Michael Covel's Trend Following Sep 21, 2026 42m 26m saved
With James Dailey, CEO of DUNN Capital Management
Measured on its own, trend following scores worse on risk-adjusted return than the alternative strategies it competes with. Added to a 60/40 portfolio of stocks and bonds, it improves that portfolio more than any of them.
That contradiction is the argument of DUNN Capital's new white paper, and James Dailey's point is that the standalone number is the wrong number to look at in the first place.
"But anyway, from a standalone basis, trend following doesn't look like the best alternative. But from a complimentary basis and what it does to add it to a 60/40, it creates the best improvement of all these other alternatives. That's the bottom line."
James Dailey, CEO of DUNN Capital Management, on Trend Following Radio, runs a firm whose composite record starts in 1974 and which has never traded anything but trend following.
The full interview is covered here so you can skip it. 42 minutes of audio, 16 minutes of reading.
Here are the 12 insights that matter.
Key Takeaways
Trend following has the lowest Sharpe ratio of the alternatives in DUNN's paper and still improves a 60/40 portfolio more than any of them — his argument is that the Sharpe ratio belongs on the whole portfolio, not on one line in it
There have been five equity drawdowns of 20% or more since 1984 and the firm's flagship program made money in all five
Commodities, energy, metals and agriculture are a little over 40% of the portfolio — the part of a 60/40 investor's exposure that does not exist anywhere else
The firm has been short fixed income for most of this year, a position Dailey says he sees almost no individual investor holding
DUNN runs at 22% to 23% annualized volatility, roughly twice the big-name trend followers, on the argument that a diluted version is easier to hold and useless when it is needed
A rival's study of 7,400 mutual funds put DUNN's sub-advised strategy second for improving a 60/40's return against drawdown, with a high-volatility competitor a basis point ahead
AI is allowed to convert and tidy code and nothing else — no creative research, no backtests
He expected a large US equity correction several times in ten years and was wrong each time, which is his own case for being fully systematic
1. A Story, Not Math
The interview opened on the problem Dailey says is the industry's biggest, which is not performance.
The hardest part of the job is explaining what the job is
And somebody asked me from the audience, what's the biggest challenge? And I said, educating people. It's so abstract for a lot of people.
James Dailey
He credited the host and Niels Kaastrup-Larsen with doing the educational work, and Cliff Asness with publishing the papers, and said the field still has not sold itself. A member of the audience told him the reason: everything he had said that day was math, and what he needed was a story.
What happened next was accidental. Dailey saw a line on Instagram, screenshotted it, cropped it and posted it to X with two words of his own.
I heard someone say, quote, "If you knew you were 100 rejections away from your dream, think how excited you'd be every time someone told you no." End quote. And that stuck with me.
James Dailey
A throwaway post outperformed everything he has written about the strategy
I woke up the next day 750,000 views.
James Dailey
He has about 45,000 followers and had never had anything approach that. His own reading of why is the point about storytelling: the line describes the shape of a trend follower's year — a long run of small losses that only makes sense if one of them turns into the trade that pays for the rest.
2. What Alternative Means
The host asked what an alternative investment means today, and Dailey went back to the original definition.
The word was about the timing of returns, not about the wrapper
Alternative investments are supposed to be investments, the original idea were investments that the timing of the return streams from those investments were different than the timing of the return streams from traditional investments, stocks and bonds.
James Dailey
What is sold under the name now, he said, frequently fails that test. A fund can look different in every visible way, being offshore, accessed through a swap, charging a performance fee and using derivatives, and still be the same exposure.
The structure is not the thing; the underlying driver is
But when you boil it down, they're still very correlated to equities because they're profiting off of the growth of the US or the global economy.
James Dailey
He was careful not to call those products bad. His objection is to where an investor files them.
It's not really a true alternative. If it's highly correlated with equities, if it's going to go down when equities go down, it's not working for you the way that maybe you envisioned an alternative would.
James Dailey
Asked whether he means private equity and venture capital, Dailey said both.
3. Opaque Versus Marked Daily
The host's follow-up was about visibility. In managed futures you can see the positions day by day; in private equity and venture the valuations are estimates and the money is locked up.
Every position has a price that somebody else set today
In managed futures, you have daily liquidity and all of the underlying investments. All the underlying futures contracts are marked to market every day.
James Dailey
There is no guesswork in the net asset value or the day's performance, he said, because both are arithmetic on observed prices.
The host's point was about scale rather than principle: private equity and venture capital are not a rich man's hobby but the destination of pension, endowment and foundation money, and of money belonging to people who may not know they are in it. Dailey agreed, and noted that the products have now been approved for 401(k) plans and that some pension plans carry very large allocations.
4. Why He Stopped Predicting
The host raised the usual objection — that US equities go up over time anyway. Dailey's answer was that history does not support treating that as a rule, and then he made the case against himself.
His own market calls have been wrong for a decade
No, honestly, Michael, I have thought that US equities would have a massive correction several times in the past 10 years. I thought we were really close. That's why I am fully committed to following the data and being systematic.
James Dailey
Nothing in DUNN's investment decisions is discretionary, he said, and he is more grateful for that in a year with this much geopolitical and market volatility than in a quiet one. The host offered an example of the failure mode: looking at an AI product, finding it unimpressive, and concluding the stock cannot work.
A rules-based process is protection from the manager, not from the market
And I really look at a rules-based process like what we have in our fellow trend following peers as being something that protects you from yourself because we all have these cognitive biases whether we're aware of them or not.
James Dailey
The host raised his recent conversation with Dimensional Fund Advisors' founder David Booth, who came down on the efficient-markets side of the old argument with Daniel Kahneman while admiring Kahneman's book, and who thought there was no practical way to implement the behavioral insight. The host's suggestion was that a firm like DUNN has implemented it, by reducing it to arithmetic.
What the firm controls is exposure, not outcome
The way I describe it to people is that we don't have any control over returns and over markets. We really don't.
James Dailey
The models, he said, are trying to take what the market offers and to manage risk systematically. An opportunity, in that vocabulary, is a market moving in a sustained direction long enough for the model to adjust and stay with it.
5. Five Dislocations Since 1984
The benefit a 60/40 investor is buying is usually labeled crisis alpha, and Dailey gave it a threshold rather than a feeling. A 10% correction is not what he means.
The bar is a 20% fall, and there have been five
So I'm talking about 20% or more.
James Dailey
The firm's record has two starting points, and the host pushed for the distinction.
The composite track record for our firm, which has always been trend following, starts in 1974.
James Dailey
The flagship investable program starts in 1984. Since then there have been five equity dislocations of 20% or more, and Dailey said the program made money in all five, as did trend following generally. His explanation is mechanical rather than prophetic: large equity declines take time to play out, which gives a model enough data to reposition and stay repositioned.
6. Where The Returns Come From
The other half of the diversification is what a trend follower holds that a 60/40 investor does not.
Over 40% of the book is outside financial assets entirely
So, we're going to have currencies, we're going to have all the commodities as well. So energies, metals, and all the agricultural products and commodities make up a little over 40% of our portfolio.
James Dailey
Asked whether the firm cares which market pays it, Dailey said it does not, and that the answer is awkward to give to a prospect without sounding flippant.
No, none of that stuff matters to us.
James Dailey
But the models don't care at all about any of that stuff.
James Dailey
He said he follows the macroeconomic picture so he can explain to investors after the fact where performance came from and why the models were positioned as they were. The models themselves read prices. In a typical calendar year, he said, a handful of markets produce almost all of the return while the rest generate small gains and losses that cancel out — sometimes as few as five markets carry the year.
That is also his answer to investors who ask for a version of the program with equities and bonds stripped out, on the grounds that they already own those.
The diversification is in the process, not only in the markets
We've been short fixed income for the vast majority of this year. I don't know any individuals that are shorting bonds.
James Dailey
Pressed that this amounts to a bet on rates rising, Dailey accepted the position and rejected the forecast.
The models have determined that right now the rates have a higher propensity to go up than to go down. And so we're riding that trade until proven otherwise.
James Dailey
7. Smoothing A 60/40
Asked to explain the correlation benefit in plain terms, Dailey gave the two-part version.
Equity-like returns with none of the equity correlation
Trend following has a very low and in some cases negative correlation to traditional assets. Trend following can generate equity-like returns with zero to negative correlation to equities.
James Dailey
The second part is what he called conditional correlation: the average is near zero, but the relationship is not evenly distributed. The strategy tends to do best exactly when equities do worst.
It offsets equity losses during those big dislocations, like the global financial crisis in 2008, like the dotcom bubble in the early 2000s, like 2022 for a recent example when inflation was going crazy and there were numerous rate hikes and equities went down 20% over a 9-month period.
James Dailey
He also described what the start of a dislocation actually looks like from inside the portfolio, which is less flattering than the finished chart. Trend followers are usually long equities going into a turn, because that is where the momentum was, so the first leg costs them money. What offsets it is that other sectors often move first, in gold, fixed income or oil, and the models are already positioned there. By the time investors are selling in earnest, the equity position has flipped short.
The host's addition was that money leaving equities has to go somewhere, and often goes into markets where a firm like DUNN already has a position.
8. Improving A 50-Year System
The foundation, Dailey said, is Bill Dunn's work from more than fifty years ago, and he was plain that a fifty-year-old method is not a finished one.
Age is not an argument against revision
Just because it's been around for 50 years doesn't mean you can't get better at it.
James Dailey
Roughly fifteen years of research and development have gone into it, and all of it points the same way.
And everything we've done has been to make the program more systematically adaptive to what's happening from all levels.
James Dailey
The three phases he named are trade entry, meaning trend identification; risk management, meaning the size of the position through the life of the trade; and the exit. All three are now more adaptive. The host's question was whether this amounts to a different strategy, and Dailey's answer was a measurement: if the changes had been cosmetic, the correlation between the current program and the old one would still be very high.
9. Positive Skew, Plainly
The host translated the academic term into the old trading rule, let winners run and cut losers short, and asked whether that is all positive skew means. Dailey agreed, with one correction.
Big losses happen; they are outnumbered
There are times when trend following has large negative outliers as well, but the number of large positive outliers far outweigh the number of large negative outliers.
James Dailey
10. Why DUNN Stays High-Vol
Several trend following firms have reduced the volatility of their programs, which also reduces the return. The host's question was whether an investor who bought a diluted version in the past twenty years has actually bought the thing they wanted.
Dailey's answer was that he understands exactly why firms did it.
A diluted product is easier to defend in a committee meeting
The watered down version is easier to hold and to explain to your investment committee.
James Dailey
But when you really need it, you're not going to have enough of it
James Dailey
To offset an equity drawdown with a diluted version, he said, an investor has to allocate a much larger share of the portfolio to it. DUNN's annualized volatility runs at 22% to 23%, which he put at almost twice the typical large trend follower, and the firm does a lot of investor education to keep people through the bad stretches. The payoff he offers in exchange is capital efficiency: a smaller allocation doing the same job.
The host's own guess, which he was explicit he could not prove, is that the watered-down programs did not start that way, and that some firms still run the full-strength version internally. Dailey's reply was two words: for their internal money.
11. 2nd Of 7,400 Funds
The concrete evidence Dailey offered came from a competitor. Eric Crittenden of Standpoint, who sells a pre-blended mixture of equities, trend following and cash, ran a study and sent it over with permission to use it.
Crittenden took Vanguard's balanced 60/40 fund as 90% of a portfolio, then tested each of the 7,400 remaining US mutual funds as the other 10%. About a thousand of them both raised the overall return and reduced the drawdown. He then ranked that set by return against drawdown.
A rival's screen put DUNN's sub-advised strategy second
And our strategy was number two out of 7,400 funds.
James Dailey
The fund a basis point ahead of it, he said, was another high-volatility trend following program — which is his argument in one line, since the ranking is driven by exactly the volatility that makes the product hard to sell.
Dailey's summary of the white paper is the same shape.
But anyway, from a standalone basis, trend following doesn't look like the best alternative. But from a complimentary basis and what it does to add it to a 60/40, it creates the best improvement of all these other alternatives. That's the bottom line.
James Dailey
On the Sharpe ratio specifically, the host asked why a trend following program should be judged by it at all. Dailey's position is that the measure belongs at the portfolio level, before and after something is added, and that DUNN publishes its standalone figure only because investors ask for it and refusing would look evasive.
The measurement belongs to the portfolio, not the line item
Sharp ratio should be used to measure a total portfolio of all your assets before and after different things are added.
James Dailey
12. AI As A Tool, Not A Brain
The host, who was in DUNN's office in 1996 and remembers the computer in the corner, asked what the firm actually does with large language models.
Dailey's answer was narrow. The research group uses AI for the operational side of research: making code more efficient, and translating code between the language the firm uses for research and the one it uses for production. Everything it produces gets checked.
The creative half of the work is off limits to the machine
We don't use AI for any creative part of the process. All of the actual research and development is done by human beings and our simulation environment and AI is just used to make that easier or more efficient.
James Dailey
Asked what he would tell someone who intends to ask a chatbot for a backtest, he gave the unglamorous answer.
I would say keep it simple. Focus on fundamentals. The blocking and tackling of research is clean data that you trust and that you've gone through and cleaned yourself.
James Dailey
I certainly wouldn't just instruct an AI model to backtest.
James Dailey
The host reached for Reagan's line about trusting and verifying, and Dailey took it.
Use it to automate operational processes. I think that's great, but research and development I would keep pure.
James Dailey
Bonus Insights
The firm is 25 people and almost nobody leaves
We have about 25 employees, the vast majority of which have been here for a long time. I've been here since 2003, and there are at least five people that have been here longer than me.
James Dailey
Dailey described the place as a hybrid: the relationships of a family office with a corporate approach to running money and servicing investors. The firm marked its fiftieth anniversary a couple of years ago.
The succession was planned, and it has already happened twice
Bill Dunn cared about his legacy and wanted the firm kept, in Dailey's phrase, in the family. Marty Bergin, Dunn's protégé, joined in 1997, became a partial owner in 2010 and took the reins fully in 2015; he has been the owner, president and leader for more than a decade.
We are one of the few of kind of the original old school trend following firms who has successfully transitioned to a 2.0
James Dailey
Bergin, Dailey said, is now thinking about the third generation.
The strategy is missing from financial television and present in sovereign wealth funds
It's not generally discussed on Bloomberg or on CNBC. It's not generally accepted in the financial media, but certainly it's appreciated by some of the world's biggest investors, sovereign wealth funds, pensions, endowments, and I think now advisors and high-net-worth individuals are really starting to appreciate what it can do.
James Dailey
Wall Street's own people are personal clients where they cannot be institutional ones
Dailey said the firm has a number of Wall Street professionals who invest their own money with DUNN while being unable to place institutional money there, and that this happens all the time. The reason given is usually volatility.
Crowded trades are the fuel, not the threat
Asked whether he worries about crowding in trend following, Dailey gave the opposite reading.
And I kind of think crowded trades are what's driving momentum. That's making money for us. We're not afraid of crowding.
James Dailey
The hardest position to hold is the one that is working as designed
The host's closing observation was that the difficulty is psychological and structural at once: trend following appears as its own line on a statement, so during a period when stocks and bonds are both working the investor sees an underperforming line item and asks why it is there. That is the period it is supposed to underperform.
Dailey's bottom line is that the number an investor should be judging is the portfolio's, not the product's — and that the product most worth owning is the one that looks worst on its own.
Products, Companies & Tools Mentioned
DUNN Capital Management (Composite trend following record from 1974, flagship investable program from 1984, about 25 employees, 22% to 23% annualized volatility and over 40% of the portfolio in commodities)
Standpoint (Eric Crittenden's firm, which sells a pre-blended mixture of equities, trend following and cash; his 7,400-fund study is the evidence Dailey cites)
Vanguard (Its balanced 60/40 fund was the 90% base portfolio in that study)
Dimensional Fund Advisors (David Booth's firm; the host's recent interview with him on efficient markets is the counter-position in the episode)
AQR (Cliff Asness's firm, credited with publishing the white papers that make the case for the strategy)
Bloomberg and CNBC (Named as the places the strategy is not discussed, against the sovereign funds and endowments that use it)
Books & Resources Mentioned
High Vol Trend Following: Most Valuable Alternative Investment (DUNN's white paper, the subject of the interview: it compares trend following against other alternative strategies standalone and then as a 10% addition to a 60/40)
Top Traders Unplugged (Niels Kaastrup-Larsen's show, which Dailey named alongside this one as the educational work the industry has had)
Daniel Kahneman's book on decision-making (Not named on the episode. The host says David Booth admired it while doubting there was a practical way to act on it, and suggests DUNN found one by reducing it to arithmetic)
Listen to the full episode
🔴 YouTube | 🔗 Episode page
Watch the full episode:
If this was worth your time, send it to someone closer to the industry than you are.
Get the latest market chatter as it happens: