you think that an index fund is the height of sophistication, and really aren`t into working hard picking individual stocks, then read on. There might just be a way to do just a tiny bit better. My starting point was the following ideas: - There is no doubt that bull and bear markets exist. Sometimes the market goes roughly upwards for several year, and sometimes it goes roughly downwards for several years. - The sort of folks who use index funds are not the sort of folks that are interested in sophisticated daily trading systems. So, if we`re going to improve on things, it had better be simple. So, lets start with the basic idea that we`d like to be in the stock market when it`s going up, we want to be out of the stock market when it`s going down, and we don`t want to switch between the two all of the time. This method has a signal about once a year on average, which isn`t exactly "active" trading. What is a rising market? Obviously prices don`t go up every day, they go up in a zig-zag. The main property of a rising market is that the zigs and zags get higher on a fairly steady basis. Thus, we have a workable definition of a bull market: the price of stocks is hitting new recent highs fairly regularly, with not too long a gap between them. Conversely, a bear market is when this isn`t true: no recent high has been made in the last while. So, how "recent" is a recent high? and how long should we wait before calling it a bear market? And, why would this help? Let`s start with the last question. My theory is that whenever the US stock market hits a new recent high on the index which is most watched, it has a reinforcing effect of pleasure on the greed portion of the brain. People will tend to want to invest for a while, but then the buzz will wear off and they`ll need another dose to keep their greed active. Without that reinforcement, they gradually become fearful and start wanting to get out of the stock market, which is of course the sentiment that generally causes falling stock prices. But, whether this theory is true or not, certainly our definition of a rising market is pretty simple: higher recent highs. Or, put in reverse, a bear market is when we haven`t set a new recent high in quite a while. So, I looked at a graph of the US stock market. The main zigs and zags are reasonably short duration, so it appears that if you go 100 trading days (around five months) without a new recent high, it`s a worrying sign. Looking at those times on the graph, nobody would describe it as a steadily rising market, and that`s likely to get people worried. So, we have a tentative definition of a bull market: if a bear market is when we have gone 100 days or more without a new recent high, then a bullish time is when market has hit a new recent high in the last 99 trading days or less. Well, we haven`t defined "recent" high, so for the sake of simplicity, I`ll use the same 99 trading days. We`re more or less assuming that the investing population rarely has a memory more than five months long! This gives rise to our signal. Imagine a graph of the US stock market. Most investors follow the S&P 500 these days, though in the early part of the century the Dow Jones Industrial Average was the fashionable thing. I`ll stick with the S&P 500 for the first part of the discussion. Now, overlay on that graph a line which represents the highest daily high price in the last 99 days. This sort of skims across the top of the market price, then falls off after a while if it doesn`t get pushed upwards again. The signal is simple: if the last change in that line was upwards, stay in the stock market. If the last change in that line was downwards, sell all your stock and hold cash. (for cash, I`ll assume the usual choice of investors is 3 month treasury bills). Pretty simple. So, how well does this do? We want higher returns and lower risk. But, we have to define our terms. I have a boring appendix* on how I calculated these at the end of this message. But basically the returns are the annual compounded average, and the risk is a single number where 0% is perfect and 15% is really bad. Since 1930, had you invested in the S&P 500 index and received all the dividends from those companies and reinvested them, you would have made a return of 9.51% per year compounded. However, it would have been a very rough ride. My overall risk metric comes out to 11.80%. Is that a good number? No. Zero is a zero risk result on this metric: something that made a perfectly steady 10.0% per year compounded. A risk score of 2% would be a really nice hedge fund, around 5% would be very steady returns, and over 15% is the sort of thing you get investing in internet bubbles. By using the rule that I described above, you would have had a total return of 11.80%, with a total risk of 6.71%. So, you would have had 2.3% per year higher returns on average, while experiencing only about 56.9% of the risk (because 6.71/11.80 = .569). Higher returns, lower risk. That sounds good, right? But, is this one of those iffy timing systems that really only works because it managed to avoid the crash of 1987 of something? No, actually it has you in the market in the 1987 crash: it`s not perfect. But does it really add value on average through the years? Not always, but pretty darned well. Sometimes you do a little worse, but never a lot worse. And when you do better, you do a LOT better. So, overall, it`s worth it. Here is a table of returns for different five years periods. Period S&P 500 Switching Improvement 1930-1935 -12.2% 4.3% 16.4% 1935-1940 10.3% 16.2% 6.0% 1940-1945 7.4% 5.6% -1.8% 1945-1950 10.2% 7.4% -2.8% 1950-1955 23.6% 22.5% -1.1% 1955-1960 15.2% 13.8% -1.4% 1960-1965 10.7% 11.7% 1.0% 1965-1970 5.0% 7.1% 2.2% 1970-1975 -2.4% 6.8% 9.2% 1975-1980 14.8% 12.0% -2.8% 1980-1985 14.8% 20.1% 5.2% 1985-1990 20.4% 18.8% -1.7% 1990-1995 8.7% 6.7% -2.0% 1995-2000 28.6% 27.9% -0.7% 2000-2005 -2.3% 4.3% 6.6% 2005- 2.0% 4.5% 2.5% At first glance, it doesn`t seem to improve things a whole lot. But, look at how steady the returns are. For example, note that five-year periods that things get worse are only worse by -1.8% on average, while the periods that things are better are better by +6.2% on average. Also, note that the times that things get worse, you`re still doing very well. The average return during a 12 month period that the system "didn`t work" (returns were lower than just holding the S&P) are still actually above average at +12.41% per year including dividends. So, in short, on average you`re giving up a little bit of return in the very good years but gaining a very large amount of return in the very bad years, for a small net gain in returns and a very large net gain in steadiness. As one very simple measure of steadiness, the standard deviation of rolling-year returns is 14.9% instead of 19.8%, and with a higher average. (a standard deviation of zero would indicate perfect steady returns). Overall, the system is in the market 73.9% of the time, and has just over one signal per year on average (1.02, actually). So, although it can not save you from a totally random unpredictable market crash, it could save you from some of them, since you`ll be in cash over a quarter of the time. Any crash that happens while you`re holding cash is not going to hurt you. Plus, of course, you are out of the market for a large part of any long decline. The longer it lasts, the more of it you miss, which is why the system works to ensure you don`t have a really bad few years. Of course, it would be nice to test the system a bit further. The Dow Jones Industrials index is older, and I was able to test it before 1930. It seemed to work just fine: For 1900-1929, the average return while it said to be in the market was +8.1% per year (not counting dividends), and the average market return when it said to be in cash was -0.8% per year. From 1930 to 1949, "buy" periods returned +3.5%, and the market returned -9.0% per year in the "cash" periods. From 1950 to 1969, "buy" periods returned +9.5%, and the market returned -0.5% per year in the "cash" periods. From 1970 to date, "buy" periods returned +8.3%, and the market returned +2.8% per year in the "cash" periods. This is a nice confirmation, but I recommend using the S&P 500 "daily high" values. All of these numbers are pretty confusing, so here are a couple of nice pictures to convey the message simply. Here is a graph of the S&P 500 index including dividends since 1930 (the lower line), compared to the money you would have if you had followed this system (the upper line). http://www.stonewellfunds.com/BearSignalLongTerm.jpg Green is "in the market", red is "what the market did during periods the system recommended cash", yellow is the times of cash returns, and the hot pink line is the "99 day high" line that determines if you are in the market (pink rising) or not (pink falling). Sorry for the tiny text, but it`s a long history to get into one pic! Which line would you like your portfolio to be? Here`s what the last couple of years looks like graphically: http://www.stonewellfunds.com/BearSignalShortTerm.jpg This only shows the S&P index, not the portfolio which also gets out of the market. It simply colour-codes the index based on the system`s signals: green for "buy", and red for "cash". The yellow line is the "99 day high" signal line. You can see how it works: if the last move in the yellow line was upwards, the index is green because you should stay in the market. If the line`s last movement is downwards, the index is shown in red because it`s time to go to cash. The system had you in the market for a long time (since November 4 2004), then triggered a move to cash on March 5th 2008 after a gain of 24.1%. Obviously it would have been better to trigger it earlier (nearer the market top last fall), or later (after the bounce in May), but the main point is the long run. Even though it picked a poor week to go to cash, the market has gone down another 9.1% since then---a loss that this system avoided---so over the longer cycle it seems to have it right. In fact this cash period is now guaranteed to be profitable, since the "99 day high" line is below the point that we left the market---we will get back in again lower than where we got out in May, though we don`t yet know when that will be, so we will definitely be ahead of the people who stayed in the market the whole time. Note that you don`t have to calculate this every day. The last point in red is last Friday`s close. If the market rises above the yellow "99 day high" line at any time in the next while, we`ll know it`s time to get back into the market. We know pretty much what that line looks like for quite a while into the future. Plus, once it is back in the market, it is guaranteed to stay there for at least 100 days, so that`s 100 days that you don`t have to even look. Now, I admit that this is not a very sophisticated signal. But, it has the huge benefit of simplicity, and extremely few signals to pay attention to. Every time you hit a new recent high, you`re guaranteed to be in the market for the next several months, and even in times of cash like now, you know about where the market will have to go in future for you to have to look more closely. Right now it`s extremely unlikely that this system will recommend the stock market before the middle of November, and even then it will have to rise at least 8.25% from here for that to happen. What about the choice of 100 days? Is that selected with hindsight? Yes, of course it is. But all sorts of different numbers work just fine to reduce risk and give quite reasonable returns. Only the numbers around 100 days actually improve average returns a bit, though all kinds of numbers from 60 days to 120 days will improve risk-adjusted returns quite a bit. In other words, even if you use the "wrong" number the worst you`re likely to experience is a small reduction in long run average returns in return for a large reduction in risk. Will it keep working? No, not every year. Some years you`ll be a couple of percent worse off. But, you`ll sleep much better, and you will take pity on your friends who blindly stay in the market all through those nasty bear periods like this year. There is a big difference between never trading, and standing on the sidelines just occasionally when things are going steadily downwards. Happy investing! Jim * Appendix: how the risk and return numbers were calculated. Feel free to skip this paragraph---I`m just describing how I measure risk and returns, so you`ll know what I mean by "risk". For returns I`ll use the usual compound annual rate of return including interest and dividends. The "long term buy and hold" approach has dividends all the time, and never any interest. The "switching" approach has dividends when it`s in the market, and interest otherwise. There is no provision for income tax in either scenario. For risk, I use a metric called a downside deviation. This is based on the ideas that (1) you really want to make a certain amount of money every single year---I assumed 10% per year, (2) that any twelve month period which is lower than that is a "failure" because it had a shortfall from your goal, and (3) that a shortfall of twice as much is four times as bad. i.e., getting 2% return (which is 8% less than your 10% per year goal) is four times as bad as getting a 6% return (which is a shortfall of only 4%). So, each doubling of the shortfall from your goal is given four times the penalty. You can ignore the details, but what it means is that I`ve come up with a single number representing risk: it`s the average of the penalties from (a) 3 month periods that didn`t even break even, from (b) 1 year periods that didn`t make 10%, and from (c) 2 year periods that didn`t make 21% (which is 10% per year compounded).