Last week, I posted the above chart from the NY Fed’s Liberty Street Economics. This morning on Squawk Box, David Tepper of Appaloosa discussed it — and his comments reversed the futures from negative to positive.

Here is a brief explanation of what this chart — a compilation of 29 valuation models — means:

The equity risk premium is the expected future return of stocks minus the risk-free rate over some investment horizon. Because we don’t directly observe market expectations of future returns, we need a way to figure them out indirectly.

That’s where the models come in. In this post, we analyze twenty-nine of the most popular and widely used models to compute the equity risk premium over the last fifty years. They include surveys, dividend-discount models, cross-sectional regressions, and time-series regressions, which together use more than thirty different variables as predictors, ranging from price-dividend ratios to inflation. Our calculations rely on real-time information to avoid any look-ahead bias. So, to compute the equity risk premium in, say, January 1970, we only use data that was available in December 1969.

If you look at the chart, we are at levels that are similar to 1975 and 1982.

An emailer last week asked me why I “waste space” occasionally publishing Fed research, academic papers and other such works. Now you know why . . .


Are Stocks Cheap? A Review of the Evidence (May 9th, 2013)

Appaloosa’s Tepper Bullish on Stocks (Still) (May 14th, 2013)


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.

46 Responses to “Equity Risk Premium is High (this is bullish)”

  1. Bob K. says:

    Barry, can you do a post on Stock market growth to GDP growth? My understanding is they are aligned, and we are already above trend.

    I also know the stock market is 85% correlated to the FED balance sheet, I just don’t know how far out traders have extrapolated that growth into the current price.

    Thank you

    • carchamp1 says:

      Equities do grow in line with nominal GDP growth, over time. I don’t think comparing these two things straight-up is useful though. I make certain adjustments in my modeling.

      GDP, money growth, and the stock market simply go up most of the time. To that extent, they will be highly correlated. 85% sounds about right. On whether or not this, alone, is useful information for investors, the answer is no. It’s just a product of our monetary system.

  2. Bob K. says:

    A note on Tepper, didn’t he go on TV and pound the table bullish while he started selling into that move higher?

  3. adbutler007 says:

    Oh my. Here I thought you published that note to demonstrate how Fed induced low rates are driving the entire equity risk premium, and that eventually interest rate normalization will destroy said premium. I never in my wildest nightmares would have guessed that you posted this as an argument for cheap stocks.

    To be clear then; so long as the Fed holds rates at artifically low levels, the equity risk premium is very high. So if the Fed holds rates here forever, then investors can expect 5.4% per year in excess returns from stocks.

    This carries the same logical fallacy as the Crestmont model, which suggests how PE ratios relate to inflation; low (but not negative) inflation has historically resulted in high PE ratios, while negative or high levels of inflation have corresponded with low PE ratios.

    The Fed model and the Crestmont model are flawed because they require investors to make a different estimate about something which can not be estimated. The Fed model requires us to forecast rates, and the Crestmont model requires us to forecast inflation. Which basically means the models are useless.

    Further, Hussman has demonstrated (see Greenbakd http://bit.ly/18IxEie) that the Fed model carries almost no empirical explanatory power.

    Frankly, I can’t understand your take on this…

    • The Fed model i something entirely different — it uses 2 variables, and its conclusion is in large part dependent upon analysts estimates, which I have demonstrated are invariably too high (except at bottoms)

      See: -The Flawed Fed Valuation Model, February 5, 2008

      • adbutler007 says:

        Without getting into the relative mechanics of the Fed model vs. the model presented, the implication is still that stocks are cheap relative to Treasuries, and that equity investors should expect 5+% excess returns per year over Treasuries over a variety of horizons.

        Then the paper goes on to show that the _normalized_ risk premium is less than zero once you extract the effect of abnormally low rates.

        Logically then, in order to harvest the implied risk premium from current rates, these rates would have to persist ad infinitum. If rates were to normalize, this would imply a NEGATIVE risk premium to stocks.

        So you should only be bullish if you believe that we will have low rates and QE forever. Which may or may not be true, but I would argue that if it is true then the theory underlying the concept of a risk premium is defunct.

      • cdwight says:

        People tend to extrapolate current trends too far into the future (rather than expecting revision to the mean). As long as the Fed holds rates artificially low and people believe that will continue the implied equity risk premium is real. At some point it will reverse when expectations shift and there is a belief the Fed increases rates and/or reverses QE. Keep in mind the Japanese have had low rates for about a couple of decades now. I think the US is in better shape and has avoided deflation so rates should rise sooner. Until capacity utilization rises into the low to mid-80s I don’t think there will be significant upward pressure on rates.

      • Nick in AZ says:

        You are absolutely correct.

        E(r) = rf + β(RM – rf)
        Risk Premium = (RM – rf)


        rf is the risk free rate (treasuries)
        RM is the expected return on the market

        As long as the fed keeps rates close to zero, the expected return on the market IS the risk premium.

  4. agronox says:

    And the bear would say you’re looking at this wrong; that it isn’t that equities are poised to return so much, but that risk-free assets are poised to return so little that the equity risk premium is so comparatively high.

    I’m not a bear in general, but we’re in an outlier time for risk-free (lack of) returns.

  5. quantland says:

    Why is this chart bullish? I appears to me to be the exact opposite. The prior peaks of this metrics were during the great roaring late 70′s. Oh, that’s right, the return were dismal suring that time and just before the double recessions of 1980 and 81. This looks as a better contra-indicator. Can you explain otherwise?

    • This is long term bullish — unless you think the risk premium in 1982 was bearish?

      • quantland says:

        What is “Long term”? and what is its useful purpose? Are you telling me that the metric nailed it in 1974 where the 7 year return was in the neighborhood of 4% or it was really worth paying attention to 1990 when it told us to be bearish on stocks for the long-term?

        The authors seem to be selectively picking which times should serve as evidence ans which times we should ignore. The problem is they are passing it off as quantitative research mean it would never stand against a peer review…..

      • rd says:

        Most of the area under the risk premium line on the chart was between 1974 and 1981 until the 2008 financial crisis caused another graph mountain. That 1974 to 1981 period had ok nominal returns but lost quite a bit of money when inflation was factored in.

        The great bull market occurred with much lower risk premiums between 1983 and 2007.

        I don’t see much correlation between inflation-adjusted stock market returns and the risk premium. I do see a lot of correlation between periods of high risk premiums, periods of poor economic growth, and controversial Federal Reserve interest rate policies (late 70s-early 80s and late 2000s).

        Possibly a message from this graph is to expect a decade of economic pain when the risk premium goes over 4%.

      • len613 says:

        It is hard to line up dates with the value.
        But it appears that the risk premium was at record lows after the 1987 crash and that woulod have been a great time to invest.

  6. carchamp1 says:

    I think stocks are still a great buy here, but I’m wondering if this indicator is being thrown off by the low interest rates resulting from Fed policy. I don’t think stocks are currently as good a value as they were in 2009. There’s no way.

  7. GregP says:

    This is the wrong chart to select from the Liberty Street study.
    The meaningful chart is the third one (or second-to-last), titled:
    “The Equity Premium is high because Treasury yields are low”.
    For today’s high ERP to be (theoretically) bullish, you must assume that
    Treasury yields will not regress (move higher) over the next 15 years or so.
    Because the duration of stocks is even longer than the 30y Treasury.

    Or we could take the empirical approach:
    Overlay Tepper’s/Barry’s chart with SPY. Lag it a year if you like.
    Did that look bullish for the decade 1987-1997?
    Did it look bullish in 1982?
    Did it look bearish at the appropriate time in the 70′s?

    • I keep it much simpler than that: What happened in the past when the risk premium got this high?

      • GregP says:

        Um, your selection bias is showing- we’ll only use this indicator when ERP is high, not low?
        Using your narrow criterion:

        A. The chart shows a very high ERP in 1975.
        Stocks unperformed their long term average for the next five years.

        A. The chart shows a very high ERP from 1978 to 1981.
        Again, years of below-average performance until takeoff in 1982.

        Tepper is a hedgie- he’s not buying now for the prospect of good returns
        beginning four years hence. He’s using the Liberty Street article, inappropriately,
        to justify a further short term runup. Interestingly, the authors also used the word
        “justify”. Apparently, the Fed back office is onboard with the stock pump.

        (FWIW, I remain 40% long because, well, it’s all about perception at this point.)

      • sdw4srcp says:

        Am I reading this wrong? Wasn’t the ERP equally high in late 2007 early 2008? (maybe I need gridlines) An interesting difference between now and 1982 is that the CAPE was about 7 then and 23 now… Not that markets can’t go higher, but I dont see how this makes a clear 1982-like bullish case. 1982 was clearly valuation driven too (cheap!) while the best you can say is that stocks are fairly valued now (looking at 12mo trailing or forward looking PE) if not at about the 90th percentile in richness (if looking at CAPE).

      • Yes, that is 2009 you are looking at . . .

  8. Smokefoot says:

    The link doesn’t seem to give details of what the 29 models are. Hussman writes about 4 models (Forward Earnings, Dividends, Shiller PE, and Market Cap/GDP) that all show the expected 10 year return at well below 5, and nowhere near 1974 or 1982. Why are these so different?

  9. VennData says:

    Equities up 20% since the GOP-hysteria Sell off after the election.

    GOP investors, the reverse indicator.

  10. Smokefoot says:

    If extremely low Treasury yields predict large excess stock returns, wouldn’t that also apply to Japan? Japan’s government long term yields went below 2% around 1998 and have continued to go down, but that didn’t predict excess returns (except since Abenomics was introduced recently).

  11. crunched says:

    I forgot the Fed was printing 85 billion dollars a month in 1975 and 1982 to keep the stock market propped up.

    ‘The equity risk premium is the expected future return of stocks minus the risk-free rate over some investment horizon’

    –one of the two variables in this information set has become completely meaningless. Funny to watch how as the market goes higher this doesn’t become an issue anymore. Bottom line, there has to be some form of mis-placed euphoria in place for the market to turn and it sure feels like we’re almost there.

    • The greatest tell about the measure of Euphoria is a persons’s investment posture.

      If they are fully invested, they don’t see any Euphoria! Ther less exposure they have to equities, the more Euphoria they seem to notice.

      I am beginning to suspect some form of bias is involved…

  12. dctodd27 says:

    GMO says 7-year S&P 500 return expectations are now negative, as does Hussman. Baupost may return cash to investors soon due to lack of investment opportunities. Somebody is spectacularly wrong here. I’ll listen to the guys whose long-term forecasts have been spot on, not Ben “Subprime is Contained” Bernanke and his cohorts at the Fed…

  13. SteveC says:

    We’re at levels similar to 1975 and 1982. 1982 was a great time to own stocks. In 1975, was good, but we were still in the grips of the bear for 7 more years. I suspect this bear isn’t done with us yet, and we’re more like 1975. The Fed is using monetary policy to inflate stock prices, and no problem, so far. The issue is that they can’t always control where all this liquidity is going. We could easily inflate another bubble, if they’re not careful. And when they take their foot of the pedal, market dislocations may occur.

  14. Robert says:

    BR — there is an accompanying chart that shows where the ERP would be IF rates were “normal”. Given the massive impact the FED has had on rates, the ERP appears overstated………VERY overstated.

  15. Pantmaker says:

    This chart is beyond ridiculous. Holding period? interest rate variability? earnings? profits? Tepper’s gotta be selling everything that isn’t nailed down.

  16. Sasquatch says:

    In November/December, you mentioned getting cautious/defensive ahead of a potential 20-30% correction. Now 22% off the Nov. low you are using this study as justification to get even more bullish. Where was this study then? “Because it worked in the 70s-80s” is not a good enough reason to get hog-wild bullish. Interest rates were higher then and it was before the global debt boom that kicked off on in the 80s.

  17. arm50 says:

    It looks like we are also at mid-2007 levels.

  18. mdanda says:

    Aside from the data and graphs, wasn’t he also making the point that all this money floating around (from a variety of sources, domestic and abroad) has to go somewhere, and this simple fact is going to keep the market propped up? Is he saying we really can’t have a crash, because there is just too darn much money floating around, looking for a home? Is that a fair interpretation?

  19. Gonzop says:

    The Nikkei is a few months into QE, and (ex) Japanese investors have gone full Pavlovian on it because of the American example i.e. the Fed impact on stocks performance whereas US equities have already been propped by 4 years of QE – different phases imho, especially when Fed is already ‘pondering’ the famous exit, and ‘monitoring’ ahem, excess risk-taking

  20. crunched says:

    BR -

    ‘if they are fully invested, they don’t see any Euphoria! The less exposure they have to equities, the more Euphoria they seem to notice. ‘

    Perhaps I shouldn’t have said euphoria. Let me put it this way… most people are beginning to see no way this market could ever go down (whether they’re invested in stocks or not). And let’s say the market continues to rise… Will it not become likely that investors will never fear the possibility of a bear market again? Is that realistic? Have we really found the cure to a ‘bear market’? Here, in the early part of the 21st century… will this decade go down in history as the era where humans finally figured out how to employ Quantitative Easing and avoid a bear market? That’s basically where we’re headed.

    And I disagree with your view of an investor’s perception on euphoria. You insinuate those who aren’t invested in stocks actually are suffering from performance anxiety and the fear of ‘missing out.’ So they see the whole rally as ridiculous and attack it. That may have been true in past bull markets, but when you factor in demographics, etc… many people are desperate to protect their money and recognize this QE inspired bull market as the farce that it is. The higher it goes, the more conviction they have to not be involved. So sooner than later all these banks and hedge funds buying overvalued shares are going to want to unload to someone and there isn’t going to be anyone.

  21. Woof says:

    What’s it called…confirmation bias? This chart seems to say whatever you want it to say.

    Would be very useful to see the two components (expected equity returns and risk free rate) also charted in the same time frames. In 1975 and to lesser extent 1982, the risk free rate only had one way to go: down. That fed booms, didn’t it?


    But the vehement opposition to the post’s point suggest something, too.

  22. Willy2 says:

    1. REAL interest rates are becoming less negative and on their way to becoming positive. Like in 1975 & 1982. And positive REAL interest rates are a KILLER for the corporate sector.
    2. Just see what happened to the “risk equity premium” after 1975 & 1982. It went (sharply) down. And that doesn’t bode well going forward.

  23. Pantmaker says:

    keep in mind folks …today there were just as many sellers as buyers…it’s easy to get swept up in this “to infinity and beyond” silliness.

  24. Alananke says:

    The piece ends up illustrating that the high equity risk premium in our environment is entirely a factor of low Treasury yields. In order to argue that this is bullish, you must first make the case that Treasury yields are reasonable. A more appropriate title might be “Sell bonds.”

    For giggles, I took Shiller’s data and calculated an equity risk premium using the nominal earnings yield (not 10yr avg) minus Treasuries. ERP is a little above average. When I do the same calculation using the median yield on Treasuries over the same period, it’s in the bottom 20% of historical readings.

    Valuations are attractive – in the respect that low fixed income yields make stocks attractive enough to encourage corporations to reduce float – but not equities are not as attractive as this chart implies.

  25. MayorQuimby says:

    I find nothing attractive about US equities in general but still own a handful happily. I just started selling a few things here and will continue to do so should higher prices come. My biggest longs were taken a few years ago, last autumn and this winter. All Bernanke PUT buys.

    When I see SBUX paying a 1.4% divvy at such a high PE and many other stocks paying 1% or 2% I get concerned. But when even rosy growth estimates only take those divvies to 2% then I get alarmed. Simply put – if you want to buy in here – sold to you – I’m not interested.

    And not everyone’s portfolio is as bearish as their forum comments….just sayin’.

  26. Scott Frew says:

    Barry, I’m sitting with a paper I’ve just reread, written probably 5 months ago, by Andrew Smithers, the most data-driven, agenda-less analyst of whom I am aware, and a profound thinker about markets and economies. The title: “The Equity Risk Premium: Useless in Practice and a Source of Confusion and Error.” He notes that equity returns have over more than 200 years of data fluctuated around a stable mean, roughly 6% real annualized. Because this is the case, and because the U.S. equity market can be valued, its probable return over a longer time period can be roughly estimated without any reference to estimated future returns on bonds, or without any reference to bonds at all. Note that GMO, whose estimated 7 year returns for a variety of asset classes have been stunningly accurate over a long period of time, uses Smithers’ methodology as the basis for their work. The returns on bonds, unlike those on stocks, have shown no such tendency to fluctuate around a stable mean. Bond and equity prices are determined entirely independently of each other, and are unrelated. I have no doubt that periods can be found when, ex-post, the ERP appears to have “worked” in predicting future returns, but there is neither theory nor empirical evidence to justify its use as a predictor of future equity market returns, particularly given that those returns can be very adequately predicted entirely without reference to bond prices or yields.

  27. socaljoe says:

    Stock prices and corporate profits are so distorted by ZIRP and QE that I would not rely on a model that is based on historical data where these distortions did not exist.

  28. DUAR2 says:

    High ERP is generally a bullish sign for stocks, but it is definitely never a foregone conclusion. Remember that ERP and the Treasury yield (aka the risk free rate, denoted as “Rf”) are the two components that make up the discount rate. The real issue is the relative changes in the two rates as we move forward. Assuming that Rf will rise as the Fed tapers and ends QE, the real question is whether ERP will fall even faster than the rise in Rf? If ERP falls faster than the rise in Rf, then the discount rate falls and the S&P500 will rise (all other things being equal). If ERP does not fall as fast as Rf rises, then the discount rate rises and the S&P500 will fall (again, all other things being equal).

    I personally think that ERP will fall faster than the rise in Rf as the year progresses. The current discount rate is high by historical standards (even after taking into account the extremely low Rf) — a likely legacy of the heightened fear in the market due to the painful memories of the GFC. As the economy continues to recover, more of that fear will probably dissipate, and the discount rate will reduce (or, in technical terms, ERP will fall faster than the rise in Rf).