Index Profits: No Sure Bet

A commonly held belief among some investors is that most five-year holding periods for the S&P 500 are profitable, and that 10-year holding periods are essentially all profitable. We are reluctant to accept such rules of thumb without verification, and prefer to reject them until proven true.

Our study of S&P 500 rolling periods since 1926 shows that only three in four five-year periods are profitable, and that only about seven in eight 10-year periods are profitable. In fact, over the 82-year period (1926-2008) studied, even 20-year holding periods were profitable in only 19 in 20 periods. Twenty-five-year and 30-year periods were profitable all of the time. (See charts below.)

While the probabilities favor a profit over a loss for all holding periods, the risk of losing money over a multiyear holding period is still quite significant.

For young people in the early accumulation stage of life, the risks are not very important because of the dollar-cost-averaging aspect of their periodic savings, and the fact that they will not be drawing on their portfolio or be relying upon it for their standard of living for many years, perhaps decades.

However, for those investors with mature portfolios—meaning those who have essentially completed their asset accumulations, who cannot replace losses with new earned income and who do or soon will rely on their portfolios for their standard of living—10 to 20 years is too long to wait for a portfolio to break even.

Importantly, even for those who do not have the need to rely on their portfolio, we never met an individual who had the psychological attributes that would enable them to remain committed to a portfolio that lost money for multiple back-to-back years, without exiting substantially or totally, statistics and probabilities notwithstanding.

For these reasons, we believe that the statistical assurances from asset-gathering organizations are to a certain degree in disregard of both the life span of investors with mature portfolios, and the financial needs and emotional composition of those investors.

It is common, but incongruous to hear advisers recommend "sell until you can sleep" on the one hand, and on the other hand recommend against moving assets in and out of stocks based on long-term cyclical patterns of gain and loss.

Of the 27 loss years within the 82 years from 1926-2008, there were 12 clustered in multiyear losses:

Sometimes the losses in individual years are so severe that they have a long-lasting carryover effect on consecutive multiyear rolling returns. For example, the five consecutive seven-year periods from 1933-1937 were negative. Just imagine yourself measuring your investment returns and remaining fully invested when each of the past five periods of seven years were net losses for you. It would be the exceptionally unusual person who would not have bailed out long before that point.

There are more back-to-back rolling periods of loss than many might think. In quite recent times, there were three consecutive two-year rolling periods of loss from 2001-2003; three consecutive three-year rolling loss periods from 2001-2003; and four consecutive five-year rolling periods of loss from 2002-2005. These are examples, of which there are more.

The 1970s had several similar bad spells, as did the periods following the Great Depression and WWII. For example, there were four consecutive 10-year rolling periods of loss from 1937-1940. There were also two consecutive 10-year rolling periods of loss from each of 1974-1975 and 1977-1978.

If 2009 is a loss year, that would be a two-year loss period. The recent 2000-2003 loss was a three-year period, and the 1973-1974 loss was a two-year period. Is the current situation better or worse than in those two periods? Is the near future brighter or dimmer than the near future in 2000 or in 1973?

Is the fact that we are in a global growth slowdown and a nearly global asset deflation more or less of a problem than we experienced in the last two multiyear loss periods? Do the concerted efforts of the U.S., Chinese and other governments moderate the impact of the current economic problems to the extent that recovery will be swifter than in the periods beginning in 2000 and 1973?