Over the years, I have debated Wharton (University of Pennsylvania) Professor Jeremy Siegel numerous times. He is a very nice fellow who wrote the widely read book Stocks for the Long Run (aka SFTLR).

If I were to take the other side of the SFTLR argument, I would focus on 3 things:

My main critiques are:

1) Buy & Hold delivers inferior returns. Even worse, most Humans have a hard time sticking to it.

2) A simple system of either Valuation — think Shiller’s 10 Year Cyclically adjusted P/E — or Tactical application of Moving Average — Mebane Faber’s 10 Month moving average — significantly improves returns by reducing equity exposure and volatility as markets crash or have major corrections;

3) The current Fed driven markets (indeed, from 1981-2011) is an aberration that STFLR does not (and indeed, can not) anticipate. To quote either Jan L. A. van de Snepscheut or Yogi Berra: “In theory, there is no difference between theory and practice. But, in practice, there is.”

4) There are a myriad of data issues with SFTLR, particularly Survivorship bias, as described by jason Zweig in the WSJ and Birinyi Associates.

Regardless of my views, I want to crowd-source the arguments pro and con for SFTLR — What does Siegel get right, what does he get wrong? What is the weakest and strongest parts of his viewpoints?

What say ye?





Bonds for the Long Run (August 9th, 2012)

140 years of Equity Yield vs US Bond Yield (September 4th, 2012)

Bonds Beat Stocks: 1981-2011 (October 31st, 2011)

Revisiting Stocks For The Long Run (August 20th, 2012)

Jeremy Siegel is not having a good year (July 11th, 2009)

Category: Investing, Philosophy

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.

32 Responses to “Debating Jeremy Siegel & Stocks for the Long Run”

  1. A7LB says:

    There are some excellent, unanswered, questions regarding charting methodology at


    Hopefully, you can help answer them.

    Many thanks…

  2. craig.r.jackson says:

    Companies with highly capable management, with access to plentiful capital, that sell products with competitive advantage are candidates for long term hold. When one of the elements changes for the worse, sell. The SPY is not in that category. The ten month moving average and the ECRI index are not factors in determining when to sell companies in that category. That doesn’t mean these companies are without risk. Sometimes it’s hard to know when those factors have changed.

  3. maynardGkeynes says:

    I have always been perplexed by a key element of SFTLR. His case is largely based on historical evidence that purports to show that high dividend yield stocks, with dividends reinvested, have accumulated more total return than growth stocks or index mutual funds. However, his calculations do not account for the deleterious effect of taxes on reinvested dividend. (He says in an endnote that taxes are not significant for the portfolios he chose, but does not explain why; for most common stock portfolios, taxes are significant.) Dividends are taxed yearly and until recently at a higher rate than that of capital gains and that of retained earnings, which are not taxed at all. If taxes have been paid on dividends, only the untaxed part can truly be considered “reinvested”; the part that is taxed has to be made up by a new infusions of cash from the investor. The effect of ignoring this is that his historical comparisons are not terribly meaningful because he is not calculating the returns on true (after tax) contributions to dividend stocks vs. growth stocks. Naturally, if more is contributed to the dividend stocks, there is likely to be more at the end. (BTW, this is basically the same fallacy that sunk the allegedly huge returns of the otherwise delightful “Beardstown Ladies” of yore.) Given that the magnitude of the “advantage” he posits of dividend stocks vs. growth stocks is not all that great, one cannot have confidence that he has truly made his case. That said, his advice is very useful for investors in tax sheltered 401Ks. Also, the current lower tax rate on dividends also helps lessen, though not eliminate, the effects of yearly taxation of dividends.

    The newer editions of SFTLR attempts to grapple with what seems an even more perplexing challenge to Siegel’s original stocks for the long run mantra, the vexed question of what will happen if and when the populous Baby Boom generation attempts to cash in its stock and bond retirement portfolios by selling them to the smaller demographic of Gen X and Gen Y. An entire school of catastrophe futurologists, most notably Harry Dent, but also more mainstream voices like Peter G. Peterson (The Grey Wave) have warned that this so-called Age Wave is about to wreak havoc with stock market investments. Siegel does not dismiss this issue, but he does not convincingly refute it either. At the risk of oversimplifying a complex analysis, Siegel’s bottom line is that while it is true that there are not enough younger generation Americans to absorb the Boomers stock and bond assets at current prices, investors in emerging countries, like China and India, will more than make up for that and will end up buying the Baby Boomer’s paper assets as the Boomers sell them off to fund their retirements. The upshot is that foreigners will end up owning a lot of our companies by the year 2050. A potential snag, says Siegel, is whether America will be willing to let this happen, or will pass laws or adopt polices to discourage the transfer of US assets to foreign countries. This remains to be seen….

  4. Iamthe50percent says:

    May I point out that taxation of dividends doesn’t matter if the stocks are held in a tax-deferred IRA or 401K?

    While not agreeing with buy and hold or at least buy and hold blindly, I think the Professor was speaking to the majority who do hold most of their investments in IRA’s and 401K’s.


    BR: Except 401ks and Individual Retirement Account only date to 1974 . . .

  5. jimcos42 says:

    Siegel’s “long run” is longer than the long run most people have when planning for and executing a retirement plan. It’s certainly longer than planning for a child’s college education. Most of us are not running endowments or long-enduring pension plans.

    In my lifetime, I’ve drawn the cards called secular bear, secular bull, and now secular bear, with much more wealth at stake now than 45 years ago.

    So, like most of these “sure-fire” conclusions, context counts for much more. Where you start, where you end, and the slings and arrows of life that you have to deal with in between all matter. A lot. Theory and practice are decoupled.

  6. ConscienceofaConservative says:

    I very much do not like Jeremy Siegel. He ominously leaves out valuation when making his claim of stocks for the long run or adequately discuss the failure of his Dow 36,000 call.


    BR: I dont recall Siegel making a Dow 36,000 call — that was these eejits.

  7. adbutler007 says:

    The answers to your questions really relate to the magnitude of the equity risk premium, which is addressed at great length in Eric Falkenstein’s new book, ‘The Missing Risk Premium’, which you can download here.

    Further, what really matters to investors is the distribution of possible real returns over time horizons that are meaningful, like 10, 20 or 30 years. You can find a chart here that plots the distribution of real returns over 20 year horizons using Shiller’s data back to 1871. Notice that a chunky portion of the distribution is sub 5% real.

  8. mikeinconyers says:

    I think for most humans, keep buying and don’t sell till you pull out to fund retirement would be the best choice. My reasoning is that most are so ill informed and governed so completely by the (now) well know cognitive biases that we have been reading so many good books about lately, that if they are buying and selling, chances are they are screwing it up.

  9. DiggidyDan says:

    I don’t know what the long run really means, but for me, with the lost decade, and job prospects not very good, leading to increased “productivity” (read, overworking employees to an early grave to scratch out enough to pay the mortgage and the ever increasing grocery bill) I don’t know if I will make it to the “long run” of retirement. A couple more years of this and it won’t matter.

  10. RW says:

    jimcos42 makes a solid point: context matters, a lot. Investors don’t accumulate for a century, they do the bulk of their accumulation over maybe two decades and then do most of their distributing a few decades after that if they live that long. The decades in which each of these phases occurs can clearly make the difference between a tough retirement and an easy one.

    Crestmont Research has developed matrices depicting returns over multiple stock market periods for taxpayers and for tax exempt or deferred investors as well as returns on a nominal and real (after inflation) basis http://www.crestmontresearch.com/stock-market/ but also posts summary analysis such as Waiting For Average: Why The Long-Term Average Will Never Occur For Today’s Investors.

    The upshot is that accumulating during a secular bull and distributing during a secular bear is not going to give the same results as the opposite case. In these days of tax deferred vehicles the biggest problem with the SFTLR is that it tries to be a one-size-fits-all solution in a world where roughly one third to one fourth of the time (given average length of secular trend) that size could be close to lethal.

    The inimitable Mr. Micawber advises David Copperfield, “Annual income twenty pounds, annual expenditure nineteen nineteen six, result happiness. Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.”

    For the average investor, accumulating in a bear and selling in a bull, result happiness. Accumulating in a bull and selling in a bear, result misery. You don’t get to choose when you are born or when you grow old but you can choose the saving and investing style that better fits your times but, for all that, the times may simply not favor you. Folks who insist that timing is a bad thing don’t realize that everyone is a timer and there is actually no choice about that at all; the choices have to do with horizon and style.

  11. Randel says:

    I audited some of his classes at Penn. Professor Jeremy Siegel is correct. In an overlapping generations model where trusts, corporations, and pensions are infinitely lived, his theory holds up well. Mere mortals with finite lives have to deal with matters as jimcos42 states above.

    That all said, there is a big caveat to both Professor Siegel’s and the BR approach: the history of mankind shows most if not all working till death. The idea of retirement, taking cruises, drinking daquiris, playing golf, travelling the world, living at the beach is a very modern American idea of the past 40 years. It will not be so going forward. Biblically speaking, we are here to work. It is the fool who believes a life of idleness and plenty awaits, actually sounds pathetically boring and full of enui.

    You have an excellent site BR.

  12. Joshua Brown says:

    what Jeremy Siegel gets right is that 95% of all 10 year rolling periods (each beginning at month’s end) have produced positive returns for the US stock market and 100% of all 20 year periods have produced positive returns the US stock market using only “clean” data. This encompasses all major crashes post-WWI and in the time frame that data became reliable.

    If you’re invested in indexes rather than in all 500 S&P stocks individually, survivorship bias is irrelevant, the index will add and subtract and the index product you hold will mimic it, who cares which individual company faltered along the way.

    Thus, buy and hold is great for broadly diversified portfolios with a long time horizon, specifically if costs and friction can be contained. Market timing is actually riskier than staying in the game and, thanks to inflation, being out of stocks over the long term is deadly to your purchasing power.

  13. A7LB says:

    “There are myriad data issues with SFTLR, particularly Survivorship bias…”


    If these myriad issues remain unaddressed, how valuable might any investing related conclusions be?

    The only logical answer is: zero. No value.

  14. gman says:

    I agree with your point about “homo sapiens have a hard time sticking to SFTLR”..but it may take even more skill and discipline for the human primate to to stick to the models and game plans you rightfully advocate.

    Although my expertise is not in long term asset management, I am amazed by the burst of emails from friends and extended family I received first when the mkt reached 1100 and another recently @ 1400…”could you take a quick look at what I have and tell me what you recomend?”…I look and they are 60-80% cash, missing a huge chance to average down, collect dividends and then sell upside calls all the way back up.
    Many of these people no matter what the model or adviser told them (most of whom stink and commit soft core financial malpractice w/ the product they push) would have ended up almost in the same spot.

    Given that landscape, a “set and forget” active index plan w/ annual rebalance may at least be an ok alternative. One caveat, buy at a random day mid month. Do not buy on the 31st/1st because algos have already tended to put a nice mini markup for the EOM dumb 401k money being put into the mkt out of payrolls.

  15. gman says:

    “May I point out that taxation of dividends doesn’t matter if the stocks are held in a tax-deferred IRA or 401K?

    While not agreeing with buy and hold or at least buy and hold blindly, I think the Professor was speaking to the majority who do hold most of their investments in IRA’s and 401K’s.


    BR: Except 401ks and Individual Retirement Account only date to 1974 . . .”

    When I field that type of investment question at a cocktail party I always respond by saying “I don’t expect the stock market to go down by 30-40% in the next year and that is what it would have to do for you to lose money once you consider the tax implications..so keep contributing”

  16. bart says:

    “Don’t gamble! Take all your savings and buy some good stock and hold it ‘till it goes up, then sell it. If it don’t go up, don’t buy it.”
    – Will Rogers

  17. gordo365 says:

    I occasionally portions of my Enron and AIG stock to offset gains when I sell other winners.

  18. Bob A says:

    If you like to spend your life in a stupor and you can afford to retire on half of what you’ve got… buy and hold is something you might consider. Otherwise you better sober up and pay attention to the cycles.

  19. nofoulsontheplayground says:

    I’m all for SFTLR in 17-1/2 year periods, and only if you sit out alternating periods. Otherwise, timing is of the essence.

    The next buy point if you are playing SFTLR in alternating 17-1/2 year periods would be in the fall of 2017. You’d exit those positions around spring 2035 if you followed this methodology.

  20. viningengineer says:

    You might want to review Bill Gross’ comments on Siegel and his stocks for the long run in his August PIMCO newsletter. Summary: 1) Gross concludes (as someone noted above) that Siegel’s rule of 6.6% return only works for the REALLY long run (100 years +) if at all. 2) Gross points out that stocks over the long haul should also track the Country’s GDP growth (at 3.5% for same time periods). Stock returns have been outsized in recent decades. Conclusion is that they have been getting the lion’s share of GDP growth, while labor, bonds, and government have been reducing their take. This could be bad for stock returns (and taxes) in the coming years if this all implies an ugly reversion to the mean is coming. It also implies that Siegel might have been inadvertently cherry picking his data simply by only looking at recent decades. Don’t know if I agree with all of Gross’ assertions, but they are worth contemplating and may be fodder for a couple of zingers to toss Siegel’s way for him to expound further on.


  21. socaljoe says:

    How would one apply STFTLR in practice?… Buy the SP500 and hold it forever?… Surely you are joking.

  22. Tren Griffin says:

    Answer to Ritholtz Question:

    Jeremy Siegel’s detractors include Charlie Munger:

    “Q: Ibbotson finds 10% average returns back to 1926, and Jeremy Siegel has found roughly the same back to 1802.

    Munger: Jeremy Siegel’s numbers are total balderdash. When you go back that long ago, you’ve got a different bunch of companies. You’ve got a bunch of railroads. It’s a different world. I think it’s like extrapolating human development by looking at the evolution of life from the worm on up. He’s a nut case. There wasn’t enough common stock investment for the ordinary person in 1880 to put in your eye. http://www.kiplinger.com/features/archives/2005/11/munger3.html?kipad_id=1#ixzz280PNI1QL

    The fact that Sigel’s numbers include survivorship bias makes Munger even more right.

  23. Finster says:

    A good observation regarding tax effects. I am less convinced regarding the Baby boomer to Gen X, Y doomsaying. Simply put, the smaller Gen X and Y do not overwhelmingly buy the portfolios of the boomers, they inherit them. For a cataclysmic sell off a great preference of liquidity over holding assets would have to materialise and we would have to see the insurers and pension funds to unload large amounts of equity in exchange for cash. Such an event is much more likely to be spawned by uncertainty and panic in the financial sector.

  24. larrr1 says:

    I think Jeremy Siegel goes so far as to be harmful, in obvious ways – namely touting stocks when they are expensive.

    YET, I think buy and hold indexing for most investors is the only viable option. Shiller/PE10 valuation strategies involves years of patience (years of underperforming) to work out – not even pros will stick with it, frankly. Meanwhile, I do not believe a 10 month moving average strategy is robust enough to work after taxes and expenses.

    In short, Siegel often gives a ridiculous (bullish) take on the market at any given time, yet for most people something approximating buy and hold with some annual rebalancing is about as good as you can hope for.

  25. AHodge says:

    my four to add would be
    1 you can do better with a minimum of cycle timing-real rough doesnt have to be perfect
    2 the data is bad and more survivorship-and selection bias issues than are often counted
    3 the notion of superior returns above GDP cannot work real long term,
    and blows up of you try to model it for 30-50y ears
    4 if you get a managed portfolio for this and they take say 1.5% per year managment fees
    you end up 20 years later with roughly 30% less

    also if you buy at a bad time of the cycle it can be a real disaster.
    say someone pushed all the chips in in 1999

  26. rpseawright says:

    I generally agree with you, but what is your best data source for propositions 1 and 2, Barry?

  27. Lord says:

    A really significant one would be if global population plateaus this century, doesn’t that assure slower growth and lower returns in the future than we have seen in the past though probably beyond our lifetimes. If you want a real question, ask him why low beta value stocks outperform over the long run even when we know they have, that is, why it persists. I suspect this has to do with the limited time horizons of market participants that must rely on trading when buy and hold is insufficient. There is some evidence people are better at buy and hold in their retirement accounts. Buy and hold is hard but market timing is often harder. There is some evidence of better risk adjusted returns but not absolute returns using moving averages, but I believe most of these can be captured by rebalancing. Survivorship is less of an issue with indices as they represent a crude form of selection and timing such that their calculated returns would understate actual returns. History is replete with their adding stocks after they have boomed and removing them after they have busted.

    My expectation is stock total returns are necessarily double bond total returns. This provides an equivalent income from each according to the growth of the economy while offering a growing valuation in exchange for volatility for stocks or a stable valuation in exchange for reduced volatility for bonds. Were it not so, either asset class would take over the other.

  28. [...] Can you name the many ways in which the “stocks for the long run” argument is wrong?  (Big Picture) [...]

  29. MPhelps says:

    machinehead’s comment about the period shown being a period of decreasing interest rates is obviously true, but that helped stocks even more than long bonds over that period. Estimating the duration of stocks is difficult, but using the price/dividend ratio as Hussman does (http://www.hussmanfunds.com/wmc/wmc050829.htm), stocks have at least twice the duration of long bonds. Meaning the effect of generally decreasing interest rates over the period shown above was more beneficial to stocks than bonds. And, a future rising rate environment would be a stronger headwind to stocks than to bonds.

    I am not suggesting that I think stocks will underperform 30-year Treasuries over the next 30 years, just that the interest rate argument machinehead uses cuts both ways.

    I have two grievances with the oversimplified “stocks for the long run” argument. The first is that it only pits stocks versus US Treasuries – a (credit) risk-free instrument. The universe of bonds is much broader than US sovereigns with (assumed) zero credit risk. Investment grade corporates, MBS, ABS, CMBS, non-US dollar-denominated sovereigns are all part of the Barclays US Aggregate bond index. High yield corporates and a large chunk of the mortgage market aren’t included in that index, but are well developed US bond markets. Simply using the term “bonds,” but only referring to US Treasuries is naive. (investingforaliving hits on this too. Just because the Barclays Aggregate Index was only invented in 1976 doesn’t mean it should be ignored for periods after 1976 – especially if the point of this analysis is to say something about the future. It is a much more valid benchmark for the bond market than long UST.)

    My second grievance is that most investors do not use bonds for portfolio growth, so comparing them to equities simply on the basis of total return over a given period is comparing apples to oranges. They serve different purposes in portfolios than equities do, including volatility-dampening and liability-matching, to name a couple. For these purposes, bonds greatly “outperform” equities.

  30. tippet523 says:

    I do believe a lot of this stems from firms who would prefer most folks own an asset allocation and stick to it. Rarely is that allocation 100% equities, so an annual rebalance will help a portfolio. But I believe most firms do not want the mass of folks to actively change the equity stake for fear they will be wrong more than they are right

  31. [...] Can you name the many ways in which the “stocks for the long run” argument is wrong?  (Big Picture) [...]

  32. [...] Ritholtz asks here:  “I want to crowd-source the arguments pro and con for [Stocks for the Long Run] SFTLR — What [...]