A common claim is that extreme heat will cost the world $16 trillion a year in lost work by 2050. The fact is that figure is not an annual forecast. Callahan and Mankin, in Science Advances in 2022, put cumulative losses from human-caused extreme heat between 1992 and 2013 in a likely range of $16 trillion to $50 trillion. The central estimate was more than $33 trillion over those 22 years. The total covers health, productivity, and farm output. It is not a 2050 payroll.

The ILO’s labour count is smaller, and it is dated. Under a path to 1.5°C by the end of the century, and assuming work in the shade, the ILO projects that 2.2% of working hours will be lost in 2030. That is about 80 million full-time jobs, and $2.4 trillion. In the sun, the hour loss rises to 3.8%. Agriculture is projected to account for 60% of lost hours. Construction accounts for 19%. Southern Asia and western Africa each lose about 5% of working hours, about 43 million and 9 million jobs.

Lancet Countdown counted the loss already on the books. In 2023, heat exposure cost 512 billion potential work hours and $835 billion of potential income. About two-thirds of the hours were in agriculture. Low-HDI countries lost the equivalent of 7.6% of GDP. Very high-HDI countries lost about 1%. A 2021 Nature Communications paper on heavy labour put annual losses at about $1.6 trillion in a 2°C world. That is a labour model, not a 2050 bill, and it is not $16 trillion.

The ILO uses temperature, humidity, and the pace of work. A hotter hour cuts the safe work rate. The dollar figure prices those lost hours. It is not a count of wages that failed to clear a bank. The shade assumption makes the 2.2% figure a lower case. The sun case is 3.8%.

Lancet uses the same idea of potential hours, then prices them. A country can lose a large share of farm hours and a small share of office hours. Callahan and Mankin link the hottest five days of each year to regional GDP per person, then attribute the human-caused share. Their $16 trillion is the low end of a likely cumulative range, not a yearly run rate.

It is settled that heat already cuts outdoor work, and that the ILO’s 2030 figure is $2.4 trillion under a 1.5°C path. It is not settled that the world loses $16 trillion a year by 2050. Use $2.4 trillion for 2030. Use $835 billion for 2023. Use $16–50 trillion only as a 1992–2013 cumulative range.