Finance, information and professional services now produce 16% more with 2% fewer jobs

Levanon's two rounded figures imply a large gain in output per worker. Dividing 1.16 by 0.98 gives about 1.18, so each remaining worker in these sectors accounts for roughly 18% more output than at the end of 2022 ^[5]. His explanation for the gain is a transfer of work between firms. He told Business Insider: "A growing share of the work businesses used to get done by consultants, contractors, call centers and their own back-office staff is now delivered by software, cloud and IT firms, which produce far more per worker and add almost no staff." ^[6] The account does not split jobs or output by subsector, so how much of the gain comes from software and IT firms themselves is not shown.

If Levanon is right, demand and hiring have come apart. He said there is demand in these fields, but firms can meet it with fewer workers ^[8]. Cory Stahle, an economist at the Indeed Hiring Lab, put the wider point this way: "Even a strong, continual growth in GDP may not translate into jobs for people, may not translate into a higher quality of life for people as well." ^[3] In our view, that makes sector growth a poor forecast of headcount for this group, at least over the window since 2022.

Levanon's software story is one of several explanations in the same account. Stahle said output per worker was already climbing before ChatGPT, and that "AI is part of the story here, but not the complete story. This is part of a longer-term trend." ^[9] A skeptic can cite Eric Kutcher, chair of McKinsey North America, who said: "You have people that are getting more productive as individuals, but they're just changing some of the work that they're doing." ^[12] His remark concerns work inside companies. Levanon's claim concerns work moving between them, so Kutcher's observation does not test it.

The decision and its cost sit on different clocks. A firm that holds junior hiring flat this quarter is acting on a window that opens at the end of 2022 ^[1]. The cost of a thin entry-level cohort shows up years later, when those roles would have produced people with experience. Whether the pattern lasts is not yet known, and we'd expect the gap to be hardest to see while output figures keep rising.