· ISM Non-Mfg declined to 53.5 in Apr vs 55.3 exp and 56 prev. Orders and Employment subindices were weak, while inventory subindex was stable.
· US Initial Jobless Claims declined to 365k last week vs 379k exp and 388k prev
· US Unit Labor Costs declined to 2.0% in 1Q vs 2.7% exp and 2.8% prev
- ECB left policy unchanged as expected. Draghi did not hint at additional policy moves in the near future
- UK Services PMI declined to 53.3 in Apr vs 54.1 exp and 55.3 prev
- EU PPI declined to 3.3% YoY in Mar vs 3.4% exp and 3.6% prev
- China Non-Mfg PMI declined to 56.1 in Apr vs 58 prev
- Australia Services PMI declined to 39.6(!) in Apr vs 47 prev. This was the lowest print since early 2009:
- Fri: Italy PMI Services, EU Retail Sales, US Payrolls, Unemployment
- Mon: AU NAB Business Confidence, Retail Sales, Swiss Unemployment, CPI, UK RICS House Price Balance
- Tues: US NFIB Small Business Optimism, JOLTS Job Openings
- Wed: German Trade Balance, Japan Trade Balance, Australia Employment
On the PE Ratio:
Most equity market commentary goes along the lines of: we think earnings are going to increase x, and PE ratios are undemanding at y, and so we implicitly assume that PE ratios will mean revert, thus we project that the market will go up by f(x,16-y) = z. There’s nothing wrong with this analysis, if your investment horizon is over a 40 year period when you can be reasonably confident that PE ratios are likely to revert to the mean. At shorter horizons, however, these assumptions can be misleading. Equity strategists have been talking about / assuming PE mean reversion for at least 5 years. The fact that PE ratios have been contracting for almost 13 years does not seem to trouble them.
So what drives P/E? There are various theories for long term drivers, including the demographics cycle, (SF Fed) and inflation. (Crestmont Research) However, both only explain secular changes. I wasn’t able to find much research that was useful at shorter horizons, so I decided to formulate my own hypothesis. IMO, the most likely factors that explain PE changes are National Savings + % of Savings that goes into Stocks + Debt used to buy Equities + Earnings.
Hopefully, that point won’t be too contentious. The other way to say it is that how many multiples of earnings people will pay are a function of how much they have and how much they’re willing to put into equities. This hypothesis is also supported by historical data. In the chart below, the S&P trailing P/E ratio is in red, and the green line is a non-regression function of the variables listed previously.
OK, let’s extrapolate a bit. How are those drivers likely to evolve over time given the cyclical and secular backdrop?
National Savings : The most recent print for the personal savings rate was 3.8% – close to the lowest point since 2008. In the context of weak employment growth and inevitable higher taxes, the outlook for this is factor is negative. (Note that a repeal of Bush tax cuts will take the savings rate into negative territory)
% of Savings that goes into Stocks : This is predominantly a function of target liability duration (how far the average person is from retirement) and speculative demand. Note that speculative demand is negatively correlated to liability duration. Since the baby boomer ‘bulge’ in the demographic distribution will continue to shift into retirement over the next decade or so, the outlook for this factor is also negative. (The trend of inflows into fixed income and out of stocks over the decade is NOT a fad)
Debt used to buy Equities : primarily a function of speculative demand, and tends to be correlated to recent price movements (i.e. tends to be driven by momentum followers) Also affected by credit availability via brokers & investment banks. This is a procyclical factor, so the outlook for this is marginally positive given that positive GDP prints are likely to continue. Note that this only refers to debt held overnight – HFT intraday margin does not apply.
Earnings : I see blended 12m Fwd consensus estimates of 11% earnings growth. In the context of consensus estimates of 4.7% nominal US GDP growth, slowdown overseas and margins near prior cycle highs, (see chart) this appears to be on the high side of likely outcomes. Nevertheless, if positive GDP prints do persist, earnings are likely to follow, but perhaps at a slower pace than expectations. However, it is worthwhile to note that this expansion is now 3 years old. If historical patterns persist, we are likely to see a recession within the next couple years.
The generalize conclusion here is simply that the consensus expectations for PE mean-reversion in the short term is likely wrong. Of the 4 factors identified here as the primary drivers, the outlook for 2 of them are strongly negative, with the other 2 only marginally positive. This suggests that the PE contraction we’ve seen over the past 13 years is likely to persist. While this will be offset by earnings growth over the long term, the secular bear market that started in 1999 is most likely NOT over.
On High Yield Bonds:
I’ve read a lot of commentary talking about how High Yield is cheap because the average spread over treasuries is well above where they were in prior cycles given how low defaults are. They usually include a chart like this:
Given how low treasuries yields are, however, the spread alone is probably NOT the right metric to use.
Furthermore, recovery rates for speculative debt are low – less than 30% historically:
Currently high yield debt trades at a yield to maturity of ~7% based on CDX and high yield ETF’s. As a result, expected returns for this asset class will only be positive outside recessions, and likely to be in the low single digits in real terms across an entire business cycle.