Lately, as my knowlege of investment grows, I’ve been wondering if market timing is always bad. Predicting the next month’s returns will always be virtually impossible due to the number of people pushing markets in different directions, but if investors forget some of the lessons of the last 10 years there may be a time when it’s clear that a market has to revert to the mean.

GMO‘s Jeremy Grantham has some interesting points – I started reading his quarterly newsletter just over a year ago. After many predictions in recent years that the S&P 500 would reach 1100 by 2010, what was once a very pessimistic outlook now seems a little too optimistic (of course with the recent volatility that could be a week away).

Grantham’s quarterly newsletters often talk about which asset classes he believes are overvalued and how this guides his investment decisions. This isn’t the day-to-day technical analysis type of market timing. Instead it relies on a very simple analysis, such as the fact that even after 2000-2003 the S&P 500 was well above the long-term growth trend. Faced with a growing awareness that stock markets can be stagnant for a decade or two, I’ve been thinking about how this could be integrated into an individual investor’s framework.

The first problem is how you can decide if an asset class is overvalued or undervalued. For a stock market index you can look at long-term real return if enough historical records exist (there’s a lot of studies about the US market but I’d like to know more about the canadian market; in other countries there’s not much data at all) and P/E ratios. While there can be market-changing factors – for the last two decades there have certainly been more investors in the market than a century ago, which probably bids up prices a bit – it’s hard to forget the lesson of 2000: fundamentals always count. Of course it’s still hard to know which ones are coincidences and which ones matter. A real return trend of over 100 years should be fairly reliable – if there really is a deviation it should be acompanied by something like a big change in the GDP, although that can be temporary too.

The second decision is what to do if an asset class is overvalued or undervalued. If all equities around the world are overvalued, parking cash in bonds for a long time may not be a gain unless the markets truly are flat and the dividend yields are low. If it’s a bull market or a bubble there may still be gains to be had from sticking to an asset allocation and steadily moving profits into other asset classes until the correction.

Asset allocation may be the key to this idea. One of the reasons to rebalance regularly is that a growth in one asset class increases risk – if you think your risk tolerance is for 60% equities, you don’t want to be at 75% because it’s more risky. If there are obvious signs that a certain asset class is overvalued it would become more risky, so it makes sense to reduce the allocation. If it’s severly undervalued it becomes much less risky, so the allocation could be increased. This only matters in relation to other asset classes though – if everything goes down by 10% there’s no reason to change your allocation. 

This is a difficult decision, but decisions like this shouln’t be made lightly so there’s always time to consider the relative merits of different asset classes.

Even with the extended periods of slow growth that have happened before and will happen again I firmly believe that dollar-cost averaging into a large market index over a long period makes for a very reliable investment. However the history of the markets shows that there are times when the market is obviously inefficient. If there’s a way to make some adjustments for that without paying a higher price it would be very useful.

Have you changed your view of any investing risks recently? Do you think you can know when to change your allocation to get better returns?