The Digester

A replication of George Borjas's paper finds major data and geographic errors that inflate the H-1B wage gap; after corrections, the gap falls to about 5 percent and reverses for younger workers, calling into question a $100,000 H-1B fee policy based on the original results.

Borjas reported a 15 to 16 percent H-1B wage gap, but replication found major errors that biased the result downward.
The largest errors were a time mismatch between H-1B and native samples and use of residential PUMAs instead of worksite geographies, plus inconsistent sample filters.
Restricting to 2023 H-1B observations and using place-of-work controls cuts the estimated wage gap to about 5 percent, roughly one third of the original estimate.
Wage differences vary strongly by age: H-1B holders aged about 22 to 33 earn more than comparable natives, while older H-1B workers earn less.
A $100,000 fee proposal based on the flawed estimate would likely mis-target workers and employers and penalize young, high-value H-1B hires.

Correcting the record on the H-1B wage gap: it is smaller and age dependent

Sources