It should go without saying that stocks are far riskier than most people assume, and there are no guarantees Investors are drawing the wrong lessons from Japan’s Nikkei 225 Index JP:NIK topping its 1989 high last week.

Many investment professionals claim this proves that stocks eventually produce a profit, if you just hold on long enough. In truth, the correct lesson to draw from Japan’s experience is that stocks are far riskier than most assume, and there are no guarantees.

Most investors ignore Japan altogether when concluding that stocks should come out ahead over investment horizons as short as a decade. There’s a belief that Japan’s multi-decade dry spell is an exception to the rule.

But it isn’t. Seven countries in addition to Japan have, at some point in their century or longer histories, produced a 30-year total stock-market return below inflation, as you can see from this chart.

The exceptional U.S. experience Some investment professionals respond with what I call U.S. exceptionalism — in essence, the notion that the U.S. is different and is therefore immune to the historical precedents from other countries.

But the U.S. has its own experience with stocks failing to produce a gain over not just years, but decades. According to Edward McQuarrie, a professor emeritus at the Leavey School of Business at Santa Clara University in California, in one 75-year stretch of U.S. history — from 1909 through 1984 —the price-only S&P 500 SPX, in inflation-adjusted terms, was no higher at the end than at the beginning.

Even if you believed in U.S. exceptionalism, therefore, you’d have to allow for the possibility that the price-only S&P 500 at the end of this century could be no higher than today in inflation-adjusted terms.

Where’s the risk? In any case, there’s a theoretical reason to question the blind faith investors place in the argument that stocks always win if you hold on long enough, McQuarrie told me in an interview. According to that argument, of course, stocks make money over the long term in order to compensate for their riskiness. But, as McQuarrie asks, where’s the risk if stocks always make money over the long term?

You can’t have it both ways, he says. On the one hand, if stocks held for the long term are indeed risky, then that means there’s a possibility they’ll lose money over the long term. On the other hand, if you deny the possibility of a long-term loss, then stocks’ long-term risk is a lot lower — in which case, from a theoretical point of view, their expected return would need to be a lot lower as well.

“Cushion investment risk with a combination of stocks and U.S. Treasury Inflation-Protected Securities.”

As Japan’s experience teaches, stocks held for the long term are still risky. Consider simulations conducted by Elroy Dimson, Paul Marsh and Mike Staunton, all finance professors at the London Business School. Assuming stocks’ historical average returns and the standard deviation of those returns, they found that there is a considerable probability the stock market can lose over very long holding periods. In one simulation, for example, they calculated that, over a 40-year investment horizon, the probability of an inflation-adjusted loss is 15%.