Leverage — investing with borrowed money, or through leveraged ETFs — looks simple on paper. If the market returns 9% a year, a 2× position should return 18%. In practice, it almost never works out that way. Two forces get in the way: volatility drag and the cost of borrowing. And leverage makes drawdowns so deep that many investors are forced out before the long-run maths can work.

Volatility drag in one example

Suppose an index rises 10% one day and falls 9.09% the next. It ends exactly where it started:

100 × 1.10 × (1 − 0.0909) = 100

Now look at a fund that resets to 2× the index every day. It rises 20%, then falls 18.18%:

100 × 1.20 × (1 − 0.1818) = 98.18

The index is flat, but the 2× fund has lost 1.8%. A 3× fund would rise 30% and fall 27.27%, ending at 94.55 — down 5.5%. Repeat that back-and-forth and leveraged funds steadily lose value even when the market goes nowhere.

The reason is the same asymmetry that drives drawdown recovery: a percentage loss has to be followed by a larger percentage gain to recover, and leverage makes the losses larger.

The rule of thumb

For an investment with average (arithmetic) annual return a and volatility σ, the long-run compound growth rate is roughly:

compound growth ≈ a − σ² / 2

The σ²/2 term is the volatility drag. Leverage of L multiplies the average return by L but multiplies the drag by L² — and you also pay interest on the borrowed part. Using our calculator's "Aggressive" assumptions (9% average return, 17% volatility) and a 5% borrowing cost:

Leverage Average return Volatility Approx. compound growth after borrowing costs
1× 9% 17% about 7.6%
2× 18% 34% about 7.2%
3× 27% 51% about 4.0%

Under these assumptions, doubling the risk produces no extra long-run growth, and tripling it produces less. The exact numbers depend on the assumptions, but the shape — rapidly diminishing, then negative, returns to extra leverage — is general.

What history says

To test this against real data, we simulated portfolios holding 1×, 2× and 3× the S&P 500 (with dividends), rebalanced to the target leverage every month, paying interest on the borrowed amount equal to the 10-year Treasury yield. The results:

Period 1× 2× 3×
Aug 1929 – Jun 1932: total change −81% −98% −99.9%
Oct 2007 – Mar 2009: total change −49% −78% −92%
Jan 2000 – Jun 2023: annual return 6.8% 8.4% 7.2%
Jan 2000 – Jun 2023: worst drawdown −49.0% −80.7% −95.4%
Jan 1950 – Jun 2023: annual return 11.3% 15.4% 17.6%
Jan 1950 – Jun 2023: worst drawdown −49.0% −80.7% −95.4%

There is an honest case for leverage in this table: over the unusually strong 1950–2023 period, the simulated 2× and 3× portfolios compounded faster than the index. But look at what they had to survive to get there. A 2× investor saw their portfolio fall more than 80% at least once; a 3× investor saw it fall more than 95%. From 2000 onwards, 3× leverage earned only slightly more than holding the index, despite carrying several times the risk. And anyone who was 2× or 3× leveraged going into 1929 was all but wiped out.

Why the real-world results would probably be worse

This simulation is generous to leverage:

  • Monthly data hides daily swings. It uses monthly average prices, which understate volatility. Real daily-reset leveraged ETFs experience the full daily volatility, and with it more drag.
  • No margin calls. A real leveraged account that falls 80% would usually have been closed out by the broker long before the bottom, locking in the loss.
  • No fees. Leveraged ETFs charge higher fees than index funds, and borrowing costs for individuals are typically above the Treasury yield.
  • Perfect discipline. The simulation assumes the investor never sells in panic during an 80–95% loss. Very few people could.

Practical takeaways

  1. The biggest risk is not low returns — it is being forced out. Leverage can work over a long period only if you can survive the drawdowns without selling or being margin-called.
  2. Volatility matters more as leverage rises. On a volatile asset, even modest leverage can reduce long-run growth.
  3. Leveraged ETFs are designed for short-term trading. Their daily reset makes long-term results depend heavily on the path of the market, not just where it ends up.
  4. Test it before you do it. Move the leverage slider in our calculator and watch the 10th-percentile outcome. It falls much faster than the expected path rises. Then check how big a gain you would need to recover from a leveraged drawdown with the drawdown recovery calculator.

Notes

Simulation data: Robert J. Shiller, U.S. Stock Markets 1871–Present, via datasets/s-and-p-500. Monthly total returns, rebalanced monthly to constant leverage; borrowing cost is the 10-year Treasury yield, used as a rough proxy. Approximation table uses a·L − (σL)²/2 − (L−1)·5%. Calculations are our own.

This guide is educational and not investment advice. Leveraged products can lose most or all of their value.