The Quiet Reason the Greatest Investors Almost Never Use Leverage

The greatest investors avoid leverage because it turns a temporary price decline into a permanent loss. Borrowing adds a small and uncertain gain to outcomes that were already good, and it attaches a real chance of total ruin to outcomes that would otherwise have been survivable. Wealth compounds multiplicatively, so a single wipeout erases decades of gains and no later recovery can repair it. Berkshire Hathaway is the usual counterexample, yet its insurance float has no margin call, often costs less than nothing, and never has to be refinanced at a bad moment. A retail margin loan fails all three of those tests.

Why Does Leverage Destroy Good Investors?

Leverage destroys good investors because losses hurt more than equal gains help, and because a forced sale locks in the loss before a recovery can arrive. A portfolio that falls 50% needs a 100% gain to break even, and a leveraged portfolio that falls far enough has no equity left to recover with.

The deeper reason comes from an idea called ergodicity. Imagine a game that pays 50% on heads and costs 40% on tails. Averaged across a large crowd of players in a single round, each flip is worth a gain of 5%, and the game looks attractive. A single player who flips a hundred times in a row multiplies by 1.5 and by 0.6 in turn, and the typical outcome is a fortune that shrinks to roughly half of one percent of its starting value. The crowd average describes a population at one moment, while the time average describes one person living through one sequence. Physicist Ole Peters has spent years arguing that finance leans too heavily on ensemble averages for exactly this reason.

An investor lives only one path through the market, so what matters is whether that single path survives, whatever the average of all possible paths looks like.

Warren Buffett made the same point in plain language in his 2010 shareholder letter. He wrote that leverage magnifies gains when it works, that it becomes addictive once it has worked, and that any series of positive numbers, however impressive, evaporates when multiplied by a single zero. History, he added, shows leverage often produces zeroes, even when very smart people use it. He also described credit as being like oxygen, unnoticed when abundant and the only thing anyone notices when it vanishes.

What Does the Math Say About How Much Leverage Is Too Much?

The math says the growth-maximising amount of leverage is surprisingly low, highly sensitive to inputs nobody knows, and far below what most leveraged products offer. A common approximation puts the long-run growth rate of a leveraged portfolio at leverage times the expected excess return, minus half of leverage squared times the variance of returns. The second term grows with the square of leverage, so volatility drag eventually overwhelms the gain.

Consider an illustration built on round assumptions rather than historical estimates. Suppose stocks earn an average of 5% a year above the borrowing rate, with annual volatility of 16%. Unlevered, the approximate growth rate is 3.7%. At 2x leverage it rises to about 4.9%, and at 3x it falls to about 3.5%, below the unlevered figure. The peak sits near 2x, which means a long-run investor would endure drawdowns of ruinous size to collect an extra gain of barely one percentage point. Lower the assumed excess return to 3%, a plausible estimate when valuations are high, and 2x leverage grows more slowly than no leverage at all.

That sensitivity is a significant limitation of any leverage case. Nobody knows the forward equity premium to within a percentage point or two, and overestimating it while using leverage costs far more than underestimating it. Real markets add fat tails, financing spreads above the risk-free rate, and lenders who change terms during a panic, all of which push the true optimum below the tidy formula. Practitioners who rely on these formulas commonly run only a fraction of the theoretical maximum.

How Bad Can a Drawdown Get for a Leveraged Portfolio?

A drawdown can get bad enough to end the game well before the market recovers. The verified record of major S&P 500 declines shows 48% in 1973-74, 34% in the Black Monday crash of 1987, 49% in the dot-com bust, 57% in the global financial crisis, 34% in the COVID crash, and 25% in the 2022 bear market. The 1987 decline bottomed within 14 weeks and the COVID decline within 5, which left little time to react.

Margin rules turn those figures into forced selling at the worst moment. As of 2026, standard US rules let an investor borrow up to half the purchase price, and brokers typically require equity of at least 25% of the account’s value afterward, though many apply stricter house limits. Take $100,000 of savings combined with $100,000 of borrowing, so the account holds $200,000 of stock. A decline of about one third, to roughly $133,000, leaves equity at 25% of the account and triggers a call, a depth both 1987 and 2020 reached. A 50% decline erases the equity entirely. The 1973-74 and dot-com declines came within a point or two of that line and the 2008-09 decline went well past it. The COVID decline recovered in 21 weeks and the 1987 decline in 98 weeks, but a margin account does not get to wait for them.

The present market makes the point concrete. As of mid-2026 the S&P 500 trades near 7,792, roughly 37% above its 200-week moving average of about 5,706 and close to its 52-week high of 7,845. A retreat to that long-term average would mean a decline of about 27%, similar in size to the 2022 bear market. An investor at 2x leverage would lose roughly 54% of their own capital on that move alone, before interest, and price has briefly undershot the average before, as in 2009. Readers who want background on the indicator can start with the 200-week SMA explainer, and the S&P 500 200-week SMA history shows how often and how deeply the index has visited it.

How Did Buffett Use Leverage Without Breaking His Own Rule?

Buffett used a form of leverage that lacks the features that make margin debt dangerous. Berkshire’s main source of borrowed-style capital is insurance float, which is money collected as premiums and held until claims are paid. Buffett has described float as a revolving fund, and it grew from $16 million in 1967 to $62 billion by the end of 2009. Claims arrive on the insurer’s timeline, so no lender can demand the money back early.

The cost side matters just as much. Buffett has noted that float is wonderful only if it does not come at a high price, and its price is set by underwriting results. His 2005 letter recorded an underwriting profit in about half of the 39 years Berkshire had been in the business, and his 2018 letter recorded profits in 15 of the preceding 16 years. In those years Berkshire was paid to hold other people’s money. He also warned that a major catastrophe or a long-hidden error, with asbestos as his example, could turn float into an expensive liability. As with any Berkshire example, results depended on valuation and discipline at the time of each investment, and the same float deployed at peak multiples would have produced a much weaker record.

The rest of the balance sheet is built so that float never forces a sale. Berkshire has pledged to hold at least $10 billion of cash and customarily keeps $20 billion or more, and its 2023 letter states that during the 2008 panic the company relied in no way on commercial paper, bank lines, or debt markets. Debt at subsidiaries such as MidAmerican was described in the 2005 letter as that company’s obligation alone, not guaranteed by Berkshire and serviced by diversified utility earnings. Buffett likened the extreme conservatism to an insurance policy on a fortress-like building. A deeper look at the temperament behind this approach appears in Buffett’s Real Edge Wasn’t Stock Picking, It Was Patience.

What About AQR and Systematic Leverage?

The systematic exception is real, and it rests on a specific economic argument. In a paper titled “Buffett’s Alpha,” Andrea Frazzini, David Kabiller, and Lasse Pedersen argued that a large part of Berkshire’s record can be explained by owning cheap, safe, high-quality stocks and scaling them up with leverage. The logic holds that many investors, such as pension funds and mutual funds, cannot or will not borrow. They chase higher-risk stocks instead, which leaves lower-risk stocks relatively cheap, and a patient borrower can harvest that gap.

Firms that run this idea at scale, AQR among them, differ sharply from a retail account with a margin loan. Institutions of this kind typically finance through futures and term arrangements rather than overnight callable loans, spread exposure across many return streams, reduce positions when volatility rises, and employ full-time risk teams with rules written before a crisis starts. Even so, these strategies fail in specific environments. Risk parity approaches suffered in 2022 when stocks and bonds fell together, and the collapse of Long-Term Capital Management in 1998 showed that a firm staffed by renowned academics does not repeal the arithmetic. Leverage in professional hands is a managed hazard with institutional shock absorbers, and it still occasionally kills.

Why Does Retail Almost Never Qualify?

Retail investors rarely qualify because a margin loan reverses the advantages that make Berkshire’s float and institutional leverage work. The loan can be called when prices are lowest, the borrowing rate carries a spread over the risk-free rate that eats into expected returns, and the collateral is the same asset whose price is falling. A household also has a job, an income, and emotions that tend to deteriorate at the same moment the market does.

Research on retirement leverage makes the sequencing problem vivid. The Safe Withdrawal Series by the researcher known as Big ERN examined borrowing against a portfolio instead of selling assets in retirement. For a retiree starting in 1965, the leveraged strategy would have run out of road after about 15 years and forced liquidation at the bottom of the 1982 recession, just before the great rally of the 1980s. The analysis found that leverage worsened sequence risk and concluded the approach suits households so wealthy that a draw rate near 1% covers their lifestyle. Everyone else cannot wait out the trough because the lender will not allow it.

Leveraged ETFs remove the margin call and add a different problem. They reset leverage daily, so in choppy markets they lose ground even when the index ends roughly flat, and their fees add to the drag. Commentators focused on long-term index investing have long called them speculative products with no place in a buy-and-hold portfolio, and the volatility arithmetic above explains why. The one common form of household leverage that avoids a collateral call is a fixed-rate mortgage, whose risk comes from cash flow and job security rather than from a price feed, and it deserves a separate decision from borrowing to buy stocks.

Leverage rewards the investor who can wait out any drawdown, and borrowing is a decision that removes the ability to wait.

The alternative is less exciting and much sturdier. A higher savings rate, a longer holding period, steady purchases through market cycles, and a willingness to add more when valuations are low do the work leverage promises, without a path to zero. The Buy the 200 strategy follows that logic by using the 200-week average to lean into weakness with unborrowed money, which keeps the investor in control of the timeline. The calculations here draw on the verified drawdown record, Berkshire’s own shareholder letters, and standard margin rules, applied to a simple illustrative model rather than any promise of returns.

Frequently Asked Questions

Q: Does Warren Buffett use leverage?

A: Berkshire uses insurance float, which works like leverage but cannot be called by a lender and often costs nothing. Buffett has warned that conventional leverage can be lethal even for smart people, and Berkshire holds large cash reserves so it never has to borrow in a crisis.

Q: What happens to a margin account when the market falls 33%?

A: An account bought with 50% borrowed money typically faces a margin call after a decline of about one third, because equity drops to the 25% maintenance level. Both the 1987 crash and the COVID crash fell about 34%, and brokers with stricter house rules would have issued calls sooner.

Q: Is 2x leverage on the S&P 500 ever a good idea for long-term investors?

A: Under optimistic assumptions the growth-maximising leverage is close to 2x, but the gain over an unlevered portfolio is small and the drawdowns are severe. If the true equity premium is lower than assumed, 2x leverage can grow more slowly than no leverage at all.

Q: What is ergodicity in investing?

A: Ergodicity describes whether the average outcome across many people matches the outcome one person experiences over time. Investing departs from that match whenever ruin is possible, so one investor cannot rely on the crowd average and must protect against a path that ends in zero.