- While second-quarter earnings came in above expectations, the third quarter may well turn negative.
- Against the backdrop of possibly ongoing valuation pressure (per the Brexit roadmap), I see little upside for stocks until we get past October's seasonal trough.
- The silver lining is that equity valuations relative to bonds (the equity risk premium) are back to levels seen at the Q4 2018 low, when the S&P 500® index was down 20%.
- Over the long run, a well-diversified mix of stocks, bonds, and cash should continue to serve investors well.
About the expert
Jurrien Timmer is the director of global macro in Fidelity's Global Asset Allocation Division, specializing in global macro strategy and active asset allocation. He joined Fidelity in 1995 as a technical research analyst.
On the heels of a solid first quarter, Q2 earnings season has been better than many expected; nevertheless I fear the worst may still be yet to come for earnings growth. While 2019 has been exhibiting the same kind of "V-shaped" to "L-shaped" progression that we saw back in 2016, this year so far has produced a series of higher lows for the estimated growth rate versus the lower lows of 2016.
But when I do the math of the typical progression of the expected earnings growth rate—which shows a decline of 1000 basis points (bps) during the 9 months leading up to the second quarter, followed by a 300 bps bounce during earnings season—I wonder whether I am being too optimistic. Currently, the Q3 estimate sits at −2.3% with 7 weeks to go before the end of Q3. At the beginning of the year, that estimate was +5.7%.
If we apply the typical downward drift, then the Q3 estimate should bottom at −4.3% and then bounce to −1.3%. That would be a new low for this cycle and its first negative print (since Q1 and Q2 both started negative but bounced to positive). And given that both Q1 and Q2 underwent a larger-than-normal drift of around 1200 bps, I find it plausible that the growth rate could finish even lower.
This suggests to me that perhaps the market will remain under pressure until we get past the normal seasonal low in October (i.e., when Q3 earnings season begins). It's easy to see how this could happen by analyzing the 3 components of the market's total return: earnings growth, dividends, and valuation.
The story for 2018 was that the price/earnings (P/E) valuation was getting de-rated (from 19.7x in January to 13.7x at the December low) at the same time earnings growth was accelerating (from +12% in 2017 to +22% in 2018). The net effect was a 6% decline in the S&P 500® for the year. For 2019, this dynamic has reversed: Earnings growth has been slowing (from +22% to +3% or lower) while P/E multiples are up. But if earnings growth is heading even lower and valuation multiples contract from here (following the Brexit roadmap), by my calculation it will be near mathematically impossible to avoid a negative-return outcome for the year, or at least the third quarter.
Meanwhile, earnings are not the only storm clouds out there. Trade disputes remain front and center for the economy as well as the markets, especially now with slowdowns wearing on many major foreign countries. It's difficult to sugarcoat the possible, unintended consequences of a Chinese currency devaluation weakening past the politically sensitive 7.0 yuan/USD level. This always seemed a worst-case nuclear-option type of action on China's part, and yet all of a sudden, here we are.
With the S&P 500 down 7% from its recent high of 3020 and the forward P/E still a relatively high 16.8x, the equity risk premium (defined here as the earnings yield minus the bond yield) has reached 4.5%. The premium is now as high as it was at the December 2018 bottom when the S&P was down 20% from its 2018 high (with the forward P/E a mere 13.7x), or about twice what it was a year ago.
All of this is compliments of a tumbling bond yield. It's certainly a good reminder that valuation should always be considered within the context of interest rates and that perhaps equities have de-rated more than some will give them credit for.
That leads me to be hopeful that perhaps the US market is further along the Brexit roadmap than meets the eye. To recap, my Brexit roadmap traces the effects emanating from the June 2016 UK referendum setting up the country's break from the EU. In the 3 years since, the UK stock market has moved sideways while earnings have grown 30% and the forward equity P/E has fallen from 16x to 12x. For the US analog, that would suggest a valuation de-rating from 19.5x (Jan 2018) to 15x (August 2019). We are halfway there.
Another silver lining for investors is diversification: Over the long run, a well-diversified mix of stocks, bonds and cash should continue to serve investors well.
Will this continue? I don't know. Diversification does not ensure a profit or guarantee against a loss. But with the federal funds futures curve now pricing in 4 more cuts (heading down to 1%), and with about $16 trillion worth of global debt trading at negative yields (compared with a mere $13 trillion just a few months ago!), I find it completely plausible that yields on US 10-year Treasuries could fall below the 2016 low of 1.32% and end up somewhere around 1%.
So, can yields actually go much lower?
I think it's quite possible yields could continue downward, but I also think it would require a full easing cycle on the part of the Fed. We tally nearly $16 trillion of negative-yielding debt out there—up from zero (indeed, unthinkable) just 10 years ago and rising fast. What if the share of global debt trading at negative yields continues to expand as the world economy continues to slow? What if a few years from now, the amount were to reach $20 trillion or even 25 trillion? Where would our 10-year Treasury yield be then? 2%? 1%? Zero?
The thought is not all that implausible, in my view, given that the FTSE World Investment Grade Bond Index (WorldBIG®) ex-US already yields close to nothing—and its yield could well go lower if the world economy keeps slowing and central banks everywhere keep easing and global QE (quantitative easing) resumes. Heck, if Greek bonds can go from 12% to 2% in 5 years' time, what's to stop US Treasuries from falling from 3% to 1% or all the way to zero?
As I see it, for US yields to decline further, all we need is a Federal Reserve easing cycle that weakens the dollar and changes the relative shape of the basis swap* curve to make it once again attractive for foreign investors to buy Treasuries on a currency-hedged basis. (A basis swap is a contract to exchange cash flows based on different variable interest rate references and may also involve different currencies as a way to accommodate liquidity needs.) They haven't done so lately because the US 10-year is yielding -0.60% on a hedged basis, below the yield of German bunds (−0.38%) and Japanese government bonds (−0.15%).
How can we know where yields go from here? Maybe it's as simple as demography. I did a simple exercise comparing the buildup of negative-yielding debt with the percentage rate of change of the 65+ US age cohort. The similarities should come as no surprise: The demographic tidal wave sweeping across large swaths of the world's population is bound to have an effect on interest rates as more and more of society shifts from solving for growth to solving for income.
What will a decline in Treasury yields to, say, 1% or less due to equity valuations, especially those "bond proxy" equities that offer a compelling alternative to debt instruments? I find it no accident that both the secular growers and the stable, bond-proxy growers have been leading the market in recent years. If rates keep falling, this odd couple of market leadership may well continue. That suggests to me that the generational lows seen in the relative performance of value versus growth equities might well be a value trap for now.
Given that the relationship between the yield of a bond and its “P/E” (its price-to-yield or price-to-cash flow) becomes increasingly non-linear as yields approach zero (aka convexity), one has to wonder if the path for some equity valuations might also become non-linear if bond yields end up declining much further. For instance, the “P/E” of a 5% bond is 20x, and the P/E of a 2% bond is 50x. A bond yielding 1% has a P/E of 100x; a bond yielding 0.1% has a P/E of 1000x.
Given the above, will the disappearance of bond yields be the thing that solves the riddle of how we can have a stock market that promises double-digit returns and mediocre returns at the same time? The double-digit promise derives from the analog of past secular bulls (see my report from last month); the projection for mediocre returns stems from the strong inverse correlation between the 10-year CAPE (cyclically adjusted P/E) and the 10-year forward equity return.*
If ultra-low yields force more and more investors into alternative sources of income—especially equities of companies that pay stable dividends (or at least buy back their shares)—will investors increasingly use price-to-cash flow instead of price-to-earnings to value their investments? I am not predicting this shift, but it also wouldn't surprise me in the least. As I pointed out last time, on a price-to-cash-flow basis, the S&P 500 is technically cheap versus history (in the 28th percentile), whereas measured by the 10-year CAPE, the market has hit "nose-bleed" levels (92nd percentile).
From the 1960s through the 1990s, equity and bond P/Es exhibited a tight correlation. This relationship served as the foundation of the so-called Fed Model, on which former Fed Chair Alan Greenspan relied heavily and which foreshadowed the 1987 stock-market crash. The Fed Model simply tracked the spread between the S&P 500’s earnings yield and the 10-year Treasury yield (or, inversely, the ratio of the equity P/E to the bond P/E). Beyond certain upper and lower bounds, the spread was taken to signal that stocks were cheap or expensive, but historically the spread always kept to a mean-reverting range.
The model stopped working when the market entered the deflationary era of the 2000s, and today the 48x P/E on long-dated Treasuries far outstrips the measly 18x P/E multiple on the S&P 500, or even the 27x calculated using the CAPE.
What if the bond P/E rises to 100x or 500x? What will that do to the equity P/E? Sorry, I don't know—but it's worth thinking about. If the equity P/E multiples of those sectors, styles, and factors that generate a compelling cash flow expand in order to keep up with the rising bond multiple, then maybe that will form the narrative by which, a decade from now, we can all explain how the market just kept on going up and up and up despite the myriad reasons why it shouldn't have.
One major counterpoint to that line of thinking is the simple fact that we already have zero or negative rates on bonds in Japan and parts of Europe, and their equity markets have hardly gone convex or parabolic. The MSCI Japan Index has a forward P/E of 12.7x, and MSCI Europe is at 13.7x. So, no valuation bubbles there, even though, compared with the United States, their rates are lower, their age waves more advanced, and their dividend yields higher (2.5% in Japan and 3.8% in Europe versus 1.9% in the US). It's a point that's difficult to dispute.
The only thing I can think of in terms of why the US cycle might be different is that Europe and Japan have pension systems whereby most pensioners simply need not rely on their own asset allocation to solve for their retirement. The onus to deliver pensions is on the State, not on the citizens themselves. Moreover, the European and Japanese systems also are very bond-centric: If a yield shortfall for those pensions were to develop due to zero or negative yields, I think it unlikely that the State would load up on equities to make up the difference.
The US is a different story: While the defined-benefit system is not totally dead, Americans for the most part must fend for themselves via their 401(k)s and IRAs. So, if bond yields run dry here, I believe it at least theoretically possible that American savers might add more (income-generating) equities to their portfolios in order to make up the difference. I have no idea whether such a dynamic would be big enough to create a different outcome in the US versus Europe and Japan, but it's something worth considering. Of course, in terms of secular growers, US sector weights also are quite different from those in Europe and Japan, so that too could prove a swing factor in terms of valuation. Indeed it already has.
One final point: It's interesting to me that the demo-graphic tidal wave currently underway peaks in 2026. That is the same year in which the secular bull-market analog that I have displayed so many times also peaks out, based on the experience of 1982–2000 and 1949–1968. Coincidence? It makes one wonder whether the former, the wave, can become the narrative for the latter.
As I said in my last report, I don't know any better than anyone else what the coming years will bring or whether or not we are in a secular bull market. But I am convinced of one thing: If this is a secular bull and we indeed have another 5-plus years of strong valuation-driven returns ahead of us, it likely will be because of the yield phenomenon described herein.