What They Carried

You take the forecast and build your plan around it. The number has a source, a date, perhaps a careful explanation. You still cannot tell what being wrong would cost the person who supplied it.

In November 2007, the Philadelphia Fed surveyed 48 professional forecasters. Their median forecast put real GDP growth for 2008 at 2.5%.

The New York Fed's staff expected 2.6%. In its own 2011 account, the bank put actual growth at minus 3.3%. Its forecast had missed by 5.9 percentage points.

The forecast had been clear enough to plan against. It was not clear enough to tell you who would answer for the gap.

But Michael Burry had done more than publish an expectation. In 2005, he began buying credit default swaps on subprime mortgage bonds. He was taking a position against the mortgage market.

In letters to clients in 2005 and 2006, he argued that the mortgage market would melt down in the second half of 2007. That is his account in a 2010 New York Times op-ed, carried by Reuters.

Some investors demanded their money back. Burry restricted withdrawals and held the trade. His judgment now sat between those people and money they wanted to take home.

John Paulson also had to persuade people to commit money. Gregory Zuckerman told Newsweek that most investors Paulson approached declined. He said the dedicated fund raised about $147 million.

Paulson & Co. made roughly $15 billion in 2007. Yet some investors were nervous even while the trade was paying. Success had not removed the need to answer them.

The familiar story makes these men lone believers who staked their fortunes. Both were managing mostly other people's money. Burry's inheritance and family loans had helped start his fund, not financed the whole mortgage trade.

What they carried was an answer owed to people who would share the loss.

Answering to those people is closer to Nassim Nicholas Taleb's test in Skin in the Game. He describes shared harm: “paying a penalty if something goes wrong.” A share of the gains alone does not pass that test.

Managing other people's money does not make the downside equal. I could not find what either man would have lost personally. But a position tied to a consequence gives me information that a forecast alone cannot.

Warren Buffett made the limit of his knowledge plain in his 2008 New York Times article: “I can't predict the short-term movements of the stock market.”

In his 1992 shareholder letter, he wrote that stock forecasters had only one use: “to make fortune tellers look good.” Yet his own finances remained tied to his judgment.

In June 2025, Buffett reported that well over 99% of his net worth was in Berkshire Hathaway. His wealth was exposed to the same shares whose performance he reported.

Berkshire's 2024 letter reported compounded annual gains of 19.9% from 1965 to 2024. The S&P 500, with dividends, returned 10.4% a year over that span. Those are returns on shares, not a score for predicting next month's market.

Buffett could admit uncertainty without withdrawing his commitment. His position told you what he was prepared to carry while he waited. The admission and the commitment belonged together.

But Burry's doubting investors might have been right. A costly mistake remains a mistake, and a free forecast can be sound. Having something to lose does not settle the question.

CXO Advisory graded 6,582 forecasts from 68 market gurus, collected from 2005 to 2012. They were right 46.9% of the time, close to a coin flip. That does not prove that people with money at risk predict better.

I have my own reason to insist on a record. For six months, I wrote my calls down and checked them. My 90%-confident calls were right 58% of the time.

The record told me what my certainty couldn't. What being wrong would cost someone shows what they believed enough to act on, and who must answer if that action fails.

A business's promises can be checked against its budget, calendar and promotion list. I want the same check on a forecast: the stated belief beside the consequence its maker accepts.

When you plan against a number, you accept consequences whether its maker does or not. You cannot pass those consequences back simply because the forecast came from someone else.

The 2.5% growth forecast told you what the panel expected. It did not tell you how much of your downside anyone on it shared.

What happens to this person if they're wrong?

A claim belongs beside what happened to the person who made it. My free letter arrives on Sunday mornings, most weeks.