The Single Greatest Stock Market Predictor Is… What?

Last time we waved CAPE around at our aging equities bull market and formed some tentative conclusions about what current CAPE readings mean for long-run future stock returns.

We also identified some shortcomings of the CAPE approach. It does provide useful forecasts of future returns. But our confidence in this single measure has to be muted since there are lingering concerns about how CAPE deals with certain accounting issues, evolving payout ratios, market structure and investor sentiment about “paying for earnings.”

These factors limit how much emphasis we can put on CAPE’s predictions – which already may statistically explain a little less than half of long-run stock market performance, with a fair amount of play in the prophesy.

So now we toss our CAPE into the dirt and get friendly with a testy statistic that has perhaps twice the explanatory and predictive power of CAPE. And which can help us refine our CAPE-based predictions.

And actually we’re gonna get more than just friendly with this statistic. We’re gonna get downright infected with it. With HEP, that is.

 

Getting Tested

There’s some good evidence that HEP is twice as powerful a predictor of long-run market performance as CAPE. And HEP isn’t susceptible to some of the limitations of CAPE, meaning its predictions on their own may be more infectious than even PE10. So we’ll talk about what HEP is and what it can tell us.

But most importantly we’ll see that, when we combine forecasts using the HEP measure with those of CAPE, we can say some very strong things about future returns and what we can expect from our portfolios going forward.

In part, that’s because the two measures view things so much differently. We’re talking Western medicine versus Eastern medicine here. Diametrically different. And yet incredibly powerful when combined.

HEP is our acronym for average “household equity percentage.” And just like its name implies, HEP gives us a measure of what share of household finances are invested in equities. And that figure can be used to render some statistically solid conclusions about the long-term future of stock returns.

The figure is calculated by taking the dollars held by U.S. households in stocks and dividing by the quantity of household wealth held in stocks plus bonds plus cash.

Which implies a few moving parts. First, you’ve got a mobile target in the total household financial holdings denominator, which fluctuates as invested dollars gyrate with the markets and as household savings slosh around. And then you’ve got the invested sum numerator, which can be altered on the basis of both market movement and increased (or decreased) household allocations to equities.

Which makes HEP kinda funny. It’s got more wobble built into it than a Jell-O mold. And so, yes, there’s lots to not love about HEP.

But that’s just the best part. It gets worse.

For one, there’s a foreign investor problem that I won’t bother discussing. But it’s there and there’s no real easy way to deal with it. So just be aware of that.

And you’ll also notice a conspicuous absence of any mention in the HEP description of stock valuation, volatility, interest rates, inflation rates, dividend yields, earnings or, well, pretty much anything else that has any academic claim to explaining stock market activity on an intellectual basis.

Which is weird since HEP can predict around 91% of long-run returns in the S&P 500. Oh, and that’s going back to 1951. Oh, and the relationship between HEP and returns has arguably tightened over time. Oh, and that’s your jaw on the floor. You should pick it up.

To recap: HEP has the ability to explain more than 90% of variation in long-run stock market returns. CAPE can explain around 45%.

Here’s An A, Two Bs, And One C

Which leaves us nonplussed.

What do we do when we see explanatory power like that, coupled with questionable logical underpinnings?

As an econometrician, I’d have told any dim-eyed analyst who came to me with this garbage to kindly get the f out of my office and go pick up my lunch. (I’m a jerk.) Because stuff like HEP violates the very first rule of economics club, which is: No matter what the statistics say, there’s gotta be a good logical/economic argument for why it works – or else it’s probably crap.

Which is another way of saying: If you back-test enough crap, you’ll eventually find something that’s statistically amazing…for a while.

Except, hmm…

More than 65 years is quite a while. And there actually is some homespun wisdom baked into its numbers that makes HEP hard to ignore once you get it: As soon as everybody’s doing something, it just ain’t cool anymore.

Which is a nice way of saying that, at a certain point, the marginal investors plunking money into equities are almost sure to be the “greater fools.” It’s the same effect seen in how fools spend their money – they follow the herd. So once we see loads of “me-too” dumb money pouring into equities, we can be pretty sure the party’s about to get lame. And long-run future returns are going to be below average.

It works like this: As soon as every Joe living on a cul-de-sac somewhere catches wind of this great thing called a stock market that’s making everybody rich, Joe decides to take some of his cash and fire up the trusty E-Trade machine. And so the share of Joe’s household financial portfolio in equities goes up – usually just around the time equity prices have already gotten peaky. (Joe likes buying expensive stuff.)

As more and more Joes inject the E-Trade machine with ever increasing doses of cash, the average household equities share climbs. And at some point, the Joes run out of free liquid assets to funnel into stocks. Or: Household aggregate demand for equities isn’t able to shift outward anymore. And so the dumb money effect stalls. Which causes the price action that got Joe all excited in the first place to stall. Which means…

At around the moment stocks become least likely to provide the greatest future returns, they’re most loved by Joes and other cul-de-sac suckers. And vice versa.

Which gets us now to the feedback effect built into HEP. It goes like this: While individual investors are price takers, aggregated investors are price makers. If there’s a shift in investor sentiment toward (away from) stocks, prices will rise (fall) as a consequence of aggregated household investment demand trends. And this effect is most intuitively pronounced at the margins, when HEP is much above or below average.

And all this has zero to do with valuation. Instead, it has everything to do with price action itself and human psychology and expectations biases and risk aversion and FOMO and other stuff Danny Kahneman probably didn’t know he already knew.

A Little HEP Here

Which now brings us to the part where our HEP infection doesn’t get cured. It becomes fiercely untreatable. Because what makes HEP maybe a little stupid on its own is exactly what makes it un-stupid when coupled with more traditional valuation metrics.

What makes HEP stupid on its own: It ignores valuation and rationality and prices and earnings and yields and everything else reasonable and informed investors care about when buying/selling stocks.

What makes HEP smart when paired with traditional valuation metrics: It addresses the behavioral and psychological components of household investment decisions that are essentially ignored by traditional valuation metrics and forecast tools. And it provides an alternative perspective on equities price action that, it turns out, is rooted in good economics.

And that, patients, is simple supply and demand.

Which means HEP is based in something arguably economically defensible (i.e., household investment budgets and behavioral finance biases). So it’s like all the aggregated savings, employment, corporate investment, consumer sentiment and Fed minutes data balled into one aggregate-digested clump of everything-ness: What regular Joes do with their money is pretty much the wrong thing to do at exactly the wrong time. And we should be especially worried/excited when all the Joes seem to agree with each other.

As in: The way most people interpret all the signal and noise out there about managing their finances, over the long-term, is exactly stoopid. And we should zag while they’re busy zigging all over themselves.

If you’re interested in reading more about some of HEP’s underlying mechanics, you can head over to Philosophical Economics, where this HEP thing first infected patient zero. Note that, while I have healthy respect for Philosophical Economics, I don’t much love the post about HEP. It reads like a Unabomber manifesto, and it’s less measured than a cheap suit off the rack. So, um, you know. Proceed with caution.

Nevertheless, Philosophical Economics deserves props. And so props are given. Because this is an interesting phenomenon that, despite my best efforts to hate HEP, I just can’t help but think it’s a pretty smart finding in need of a better explanatory argument.

For additional background, Ned Davis Research has taken up the HEP thing for subsequent study and periodically generates updated materials about HEP. At a high level, the research implies continued robust predictive power of HEP.

So let’s get fevered and sweaty as we explore HEP’s figures, forecasts and a little bit about how it works.

Super HEP

HEP compares, on one axis, “average investor equity allocation percentage” with subsequent 10-year S&P average annual total returns, on the other axis. When the equity allocation percentage increases, that implies lower future S&P returns, and vice versa.

Historically (on the HEP measure), U.S. households have held an average of about 45% of assets in equities. Right now, the allocation percentage is around 53%. Which is pretty high. It’s not the highest ever, but it’s hovering around fourth or fifth place over the last 65 years.

Ned Davis Research updates the analysis periodically, and you can find various versions on the web, including at Advisor Perspectives. And, for purposes of critical commentary and thus under the auspices of fair use, I’ll show a recent Ned Davis Research / Advisor Perspectives graph below:

HEP - Ned Davis Research / Advisor Perspectives
HEP – Ned Davis Research / Advisor Perspectives

Let’s just appreciate the closeness of those two lines. But let’s also note some instances of divergence. In the mid-1980s, HEP forecasted 18% annual returns, but only 14% average returns were realized over the next decade. And, in the run-up to the Mortgage Crisis, HEP was forecasting returns of just 3% or so. And yet actual returns were more like 7% in the following 10 years. The 1950s also look loose.

So there’s some of that Jell-O wobble every once in a while. Which is why we maintain healthy skepticism and temper our consideration of HEP by using it with CAPE.

As for investments today, what does HEP tell us? The above chart was done mid-year 2016, and it’s still relevant now (though HEP has increased slightly in the last 6 months). In general, it tells us we can expect average annual returns of around 3% from the S&P over the next decade.

Note, though, that HEP provides nominal total returns estimates (i.e., dividends reinvested, not adjusted for inflation) rather than real total returns like those given by CAPE. Which means if we make any adjustments for expected inflation over the next decade – say an average 2% per annum, in accordance with the Fed’s target valuewe’d anticipate total real returns according to HEP of around 1%.

Which is in line with our qualitatively textured CAPE forecast. From CAPE, we anticipate real returns in the 3% to 4% range, but with downside risk of something like zero real return over a decade. Our HEP forecast of 1% or so is smack dab in the middle of that range. So, with HEP’s help, we’d now refine our 10-year estimate of the S&P’s annual average real returns to something like this:

“It’s unlikely we’ll do better than 3% or 4%, and there’s a darn good chance we’ll do worse – on the order of 1% or so.”

Now, I won’t go on defending the intellectual merits of HEP. It’s an interesting phenomenon. It’s got some nice statistics to support it. And there’s some not-quite-horrible theoretical support, though I think it needs a better argument than it’s got. But just because it doesn’t have an ivory tower lineage doesn’t mean we should ignore it. (And in the text above there are a handful of academic studies linked that look at HEP with positive findings.) After all, the Buffett Indicator doesn’t really make tons of academic sense, yet old Warren digs it.

Moreover, if we use CAPE as the academically-oriented long-run predictor benchmark, we note that HEP has historically generally agreed with CAPE, a metric that helped win Shiller a Nobel prize. And, actually, HEP has been better in virtually all instances at rendering superior forecasts, while, in instances when the two predictors have disagreed, HEP has been right(er) and CAPE has been wrong(er). But when they do agree…well, look out. And, yeah, now they agree.

Getting Around

So it’s an interesting moment for equities. Whether we look at arguably the best “academic” model of forecasting (i.e., CAPE) or at arguably the best “non-academic” model (i.e., HEP), we’re left scratching around the cul-de-sac wondering what we ought to do with our money.

To be sure, whether the next ten years’ average annual real returns hit 1% or 3% or 4% or -0.5%, they’ll be well short of the 8% trends we’ve gotten used to. And, around these here parts, we like low returns about as much as dirty needles.

What’s more, no matter how hard we might look for a different story in other indicators, the forecast doesn’t seem to change all that much. Price-to-book ratio is another reasonably good indicator, but it’s not any help at providing a silver lining (depending on a handful of assumptions, marketwide price-to-book ratio is around or a little above the 75th percentile right now, implying mean reversion and lower future returns). Nor is the Buffett Indicator (which shows we’re well above trend, suggesting the future is darker than the past).

So here we are.

Which leads us to where we’ll go next: With all that nasty, infectious negativity about S&P returns safely triaged away, we’ll turn our sharps to the question of what the heck we can do to take advantage of our penetrating knowledge and actually make some money. Don’t worry. Next time’ll be here even before your HEP infection’s had time to clear up.

Luchadores, I’m eager to hear your thoughts about HEP! Do you think it’s worthwhile and useful? What do you make of its historical forecasting accuracy?

22 Comments

Libre Your Mind