One of the key facts repeated by the investment industry (although I’m sure some people selling mutual funds say it under their breath) is that past performance doesn’t predict future results. Although there are some time periods that are a bit more predictable, there’s uncertainty at every level from the random day-to-day variations to the long-term trends shaped by the current environment. The reasons vary at different time scales – recently I came across a new reason for long-term uncertainty.
Clearly short-term moves in the next 1-2 years can easily go in any direction. In the 7-10 year range it’s a bit easier to form a guess on what’s likely to happen, but the end result will still be unpredictable. Over 20-100 years people generally expect significant economic growth.
The recent “A Night With The Bears” event highlighted one speaker described his theory about “long waves”, or economic cycles that are about 60 years long. According to this, we would be at the start of a long decline or stagnation. Although there does seem to be more predictability in longer time periods and nothing is immune from cyclical effects, the idea that we’re reliving the past isn’t right either.
One of the obvious reasons is one-time events. Many people speak of the “post-war” economy that’s developed in North America and Europe in the last 60 years. That particular course of events was most likely heavily influenced by the Great Depression and World War 2, so we’re already on a different course. That’s not the only reason things will turn out differently - there are also fundamental changes in the nature of the markets.
While reading Geroge Soros’ book The Crisis of Global Capitalism, I noticed that one of the points he brings up repeatedly is how the market eceonomy has changed significantly since 1980. I already knew that the last 20-25 years have been an unusual and very long bull market for stocks, which had to result in either a decline or a reduction in future returns since stocks were growing at an unsustainable rate. Add this to some historical studies that show a long-term real return of around 6% from stocks and you can guesss the specific adjustments.
However with the new information from this book it seems that the bull run may have simple explanations, rather than being a random variation that has to return to the mean. According to Soros, the policies of Regan and Thatcher in the early 80s played a big role in opening up markets to free competition and innovation with less regulation. This change may have been a major factor in the rise in stock prices and the eventual speculation and instability.
Although it doesn’t explain everything, being able to put the recent bull market into context has interesting implications. It’s also a great demonstration of why it’s difficult to predict the future. Everyone knows that over the centuries the dominant countries, empires, and cultures change drastically. But there can also be big shifts in a times as short as a decade or two. The specific changes are different each time, making the future highly uncertain.
Current events could result in more regulations and lower returns for some time before market participants figure out the best thing to do in light of recent lessons. Or quick fixes could encourage people to restart an unsustainable bull market, taking stock prices even higher before a change is forced on them. What’s most worrying is that there’s still a possiblity that people as a whole will forget common sense and lessons from the past and take a big step back to policies that restrict economic growth for a long period.
The long-run trend in stock prices may not be affected much when the last 20 years are combined with the next 20, but that’s still a historical fact rather than a fundemental law. In The Black Swan, Nassim Taleb mentions that if you look at a stock chart with the dates removed it’s hard to tell whether it’s for one day or 5 years (you can even turn it upside-down and get the same effect). This illustrates how unpredictable markets are on different levels but if you want to understand what’s actually going on and possibly profit from it you need to know the different reasons.
In the long-term we can make predictions about where current pressures will drive the markets, but we have to remember that the markets we’re talking about are always changing. Trying to predict the future based on the past at any level is likely to be a losing game because the conditions and participants are always changing. Understanding the past is a great way to keep up with future developments though.
April 23, 2009 at 10:36 pm
Trying to predict the future based on the past at any level is likely to be a losing game because the conditions and participants are always changing.
I think that things are more predictable than this comment suggests, Silicon. But I also think that your words here are one of the best arguments I have read as to why perfect predictability will never be possible.
The thing that suggests long-term predictability to me is that the long-term return in the U.S. has been about 6.5 percent for a long time. It always comes back to that. That suggests that there is some force that is bigger than the short-term developments that pulls things back to a long-term norm.
I certainly don’t say that there is any law of the universe that says that yesterday’s norm will always apply. It could be that the norm is in the process of changing. But my thought is that we should start with a belief that a return to the norm is the most likely long-term outcome. That should be the default position. Then we should take into consideration factors that might pull us from the norm but always with a bit of skepticism thrown in.
I think it’s a mistake both to overestimate predictability and to underestimate predictability. I think that some view it as “safer” to assume no predictability. But assuming too little predictability can get you into as much trouble as assuming too much predictability. The goal should be to get it as right as possible (which if difficult, to be sure).
Rob
April 24, 2009 at 12:57 am
You make some good points. Some things are consistent, such as getting a higher return from stocks than from bonds. They could be interrupted temporarily but there are good reasons for them to return. The problem is mainly with taking market cycles and assuming they’ll repeat themselves consistently – you could call it long-term technical analysis. In this case I believe the trend towards economic growth and the value of investing in good businesses will override the similarity to past cycles, although it’s good to remember that we’re only talking about the markets that have survived this far.
April 24, 2009 at 11:37 am
Some things are consistent, such as getting a higher return from stocks than from bonds.
I hope you won’t consider it argumentative if I offer a link making a bit of a case otherwise:
http://www.indexuniverse.com/publications/journalofindexes/articles/149-may-june-2009/5710-bonds-why-bother.html
I agree that there are good reasons why stocks should GENERALLY offer a better long-term return than bonds. But I also think that it depends on the price at which stocks are selling. When we get into one of those time-periods in which millions come to believe in Passive Investing, the price of stocks is going to go so high and the long-term return on stocks is going to go so low that the less risky asset class (bonds) will offer the better long-term return. Human emotion is capable of overcoming the logical “requirement” that risk be correlated with long-term return.
you could call it long-term technical analysis.
“Long-term technical analysis” is a good term for what I believe in. I believe that timing MUST work because changes in valuations obviously affect long-term returns and timing is just a way to take advantage of that reality. But I am personally convinced that short-term technical analysis does not work because there are just too many variables that need to be taken into account for the investor to be right often enough to be able to “beat the market.”
The edge that you get with long-term timing is that you don’t need to buy at the precisely right time or sell at the precisely right time. You just need to get things generally right to be able to “beat the market” because the market is pricing stocks pursuant to emotion and you are pricing stocks pursuant to the economic realities (you always factor out the emotional part of the market price in forming your assessment of whether stocks offer a good value proposition or not).
Rob
April 25, 2009 at 3:45 pm
Of course there are sometimes temporary changes – but if stocks stop offering higher returns than bonds eventually many people will stop investing and drive down the price. This is one of the reasons that it tends to be a consistent fact over a long enough time period. It’s interesting to see that in this analysis there were indeed very long periods of stock underperformance – it’s possible that they’re based on a starting time when bonds are strongly underpriced and stocks are strongly overpriced. I don’t know if this is enough to change my view that stocks should be the foundation of a long-term portfolio, especially since the extensive research in Stocks for the Long Run shows different results. The bond investing method in this article isn’t clear (in particular, whether reinvestment is done by choosing the 20-year bonds that expire soonest or the new ones that will expire in 20 years) – it seems very limited and may be a very specific method that happened to give the best returns for bonds during those periods.
I would say that what you’re talking about and what I hope to do is simply fundamental analysis. There are certain basic facts that can’t be ignored – if an investment would take 50 years to return the money put in I’ll never be interested. Thinking in terms of 60-year cycles is more like technical analysis because it tries to predict where the market will go based on where it’s gone in the past using similarity to other time periods rather than the realities of investment. In that sense any analysis is partly technical, but the ones based on something constant are more likely to assert themselves given enough time. Technical analysis only works when it happens to match the fundamentals – such as guessing that any market that triples within a year is likely to come back down.
I agree that if you need to buy and sell at exactly the right time you’re likely to fail. Over sufficiently long time horizons you may be able to do this more easily, but I would prefer incremental adjustments rather than buying and selling everything. This helps break up the decision into many pieces and factors in the uncertainty (a bubble could keep rising for a few more years, or high P/E ratios could turn out to be justified – although it’s not as likely for a whole market as it is for a single stock).
April 25, 2009 at 3:59 pm
but if stocks stop offering higher returns than bonds eventually many people will stop investing and drive down the price. This is one of the reasons that it tends to be a consistent fact over a long enough time period.
That sounds right to me, Silicon.
I don’t know if this is enough to change my view that stocks should be the foundation of a long-term portfolio, especially since the extensive research in Stocks for the Long Run shows different results.
I also share your view that stocks should be the foundation of most people’s long-term portfolio. And I am not at all crazy about the conventional type of bonds as my alternative to stocks. I prefer Treasury Inflated-Protected Securities (TIPS) for times when stocks do not offer an appealing long-term value proposition.
I have mixed feelings about “Stocks for the Long Run.” It’s a wonderful book that offers many insights. However, it does not explore the effect of valuations on the long-term returns. I view that as a big failing.
I would prefer incremental adjustments rather than buying and selling everything.
I strongly agree.
Thanks for engaging in the back-and-forth, Silicon. It’s helpful to learn where other smart people are coming from re these matters. I have hopes that there will be many more discussions of these sorts of matters in coming days.
Rob
April 25, 2009 at 4:06 pm
Real-return bonds are a nice addition to anyone’s investing options
It’s possible that they could underperform regular bonds at some times but they offer the ultimate security as long as the government can measure inflation.
Thanks for your comments – it’s always good to see a different view on this.