Advanced concepts: market "timing" using valuations; how to incorporate very strong convictions into your investment portfolio allocations.
Overall I would discourage attempts to time or finesse your target weightings. Once you've decided your risk tolerance, generally it is best to stick to it, and rely on rebalancing to see you through drawdowns.
However, there are rare occasions where the is a clear valuation discrepancy in the markets. I definitely do not subscribe to the pure efficient market school of thought. In my opinion, it is an indisputable fact that on occasion asset markets sometimes go to absurd speculative excess valuations; equally, markets also periodically go to panic-driven crisis lows, with incredibly attractive values based on any rational economic measure. Perhaps the most obvious example of a valuation discrepancy is the 1999-2000 bubble in the S&P and technology stocks. Japanese stocks in 1989 were another example - NTT, a single telecoms company, at the time was worth more than the entire Australian stock market. On the cheap side, the Asia crisis caused incredible undervaluation in those markets, ditto the Russian default in 1998. The Brazil and Argentina financial crashes at the start of this decade caused a similarly extreme level of distressed value. US stocks in 1932, or UK stocks in 1974 (where the FTSE was at a P/E of 3.8 and a dividen yield of 12%), were perhaps the most egregious examples of sound blue chip stocks virtually being given away
In each of these cases, no rational observer could conclude anything other than that the valuations were way out of whack, driven by insane sentiment excess or panic-driven risk aversion at absurd levels. Therefore, I personally do not think it makes sense to maintain target allocations if one of your asset classes ventures into either insane bubble territory, or ridiculously cheap fire-sale levels.
My solution to this is simply to maintain target allocations, but then to slowly reduce (or increase) them, only if the asset class gets to truly ridiculous bubble/bust levels. You can actually manage a lot of this simply by shifting from the growth portfolio to the conservative portfolio. As we saw, 2000-2003 was actually slightly positive for the conservative portfolio. However, I personally would have made stocks an even lower allocation in 2000 than the conservative portfolio's 35%. I think it would even have been acceptable to have no stocks at all, although that is quite an extreme position.
If you are going to follow this "bubble avoidance" approach, fading valuation extremes, I would give the following advice. First, make sure it really is a genuine bubble/bust, and not just a slightly expensive or slightly cheap market. Stocks were pricey in 1996, 1997, 1998 and 1999 - selling out then would have been very costly. Asia was cheap in 1997 but still cratered for 1 more year. In the early 1990s, US stocks were expensive on a P/E basis, fooling many into selling. But earnings were at recession multiples, thus the true P/E was misleading - once earnings bounced back, stocks become cheap quickly, and had already moved higher. In 2000 by contrast, earnings were at cycle peaks after a multi-year boom, margins were at record levels, corporate profits as a % of GDP, and market cap as a % of GDP were all at historic highs.
Second, scale out (or in) slowly, in steady increments. Don't go from 50% to 0% overnight, like some value managers did in the mid to late 90s. Go from 50% in 1996 to 45% in 1997 to 40% in 1998 to 35% in 1999 and finally 20% or less in 2000 when things went nuts.
Third, check sentiment. Is it really a raging bubble, with taxidrivers being multimillionaires, your grandmother raking it in? Or is it just a typical bull market? Do not try to time normal bull markets. Stick to bubbles only.
Fourth - try to estimate long-run returns using valuation metrics. For example, at a P/E of 30-35, the earnings yield of the S&P in early 2000 was 3%. That is not exactly a great long-run return for risky stocks, when long-term US treasuries were yielding 7% with much lower risk, and long-term TIPS were yielding 4% *plus inflation*, with virtually no risk at all. It should have been no surprise that bonds performed very well from 2000-2003, whilst stocks got annihilated in the worst collapse for a generation. In 1989-1990, real estate yields were 4-5% - pathetic compared to stock earnings yields, or bond yields at the time. Sure enough, real estate sucked donkey balls in the early 1990s.
So, that's my advice for incorporating valuation - use it to reduce exposure and risk during the blowoff period of true valuation bubbles. And you can use it to try to make Buffett-style deep value purchases in truly depressed markets and asset classes. Just avoid "all or nothing" bets as much as you can, make your shifts gradual and conservative, and remember that most of the time, being long investment assets at your target allocations is the best play. Don't be one of those people who sold stocks in 1990 and then never bought back again, still waiting for the next Great Depression and making 2% in T-bills. Remember, even the conservative portfolio has 6% expected returns, and superb bear market protection.
Overall I would discourage attempts to time or finesse your target weightings. Once you've decided your risk tolerance, generally it is best to stick to it, and rely on rebalancing to see you through drawdowns.
However, there are rare occasions where the is a clear valuation discrepancy in the markets. I definitely do not subscribe to the pure efficient market school of thought. In my opinion, it is an indisputable fact that on occasion asset markets sometimes go to absurd speculative excess valuations; equally, markets also periodically go to panic-driven crisis lows, with incredibly attractive values based on any rational economic measure. Perhaps the most obvious example of a valuation discrepancy is the 1999-2000 bubble in the S&P and technology stocks. Japanese stocks in 1989 were another example - NTT, a single telecoms company, at the time was worth more than the entire Australian stock market. On the cheap side, the Asia crisis caused incredible undervaluation in those markets, ditto the Russian default in 1998. The Brazil and Argentina financial crashes at the start of this decade caused a similarly extreme level of distressed value. US stocks in 1932, or UK stocks in 1974 (where the FTSE was at a P/E of 3.8 and a dividen yield of 12%), were perhaps the most egregious examples of sound blue chip stocks virtually being given away
In each of these cases, no rational observer could conclude anything other than that the valuations were way out of whack, driven by insane sentiment excess or panic-driven risk aversion at absurd levels. Therefore, I personally do not think it makes sense to maintain target allocations if one of your asset classes ventures into either insane bubble territory, or ridiculously cheap fire-sale levels.
My solution to this is simply to maintain target allocations, but then to slowly reduce (or increase) them, only if the asset class gets to truly ridiculous bubble/bust levels. You can actually manage a lot of this simply by shifting from the growth portfolio to the conservative portfolio. As we saw, 2000-2003 was actually slightly positive for the conservative portfolio. However, I personally would have made stocks an even lower allocation in 2000 than the conservative portfolio's 35%. I think it would even have been acceptable to have no stocks at all, although that is quite an extreme position.
If you are going to follow this "bubble avoidance" approach, fading valuation extremes, I would give the following advice. First, make sure it really is a genuine bubble/bust, and not just a slightly expensive or slightly cheap market. Stocks were pricey in 1996, 1997, 1998 and 1999 - selling out then would have been very costly. Asia was cheap in 1997 but still cratered for 1 more year. In the early 1990s, US stocks were expensive on a P/E basis, fooling many into selling. But earnings were at recession multiples, thus the true P/E was misleading - once earnings bounced back, stocks become cheap quickly, and had already moved higher. In 2000 by contrast, earnings were at cycle peaks after a multi-year boom, margins were at record levels, corporate profits as a % of GDP, and market cap as a % of GDP were all at historic highs.
Second, scale out (or in) slowly, in steady increments. Don't go from 50% to 0% overnight, like some value managers did in the mid to late 90s. Go from 50% in 1996 to 45% in 1997 to 40% in 1998 to 35% in 1999 and finally 20% or less in 2000 when things went nuts.
Third, check sentiment. Is it really a raging bubble, with taxidrivers being multimillionaires, your grandmother raking it in? Or is it just a typical bull market? Do not try to time normal bull markets. Stick to bubbles only.
Fourth - try to estimate long-run returns using valuation metrics. For example, at a P/E of 30-35, the earnings yield of the S&P in early 2000 was 3%. That is not exactly a great long-run return for risky stocks, when long-term US treasuries were yielding 7% with much lower risk, and long-term TIPS were yielding 4% *plus inflation*, with virtually no risk at all. It should have been no surprise that bonds performed very well from 2000-2003, whilst stocks got annihilated in the worst collapse for a generation. In 1989-1990, real estate yields were 4-5% - pathetic compared to stock earnings yields, or bond yields at the time. Sure enough, real estate sucked donkey balls in the early 1990s.
So, that's my advice for incorporating valuation - use it to reduce exposure and risk during the blowoff period of true valuation bubbles. And you can use it to try to make Buffett-style deep value purchases in truly depressed markets and asset classes. Just avoid "all or nothing" bets as much as you can, make your shifts gradual and conservative, and remember that most of the time, being long investment assets at your target allocations is the best play. Don't be one of those people who sold stocks in 1990 and then never bought back again, still waiting for the next Great Depression and making 2% in T-bills. Remember, even the conservative portfolio has 6% expected returns, and superb bear market protection.
