Nassim Taleb has a concept he calls the barbell strategy. Put most of your money somewhere extremely safe, a small slice somewhere extremely aggressive, and nothing in the middle. The safe side means nothing can wipe you out. The aggressive side gives you exposure to big upside. The middle is where people think they’re being moderate, but they’re carrying risk they can’t see and not getting paid for it.
David Swensen is a good example of knowing which side you’re on. He ran Yale’s endowment from 1985 until his death in 2021 and took it from $1.3 billion to $42.3 billion, one of the most aggressive institutional strategies of his era. Then he wrote a book for regular people, Unconventional Success, and told them not to try any of it. Buy low-cost index funds and avoid high-fee products.
My version
Most of my money sits in boring, low-fee S&P index funds. No mutual funds, nothing proprietary, nothing clever. The rest goes into startups, where most will fail and a couple could return many times what I put in. If one or two hit, I’m more than covered. If they all go to zero, I’m fine. I don’t like anything in between.
Same with how I live. I run ultras: long runs, sometimes hours or days at a time. But I won’t ride a bicycle on a public road. The risk of getting killed is too high, and I’d like to be around for a while.
The principle under all of it: cap the downside you can’t come back from, then go hard everywhere else.
In construction: the job that can kill you
Bid aggressively. I mean really aggressively. But avoid taking on a job that can kill you.
Here’s the test. Take your biggest job and imagine losing 20% on it. If that loss would bankrupt you, don’t take it.
Run it on a $100 million heavy civil contractor. According to CFMA, heavy highway contractors average roughly 7 to 8% pretax. Call it $7.5 million a year in profit and somewhere around $25 million in equity.
A $100 million contractor can take on $10 million jobs all day. Put everything into a $200 million job, lose 20%, and you’re probably done.
If 20% sounds pessimistic, it isn’t. We had some dog jobs. We once lost $4 million on a $12 million job. That’s 33%. We broke even on a $40 million job we expected to clear 18% on. What saved us was that neither job was big relative to the company. Our largest job was about 30% of annual revenue. We never had a really big one go bad. Maybe we got lucky.
Size matters more than markup. I’ll bid a $5 million job at minus 2% without blinking. That’s a known $100,000. A $150 million job bid aggressively at 10% markup scares me a lot more, because the range of outcomes on a job that size is far wider than the markup. Swing from +10% to -15% and you’ve lost $22.5 million.
And the damage doesn’t stop at the loss. Sureties bond off your working capital, so a bad job also shrinks what you’re allowed to bid next year.
This is why most routine civil and paving work is forgiving. You can chew through backlog fast, and bad jobs end. On a multi-year monster, problems compound for years and often show up after the money is gone.
FMI has been saying this for a long time. Their research finds that contractors’ most profitable years are often the ones without a major loss, and that the most selective contractors are consistently the most profitable. They put it best: contractors “do not die of starvation but of gluttony.”
Going big anyway
None of this means never take a big job. We could have gone bigger than 30% under the right conditions. But if you’re taking on a monster, be very careful and prepare for the worst:
One thing to try Monday
List every job in your backlog. Next to each, write what a 20% loss would cost you. If any single number is bigger than a year’s profit, that’s the one to watch.
Have a good weekend.
Tristan