Portfolio and risk · Managing a portfolio
Leverage and volatility drag: why 2× risk is not 2× return
How borrowing multiplies returns and losses, why volatility lowers compound growth, and what a 2× S&P 500 fund did from 2006 to 2026 against the index.
On this page
Leverage means controlling more assets than one's own money would buy, by borrowing or by using instruments that have borrowing built in. It multiplies each period's gain or loss. What it does not do is multiply long-run growth by the same number. Twice the exposure gives twice the average return, twice the volatility, and something less than twice the compound growth, sometimes much less and sometimes less than none.
The gap is called volatility drag. It is not a fee or a flaw in any product. It is arithmetic: a loss needs a larger gain to recover from, and the bigger the swings, the more that asymmetry costs. This lesson explains the arithmetic and measures it on a real fund.
The idea
Start with an index that goes up 10% one day and down 9.09% the next. It is back where it started: 100 → 110 → 100.
Now hold it with 2× leverage, so each day's move is doubled: up 20%, then down 18.18%. 100 → 120 → 98.18. The index is flat and the leveraged position has lost 1.82%. With 3× leverage the loss is 5.45%.
Nothing went wrong. The second day's loss was taken on a larger amount than the first day's gain was earned on. Every round trip of this kind costs a leveraged holder something, and the cost rises with the square of the leverage.
The figure shows the same thing over twenty years. The 2× fund finished well ahead. On the way it lost more than four-fifths of its value, and an investor who needed the money in 2009 or 2010 had a very different experience from one who did not.
How it works
The first line connects the two kinds of average described in CAGR: compound annual growth rate, with the formula and its limits. The arithmetic average is what a typical single period returns; the compound rate is what an investor actually ends up with; the difference is about half the variance.
The second line shows why leverage disappoints. The return term grows in proportion to L. The drag term grows with L squared.
A worked example: SPY and a 2× fund
The data: month-end prices of SPY and SSO, August 2006 to October 2026 (243 months), price only. SSO aims at twice the S&P 500's return each day.
| Measure | SPY | SSO (2× daily) | Ratio |
|---|---|---|---|
| Average annual return (arithmetic) | 10.0% | 18.3% | 1.83 |
| Volatility | 15.3% | 31.1% | 2.04 |
| Average − ½ × volatility² | 8.8% | 13.5% | |
| Compound growth actually earned | 9.2% | 14.2% | 1.55 |
| $1 became | $5.91 | $14.68 | |
| Deepest fall | −52.2% | −82.5% | |
| Back to the earlier high | March 2013 | December 2013 |
The approximation is close: it gives 8.8% and 13.5% against actual growth of 9.2% and 14.2%. Twice the volatility delivered about one and a half times the growth. The average return did not quite double either, because the fund pays a financing cost and an expense ratio inside its price.
Individual years show the path dependence.
| Year | SPY | Twice SPY's year | SSO actually |
|---|---|---|---|
| 2008 | −38.3% | −76.6% | −68.3% |
| 2009 | +23.5% | +47.0% | +45.7% |
| 2013 | +29.7% | +59.4% | +70.0% |
| 2020 | +16.2% | +32.3% | +21.2% |
| 2022 | −19.5% | −39.0% | −39.2% |
| 2023 | +24.3% | +48.6% | +46.3% |
A fund that doubles each day does not double each year. In a steady year such as 2013 daily compounding helped and the fund returned more than twice the index. In 2020, when the index fell by a third and recovered inside the year, the fund returned 21.2% against the 32.3% that a simple doubling suggests. In a one-way decline such as 2008 it lost less than twice the index, because each day's loss applied to a smaller amount.
The daily target itself was met. Over November 19, 2024 to November 17, 2026 the fund's daily returns had a beta of 1.93 to SPY's with an R² of 0.996. Over those two years SPY rose 24.5% and the fund 34.8%, about 1.4 times.
How much leverage is too much?
Putting numbers into the second formula gives a curve that rises, peaks and falls.
| Leverage | Growth, 2006–2026 inputs | Growth, cautious inputs |
|---|---|---|
| 0× (all cash) | 1.6% | 3.0% |
| 1× | 10.7% | 6.4% |
| 1.5× | 14.3% | 6.9% |
| 2× | 17.4% | 6.5% |
| 3× | 21.8% | 3.4% |
| 4× | 23.8% | −2.9% |
The first column uses what SPY delivered with dividends over the same 243 months: an average return of 11.8%, volatility of 15.2% and a bill rate of 1.6%. On those figures the peak is at 4.4× leverage. The second column uses a borrowing rate of 3%, a premium of 5% over it and volatility of 18%. There the peak is at 1.5×, and by 3.1× the investor does worse than holding cash.
The level of leverage that maximizes growth (the excess return divided by the variance) is the Kelly criterion applied to investing. The two columns show why it cannot be relied on: the answer moved from 4.4 to 1.5 with a modest change in assumptions, and the 2006 to 2026 inputs describe an unusually good period with near-zero borrowing costs. Real borrowing costs more than the bill rate. And the curve says nothing about what the journey is like: at 4.4× a fall of 23% wipes out a position that is not cut on the way down.
Using it
- Recognize leverage in its different forms. A margin loan, a leveraged fund, options, futures and a mortgage on an investment property are all leverage. So is a concentrated position in a company that itself carries heavy debt. See Buying on margin: how margin loans, interest and leverage work.
- Size by the loss, not the gain. With leverage L, a market fall of 1 ÷ L wipes out the position if nothing is done on the way down: 50% at 2×, 33% at 3×. Both have happened to the S&P 500 within the period shown here.
- Expect less than the multiple. Over years, plan for a fraction of the stated leverage in compound return and the full multiple or more in drawdown.
- Volatility drag applies without borrowing too. Two portfolios with the same average return and different volatility compound at different rates. This is one reason diversification, which lowers volatility without lowering the average return much, can raise long-run growth.
- Know the product. Daily-reset funds are designed for short holding periods and say so in their prospectuses. Leveraged and inverse ETFs: daily reset math and volatility drag covers how they work.
Where it fails or misleads
- The formula is an approximation. It assumes returns without jumps. Real markets gap, and a leveraged position can lose more in a day than the formula allows for.
- Forced selling is missing from it. A margin loan comes with a maintenance requirement. A fall can trigger a margin call that forces sales at the low, turning a temporary loss into a permanent one. The arithmetic assumes the position survives.
- "Decay" is not automatic. Leveraged funds do not inevitably trend to zero. Over this rising 20-year period the 2× fund beat the index. The drag is the price of the volatility; in a strongly trending market the leveraged gain outweighs it.
- A period that includes a recovery flatters leverage. Had the sample ended in February 2009, the 2× fund would show a loss of more than 80% from its high.
- Borrowing costs change. The period shown includes more than a decade of interest rates near zero. Leverage financed at 5% faces a higher hurdle than leverage financed at 0.5%.
Compared with its neighbours
| Idea | What it describes | Effect on long-run growth |
|---|---|---|
| Leverage | More exposure than capital | Raises it up to a point, then lowers |
| Volatility drag | The gap between average and compound return | Always a subtraction |
| Sequence risk | The order of returns when money is added or withdrawn | Matters only with cash flows |
| Concentration | Dependence on one holding | Raises volatility, so raises drag |
| Diversification | Spreading across imperfectly correlated assets | Lowers volatility, so lowers drag |
How Finabase does it
Finabase does not calculate volatility drag, an optimal leverage level or the leverage of a portfolio.
The comparison view can show the effect directly. Overlaying SPY and SSO as percent changes over six months, one year and two years shows how far the leveraged fund's result is from twice the index's over each window. Each symbol's History tab lists daily prices, from which a single day's moves can be checked against the 2× target. A symbol's key statistics show the provider's beta, which for a leveraged fund should sit near its stated multiple. The paper-trading account allows a leveraged fund to be held with simulated money to see how its value behaves through a volatile stretch before any real money is committed.
Frequently asked questions
What is volatility drag in one sentence?
It is the amount by which compound growth falls short of the average return, roughly half the variance of returns, caused by losses requiring larger gains to recover.
If the market returns 10% a year, will a 2× fund return 20%?
No. It will return about twice the market's daily move each day. Over a year the result depends on the path, and over many years it has been well below twice the index's compound rate.
Do leveraged ETFs always lose value over time?
No. In sustained rising markets they have outperformed their indexes by a wide margin. They lose badly in volatile sideways markets and in declines, and their falls are much deeper.
Is a mortgage leverage?
Yes. A house bought with 20% down is held at 5× leverage: a 10% fall in its price is a 50% fall in the owner's equity. It differs from a margin loan in that the lender cannot demand repayment because the price has dropped.
Why does leverage have an "optimal" level?
Because return rises in proportion to leverage and the drag rises with its square. Eventually the second overtakes the first. The level at which that happens depends on future returns and volatility, which are unknown, so in practice it is a warning and not a target.
Summary
Leverage multiplies each period's return and does less for compound growth, because volatility subtracts about half its square from growth and leverage raises that drag fourfold when it doubles the exposure. A 2× S&P 500 fund from 2006 to 2026 had twice the index's volatility, 1.55 times its compound growth and a fall of 82.5%. More leverage raises growth only to a point that cannot be known in advance, and the arithmetic leaves out financing costs and forced sales. Judge a leveraged position by the loss it can survive.
Check your understanding
5 questions on this lesson. Each answer comes with the reason, and nothing is scored.
An index rises 10% one day and falls 9.09% the next, ending unchanged. What happens to a position with 2× daily leverage?
- It also ends unchanged
- It ends down about 1.8%
- It ends up about 1.8%
Show the answer
B. It ends down about 1.8% 100 → 120 → 120 × (1 − 0.1818) = 98.18. The doubled loss applies to a larger amount than the doubled gain was earned on.
An asset has an average annual return of 12% and volatility of 20%. Approximately what is its compound growth rate?
- 12%
- 10%
- 8%
Show the answer
B. 10% Compound growth ≈ average − ½ × volatility² = 12% − ½ × 0.20² = 12% − 2% = 10%.
When leverage is doubled, what happens to the volatility drag?
- It doubles
- It quadruples
- It stays the same
Show the answer
B. It quadruples The drag is ½ × L² × volatility². Doubling L multiplies it by four, while the return term only doubles.
In 2020 SPY gained 16.2% (price only) and the 2× fund gained 21.2%, not 32.3%. Why?
- The fund failed to deliver twice the daily return
- The index fell sharply and recovered within the year, and daily compounding through that swing cost the fund
- The fund paid a large dividend
Show the answer
B. The index fell sharply and recovered within the year, and daily compounding through that swing cost the fund The fund met its daily target. A deep fall followed by a recovery is the path on which a daily-reset fund falls furthest behind a simple doubling of the period return.
What does the growth formula leave out that matters most to someone borrowing on margin?
- The possibility of being forced to sell after a fall
- The average return
- The volatility
Show the answer
A. The possibility of being forced to sell after a fall The formula assumes the position is held throughout. A margin call can force a sale at the low, making the loss permanent.
Sources and further reading
- Minder Cheng and Ananth Madhavan, “The Dynamics of Leveraged and Inverse Exchange-Traded Funds,” Journal of Investment Management, vol. 7, no. 4 (2009)
- Edward O. Thorp, “The Kelly Criterion in Blackjack, Sports Betting, and the Stock Market,” in Handbook of Asset and Liability Management, vol. 1 (North-Holland, 2006)
- U.S. Securities and Exchange Commission, Investor.gov — Updated Investor Bulletin: Leveraged and Inverse ETFs
- ProShares — Ultra S&P500 (SSO), fund page: 2× the daily S&P 500, inception June 19, 2006
- FINRA, Regulatory Notice 09-31: Non-Traditional ETFs
- Finabase chart data — monthly bars of SPY and SSO, August 2006 to October 2026, and daily bars, November 19, 2024 to November 17, 2026; statistics computed by Finabase
