Barbell Strategy

Thinking Tools4.4 Refining Ideas Into Practical SolutionsAlso known as Taleb's Barbell

Developed by Nassim Nicholas Taleb (Antifragile, 2012), who took the name from fixed-income investing.

Puts most resources in very safe positions and a small share in high-upside bets with capped losses, while avoiding the middle that carries risk without protection or upside.

Splitting a $10 million reserve: safe, middle, and long shots across a bad, a normal, and a great year

Safe holdings earn 3 percent in any year. Middle-risk holdings earn 6 percent in a normal year, 12 in a great one, and lose 20 in a bad one. Long shots lose everything in a bad year and half in a normal one, but return four times their stake in a great one. The bars are your split; the black line is the whole reserve in the middle. Set the share in long shots and the share of the rest in the middle: with nothing in the middle, the worst loss can never exceed the long-shot stake. Illustrative.

10 percent
0 percent
Reserve after a bad year, $M
$9.27M
After a normal year
$9.77M
After a great year
$13.27M
All in the middle: most it can lose in a year
$2.00M
Most the reserve can lose in a year
$0.73M

Your split: reserve at year end, $MAll in the middle

0.05.010.015.020.09.39.813.3Bad yearNormal yearGreat year0.05.010.015.020.09.39.813.3Bad yearNormal yearGreat year

An example

A family-owned food manufacturer held a $10 million reserve, and its treasurer had spread it across medium-risk corporate bonds and a balanced fund. The board's new chair asked what the reserve was for. The answer was surviving a bad year without layoffs and having money to move when a chance came.

The treasurer modeled three kinds of year. In the middle, the reserve earned about 6 percent in a normal year and 12 in a great one, but a credit crunch could take 20 percent, $2 million, in exactly the year the company would need it. She proposed a different split: 90 percent in short-term government bills earning 3 percent in any year, and 10 percent in small stakes in packaging and ingredient startups, each of which could go to zero.

In a bad year the split could lose at most about $730,000. In a normal year it trailed the bond portfolio, but in a great year it could reach $13 million or more. The board adopted it, with a rule that no single startup stake would exceed $250,000.

Why it works

Middle-risk positions look prudent, but their risk is hard to see and hard to cap: they earn modest returns in good times and lose heavily in exactly the conditions when everything else is going wrong. A barbell separates the two jobs a reserve has. The safe end is truly safe, so the worst case is known and survivable. The risky end is small and made of bets whose loss is capped at the stake and whose upside is open.

The result is a portfolio whose worst year is bounded by the size of the risky share, however bad conditions get, while it keeps exposure to the rare large gain. It gives up some return in ordinary years to buy that shape. The trade makes sense when surviving the bad year matters more than winning the average one.

The thinking behind it

Nassim Nicholas Taleb Nassim Nicholas Taleb, a former options trader who writes on risk and probability, set out the barbell in Antifragile: Things That Gain from Disorder (2012), taking the name from fixed-income investing, where a barbell portfolio holds very short and very long bonds and little in between. His broader argument, begun in The Black Swan (2007), is that rare, high-impact events dominate outcomes, and that a barbell of very safe and very risky positions beats the middle. His books are essays for a general reader, with technical appendices; some statisticians have faulted his treatment of statistics.

How to run it

  1. Name what the money is for. State what the reserve or budget must survive, and what chances it should be able to seize.
  2. Size the worst case you can bear. Decide the most it may lose in a year; that sets the risky share's ceiling.
  3. Make the safe end truly safe. Put most of it where losses are close to impossible, even at a low return.
  4. Make the risky end small and capped. Use bets whose loss cannot exceed the stake, such as small equity stakes, pilots, or options.
  5. Spread the risky end. Make many small bets, so one large winner can pay for the losers.
  6. Empty the middle. Move money out of holdings that carry hidden losses in bad years without a large upside.
  7. Rebalance on a schedule. Each year, restore the split, taking winnings off the risky end and refilling it only to its share.

Where it goes wrong

The safe end is not safe. Money parked in something called low risk that falls in a crisis turns the barbell back into the middle; spot it when the safe holdings dropped in the last downturn. Check the safe end against the worst year on record.

The risky bets have uncapped losses. A position that can lose more than its stake, through guarantees or follow-on commitments, breaks the bound. Read each bet's terms for any obligation beyond the money staked.

The risky end grows unchecked. After a good run the long shots become a third of the reserve, and the worst case quietly grows more than fourfold. Rebalance to the set share every year, whatever the recent returns.

When to use something else

  • Convexity Reasoning. When the question is the shape of one commitment's payoff, before any portfolio is built.
  • Tail-Risk Hedging. When the existing portfolio stays as it is and the job is buying protection against a crash.
  • Real Options Analysis. When the risky bets are projects whose next stage can be bought or abandoned as results come in.

Where it sits in SOLVE eX

Go to the source

Back to the catalog