The Cato Journal’s new issue is all about immigration, and it includes an article from Gordon Hanson, who is one of the leading economists on this. Hanson repeats a claim that is common, and I think worth debating: that the fiscal externality of low-skilled immigrants should be internalized by their employers. Here is how he puts it:

Another source of unequal burden sharing is that U.S. employers enjoy benefits from immigration, in terms of higher productivity for their operations, while taxpayers pay for the education and health services that immigrant households receive. Taxpayers thus subsidize employers in agriculture, construction, meatpacking, restaurants and hotels, and other sectors that have high levels of employment of low-skilled immigrant labor. A reasonable solution to the current predicament is to eliminate such subsidies by making employers internalize the fiscal cost of immigrant workers. One way of achieving internalization is to subject employers to an immigrant labor payroll tax that would fund the benefits that their immigrant employees, and their family members, receive. Such a tax would make employers bear the fiscal consequences of immigration, releasing taxpayers from the burden and perhaps easing political opposition to immigration.

The notion here is that the employment of low-skilled immigrants is generating an externality, and that the employer who make up one side of these exchanges are not internalizing the cost of this externality. This leads them to hire them more low-skilled immigrants than they would if they had to pay the full cost. This treats low-skilled immigrants as being comparable to pollution generated by the employers, and proposes a Pigouvian tax to internalize the costs. But is this logic correct?

For one thing, I am surprised when I read liberals accepting that welfare and subsidies to low-income people are externalities to the Americans who pay for them. Is there some economic reason why these externalities should be internalized when foreigners generate them and not when natives do so?

Typically, transfers to the poor are justified on the economic grounds that people care about other people, and so improving the welfare of the poor is a public good to some extent. This isn’t just a liberal argument, here’s Greg Mankiw embracing it. In this case the fiscal burden of the low-skilled is not an externality, because everyone else receives positive utility from seeing them less poor.  Some might argue that by this logic what separates subsidies to immigrant from subsidies to natives is that Americans in general don’t care about the welfare of immigrants enough to justify the subsidies, so they are in fact externalities. This, to me, suggests that it is the policy that is generating the externality, and not the employers. If we wish to correct this is it not better to change the law so we stop giving subsidies to immigrants?

Another problem with thinking of employers as generating the externality is the fact that non-working immigrants surely generate a greater net fiscal burden than working immigrants. A tax on low-skilled immigrant working hours will result in less low-skilled immigrant working hours. In some cases this will result in immigrants returning home which will reduce the fiscal burden. But in other cases it will just result in immigrants working less hours, with perhaps less family members per household working, which will increase the net fiscal burden. Given this it seems to make more sense to tax immigrants themselves rather than those who hire them.

Gordon Hanson is just trying to find ways to make immigration more acceptable to Americans so that there can be more of it, so I don’t begrudge him for proposing an unpalatable solution. After all, most pragmatic solutions to convincing Americans to tolerate more immigration seem to have some unpalatable aspect to them. But I think accepting that the fiscal burden of low-skilled immigration represents a real externality generated by employers is both untrue and an unproductive framing of the problem.