The Quantitative Momentum Investing Philosophy

About the Author

Jack Vogel, Ph.D.

Jack Vogel, Ph.D., conducts research in empirical asset pricing and behavioral finance, and is a co-author of DIY FINANCIAL ADVISOR: A Simple Solution to Build and Protect Your Wealth. His dissertation investigates how behavioral biases affect the value anomaly. His academic background includes experience as an instructor and research assistant at Drexel University in both the Finance and Mathematics departments, as well as a Finance instructor at Villanova University. Dr. Vogel is currently a Managing Member of Alpha Architect, LLC, an SEC-Registered Investment Advisor, where he heads the research department and serves as the Chief Financial Officer. He has a PhD in Finance and a MS in Mathematics from Drexel University, and graduated summa cum laude with a BS in Mathematics and Education from The University of Scranton.

  • TheChrisp1231

    Looks good guys , I’m keen to get some access to the momentum premium ! , I thought you guys had mentioned that you can add drawdown reducing techniques/hedging to momentum portfolios but not to value .So i was surprised to see none applied above .

    Also do you guys have the yearly returns ? , be interesting to see the differences in generic , low quality and high through different periods ( the dot com,)

  • Hello Jack,

    Aren’t the high drawdown levels in excess of 60% of concern and maybe this is an answer to the question: Why aren’t all investors doing it?

    Add slippage and other friction effects and drawdown can get even larger.

    It appears that these strategies are too sensitive to volatility spikes and clustering.

  • Jack Vogel, PhD

    Yes, the portfolio above is 100% invested, so there are no stop-loss rules. The stop-loss rules can help to attempt to lower drawdowns but we prefer to use these outside of the security selection process.

    Here are the annual returns:

  • J K


    This is very interesting. Thanks for the findings!

    But how does the “frog-in-the-pan” phenomenom relate to the findings of “momentum acceleration” studies, e.g.

    In addition to that study, there are also additional studies suggesting that stocks with high price acceleration outperform stocks with low price acceleration.

    Isn’t that somewhat in conflict with “frog-in-the-pan” anomaly?

  • TheChrisp1231

    Thanks Jack , one more question: was a similar premium ( quality momentum over generic) observed in international stocks ?

  • Jack Vogel, PhD

    In our tests, yes.

  • Jack Vogel, PhD

    I can’t view the paper on SSRN, can you send the link?

  • JAK78

    Are you at all concerned about excessive data mining and overfitting your data by having multiple screens?

  • Jack Vogel, PhD

    No, we are not optimizing a model and loading on factors that worked the best. We use a screening methodology takes 3 steps, all backed by a sound theory and then academic evidence. Step 2 sorts to the top 10% of firms on momentum — there appears to be a continuation of returns for past winners and losers (Jegadeesh and Titman 1993). Step 3 examines only the top half of the top 10% of momentum firms based on the path of their momentum, where we look for high-momentum firms with a “smoother” path. The theory is that these firms are more likely to be overlooked, and have a larger continuation in returns, as in Da, Gurun, and Warachka 2014. Step 4 examines the seasonality component of momentum. The theory is that window dressing causes mutual fund managers to purchase the best-performing stocks to show this in their portfolio at the end of the quarter. The evidence in Sias 2007 shows that the momentum portfolios perform the best in quarter ending months — so we use this to time our rebalance.

    I hope that helps!

  • Doug

    Nice work! A few questions/comments from the peanut gallery:

    1) Why exclude REITs? To make the overall asset allocation decision cleaner? (i.e. if you already have 10% in VNQ, then you won’t have an extra 1%-3% sneaking into your portfolio via QMOM?)
    2) I’m concerned that the seasonal anomaly is ripe for the picking by specialized funds, and may get arbed away in the future.
    3) Right now, Large Cap Growth Funds (NOT growth “stocks” but “funds”) have a big momentum component. MTUM has closely tracked VRGWX (Vanguard Russell 1000 Growth Fund) since inception. That may change, but Large Cap Growth has been a cheap substitute for momentum over the past 2-3 years.

    Despite the quibbles, I’m looking forward to QMOM/IMOM launch!


  • Great work, as always!

  • J K

    Helli Jack, seems that the paper has been removed from SSRN, I can’t find it anywhere either. Sorry!

    But here are two similar papers on the same (momentum acceleration) topic. And it looks to me as if they don’t really support the “frog-in-the-pan” theory. Maybe you could comment these? Thanks.

    Momentum acceleration:

    Investor attention, visual price pattern and momentum:

  • Jack Vogel, PhD

    Yes, this strategy has a large tracking error, and in the past had a larger drawdown than the market.

  • Steve

    Jack, I think I’ve sent to Wes before – will email again. See the post you guys did a couple days ago on momentum, I talk about this in one of my comments to that thread.

    To J K – I don’t believe they relate (directly). Acceleration is different to return consistency.

    Also, the way Wes/Jack prefer to rebalance (ie. more frequently), they might not find as much benefit to acceleration. There’s a paper (Ibbotson) from memory (recent one)…that shows acceleration as a bad thing, but that’s because of the monthly rebalance – and the paper uses it as an explanation for short term reversal.

    However, the Marks paper shows Acceleration as useful on a longer holding period (because the benefits come later. i.e. Accelerating momentum stocks don’t actually do better in the shorter term, its later on that they do). That’s what makes it interesting for the investor trying to hold on to that magic one year mark (also a reason to use 3 or 6 month momentum rather than 12 month).

    Also – my opinion – I think the Moving Average ratio paper (Parkes, from memory) inadvertently also shows the acceleration idea.

    There is also one other paper on this that goes really well with the original Marks paper – trend salience, by some Aussie researchers.

    Jack – I’d love to see the AA team do your thing to examine this idea from the real world perspective of investors! Can we twist your arm? I’ve not yet seen any papers do a back up study of trend salience / acceleration on any other markets – so I’ve still got the usual lack of confidence when only tested on one market.

  • Steve

    Monster post! Congrats guys, on pinning down your strategy after all the research – and presenting it in such a methodical manner. Love the symmetry with QV!

  • Jack Vogel, PhD

    Both those papers are interesting, however the frog-in-the-pan paper is in one of the top 3 Finance Journals (Review of Financial Studies). We did test a related paper to the ones you listed ( ) and found frog-in-the-pan momentum was better (when comparing the long-only strategies on mid/large cap stocks).

    All else equal, just being in the top decile of stocks formed on momentum was a good bet in the past. All of these papers are trying to find a better way to sort “within” the top decile — a noble goal! We like the frog-in-the-pan measure as it is backed by a behavioral theory (people underreact to certain stocks, those with less attention).

  • Eric Darnel

    Even according to your own research, value stocks offer higher returns with lower standard deviation. So I wonder, why bother with momentum?

  • Mark

    What if you start with step 3, followed by step 2?

  • Jack Vogel, PhD

    In general, value and momentum tend to work well at different times, so the overall volatility can be decreased in a portfolio by combining the two (however still slightly higher than standalone value). Here is a blog we wrote on combining value and momentum

  • Jack Vogel, PhD

    It is better to do 2 and then 3

  • J K

    Jack, thanks a lot for your comments, and Steve, for your additional insights (as well as the interesting papers). Would be interesting indeed to see a back up study of trend salience / acceleration.

    Thanks again, AA team, for the great work!

  • Steve

    Just pushing the question – not sure if you looked at it, but if you did…what about the seasonality in international markets? I wonder, because not everywhere has quarterly reporting.

  • Jack Vogel, PhD

    in our tests, yes on the end of quarter effect. No on the January effect.

  • Steve

    Thanks Jack – that’s really interesting. I knew about the Jan effect being US centric…but I didn’t expect that answer re: quarter effect. Hmmm…

  • Anthony Ialeggio

    Great post! The drawdowns seem to be the biggest concern here, as others have pointed out. I use a dual momentum approach in my portfolio to limit downside in each asset class. Do you think that would work well if I employed this approach in my equity allocation? Thanks.

  • If you want to avoid equity-like drawdowns on a long-only strategy you need to deploy some sort of market-timing mechanism as an overlay. Dual momentum is certainly a reasonable approach, but there are also others:

  • Thomas

    Have you looked into applying dual momentum to QMOM and IMOM ? I.e., simply focusing on equities, e.g., QMOM or IMOM / BND or BNDX ?

  • Simon Fung

    Hi Jack,
    I read with great interest your good book : DIY Financial Advisor and learn a lot. But I am not sure I understand why in calculating the cumulative return, you ignore that for the last month ( page 146 refers). Your advice please.

  • Jack Vogel, PhD


    Becasue there is a short-term reversal in returns from month to month. Here is a blog post we wrote on the topic:

    Two papers document this, Lehmann (1990) and Jegadeesh (1990).

    Since we know there is a short term reversal in returns, we exclude this from our momentum calculation. That being said, if one includes the past month, the results are similar.

    I hope that helps.

  • Simon Fung

    Thank you very much for your quick return. When I try to DIY, I ran into another problem :
    When u mentioned risk free return of treasury bill in page 110 of your good book, are u refering to the yield of 1-year US treasury bill or 10-year treasury bond yield? Also in the last column of table 7.2, are u simply multiply the monthly returns to arrive at the 12-month cumulative return?
    Your advice please

  • Jack Vogel, PhD

    We use short-term T-bills return (1 or 3 months) and yes you multiply 1+ the return in the last column in Table 7.2

  • Henning Hammar

    Great post and a great new product. The momentum effect seems really interesting to use and especially in combination with value. Momentum has seemed to work well the past year, but looking at the past 10-15 years the effect seems to disappear for US stocks, something discussed on Jesse Livermores blog and that can be easily checked using Fama-French data. Do you see that effect in your data and why should momentum continue to work if not arbitraged away? Quantitative momentum seems easier to arbitrage away than quantitative value.

  • Hi Henning,
    Generic 2-12 momentum has been “blah” in the recent past (talking about the long-only strategy here). You still see benefits to a “quality momentum” strategy, however. And as the post above highlights, as well as the original frog-in-the-pan paper, the so-called momentum effect is driven by momentum stocks that steadily grind to get their momentum. So one could argue that generic 2-12 has never really worked, especially after controlling for those quality momentum stocks that actually drive the anomaly. At least for the time being, there isn’t much evidence that the “quality momentum” stock effect has been arbitraged away. If you layer in seasonality elements, the probability of the effect being completely arbitraged away gets even lower. Of course, anything is possible, but the intense volatility of momentum strategies will tend to keep a lot of arbitrage capital at bay…

  • Henning, forgot to mention a more wonky — but valid — argument that momentum hasn’t “lost”. The argument is one put forth by Asness et. al. Basically, even if momentum excess returns were ~0, they bring so much diversification benefits to a portfolio, they should still be considered “anomalous.”

  • Henning Hammar

    Thanks for the quick reply. Interesting thoughts. It seems “quality momentum” has a certain benefit compared with regular momentum, especially in combination with value.

    Thanks as well for the paper, will have a look at it. I’m also looking forward to see your results for a risk-managed QMOM, especially due to the large drawdowns.

  • Hannibal Smith

    You still made the classic mistake of style boxitis! The opposite of value is NOT growth. The results at the other end of a screening filter is not the inverse opposite, just the lowest quality results of the filter. If you screen filter for growth properly, you should see that technical momentum is merely a sub-set of fundamental momentum. After all, what ultimately drives stock prices higher? Earnings.

  • Mark

    Hi Jack,

    Is there a good reference paper outlining why you should use a 12 month lookback rule over other time frames or other more complex measures of momentum?

    Thanks for the great post.

  • Jack Vogel, PhD

    Thanks Mark. Here is our summary of the original paper using different lookback and holding periods

  • Lucas Nogueira

    Hello Jack,

    Has a finance student I have a lot of interest for market anomalies and risk, so I’m very curious about your way to diversify your portfolio. What is your opinion about anomalie/quant base stock picking and portfólio diversification using markowitz/sharpe ratio? I’m making this question because seems so odd to me building a potfólio based on a comportamental/behavioral aspect but using a market efficiency approach to minimise risk.

    Thank you!

  • Jack Vogel, PhD

    Somewhat related — here is an asset allocation post we wrote:

  • dph

    Jack (or Wes…) How do the return differ if the re-balance was monthly vs quarterly and if you dont use 6 vs 12 month look backs?

    And would even a shorter look back, as a confirming screen, maybe used to avoid momentum crashes like 2008. Can the crashes be mitigated without killing the CAGR?

  • Jack Vogel, PhD

    Here is a summary of the original paper (JT 1993) which varies the look-back period and the holding period.

    Main story is intermediate-term price momentum shows a continuation using past data.

  • dph

    I assume in long/short momentum the extreme drawdowns are when ones shorts gets crushed on market bounces and long only gets crushed during the market fall?

    Could long momentum drawdowns be used to time/predict and scale out of the short leg? Would it make sense to scale out of shorts as the market drops every 20% or so?

  • Jack may have some additional thoughts. Here are some off the cuff remarks…

    The epic drawdowns typically occur because there is a “beta” mismatch when markets shift. So after a drawdown, low beta stocks will be the “winners” and ultra high-beta trash will be the “losers”. If the market experiences a violent recovery, the “winner” book will go up, but not much, but the ulta high-beta trash will absolutely go bananas– but you’d be short and get your face ripped off….ouch.
    See here for a good paper (you’ve probably read it at one point):


    1) Dynamically match beta?
    2) Scale based on volatility or trend
    3) stop-loss type setups…
    4) Don’t short losers…too dangerous!
    4) probably others…

    Some old posts you might like

  • janvrots

    I am not an expert at risk adjusting returns or managing a portfolio with a constant risk. I would like to see an AA piece explaining this.

    Please correct my analysis as I want to understand this better. In your analysis the sp500 generated a 10.92% return with a 15.4% stdev. The qmom portfolio generated a 17.03% return with a stdev of 25.38%.

    In order to drop the risk on the qmom portfolio to the sp500 risk the qmom return falls to 10.2% (17.03% * 15.4%/25.38%). This means that on a risk adjusted basis qmom does not beat sp500. What am I missing?

  • The Sharpe ratio suggests that this is approximately correct, although there is a slight edge to the quant mom system. But let’s just assume their Sharpe ratio’s are exactly the same. The real question is how much of the excess risk associated with momentum is systematic vs idiosyncratic (diversifiable). The evidence suggests that systematic momentum strategies are high volatility, but a lot of that vol is diversifiable, and therefore “washes out” when included in a broader portfolio.

    See the following articles for insight into these ideas:

    Bottomline: You are correct and if quant mom was your only risky asset holding it would be arguably equivalent to the S&P 500. However, most people would presumably use quant mom in a broader portfolio set where they can more effectively harvest the higher expected returns without pushing up their overall risk

  • Hannibal Smith

    It would be more like 10.40%, 10.13% and 8.56% using downside SD. No point in including upside SD as “risk”.

    Interestingly, correlation is included in the SD calculations, but you’ll only see that effect by combining individual systems into a single portfolio not by relative comparisons of said individual system’s SD values.

    SD is really a flawed and terrible metric, but I digress.

  • Agree.
    I like the appraisal ratio. Sortino is good for accounting for downside SD. Probably best to view a strategy across a variety of metrics and to understand the use case.

  • idwright


    Looking at your Momentum selection procedure I believe you may be prejudicing your Momentum results by making the min Market Cap liqiudity criteria too high. I have noticed that PEAD and Smooth Momentum tends to be more prevalent in under researched smaller companies. Would be interesting to test again with say $1B min Cap. Ian

  • Great question. We always look at the robustness of our various systems across size so we know what we are dealing with and the tradeoffs we face between transaction costs and expected returns. Price momentum, unlike value (or PEAD/earnings based strats), is less influenced by the size effect. There are some marginal benefits to going smaller, but the additional expected costs associated with impact,frictions, etc. likely offset these benefits.

  • idwright


    I have searched for further positive PEAD data. I found a summary of a French paper published in 2015; “Post Earnings Announcement Drift, a Price Signal”? – Julien Messias. He studied the US market 2003 -2015 and concluded, “We find a strong empirical evidence of the pre-eminence of this (PEAD) bias for Momentum stocks rather than blue-chips or non-Momentum small-caps. Even by challenging the strategy, the conclusion remains strong with abnormal returns linked to such market inefficiency, with better returns for positive signals than negative ones”.

    Unfortunately I cannot access the whole paper and the English is not perfect, but it appears to say that Blue Chips and non-momentum small caps do not exhibit strong positive PEAD. Do you read it this way? Ian

  • I have not read the paper so I can’t really comment to be honest. I’ll see if I can hunt down the source document

  • anon1238

    lol…I figured this out on my own in college. It took two PhD’s to come up with this?

  • we got too much brain damage banging our heads against the wall building complex systems. So we settled on something fairly simple after the dust settled. You are obviously smarter than we are.

  • anon1238

    I know 🙂 intuition supersedes empirical research with less certainty but faster results. All models are estimation anyway.