October 25, 2019

Over the past decade, stock markets have drifted far from their typical behavior in several ways. There is a big opportunity at hand if markets should end up moving back towards normal as they generally have in the past.

This idea is known as “mean reversion” in investing jargon. It sounds complicated, but it’s pretty straightforward: if a measurement has tended to gravitate towards its historical average in the past, then there’s a good chance it will continue to do so in the future. This is especially true if there are sensible “real-world” reasons to expect the mean reverting pattern to occur.

Below we’ll describe a number of measures that have been historically mean reverting for good reason, but that have strayed very far from their typical levels in recent years.  


We’ve discussed this one often, but it’s a very important and dramatic example of deviation from historical norms.

The chart below shows stock valuations, aka expensiveness, for US, international developed, and (with a shorter history) emerging stock markets.(1)

Click for larger version

A couple things jump out. First, US, developed, and emerging valuations have all been mean reverting. They've had huge swings up and down, but after that they've tended to trend back towards middle ground. This is no surprise: investors' moods can fluctuate wildly, and stock prices with them, but valuations should eventually return to a level that makes sense given the economic fundamentals.

Second, US and international valuations have tended to be similar to one another at any given point in time. This also makes sense. In a global and competitive economy where it's easy to invest overseas, relative valuations even out over time as investors seek the best returns.

But this is not happening right now. While US stocks are near the top of their range, on par with the levels seen in 2007 before the global financial crisis, international valuations are nowhere near their pre-crisis levels. The premium being paid for US stocks is completely unprecedented. (The only other period in which valuations diverged to this level was in the 1980s, when the huge Japanese stock bubble temporarily drove international valuations much higher than the US.).

US Corporate Profits

Another measurement that seems intuitively mean reverting is corporate profit margins (which to use a simple definition means "what percentage of companies' income ends up as profits.") When profit margins are high, more competition should crop up to try for a share of those ample profits, which dilutes margins for everyone else. When margins are low, the weakest companies should fold, leaving more profits for those left standing. This is simply how generally free and efficient markets should function.

Indeed, we see that margins have been very mean reverting over time – until recently:

Source: GMO – Click for larger version

Whether margins have undergone a structural shift higher, or whether they will eventually decline to historically normal levels, is a controversial question. But there are compelling reasons to believe that they will revert at least partway to the norm, and if this happens, it will have major implications for US stock prices. (For a detailed treatment of the topic we highly recommend this article).

US vs. International Stocks

This one is really a by-product of the above two, but it’s still interesting enough to merit a chart. Unusually high stock valuations and margins in the US, at a time when the rest of the world has neither, has resulted in huge outperformance by US stocks over the past decade.

This chart shows US stock returns compared to international stock returns over nearly 50 years. A rising line in the chart (colored in orange), means the US was outperforming. A falling line (colored in blue) denotes international stocks outperforming.(2)

Click for larger version

As you can see, US and international stocks have traded leadership back and forth over the years. Although US stocks have outperformed over the whole period, many people would be surprised to learn that for the nearly four decades spanning from the early 1970s through 2008, international stocks actually outperformed US stocks. The entirety of the US market’s historical outperformance since the early 1970s has taken place just since 2008.

As with the mean reverting patterns above, this one makes sense. Various parts of the world have their booms and busts at different times, and investment fads are famously erratic. Over time, this should even out – and it has, until the past decade or so. But the chart suggests that this recent period could merely be an uncompleted half-cycle of the ongoing US-international horse race, albeit an unusually long and strong one.

Value Stocks vs. Growth Stocks

Yet another outlier this decade has been the relative performance between growth stocks and value stocks.

Growth stocks are more expensive valuation-wise because those companies are expected to grow their earnings faster. Value stock have more modest earnings growth expectations and are thus cheaper. Over the long term, value stocks have actually outperformed as investments. The reason is that while growth companies have indeed enjoyed faster earnings growth than the market as a whole, investors have historically overestimated (and overpaid for) how big that growth advantage would be.

Over the past decade, however, value stocks have fallen behind all over the world:

Click for larger version

The dotted line on this chart shows the average value discount, meaning the extent to which value stocks are cheaper than the market overall. As with everything else we’ve discussed, the relative cheapness of value stocks has fluctuated, and mean reverted, over time. A notable extreme was the tech stock bubble, when investor optimism towards growth was so high that money piled into growth stocks regardless of price. That left value stocks with their lowest-ever relative valuations, setting them up to significantly outperform in the years to follow.

Outside the tech stock bubble, the cheapest value stocks have been during this 38-year period is… right now.

A Very Rare Opportunity

During the tech stock bubble in the latter half of the 1990s, valuation measures reached unusual levels – and just kept on going right up until the 2000 peak. That made for a very unpleasant experience for value investors, but it set also up a period of dramatic outperformance once markets started to return to normal.

The whole episode can be illustrated by the experience of GMO, a longtime investment firm that employs a global value/mean reversion strategy. GMO has been very influential on our own investing philosophy, and their approach is a good proxy for our own.

As the bubble grew larger in the late 1990s, GMO lagged further behind – but the future prospects for value investors got better and better. By the time the bubble peaked, GMO’s return forecasts suggested that, if markets did mean revert, global value portfolios would beat a pure US portfolio by a long shot.

The forecasts proved to be on the mark. The bubble started to burst in the year 2000, markets began a long period of mean reversion, and GMO’s asset allocation strategy went on to outperform the US stock market by a very large margin over the following 10 years.(3)

The folks at GMO figured this was a one-time experience. Their head of asset allocation said that they had “always assumed that neither the pain from the 1990s, nor the opportunity set that the pain created, would come our way again.”

But the tech bubble experience has turned out not to be as unique as they expected. The extreme divergences we’ve discussed here have led to another rough period for global value investors, but also to an opportunity that is on par with the one seen in 2000.

This next chart gives an idea of just how big the current opportunity is. It shows GMO’s forecasts for future outperformance starting now and, for comparison, what they were forecasting in September 2000. The chart requires some explaining, but it’s worth taking some time to understand it, because it really quantifies how big an advantage we may be looking at by sticking with the global value approach.

  • “Global equity” describes a global, value-based, 100% stock portfolio.
  • “Multi-asset” describes a global, value-based portfolio that is balanced between stocks and other investments such as bonds.
  • The bars represent the expected annual outperformance (over the next 7 years) of each value-based portfolio vs. a benchmark. The benchmark for global equity is the global stock market, while the benchmark for multi-asset is 60% global stocks and 40% US bonds.
  • The blue bar is the expected outperformance that GMO was forecasting in September 2000, which was indeed followed by years of actual outperformance by the value portfolio.(3)
  •  The red bar is the current expected outperformance over the next 7 years, assuming that all the mean reverting factors we described above fully mean revert.
  • The pink bar is the expected outperformance assuming that things only partly mean revert.

Click for larger version

Right now, a pure-stock value portfolio is expected to outperform the benchmark by 4.8% per year if everything mean reverts, or 4.4% if things only partly mean revert.

The multi-asset portfolios are a better comparison for investors in a moderate risk allocation, as most of our clients are. Here, the value-based portfolio is expected to outperform by 4.4% per year. This is just shy of the projected outperformance at the peak of the craziness in 2000. Even if things only partly mean revert, the expected outperformance is still 3.5% per year.

The likely outperformance to the typical US investor’s portfolio is even higher. Most Americans have much more US stock exposure than is assumed here. Because so many of the extremes have centered on the US (as seen in the first three charts in this article), the value portfolios are expected to outperform US-heavy portfolios by a wider margin than shown above.(4)

To sum it up: if the factors described earlier in this article revert to normal, value-based global investors are likely to outperform a global benchmark by a lot, and a US-heavy portfolio by even more than that.

Will everything revert to normal? There’s no guarantee that it will. But we described how each mean reverting pattern has real-world underpinnings that haven’t gone away. Given this, and the strong tendency towards reversion throughout history, it seems like some degree of mean reversion is the most likely outcome.

The possibility that there will be no mean reversion, and that everything just freezes at the current levels, seems quite slim in comparison. But even in this case, the global value portfolio should outperform the US market because it’s starting out with higher yields.(5)

For the kind of US outperformance we’ve seen to repeat itself, the extreme moves we’ve described above would have to repeat themselves and become twice as extreme. This is hard to fathom.

It seems much, much more realistic to assume that history will repeat yet again and that these factors will return towards normal – at least partway, and maybe even all the way. If this happens, global value investors could be looking at some good times ahead.

1 – The valuation measure is price/10-year average earnings, which has been a pretty good predictor of long-term returns.

2 – Japanese stocks are excluded because their record-setting bubble in the 1980s distorts the underlying pattern.

3 – Source: Morningstar, return for GMO Global Asset Allocation fund and SPDR S&P500 ETF

4 – Source: GMO 7-Year Asset Class Forecast

5 – Source: GMO Q3 2019 Quarterly Letter