As the time horizon expands, uncertainty increases because the range of possible outcomes widens as we look further and further into the future.
So much for theory and common sense.
Today we hear that the long run in stocks is less risky than the long run in bonds; therefore, the Dow should now be at 36,000, perhaps even higher. Therefore, asset diversification is nonsense, something for wimps, but surely not for truly knowledgeable investors.
All those investors holding bonds and perhaps even a little cash are not only excessively risk averse - they are stupid! Stocks should have the lowest risk premium of all assets.
The alarming assumption in this argument, which has been receiving increasing attention in respectable publications, is that we know more about the long run than we know about the short run. This view is both fallacious and dangerous, as I shall demonstrate.
First I invoke Gottfried von Leibniz. In 1703, the great mathematician Jacob Bernoulli wrote to Leibniz that he found it strange we know the odds of throwing a seven instead of an eight with a pair of dice, but we do not know the probability that a man of 20 will outlive a man of 60. Bernoulli suggested he might be able to calculate such a probability by examining a large number of pairs of men of each age.
Leibniz was profoundly skeptical: "Nature has established patterns originating in the return of events, but only for the most part . . . No matter how many experiments you have done on corpses, you have not thereby imposed a limit on the nature of events so that in the future they could not vary."
Leibniz emphasized "but only for the most part" because of its overwhelming importance. If it were "always" rather than "for the most part," there would be no uncertainty. No model has an R-squared of 1.00. We cannot prove a hypothesis; on the contrary, we can only attempt to falsify it. That is why I emphasize that the experience of the past 75 years is only suggestive, not definitive of probabilities - we can only guess about the future.
Yet these projections of the Dow hitting 36,000 and above allow no error: they tell us we know more about portfolio values 20 and 30 years from now than we know about what values will be tomorrow or next year.
Once we recognize the models that support these projections have an R-squared of less than 1.00, we have to accept the possibility that the optimism about Dow 36,000 is misplaced. When the R-square is less than 1.00, the crucial element in the decision is the consequences of being wrong.
Suppose you put 60% of your portfolio into stocks and 40% into bonds because you bet that stocks are in fact risky over the long run. Suppose you are wrong. Your fortune will still grow over time as stocks continue to work their miracles over the long run, even if not to the sky. But suppose you put 100% of your portfolio into stocks because you bet that stocks are less risky over the long run. Now suppose you are wrong. Good bye!
The return of events - a replay of the patterns of the past 75 years of capital market history - will happen only for the most part. Most is not all. There is no certainty. Rational people do not bet the ranch on a model that works out only for the most part. And God forbid it works out only for the minor part! Consequences, not probabilities, determine the decisions that matter. Diversification is still the optimal strategy for the long run.