DFEN Fell 33% in a Month While Its Index Fell Only 11%
DFEN closed Tuesday near $61, down from roughly $92 on August 6. That is a third of the fund gone in eighteen trading sessions, and a number that size invites the assumption that something broke. Nothing broke. The fund did precisely what its prospectus says it does, and the entire result is available from one line of arithmetic.
DFEN targets three times the daily return of the Dow Jones U.S. Select Aerospace & Defense Index. That index fell somewhere in the region of 11% across the same stretch. Spread evenly over eighteen sessions, an 11% decline works out to about 0.65% a day. Triple it and the fund loses 1.94% a day. Compound 1.94% eighteen times and the result is 29.7%. The realised loss was 33%. So of the thirty-three points, close to thirty are leverage applied to direction, and roughly three are everything else.
Those three points are the decay that dominates most retail commentary about these products, and the proportion is the useful part. Decay is a function of variance, not of direction. A rough approximation for a fund with leverage L is that the drag runs at (L² − L)/2 times annualised variance, which for a 3x fund is three times variance. At 30% annualised volatility that is about 27% a year, or a little over 2% a month. At 40% it is closer to 4% a month. So a flat, choppy month in the underlying index costs a DFEN holder a few percent. A month like the one just finished costs thirty. Anyone who watched this drawdown and concluded that leverage decay is the risk has learned the wrong lesson from the right event. Decay is what kills you slowly when the sector goes nowhere. Direction is what killed this position, and it did so in three weeks.
The more interesting question is what pushed the index down 11% in the first place, because the answer has almost nothing to do with defense.
Start with what DFEN actually owns. Published holdings pages often show GE Aerospace near 14.7% and RTX near 11.4%, which describes the equity sleeve. The balance of the exposure sits in swap agreements, and on an index-weight basis GE is closer to 23% and RTX to 17%. Those two names carry roughly 40% of the fund. Boeing adds another 8%. The defense primes that give the ticker its name, Lockheed Martin, Northrop Grumman, General Dynamics and L3Harris, share what is left.
GE dropped 5% on August 19 to $356.23, against a 52-week high of $388.84, then a further 3.07% on August 27. In neither case was there company news. The commentary attached to both sessions cited the same two things: a stretched multiple at 35x to 41x forward earnings with a PEG around 2.5x, and profit-taking after a long rally. The July quarter had beaten on both lines, $2.02 in EPS against $1.86 expected and $12.63 billion in revenue against $11.87 billion, with revenue up 21.1% year over year and the full-year outlook raised. The stock de-rated into improving numbers. That is multiple compression, and multiple compression in a 23% position is enough on its own to move an index several points.
Which exposes the structural point that the ticker hides. DFEN says defense. The money says commercial aerospace aftermarket. GE Aerospace sells engines and then sells decades of servicing on each one, a recurring-revenue business tied to how much of the installed fleet is flying. RTX carries Pratt & Whitney and Collins alongside its missile business. Boeing builds airframes. An investor who bought DFEN on the view that global military budgets are rising has instead purchased a triple-levered position in the forward earnings multiple of a jet engine aftermarket franchise, priced at a growth multiple and therefore vulnerable to exactly the sort of rate-driven and valuation-driven repricing that hit it in August.
The year’s scoreboard makes the same case more quietly. Through late August, RTX was up 14% for 2026 and GE up 11%, while Boeing sat 3% in the red and the passive basket returned about 8%. That is a middling single-digit year for a sector whose spending outlook improved continuously throughout it. The structural defense thesis and the price behaviour of these particular equities are separate propositions, and they have been separate all year. Lockheed spent months working off a Q1 free cash flow problem. Northrop spent months working off a guidance cut. Neither trajectory was set by the appropriations calendar.
None of which makes $61 cheap or expensive, because valuation is not the operative variable in a daily-reset product. The operative variable is holding period measured against realised volatility. At the volatility the index is currently printing, every month of flat, directionless trading costs a DFEN holder somewhere around two to four percent before any adverse move at all. That is the number that decides whether the instrument fits the view, and it is knowable in advance in a way that the next 11% index move is not.
The series to watch from here is GE’s forward multiple. The defense budget will do what it does regardless.