Annual Forecasts Are Worthless But Go With 10% Annual Return

Nobel Prize-winning economist Ken Arrow started his career as a weather officer in the US Army Air Force during World War II. He worked with a group responsible for preparing long-range weather forecasts for military leaders. As a trained statistician, he began to wonder if the forecasts were accurate. After examining the old forecasts, he determined that the one-month predictions were no better than random chance. Based on this conclusion, Ken sent a message to the commanding general asking to discontinue the practice. Here’s the response he received:

“The Commanding General is well aware that the forecasts are no good. However, he needs them for planning purposes.”

Well, the weatherman advising General Eisenhower suggested a delay which led Ike to push the Normandy invasion back one day. D Day went forward on June 6, 1944. The weather was choppy and a number of landing crafts sank.

We won the war anyway.

Now Let’s Think About the Markets

You can understand what I am about to argue better if you think about coin flips. Betting on a single flip of an honest coin has odds of 50/50. To take the bet you need someone to offer you better odds. Whether to take those odds or not is a matter of knowing yourself and your situation. How much will you be damaged if you lose? How much will you gain if you win? Should you just sit on what you have? That’s probably the approach that most people should follow with their investments.

I should caution that the 50/50 assumption is misleading. With an honest coin the more flips you make the closer the percentage converges upon 50/50. That’s not what happens in the markets. The market coin is not honest. One internet site reverse-engineered long term results and concluded that the one-day odds were either about .53% positive for one period or 0.54% for another. I must add that the statistical approach on the site seemed dubious and those numbers seem intuitively too high. One possibility is that down days were down more than up days were up. That’s a factor in many efforts to sort out what the market tends to do.

One thing to keep in mind is that the market operates on its own schedule. It doesn’t have any special attachment to one-year intervals. Trends often turn around in the middle of a year. Major surprises also sometimes come in the middle of the year. The recent huge run-up in interest rates and upward inflation surprises took place after we were well into 2022. Who could have guessed the month?

So what’s the average annual return on stocks? That’s an easy one. Most sources say it’s around 10%. That’s 10% nominal, before inflation. Guess what? Most forecasters are predicting a 10% return for 2023. Why is that? Well, that 10% is what most of us think the market owes us. It must be pretty close to what is actually going to happen, right? If you are asked for a forecast, it’s the number least likely to make you look bad. Everybody else seems to be saying 10%, with Barron’s jumping in on December 17. If Barron’s does nothing else, it informs you of what the consensus is on any given subject. It enables you to avoid looking silly by hiding invisibly in the middle of a large crowd.

Now let’s look at a few more numbers. In his wonderful book The Psychology of Money Morgan Housel had this to say:

The historical odds of making money in U.S. markets are 50/50 over one-day periods, 68% in one-year periods, 88% in ten year periods, and (so far) 100% in 20-year periods.”

Housel’s point was to press home the argument that the way to compound wealth in the stock market is to stay in the market for a long time. That’s my approach as well. I pay very little attention to any single year. My defense against what may happen in any single year is that I have a diversified portfolio of stocks bought at good prices which always come back fairly quickly from a rough period. For this article, however, I am willing to talk about a few things that you might pull from Morgan Housel’s numbers.

Let’s start with that average 10% return. It’s a little like the guy (women have too much common sense to do this) who stood with his feet in two buckets, one filled with ice and the other with boiling water. Asked how it felt, the guy said that on the average he was comfortable. Quite a few articles on market returns have noted that prices seldom go up close to 10% in any given year. What Housel’s numbers tell us is that in about one year out of three the market actually goes down. But how much does it go down? And when? And what does it have to do to get back up to that mythic 10%. (For a year it’s 11.1%).

That’s the kind of stuff that makes forecasting so hard. Let’s just hypothesize that in years that the market does down, it goes down 10%, and also, being optimists, that the down years are spread out in a neutral way, by which I mean that it won’t go down 50% or so in either the first year or the tenth year. In either case that would be a disaster for 10-year return and is the exact sequence risk that causes professional advisors to warn people facing imminent retirement to become conservative.

Even if distributed more favorably three 10% down years in a decade knocks about 30% off of the total return that has to be made up by better years. Calculating in a linear fashion (without considering the compounding factor) stocks need to pick up about 130% over the seven up years. What we realize at a glance is that if the market goes down 10% in 32% of the time over a ten year period, it must go up just under 20% during the period when it is rising. Don’t even think about what happens if the market drops 30% in either the first or last year. That’s a milder version of what happened to people in the year 2000 when the result was that many of them decided the solution was to load up on real estate with high leverage. Real estate bought with debt became the core of the next crash.

The market puts you in the position of Sisyphus. You push a heavy stone endlessly up to the top of the hill and then the stone comes crashing down with a thud. At least the long term stats show that it lands on a higher piece of ground than the last time it crashed. Unlike Sisyphus, the market does make net progress in the long run. Chart readers can see that markets tend to go up steadily at a modest rate and fall more rapidly so that they look like a sine curve wrapped around an upward trend line. Business and market cycles aren’t that exact but that’s often the way they are depicted The 10% annual average gain is thus the slope of the trend line.

Where Does That 10% Number Come From?

The famous Roger G. Ibbotson-Peng Chen 2002 study of the stock market from 1926 to 2000 was the first rigorous study of what stocks have done over the long haul, and it became the model for what stocks should be expected to do in the future. It continues to exert a major influence. The numbers they came up with started with a real return of 8% to which one added the long term inflation rate. The nominal return this added up to was 10.7%. This suited sell-side analysts many of whom simplified it to 10%. There were a few small catches in the Ibbotson-Chen research including the positive impact of survivor bias and the fact that neither corporate profits nor dividends keep up with GDP growth. Effects like this are not as small as they initially seem. The most important catch, however, was simply the beginning and end dates of the study.

The choice of the beginning and end dates was simple and straightforward. Before 1926 Ibbotson and Chen were unable to find reliable statistics. The end date, naturally enough, was for the last year for which they had statistics before publication. For Ibbotson and Chen this was random enough. It apparently didn’t occur to them that there was anything special about those dates. The first non-random factor was that in the year 2000 the U.S. market had the highest valuation it had reached in its entire history. It was much higher than in August 1929 just before the Great Crash caused the Dow Jones Industrial average to fall 90%. After the dot.com Crash, which began in, yes, 2000, the market has not approached the 2000 valuation. The closest it has come was late 2021 when the current bear market began. Only Japan in 1989 exceeded the 2000 valuation. That’s enough to call the 8% real/10.7% projected average return into question, but there’s another, less obvious factor but more powerful factor.

The years 1926 and 2000 nicely bracket the period on which the United States was rising to economic dominance in the world. Along with the rise of American economic and military power, the U.S. stock market was undergoing a tremendously important change in character. Before the 1920s the market had gone up and down with little net progress. Stocks were a vehicle for speculation, not a means for achieving growth. When asked what the market would do, J. P. Morgan is said to have replied that it would fluctuate. He was totally serious and quite literal. You bought stocks for dividends, although bonds were the major vehicle for long term investors.

This carefully constructed 2002 paper by Rob Arnott and Peter Bernstein, despite its wonky and somewhat misleading title calling it a study on the proper “risk premium,’ debunks the outcome of the Ibbotson Chen study. It starts with what the rational expectations of investors in 1926. At the time there was no history of upward trending in the prices of stocks. The U.S. Bond yield was 3.7% and the dividend rate for stocks was 5.1%, which was more or less the total expected return. The next 75 years were filled with happy surprises. A modest risk premium compensated for the fact that stocks were risky. Stock yields did not fall below bond yields until 1958.

What common stock investors actually got for the 75 years starting in 1926 was probably a little less than the 8% real return (10.7% in nominal), pretty close to what Ibbotson-Chen calculated, but those are two of the more misleading numbers in economic history. They occurred over a period in which absolutely everything went right. Over that period governmental agencies cleaned up the problems of “robber baron” capitalism in which corporate leaders had previously diluted their shareholders. Government policies gradually developed a framework for managing a variety of economic catastrophes. In an interesting side note Arnott and Bernstein mention that valuations in the 19th Century were below those in the 20th Century despite the fact that the overall economic growth rate of the U.S. was higher in the 19th Century.

Until WWI the United States still had some resemblance to what we now call an emerging market. In the 75 years starting with 1926 it survived a global Depression, emerged victorious in World War II with relatively few casualties and no destruction of capital assets and was well positioned to pick up the chips from both allies and enemies which were severely damaged. It also emerged as a more or less unchallenged military power. The dollar became the world’s reserve currency with all the advantages that come with that status including the ability to borrow as much as it wished and pay for important resources like oil in its own currency. It came to lead the world in the rule of law and respect for capital. Love us or hate us people from almost every other country wanted to come to the U.S. and enjoy the benefits of being American. Beginning in 1982, Arnott and Bernstein noted, the valuation of U.S. equities was rerated upward so that having traded at 18 times dividends in 1926 it traded at 70 times dividends in 2000.

This huge upward rerating of all things American over the 75 years from 1926 to 2000 has just one sure consequence: it’s an event that can only happen once. It can’t be repeated because so many positive things have already happened to transform the U.S. from being one country among many to a positions of global preeminence. The best case is probably that things stay more or less the same. It’s entirely possible, however, that some aspects of what is now almost 100 years of growth may run in reverse. One element that cannot be overlooked is slowing population growth which has provided an important chunk of economic growth. That’s without any falling off of U.S. leadership in other areas.

Despite the ups and downs of bull and bear markets, the U.S. as a whole remains priced for perfection. As late as March of this year the price/dividend ratio of U.S. equities was slightly over 70, the same as in 2000. The ongoing bear market, which looks very much like a massive valuation correction in tech, has dropped the price/dividend ratio to about 55. Rising interest rates were the trigger but not the underlying cause. The concern might be that it may be the beginning of an overall rerating. That 10% number, in short, is about the past. It is not necessarily the future

Is 2023 A Year For Mean Reversion?

When markets go strongly up or down over extended periods they tend to overshoot. Sometimes that overshoot is worth a tweak in your portfolio. The trouble is that even if you see a strong argument for mean reversion it’s hard to tell when it is going to take place. Two years ago I thought U.S. stocks were expensive compared to international stocks and especially emerging market exactly two years ago. I was wrong. I still think emerging markets were cheap then and I think they are cheaper now. Emerging markets went nowhere and U.S. stocks kept rallying for about a year before techs began their harsh rerating. International stocks and especially emerging stocks were awful until the last couple of months. The cause was strength in the dollar brought on by unexpected world events. As they say of home run hitters in a slump, they “are due.” Will 2023 finally be the year?

Another mean reversion currently in play is Value over Growth. This started a little over a year ago, but it may go on for a while. Like international and emerging, value stocks and growth stocks often flip leadership from decade to decade. My best guess is that value continues to lead. It’s what I mainly own anyway, but in the form of GARP – Growth At A Reasonable Price. There is, however, something you have to bear in mind before dismissing either Growth or the market as a whole. The economy and corporate profits chug along with moderate growth and within a year or two of improving profits begin to reduce the degree of overvaluation. The total market right now seems fully valued at best, but if economic strength emerged some time in 2023, valuations might start to look better.

The Growth side of the market has a somewhat tougher hurdle, but for companies with persistent above-average growth it’s possible to make up a lot of ground in very few years of good results. Right now the start-up companies that were popular a couple of years ago look like a boneyard. Most of them are unlikely to bounce back. The FAANG behemoths probably have a better shot but they too look a bit like wounded elephants which may or may not be able to restore their earlier vigor. I’m not sure they can add much to the promised 10% this year.

The final area of potential mean reversion in 2023 is in small caps. Many are not identifiable names but as a whole their growth keeps up with large caps. If the muddle-through economic environment continues, they might begin their own seven to ten year period of leadership. If the economic situation worsens, their lower valuations might hold their price action to modest losses. This asymmetry seems favorable to emerging markets, value, and small caps for 2023.

Remember That 2023 Is Part Of A Longer-Term Future

A ten year period is probably the sweet spot for predictions. It’s long enough for fundamental factors to be the major thing driving the outcome. If you are a professional forecaster, especially an older one, it is also long enough to remember the John Maynard Keynes quip that “in the long we are all dead.” Nobody is going to stand up at your funeral and call you out for being “wrong about the market in 2023.” You will have had plenty of time to walk your wrong opinions back and insist that if it hadn’t been for a few unexpected events you would have been right. I’m speaking here of market savants like the ones you read in the financial press and watch on CNBC, but we all do it occasionally, including me.

Start by thinking about that Morgan Housel number that the market rises 88% of the time over 10 years. That doesn’t mean that it achieves a run of great returns over that period. Ten-year periods can produce the kind of minuscule positive results that make many investors get out of the market for the rest of their lives. This was true of many ten year periods starting around 1966 when stocks were moderately expensive and ending in early 1980-82 when high inflation had driven stock valuations down to single digits. Corporate profits had chugged along just fine, but investors refused to bid them up even to modest double digit valuations. The saying is that a bear market “scares you out or wears you out.” The 1970s wore many investors out – all the way out of the market. By 1982 I couldn’t talk my best friends and family into buying stocks. That was in fact one of the best five times in a century to load up on stocks.

Right now most quantitative analysts, myself among them, see stocks as selling at valuations which put probable 10 year returns in the low single digits. This has improved slightly with the current bear market and could improve a lot if the bear market ended with an intense selloff. It could also improve if the Federal Reserve panicked, cut rates, and greatly expanded its balance sheet, perhaps the worst possible thing for the long term. What is the probability of their doing that? I don’t have a clue, but they have done it for forty years, more or less, and now see the consequences.

The coin used for the next ten years is now weighted modestly in your favor. If you throw in a decent risk premium, bonds may be highly competitive with stocks simply because of how poorly they have done over 2022. Bonds always have a better probability of bouncing back from bear markets than stocks do. Except for highly risky junk bonds, they have very few factors to take into account except interest rates. The current coupons on Treasuries of all maturities make them worth consideration. The last thing I bought on a large scale was Treasury Notes. I expect them to do reasonably well in 2023 and for at least three or four more years.

The most important thing to remember about next year is that it’s the first year of the rest of your life. What you do in the next year is important because at the end of it you will have established the base from which future returns compound. It’s always important to have goals along with some tentative working model of the way things around us will unfold. The trouble is that for some of the important factors we just don’t know.

What About 20 Years? Will It Continue Its Winning Streak

Twenty years is about the event horizon for most of us, beyond which we can’t see anything or bring ourselves to care. The first two decades of the 2000s flipped from bad to good and still didn’t hit the promised 10%. It would have been worse for our lifetime results, though, if the first of last 50% crash had happened on year one or year 2019. If about to retire and fully invested you would have seen your lifetime savings drop by 50%. Many did. Avoiding catastrophes like that must therefore be your most important goal.

While I have very moderate enthusiasm about the next ten years, I think they may well be followed by a great decade if valuations get down to attractive numbers by 2033 or earlier. That may well happen. If it does so, the major stock indexes may take off again as they did in the 2010s and had done earlier in the 1980s. There’s a chance my children and grandchildren may get the kind of returns we now tell ourselves that we are entitled to. Meanwhile the most important investment goal is survival. Don’t be ruined in such a way that even great future returns can’t help you much. The best way to do this is diversification of risks by holding enough cash and bonds. Bonds are now an alternative which provides decent return while possibly diverging from more pricey stocks. I Bonds and TIPS are also good diversifiers as is income-producing real property. The next twenty years should always be in the back of your mind and beginning of a new year is a good time to remind yourself.

Conclusion

Here’s my detailed stock market forecast for 2023: stocks will end the year with 3% real return plus or minus 15%. I want to lean to the positive side although my own portfolio positioning is neutral. That’s my age and personal psychology. It also matters where the current year ends up. Here are very tentative sectors and categories:

• Value will continue to outperform Growth. My confidence level is 60 on a 100 scale. Betting on Value and being wrong does less damage than betting on Growth and being wrong.
• Small caps will outperform large caps. My confidence level is 65 on a 100 scale. It’s about valuation.
• Emerging markets ex-China, Russia, Taiwan, and Korea will outperform U.S. stocks. My confidence level is 70 on a 100 scale. Excluding China and Russia is obvious. It saddens me to have to include Taiwan and Korea because they fall within the radius of unpredictable Chinese aggression.
• High quality bonds will outperform stocks. My confidence level is 65 on a 100 scale. Bonds will be helped by their higher coupon return and a fair likelihood of a price rally. I already bought but may add from cash and maturing T Bills and Notes.

I don’t expect to sell anything in my stock portfolio and will not add unless the market takes a meaningful hit. Have a happy and healthy 2023. Good luck on your investments.