Drinking coffee and crunching numbers on Sunday morning; the fundamentals behind this bull market get worse the more I dig…
You never want me to find an error in your math, because if I do, I check EVERY number until I’ve nailed down the whole deal. What got me started on this S&P500 valuation project – if you ask 10 retail investors, and 10 professionals, “what are the earnings”, you’ll get 20 different numbers. That is a red flag. After a month of digging, I think I’ve taken this about as far as it can go. Here are the real numbers as a public service announcement. This is the most important post on AnalystForum this year dammit, so get a coffee, open Excel, reproduce the math, and make sure you know what you are holding if you own S&P500 stocks, there is a lot of statistical manipulation out there… ---------------------------------------- WHAT ARE THE REAL EARNINGS? Authoritative earnings are the key number we need for any model or valuation. If we start with a nonsense number, then all subsequent analysis is nonsense. So you can’t use numbers from “FactSet”, Reuters, Bloomberg articles, WSJ, etc…those are all made up numbers. We can put all the confusing numbers out there into categories… 1) “Operating earnings” or “adjusted earnings” are bullshit numbers. There is no standard, different companies and analysts make different adjustments, not only are they inconsistent by company, but they are inconsistent over time for the same company. If you start with these earnings as your base, you can’t compare to prior periods, nor use the 100yr history. Plus the spread between these numbers, and the official GAAP numbers, has been increasing since the start of this bull market. A huge red flag. 2) “Forward earnings” are bullshit numbers. These are always unbelievably high, and are adjusted down right before actuals come in, so they can “beat expectations”. Forward earnings, and the expectations created by them, shared by companies like “FactSet”, are useless…notice that these companies are careful to never quote trailing as reported earnings, ever. 3) “Share-weighted earnings” are bullshit numbers. The proper way to calculate any S&P500 index earnings-related number is [aggregate earnings for quarter / divisor]. Example Q2 '15 was $201.4B as reported earnings / 8831 divisor = $22.80 EPS. Let’s say Exxon reports a $10B loss, that is an absolute value, their size is already accounted for, you as an index holder take that loss. If you weight earnings to make the losses look smaller because energy is a smaller sector, like Bloomberg does, you are making up fantasy numbers…this is simply mathematically wrong. Weighted earnings: they don’t add up… and you may get burned (by S&P) 4) “As reported earnings” are the real GAAP earnings reported in the financial statements. Everyone follows the same rules (more or less), or they go to jail. These are written in stone; once they are reported, that number to the penny will never change. It’s the best we’ve got. They are reported by S&P labeled “as reported trailing twelve month EPS”. That is our authoritative earnings number; GAAP data as crunched by S&P’s senior index analyst Howard. Q3 EPS (ttm) = $90.85, 23.0X as of Nov27 '15. Any future earnings will eventually become these trailing GAAP earnings, until that happens, it’s all fantasy. If you need to forecast, use as reported trailing as your base, and generate your own REASONABLE forecast on that. ---------------------------------------- EXAMPLE OF NONSENSE EARNINGS Q3 (ttm) EPS from S&P are: as reported $90.85 , operating $104.14 , forward+operating $122.35. Forward-operating is absurd of course, but it is how they keep telling the noobs that it’s cheap P/E 17X instead of 23X. Other companies use different operating/forward numbers than S&P, Reuters estimates differ dramatically, so none of this will ever tie out. As reported is the only solid base. ---------------------------------------- BUYBACKS, A COMPLICATION We’ve still got a major problem with using trailing as reported EPS as the driver for our models and valuations though; it is manipulated by corporate America by tweaking the denominator. You can download buyback data from S&P, I modeled it this morning. Roughly speaking, as reported EPS without the last 5yrs of buybacks, would be more like 83. So P/E isn't 17X fantasy-forward, and it isn't even 23X trailing, it's more like 25X. [Bloomberg on buybacks](http://www.bloomberg.com/news/articles/2015-07-16/corporate-stock-buybacks-make-earnings-look-better) ---------------------------------------- **SO WHERE DID RETURNS COME FROM IN THIS BULL MARKET?** Disaggregating the S&P500 5YR total return, step by step, nailing it the f@# down to authoritative S&P data. Check it out yo… 5YR TOTAL RETURN 15.3% CAGR** = EARNINGS GROWTH 2.5% + BUYBACKS 2.3% + DIVIDENDS 2.4% + MULTIPLE EXPANSION 7.7% [ERROR 0.4%] Q3 2010 to Q3 2015 TR as calculated by S&P (plus adding on the index appreciation to Nov27th, ending value of 2090). In order to nail this down, we have to realize the 5YR 4.8% EPS (ttm) growth rate from $71.86 to $90.85 is partly buybacks. Aggregate as reported earnings grew from Q2 2010 $178.0B to Q2 2015 $201.4B, a 2.5% CAGR (Q3 is not available yet). Which means the 4.8% EPS growth = 2.5% earnings + 2.3% buybacks (roughly, this is hard/impossible to get perfect). Crosschecking this confirms we are close; S&P reports the Q2 (ttm) buyback yield was 3.0%, and Bloomberg reported buybacks added 2.3% to EPS growth in 2014 (link above). Since we know total return and dividends, and now we know earnings and buybacks, the remaining return is multiples. We can cross check this; P/E expanded from 15.9X at the start of the period to 23.0X, a 7.7% CAGR. I have a 0.4% error which showed up when I started trying to break out buybacks, so either earnings growth or buybacks must be understated. ---------------------------------------- BULL MARKET SUMMARY S&P500 companies are only doing ~2.5% earnings growth for the last 5yrs**(which after inflation of 1.7% is only 0.8% real earnings growth). They aren’t investing in the future, they are currently paying out more in buybacks+dividends than they actually make in earnings…essentially liquidating their companies (we saw this is late 2007 also). Over the last five years there is a10% annual return from buybacks and increasing valuations, both of which were fueled by ZIRP/QE, and are unsustainable. That’s a fragile starting place for a 10-yr forecast: risk of multiple contraction, earnings recession due to lack of investment, increased share issuance (reverse of buybacks) or increased debt expense as rates raise. Don’t say I didn’t call it hommies, and remember I called subprime in Q4 '07 also. Timing is hard, but this game is approaching check-mate. The math hits a wall. SOURCE: S&P, http://www.spindices.com/indices/equity/sp-500 “additional info” dropdown, Excel download: “index earnings”, “stock buybacks”, “monthly and annual returns”. Pull more history from http://www.multpl.com (Josh uses S&P’s numbers too, with a philosophy of “just the facts no spin”, I’ve spot checked his math, never found an error, that’s rare for me).