Consider the following hypothetical, and then ask yourself what type of investment returns you think you could make in such a scenario and how it would alter your trading and investment habits.
Consulting with a Hypothetical Market Seer
Pretend for a moment that I introduced you to a true and authenticated “stock market seer.” Assume that this seer has demonstrated an astounding ability to make incredibly accurate predictions about future stock market returns over the 12 to 24 months following each of the seer’s predictions. Assume further that there is no funny business going on with this seer, and such seer’s track record was both publicly available and fully verifiable over the last 40+ years. You can rely on this seer’s predictions to accurately foretell future events.
The Seer’s Predictions.
The seer will offer an accurate, reliable prediction any time you like and as many times as you like. And (again, for sake of the hypothetical) the seer’s prediction will turn out to be true with virtually 100% accuracy. However, the seer view of the future is not perfectly clear, and it comes with two important limitations.
- Can Only Predict Prices Exactly Two Years from Now. The seer can only make accurate predictions regarding the price of the S&P 500 on the date that is exactly two years after the date of prediction. In other words, the only thing the seer is able to foretell relates to the price of the S&P 500 on the day that is two years into the future. The seer cannot accurately foretell any market gyrations in the interim.
- Can Only Predict General Price Movements. The seer can only tell you generally about future prices. Specifically, it can only foretell about one specific situation – namely, what will happen if you buy the S&P 500 on any particular day and sell it exactly two years in the future. In such circumstances, the seer can accurately foretell only whether there is (i) a near 0% chance of you losing money (i.e., a profitable trade) or (ii) whether there is an approximate 50% chance of you losing money (i.e., a bad trade). In essence, on any given day the seer will foretell the future by giving you either a thumbs up or a thumbs down regarding what your two-year returns will be.
Now assume that each time you want to buy into the stock market, you first seek a prediction from the seer. The seer will reply with only one of the following two statements:
Statement #1 – “Buy in peace”:
“On the date that is two years from today, the price of the S&P 500 will be the substantially the same as or higher than today’s prices. Go and buy in peace; you will not lose money if you buy today and sell in two years.”
or
Statement #2 – “Buyer beware”:
“Buying today is extremely risky. On the date that is two years from today, there is a substantial probability that the price of the S&P 500 will be lower than today’s prices – and perhaps much lower. While you have about even odds of making a positive return, you are swimming in dangerous water. If you buy today and hold for two years, the odds that you will lose money is at least as high as the odds that you will make money.”
Hence, the seer can accurately foretell the risk associated with your investment at any time. And, by consulting with the seer, you can foreknow whether you are virtually certain to avoid a loss (i.e., a great time to buy) or whether your odds of a gain or a loss are essentially a coin-flip (i.e., a suboptimal time to buy).
Heeding the Seer’s Predictions Would Lead to Substantial Profit.
What type of investment returns do you think you could make if you acted on the seer’s predictions and each prediction turned out to be 100% accurate? And how would this alter your investment strategies?
By knowing when downside risk was extremely minimal, you could not only go long at the right time; you could also potentially use options to capture additional high-probability, low-risk upside. And if you knew that the risk of substantial market declines was high, you could sell off a substantial portion of your holdings to avoid capital losses and wait until the risk/reward was more favorable to re-enter. And beyond loss avoidance you could also make positive returns through the use of put options.
In brief, by heeding the seer’s advice, you could outperform the market, making bigger gains and at lower risk. You’d be invested in the good times and on the sidelines during the uncertain times.
“But no such Stock Market Seer exists, so why the pointless hypothetical?” you might be thinking.
If so, I have great news for you. Such a seer does exist, is easily accessible, and is ready to make free and highly accurate predictions for you at any time you choose. Who is this seer? As you might have guessed, it is the yield curve.
The Yield Curve as Market Seer – Useful for More Than Just Recession Forecasting
The difference in yield between the 10-year treasury constant maturity and the 13-week treasury note is widely followed by the market and is noted for its ability to predict recessions a year in advance. The Federal Reserve has published a number of papers noting the strong correlation between an inverted yield curve and a recession starting within about a year following the inversion.
The yield curve, of course, is not usually inverted; an inverted yield curve represents an extreme outlier relative to normal treasury spreads. Most investors seem to place relatively little importance on the yield curve in “normal” times. I believe this is a big mistake.
As I will argue below, the spread on the yield curve has substantial predictive power of future stock market returns over the next 12 and, especially, 24 months. While the yield curve cannot forecast future 12- and 24- month returns with high precision, it can substantially narrow the range of returns that are likely over such periods. Perhaps most importantly of all, it can make accurate predictions about the probability of negative future returns. Specifically, at least on a historical basis, an investor could have accurately forecast when the probability of substantial negative future returns was high and when such probability was virtually non-existent – all using nothing but the yield curve.
Take a quick moment to let that sink in, and then read on.
Understanding How NOT To Use the Yield Curve
Though a basic point, it is worth understanding which yield curve metrics are predictive – and which are not. For instance, the yield curve spread today holds almost zero predictive power about market returns over the next 12 or 24 months. Take a look at the charts below, which are taken from programs I’ve written to study this topic. They illustrate this point clearly.
Before diving into analysis of the charts below, I first want to briefly describe how to read them.
- The X-axis of the chart represents the treasury spread (i.e., yield on the 10-year minus yield on the 13-week) on a monthly basis for each month over the past ~40 years.
- The Y-axis represents the aggregate price appreciation in the S&P 500 over the subsequent 12- or 24-month periods.
- 0 reflects 0% total stock price appreciation.
- Positive numbers reflect positive gains. For instance, 0.2 reflects a 20% total price appreciation over the subsequent period.
- Negative numbers reflect losses. For instance, -0.3 reflects a -30% total capital loss over the subsequent period.
- Dots are color-coded by year (legend on the right) so that the time-series nature of the data is more easily seen.
- The black dotted line represents a zero return (i.e., neither positive nor negative).
- The first chart plots 12-month returns, while the second chart plots 24-month returns.
12-Month Stock Market Returns vs. Current Yield Spread
No predictive power
24-Month Stock Market Returns vs. Current Yield Spread
Little, if any, predictive power
Interpreting the Charts
As you can quickly see, these two charts offer little more than a random scattering of colorful dots on a page. While the 24-month chart does appear to show some positive correlation between the treasury_spread and returns over the following two years, it appears weak and does not appear to be terribly helpful in loss avoidance. Note, importantly, that there appears to be little, if any, relationship between the treasury_spread (x-axis) and the subsequent return in the stock market (y-axis). Hence, if given the current treasury_spread, one could not make an accurate prediction of future returns in any meaningful way. Any such indicator would be highly unreliable, likely to be inaccurate, and not much better than random chance.
But wait – we’re doing it wrong. Recall that the yield curve’s power to predict recessions is forward-looking. Its predictions today are accurate forecasts of what will begin to occur one year in the future, not what will begin to occur tomorrow.
Let’s reexamine the exact same data above in that light.
How to Harness the Yield Curve as Market Seer
I will reproduce below the same charts as you saw immediately above, except this time I will change the x-axis from today’s treasury spread to the treasury spread one year ago. This is key, because the yield curve is forward-looking. Today’s yield curve predicts future events, not today’s.
You can read the chart in the same way as described above for the previous charts. Note that the charts – particularly the two-year chart – transform from random dots into an identifiable pattern.
12-Month Stock Market Returns vs. Previous Year’s Yield Spread
Some predictive power
We now begin to see a linear relationship between the treasury spread one year ago and the market’s return over the next year. Though there appears to be a wide range of potential returns at any given yield spread, we notice the obvious and important relationship – namely, that the larger the yield spread, the higher the “floor” on bad returns becomes. Put another way, we can begin to see that if we want to avoid investing during times where extremely negative returns are most likely, we should avoid investing when the yield curve is either inverted (i.e., a negative number) or where the spread is not very large (i.e., 1.5% or less).
But can we obtain even better predictions by extending the return period to 24 months? Yes.
24-Month Stock Market Returns vs. Previous Year’s Yield Spread
Substantial predictive power
The chart above demonstrates remarkable predictive power. As you can see, in almost every single circumstance where the treasury spread was greater than 1.5%, the market made positive returns over the succeeding 24 months. Though there are a small handful of data points where market returns were negative 24 months after the referenced treasury spread was above 1.5, they are far outweighed by the enormous number of data points where a positive return was achieved (i.e., those lying above the dotted black line).
Similarly, and perhaps even more important, the chart also shows us that, over the last 40 years, every single time the price of the S&P 500 declined by more than 20% over a two-year period, it was preceded by the referenced treasury spread being 1.5 or lower. Put another way, the treasury spread sent out an advanced warning signal that the probability of substantial negative stock market returns was high prior to every single market decline of 20% or more. Every. Single. Time. This is your Stock Market Seer.
At Least Historically Speaking, There Are Times Where You Are Virtually Guaranteed to Make or Lose Money Over a Two-Year Hold Period
The historical data shows that, at least in all measured past periods, there are times when, based solely on the applicable yield curve spread, you could identify a wide range of times where you were virtually guaranteed to earn a positive return in the market using a “buy today and hold for two years” investment approach. Similarly, there were also times where the reverse was true – the same “buy today and hold for two years” strategy would be virtually certain to make a negative return. Take a look at the charts below. These charts show the odds, historically, of an investor earning a positive or negative return over subsequent one- and two-year periods based solely on the spread of the yield curve.
While the charts above are just in bar chart format, they scream out about just how powerful the yield curve spread is in predicting longer-term (one- and two-year) returns on literally any trading day over the past 40 years. Would it surprise you to know that, when the spread in the yield curve was -0.5% or lower (i.e., the rate on the 3-month bill was 0.5% or more higher than the rate on the 10-year bond), you were virtually guaranteed to lose money in the market if you bought that day and sold two years later? In fact, out of all of the historical times where this has been true, the S&P 500 has had a negative return in 97.9% of such instances and a positive return in only 2.1%. You can see that it’s not much better even if the spread is a positive 0.5%, as you would historically still lose money more than half of the time in such circumstances.
But just as remarkably, the chart also shows that, when the treasury spread is greater than 1.5%, there was almost no circumstance when you would have lost money using a “buy and hold for two years” strategy. In fact, you would have made money in 99.5% of those instances and lost money in only 0.5%. Even more remarkable, during the ~40-year measurement period, you would have NEVER lost money if you “buy and hold for two years” on any day where the treasury spread is 2.0% or greater.
So, again, there you have it. The yield curve can be your stock market seer, easily, and in real-time, identifying times when the odds of a positive two-year return are stacked strongly in your favor and other times where the odds of a negative return are very high or even virtually certain. If you had this information available to you, would you consider the use of options to take more risk and try to capture upside (whether calls from expected market gains or puts from expected market losses)?
So Where Are We Today?
What does the “yield curve seer” predict about market returns if one were to buy the S&P 500 today, February 8, 2022, at a price level of about 4,520, and sell two years from today on Feb 8, 2024? Are we all clear? Or is there potential risk ahead?
To see what the yield curve predicts about the hypothetical return one would achieve by buying today, Feb 8, 2022, and selling Feb 8, 2024, we must look at the treasury spread one year ago (i.e., 2/8/2021). The spread on that date was 1.14, which puts us to the left of the dotted red vertical line in the chart above. This means we are on the wrong side of the red line; we are within the “risky” zone where our odds of a negative 24-month stock market return are material.
To put a finer point on it, using history as a guide, we can say that one-year returns have been negative approximately 25% of the time when the spread was between 1.0% and 1.5%. Our odds of negative two-year returns are materially lower, at only about 10%. This implies that, though we have a very good chance of a positive two-year return, there is a substantial risk of shorter-term market volatility and negative shorter-term returns. Hence, it is not unreasonable to believe that the odds of a material market dip from today’s levels justify staying on the sidelines and waiting for market prices to dip further.
And, at least historically speaking, it is not out of the question that we could see substantial declines – perhaps even as much as 20-30%. While returns that negative are relatively unlikely, there is a material chance that market prices will be lower in one year than they are today and a small-but-not-insignificant chance that they will be lower in two years as well.
To put some numbers to just what our probabilities of various return thresholds are, consider the information below.
- There were a total of 548 trading days in the last ~40 years where the applicable spread on the yield curve (i.e., the yield curve spread one year prior to whatever date is being evaluated) ranged between 1.04% and 1.24%, inclusive.
- This range was picked because it is +/- 0.10% of the applicable spread for today, which, as stated above, is 1.14.
- When looking at the total stock market return over the subsequent 12 months of each of the above-described 548 trading days, the price of the S&P 500 declined more than 25% of the time.
- The median one-year return for these same 548 trading days was only 5%.
- The worst decline for any trading day in this range was -35%.
- Over a two year period, the stats are better. Note, however, that these are far inferior to the expected return profile for dates where the yield curve spread is greater than 1.5%.
Based solely on this metric, now may not be a good time to go long. I’m sure not going to do so. Other indicators that I follow, which I may detail in future posts, also suggest that now is not the time to be fully invested in the market; each of these indicators reflects that the chance of a material market decline are high.
But if you are entering the market using a “buy and hold two years even in the face of volatile pricing in the interim” then history suggests that your odds of a positive return are reasonably high.
About Brad Grounds
Brad Grounds’ bio is here.
