This model assumes that the number of positions is fixed at any given time, but that is not true. Instead, it will fluctuate based on what must be paid. So, to extend your example, without H1B or a similiar program or de-facto state that does the same thing, then we'd get roughly the 900 jobs that are willing to pay the most to be filled, while the rest are not worth the cost of employing someone in this hypothetical market. The result, then, is a higher wage for those 900 citizens.

This is not to say that such programs are bad. To consider that, we must ask questions like:

- Is citizens being able to make much more than non-citizens a desirable outcome? (it's electorally incentivized, since citizens can vote in the relevant elections but non-citizens can not but that is not the same as being desirable in a more broad sense) - How big is this effect? - How badly do we (meaning society at large) want these final 100 jobs done? - How will this affect competition with other countries? - What about outsourcing?

A labor shortage has a temporary effect of creating an upward pressure on the 900 jobs, but there’s a ceiling and the long term effects are complex. I may have to pay 900 jobs 1.5x to attract and retain the talent as a business, but that cost will be eaten by other functions where there’s a surplus. Or long term, the businesses will find ways to either automate or ship the jobs away.

Additionally there will be a loss in productivity for business and the 900 jobs will have more work on their plates, so in a way they could be underpaid for the job they do. This is just scratching the surface, but I assume there’d be a lot of second order effects.