Valuations like this have always lead to disaster.
Ben Graham, the father of security analysis, said that the stock market is a voting machine in the short run but a weighing machine in the long run. That machine weighs sales, net worth, and sustainable earnings, and the needle takes a decade to settle. When speaking about diversified indices, it is the market’s mood swings that make equities risky, not the underlying businesses, which in aggregate produce a steady and perpetual stream of cash flows.
To handicap future prices, we need to know the current fundamentals, and the distribution of historical valuations paid for those fundamentals – what is average, what is cheap, and what is expensive. When we look at 100+ years of stock returns in the US and elsewhere, we can see that valuations have a strong tendency to revert to the mean. It may take a few years, but mania and panic alike will always subside.
Once you forget about where prices might go in the near term, you can weigh competing investment options against the next decade of returns for stocks. Perhaps US stocks are priced to pay 7% annually and you’re considering real estate with a 4% cap rate. Take the stocks. Maybe stocks are priced for 5% and the 10-year Treasury is 4.5%? If volatility is a concern, consider the Treasuries. It’s really that simple if you keep your head. But as a group, investors are not hyper-rational econs, the decision-makers of economic models, but beings who rationalize. Justifications come after price moves, and catalysts are only clear in retrospect. The proof is in the numbers – you need know nothing at all about the events of the next 12 years to forecast the annualized rate of return to within a couple of percentage points.
As of summer 2017, the US stock market has reached a level of valuation from which it has always fallen by at least half. This bull market started from a modest valuation in March 2009, implying inflation-adjusted returns of about 6% for the coming 12 years, which would double its value. Eight and a half years hence, it has returned in excess of 15% per year, almost quadrupling, despite book value and dividends increasing by roughly half, GDP by about 20%, and sales and earnings by nothing at all. For every measure of fundamentals, investors are paying multiples that have only been paid briefly near major market tops like 1929 and 2000. Lured by rising prices, they have pushed the market into a hyperbolic spike. It is foolhardy to think that their mood will not change, or that anything other than mood affects stock prices on a time scale salient to your life.
(Above chart is my own, but all data is from multpl.com)
Valuation indicators simply set fundamentals like those above in the denominator and price in the numerator (price/book, price/sales, price/earnings). John Hussman provides the below chart of several such historically-reliable indicators. Note that they all agree, and that valuations are now as extreme as at the top of dot-com mania (1929 was similar, though this chart doesn’t go back that far). You may say that high valuations are the new normal, since we have been elevated for most of 20 years now, but consider that from the more muted overvaluation of the 1960s, the market still fell and required 20 years to recover. Furthermore, the rest of the world’s stock markets are far less expensive than the US today (we stand apart like Japan in the late 1980s, and the Nikkei remains far below its old high).
Below, Hussman has overlaid actual, realized 12-year nominal returns (not adjusted for inflation) onto one of the above valuation metrics, market cap / gross value added. This is about as tight a correlation as you will find in finance, or anywhere in the natural world – future returns are practically the inverse of present valuations.
Here’s another from Hussman comparing 12-year nominal returns to total US market cap / GDP, going back to 1928. This is sometimes called the Buffett Indicator, because Warren Buffett cited it in 1999 as proof that stocks were in a bubble.
Here’s my own chart of the Shiller PE (aka CAPE) *normalized for profit margins vs subsequent 12-year real returns (inflation-adjusted), going back to 1929. When adjusted for inflation, returns can be negative for 12 years after bubbles of this magnitude. The bottom chart stops in 2005, since that was 12 years ago. To see what the S&P 500 is likely to do from 2017, compare the top chart to the bottom – when the CAPE is high, subsequent returns are low, and vice versa.
Robert Shiller’s CAPE works reasonably well, but when margins are very high or very low the picture gets distorted, as corporate profit margins are one of the most mean-reverting statistics in finance. Un-adjusted CAPE today is 30% lower than in 2000 since margins have been very high of late but were below average in the 1990s. Adjusted figures bring CAPE into agreement with other established metrics, showing that stocks are as expensive now as in 2000 and 1929, implying 0% returns between now and 2029.
*Margins normalized using corporate profits after tax as a fraction of GDP.
At present every reliable metric predicts that the market will go nowhere for a decade or longer. 10-20 years of stagnation is the norm from valuations like this. The last 140 years shows that stocks are indeed for the long run, since their progress is anything but a smooth line once inflation is backed out. If you look carefully and keep in mind the time scale, you can see long stretches where stocks have gone nowhere, the secular or generational bear markets. They have always followed periods of elevated valuation, since it takes many years for fundamentals to catch up. Prices have never simply plateaued at such levels without crashing, because it is the feedback of rapid price increases themselves that sustains the mania. Once a hyper-valued market stalls, whether it is stocks, bitcoin or frozen concentrated orange juice, there is nothing to support it until it becomes attractive to value buyers who wait far below.
Here, again from Hussman, is what returns over the following four years have done from various valuation levels. There can be some upside, but it is always wiped out by a major loss within four years.
Once you step back and forget about the immediate future, there is no mystery about what the market will do, and for the next decade or more it will return practically nothing.
You can’t just download an app, sync it to your checking account, and have all your investment goals met, unless those goals are 40 years away. If you would like to pay for a home, education, or retirement on a more immediate time scale, you will have to make allowances for volatility. This can be done by (1) diversifying across many asset classes, (2) scaling in and out of stocks according to valuation, or (3) using risk controls that cut losses in a falling market. Tail hedging with put options is another strategy that has gained some respect of late, though I won’t discuss it here. Below I lay out the basics of the first three options, all of which are simple to implement.
Part 2: Better options
Note: all of the below simulations are generated with Portfolio Visualizer.
Option 1: Diversification
A portfolio of loosely or inversely correlated asset classes produces a perfectly adequate long-term return with far greater stability. Maximum losses are reduced from over 50% to less than 20%, and recovery times are far shorter. This kind of strategy is very simple to implement with low-cost ETFs and annual rebalancing.
In red is Harry Browne’s permanent portfolio (25% each in stocks, Treasury bonds, Treasury bills, and gold). Blue is a more efficient allocation (25% US stocks, 15% foreign stocks, 15% gold, and 45% Treasury bonds). Yellow is the S&P500.
Option 2: Valuation-based market timing
This is another simple strategy that only requires attention once a year. You check a historically-reliable valuation indicator and set your equity and bond allocations accordingly, scaling into stocks as they get cheap and out as they get expensive. This demonstration uses the Shiller PE, but reliable indicators all work alike so take your pick.
Blue is valuation-based timing. Yellow is 60% stocks, 40% bonds. Green is just bonds, and red is just stocks.
Option 3: Quantitative Risk Controls
Every competent trader knows to “cut your losses short and let your winners run.” There are any number of ways to implement this, but perhaps the best-known is to go long when a security breaks above a moving average and to sell or go short when it crosses below that average. This system typically results in many small losses, a fair number of small wins, and a few big wins. When applied to a stock index using a long-term moving average in the range of 6-15 months (10 is most common), it keeps you out of big trouble and catches most major rallies. It has the advantage of being agnostic about valuation, so that you don’t have to sit on your hands while your neighbor keeps booking gains late in a bull market. When the market turns, you’ll give up some of those gains, but rarely all of them. However, you can still get caught in sudden crashes like 1987.
This system is 100% in the S&P 500 when it’s above its 10-month moving average (based on monthly closes), and in Treasury bonds when below.
Bonus: Combining diversification and risk-control
Here you assemble a portfolio of diversified asset classes, but instead of just rebalancing once a year, you check each month if they are each above their respective long-term moving averages. If not, you sell them the next day and put the cash in Treasuries until they are in an uptrend again. This produces a very respectable long-term return with a ridiculously smooth growth curve, thus a high Sortino ratio. Drawdowns are short and shallow, so portfolios like this can be levered according to risk appetite.
The only strategies that I would recommend to DIY investors are those that require attention just once a year – options 1 and 2, valuation timing and static diversification, the latter being the easiest to implement. Strategies that require monthly attention or the use of moving averages require a bit more conviction and diligence. With any quantitative strategy, the hardest part is not the calculations, but following through and placing the trades exactly as prescribed, something that is hard enough for professionals, let alone individuals.
Unfortunately, there are only a handful of retail advisors out there who are interested in the science of portfolio management. Incentives are such that being different is more risky than losing money, since employers and clients focus on recent performance in comparison to benchmarks, to the detriment of long-term performance and stability. Independent-minded professionals tend to gravitate towards hedge funds, where they have more leeway and better remuneration. However, for the independent advisor who wants to provide better service to his clients, these kinds of strategies may be implemented efficiently with off-the-shelf software and low-cost ETFs. This is one of those times when doing the right thing requires you to step away from the crowd. If you are thinking ahead, it is an extraordinary opportunity.