Click to enlarge

S&P 500 Valuation Metrics
Source: BofA Merrill Lynch US Equity & US Quant Strategy



Merrill Lynch’s quant team looked at 15 metrics* that measure equity valuation (above).

Their conclusion?

Stocks are not overvalued; indeed, by most metrics, they are fairly valued.

Key bullet points:

• S&P 500 has been playing catch up to other indices;
• 14 of 15 measures say S&P 500 is fair to undervalued;
• Only the Shiller P/E suggests stocks are expensive;
• Four rotations underway are underway (that imply where best values might be found);

One worthwhile thing to add: I have been warning certain emailers/pundits/bears who seem to look at the Shiller P/E — and ONLY the Shiller P/E — that they are engaging in some confirmation bias in their metric selection process.

I cannot say that stocks are cheap here, but they also are not wildly overvalued. What happens going forward will depend on earnings — will they rise as the economy achieves escape velocity? Will they fall 30% as the economy tips into a recession.

The answer to these questions are here.




What is the Cyclically Adjusted S&P500 P/E Ratio ? (February 26th, 2010)

Why Using P/E Ratios Can Be Misleading (March 21st, 2012)

Meb Faber: Buy Cheap Cyclically Adjusted P/E (CAPE) (September 4th, 2012)

Stocks: Cheap or Expensive?
Savita Subramanian, Dan Suzuki, Alex Makedon, & Jill Carey
Bank of America Merrill Lynch, August 9, 2013



* There are additional some asterisks and caveats:

The Trailing P/E is based on GAAP EPS from 1960-1977, there have been many changes to the rules of GAAP accounting over the years.

Normalized EPS versus risk-free rate of return of treasuries is (obviously) skewed by ZIRP/QE.

Category: Investing, Valuation

Please use the comments to demonstrate your own ignorance, unfamiliarity with empirical data and lack of respect for scientific knowledge. Be sure to create straw men and argue against things I have neither said nor implied. If you could repeat previously discredited memes or steer the conversation into irrelevant, off topic discussions, it would be appreciated. Lastly, kindly forgo all civility in your discourse . . . you are, after all, anonymous.

42 Responses to “Stocks: Cheap or Expensive?”

  1. ByteMe says:

    This confirms my bias that stocks are not over-valued, but they are also definitely NOT under-valued. And the best returns don’t come from a fairly valued market… they come from an under-valued market through expanding P/Es. So I’m not expecting as much now as I did back in 2009.

  2. rd says:

    A couple of thoughts:

    1. The averages are skewed by the 1998-2000 bubble market. The valuations then were so out of whack that it takes decades of “normal” data to take out the skew. As a result, I tend to focus on the median data much more than the average data since the medan just counts the frequency of valuation ranges instead of the area under the curve so the extreme values don’t paint an artificial picture.

    2. The longer term valuation measures, like Shiller CAPE, price to book etc., divdend yields, Tobins Q, aren’t very good at market timing but are pretty good at predicitng likely long-term total return potentials. Lower total returns predicted by CAPE now can be achieved by a market crash with subsequent recovery (like 2000 and 2007) or just grinding it out for years in a relatively small trading range waiting for the economy to catch up to the valuations. I think increasing interest rates would trigger the former while continued ZIRP/QE may allow the latter to contnue. Wall Street is not know for being patient, so it will be interesting to see how psychology fits in over the next couple of years.

    3. It is interesting that dividend yield is not on the list. That would be another valuation metric that would be in the over-valued range, although presumably the bulls would argue that it is high compared to interest rates although that argument doesn’t really jibe with pre-1960 stock market history. Dividends are also one of the few metrics that can’t be gamed by accountants, unlike earnings and book value. Ultimately, you can justify the price of a stock from its ability to pay back the purchase price in dividends or its ability to be sold to a greater fool. The current low rate of dividends implies that corporate earnings and cash flow will need to be increasing at greater than the rate of inflation for a long time to pay an investor back in cash. The late 90s and early 2000s were clearly a greater fool era. If a stock’s market cap value ballooned up to a high valuation and then collapsed in bankruptcy without paying any significant dividends, did it ever really exist from a long-term investor’s viewpoint?.

    • 1. If the averages are skewed by the 1998-2000 bubble market, wouldn’t that be offset by the 2008-09 collapse?

      2. Shiller CAPE, price to book Tobins Q are useful when thinking about 7-10 year returns.

      3. I like Dividend Yield better than P/E — you cant fake a dividend check

      • rd says:

        The 1998-2000 bubble was so extreme, it took the2000-2003 AND the 2008-9 collapse just to get it back to normal valuation ranges. As a result, now we can actually have a discussion about whether or not the market is over-valued where it was pretty clearly over-valued for a decade and was driven predominantly by technical and Federal Reserve factors. It was mind-boggling to me in 1999-2000 how so few people saw the market as off-the-charts insane.

        The most fascinating thing to me over the 1998-2009 period was that the 2000 bubble actually ignored many stocks, as my focus at htat time was value oriented stock funds that had 20% or less in tech. As a result, my portfolio from 1999-2003 didn’t show the extreme ups and downs and did ok. It didn’t have any losses until 2002 when the down-cycle became all-enveloping, although value stocks just got hit at small bear market levels..

        2007-2009 took no prisoners. You had to be in government bonds to survive without major portfolio shocks. There were no corners in the rest of the investing universe that didn’t get whacked big-time.

  3. pjmason says:

    Why are they using different time frames for the different metrics? Seems to me that if you want and accurate comparison you should use the same time periods for each valuation metric.

    • Dunno, but I can guess:

      My 1st guess is that is what is available.

      My 2nd guess is they may be looking at complete bull/Bear cycles.

      My 3rd guess is it reflects bias.

      • BenGraham says:

        Let’s see: 2/3rds of these metrics measure post-1986 which contains the largest, persistent overvaluation in history. So this chart has 40% of its data points wildly over represented. FAR more applicable would be any sample that EXCLUDES those 1997-2007 years unless you are counting on a repeat for your returns. For a shop with all the data on hand, it is sad they chose to truncate most of these series which go back at least 50 years. Hussman has shown Shiller has a 90% correlation to future returns.

      • It also contains two massive collapses, plus an earnings fall of ~100%.

  4. Funny how so many of those are 1986 or later for the oldest data….blah.

  5. swag says:

    Sweet table. Thanks!

  6. Wingman says:

    I don’t know if stocks are expensive or cheap, but using average and median is quite dangerous given the distortions (i.e. fat right tail) created by the .com bubble.

    Of course, stocks could go back 44x Shiller PE. However, 95%+ of the cumulative frequency distribution is below 30x.

  7. sandden says:

    Is this article old? Barrons shows the trailing 12 month P/E for the Sand P as 19.49

  8. JoseOle says:

    Bravo on the link to the I Don’t Know post!

  9. Frilton Miedman says:

    Call me silly but using any P/E alone without referencing other metrics is too broad, especially without considering inflation.

    I’ll look at a P/E to gauge individual stocks relative to their specific sector as part of a screening process, but to me, looking at a P/E alone is like buying cars based on body style without asking about the engine, suspension, …etc..

  10. dctodd27 says:

    For all of the above calculations with an “E” in it, its worth asking how they normalize for record profit margins. Do we know what “trailing normalized PE” means?

    Also something to consider – let’s say we are “fairly valued” here, meaning a long-term investor can expect a roughly 10% annualized return from the S&P 500. With dividend yields hovering around 2%, that long-term investor is relying on 8% earnings growth over the long-term. Considering that in aggregate US companies have only annualized 6% growth over the long-term, this seems an overly optimisitic assumption. Either the market is overvalued here, or the market needs to grow faster over the long term than it has over the last 130 years.

  11. jaymaster says:

    I’ve been arguing for a couple years that the Shiller PE is going to be overly pessimistic for the next few years because of the economic downturn of 2008-9. Just about every company’s earnings took a beating in that time period, through no fault of their own (for many of them, anyway).

    Once the earnings bounced back, and every one saw that the world wasn’t ending, their prices and earnings bounced back to normal levels.

    So now we’re bouncing along with prices close to where they “should” be, but that super negative earnings blip is still being factored into the Shiller number.

  12. Seeker says:

    If you look at an inflation-adjusted chart of equity indexes and the latest economic data and flattish corporate earnings, it is reasonable to argue we are still in a secular bear era. Therefore, it would be reasonable to compare current valuation metrics to the averages/min/max metrics of all past and present secular bear eras, in which case valuations appear rich.

  13. mllange says:

    Why bother? In 6 mos, 12 mos, 18 mos, etc. it won’t matter. Rebalance asset allocation ratios when they get out of whack and move on.

    • Frilton Miedman says:

      Completely agree, observing broad sweeping P/E’s seems silly when you know to observe long term cyclical relationships, relative strengths within the business cycle between sectors & asset classes.

  14. Ted Kavadas says:

    Interesting data…

    However, I would add at least one other measure, market cap to GDP, which is periodically featured in this blog.

    As far as the metrics incorporating earnings are concerned, I believe that one needs to be mindful that various measures of corporate earnings and margins are either at or very near to record highs. This is seen in a variety of statistics and measures, including Corporate Profits to GDP.

    I believe that going forward, on an “all things considered” basis, there are many obstacles to continued profit growth. For those interested, here is my recent blog post on the topic:

    • rd says:

      I have always been cautious using total market cap vs GDP as a valuation measure as it assumes a relatively constant percentage of business ownership is in public hands versus privately held. It also assumes a relatively constant percentage of revenue and profits from domestic versus international operations.

  15. mrjohnnyt says:

    Humans desperately seeking patterns in data that is at best random and at worst deliberately manipulated, and using a non statistically significant sample size to boot.
    Worse yet humans paying other humans advisory fees for this “analysis”. (“there must be an answer, I just need an expert tea leave reader to show me”)

    Kind of like a sports program to crunch a history of team names and final scores to predict future results.
    So tell me again how housing should perform in a period when 75% of new jobs created are part time, and the velocity of money is collapsing?

    NAR: There’s never been a better time to buy! LOL


    BR: Or sell!

  16. Smokefoot says:

    It looks like 3 are showing overvaluation, not just Shiller. The ERP and normalized ERP are both well above their average and median.

    • Angryman1 says:

      based on useless historical data.

      the system changed in the 80′s and became a full blown “new paradym” in the 90′s. The economy runs more through the “market” than maybe ever before. Especially since 1980. The pros are going to invest in the market to produce returns and investment.

  17. sdw4srcp says:

    Would it not be useful to see where profit margins and productivity stand vs. historical readings? Also agree that Market Cap to GDP and Q-Ratio would be of interest (if only to support my own biases). Frankly I just see that markets aren’t excessively cheap or excessively pricey (though PE10 and Q are getting close) so not much can be drawn from any of this and we should admit that we just “don’t know” where we are headed. Is that the conclusion the ML quant team drew?

  18. skipinca says:

    As noted above, stocks are fairly valued so there are not a lot of bargains out there for a value guy

    There are 2 big problems with the market that I see:

    1) Operating earnings on S&P companies are very high. I believe they are north of 9% vs. an average of 6%. They are high because companies eliminated any excess labor and interest rates are very low. Additionally FCF is high as companies have cut back on cap ex waiting to see a rebound in the economy. (please correct my stats here if necessary)

    2) With interest rates at zero, there is only one way to go, UP. I don’t know when that will happen, but when it does a lot of bad things will occur. Interest expense will eat up a much bigger portion of Federal receipts in the US. It could be devistating to periphery European countries. It will reduce corporate profits. It will cause losses in bonds that will reduce bank capital. Japan will suffer these effects in spades.

    Seems to me that all asset classes today have a correlation of 1. Stocks bonds even gold will go down as real interest rates climb. There seems to be now where to run. High cash balances actually look attractive, relatively.

    The Fed seems to be in a box. By their own admission, it will be hard to reduce their balance sheet without raising interest rates.

    Any suggestions BR????

    Thanks for the chart

    • lucas says:

      You say if interest rates go up, a lot of bad things will happen. However, if we look at the “big picture,” interest rates need to go up. Borrowers and Savers should both have a fair shake. Right now, based on your list of consequences, corporations have enjoyed an unfairly low cost of doing business. If interests rates go up, and the prices of stocks go down, perhaps that new price is actually a fair price. Same with houses, In this atmosphere of low mortgage rates, houses (especially on the coasts) are too expensive relative to wage demographics.

      • skipinca says:

        I completely agree with your assessment. As someone living off my savings, I am outraged at Bernanke’s confiscation of my income – talk about a tax increase!

        My point is that the Fed policy has had unintended consequences, which include boxing investors in leaving them with minimal opportunities to make a fair return. I am faced with having to hold a large portion of my capital in cash and wait for the market to adjust. And the Fed is degrading the value of cash as a matter of policy. It is a difficult time to be an investor trying to preserve purchasing power and capital.

  19. seneca says:

    Scroll down the page of the link below for a graph of the S&P500′s Price to Sales Ratio, 1955-2005. You can see the paradigm shift starting in the mid-1990s. Possible reasons for the shift are the widespread adoption of the 401(k) and open immigration, which destroyed labor’s bargaining power.

    It’s easy with Google to find the current (July 9, 2013) Price/Sales: 1.54.

    At the peak in 1987 before the crash, P/Sales equaled 1.0. From 1955 to 1996, the P/Sales never exceeded 1.1. I’d say stocks are fairly valued under the new paradigm (401(k) demand for stocks and an abundant oversupply of cheap labor).

    • Greenspan = Valuation paradigm shift

    • rd says:

      I don’t buy the 401k argument. The reason that 401ks are so plentiful is because another large stock consumer, the pension fund, is effectively extinct for future contributions from the private sector. So it is just another buyer for the same stock for the same purpose.

  20. san_fran_sam says:

    Late to the comments on this one. but what struck me about this list is that it seems many of the ratios are correlated to each other. so you don’t really have 15 different ones, but only, say, five.

    and I don’t know that SP500 relative to oil or gold prices really means anything. at least not any more.

    just sayin’

  21. patslatt says:

    The linked website provides a range of valuations for the S&P 500 based on an estimate of 10 year forward EPS. The best fit EPS for 2012 based on the trendline since 1980 (see the second chart) is way below actual EPS.

    While this model is logical,it isn’t good for market timing.

  22. [...] this month, we looked at the question of whether Stocks were Cheap or Expensive. That was a follow up to our glance at how much Earnings and Equities had rallied off the 2009 [...]