Auctions, tariffs, and taxes

A draft that Alex Nowrasteh sent me to read provoked me to think in a new way about various market alternatives for regulating immigration. All of the following policies use the price mechanism, in one way or another, to ration visas by willingness-to-pay, while capturing some of the surpluses generated by immigration to return them to natives.

1. Auctions. Under this mechanism, a certain number of visas would be sold to the highest bidder. There’s a lot to be said about auction design, but for concreteness, you could just let everyone submit a bid, and then accept the top x bids, requiring them each to pay, not what they bid (which would make them bid strategically) but the lowest acceptable bid (which would make them reveal their true values). Auction winners would pay their bid, and be issued visas.

2. Tariffs. Under this mechanism, the government would set a price for a visa, and whoever was willing to pay it would receive one. There might, of course, be restrictions: knowledge of English, perhaps, or criminal background check, or some regions of the world (Pakistan?) might be excluded or treated differently on national security grounds.

3. Taxes. This is the approach I’ve long advocated: the DRITI scheme would be one way to design it. In this case, to come would be nearly free– in the DRITI scheme, I have people preimburse the government for a deportation option which they could then exercise if they were destitute and wanted to go home– but those who came to work would pay a surtax (DRITI has some other features but never mind those for now).

Now, at a highly theoretical level, the effects of all these policies could be much the same. It is not the case that, say, auctions are necessarily more restrictive than tariffs, and tariffs than taxes. If you auctioned off enough visas, that could be quite a loose immigration policy. If you set a tariff, or a migrant surtax, very high, the policy could be rather restrictive. If the government knows its own preferences and the demand curve for immigration, it doesn’t matter whether it controls the quantity (via auctions) or the price (via tariffs/taxes): the number who will come, and what they will pay, will maximize the government’s objective function. Similarly, if immigrants know what they’ll earn and are not credit-constrained, it doesn’t matter whether they pay everything up front (via auctions/tariffs) or over time (via taxes): either way, those will come whose net present value of migrating is greater than the net present value of the payments required.

There are a lot of practical differences between the three policies but I think the crucial one is how the burden of uncertainty is distributed, if, as is realistic, the government does not know exactly what the demand for immigration is like, nor do immigrants know exactly what they’ll earn in the US.

Under the auction mechanism, the government faces no uncertainty about what may the most important variable, namely, how many immigrants will come, though it faces uncertainty about how much it will raise from the auction. Immigrants, on the other hand, face a lot of uncertainty, because they don’t know if their bids will prove sufficiently high to qualify for visas. Immigrants might find it difficult to plan, not knowing whether they would be able to come or not. Also, they wouldn’t know how much they would end up paying. If an immigrant bids $500,000, he might be fairly confident of coming, but not know whether he’d end up paying $20,000 or $100,000.

Under the tariff mechanism, the government faces uncertainty about how many immigrants will come, and also about how much revenue it will receive in total, but not about what price each immigrant will pay. It knows how much revenue it will get per immigrant. Meanwhile, immigrants know what they’ll have to pay for a visa, and face no uncertainty on that front. But they face a lot of uncertainty about what they’ll earn in the US, and how much they’ll like it there. The value of immigrating is probably quite uncertain, but they have to pay the same price in any case. If they pay a lot, then are unlucky in the labor market, or have overestimated their value, or become unbearably homesick, they’re in trouble. Of course, they might also get large, unexpected surpluses if their fortunes in the US are better than they had anticipated. But it’s a risky undertaking.

Under the tax mechanism, the government faces uncertainty, both about how many immigrants will come, and about how much the immigrants will earn and therefore how much they’ll pay in taxes. The government, moreover, has to wait to get paid. Immigrants, on the other hand, face much less risk. They don’t know how well they’ll succeed in the US, but if they do badly, they can go home, or if they stay even though they’re earning little, they’ll pay little (though even a small amount might be painful) for the privilege of staying. If they do well, they’ll pay a lot, but they can afford it then. If they earn well enough but dislike life in the US, they’re not bound to stay just so they can pay off the debt they incurred for the visa: the visa was near-free. Immigrants also don’t need to have a lot of liquid assets to come.

I much prefer the tax mechanism, because the government can handle the risk easily. The government is highly “diversified” in that many immigrants will come, and more taxes from the successful will offset fewer taxes from the less successful. And the government can easily afford to wait and garnish wages rather than getting a lump sum up front. I think only the tax mechanism would (virtually) eliminate undocumented immigration, because anyone who could afford to pay a coyote could afford to get a visa. I also think the government could raise the most money through immigration taxes, because immigrants who were relieved of risk and the need to go into debt would be willing to accept arrangements that, on average, would ultimately cost them a good deal more than what they would have agreed to pay as a lump sum ex ante.

But it occurs to me (or perhaps Vipul suggested it, I can’t remember) that a policy path to open borders might involve (a) auctions first, (b) then tariffs, and (c) taxes last.

If the public appreciated the efficiency advantages of regulating immigration by working through markets, but was still spooked by the possibility of getting swamped, they might go for the auction approach. That way they’d still control how many immigrants came in.

If the auction mechanism had been in place for a while, people might start to feel they knew something about the shape of the immigration demand curve, which would make them more comfortable with the idea of tariffs, since the number of immigrants who would use them, though unknown, would be less unknown. People could make an educated guess. Moreover, the auction mechanism would have some problems which a tariff could resolve. For example, what’s the auction schedule? If it were annual, people would have to wait a long time for it, and to lose a bid could mean the disruption of a lot of big plans. If it were monthly, people would speculate on a favorable month. Prices might fluctuate in strange ways. Some might be willing to pay a lot more than the usual auction prices during the off-season, while others would overbid and regret it. A tariff would promise to be less arbitrary and raise more revenue.

Later, when the tariff mechanism had been in place for a while, people would have a clearer idea what the income profile of immigrants tended to be, which would help them project the earnings of a migration tax. There would be lots of stories about spirited, entrepreneurial aspirants to immigration who, however, couldn’t raise the money to come, though they seemed sure to succeed in the US if they could just get over that hurdle. Others would pay the tariff, get homesick, but be stuck, needing high earnings to pay off creditors. A switch to taxes would promise to relieve immigrants of risk and avoid excluding promising people merely because they had poor access to credit, while at the same time, increasing the revenue the government could take in.

Nathan Smith is an assistant professor of economics at Fresno Pacific University. He did his Ph.D. in economics from George Mason University and has also worked for the World Bank. Smith proposed Don’t Restrict Immigration, Tax It, one of the more comprehensive keyhole solution proposals to address concerns surrounding open borders.

See also:

Page about Nathan Smith on Open Borders
All blog posts by Nathan Smith

16 thoughts on “Auctions, tariffs, and taxes”

  1. What I had suggested (I think only in private correspondence; I can’t locate any place I publicly commented to that effect) is to start with DRITI + an additional upfront immigration tariff at the beginning, then gradually phase out the tariff (and even better, announce that the tariff will be phased out over time, so that people can plan accordingly). This would smooth out the immigration flow and avoid swamping at the very outset. Those who migrate initially will be those who value migration so highly they can’t wait, whereas others will wait for the tariffs to fall. It would be a bit like the price skimming practiced by innovators of new technologies: an initial high price in order to get high profits (and a very well-selected) initial customer base, then gradually lower the price and reach larger sections of the market.

    Your three-step scheme (auctions, then tariffs, then taxes) is somewhat similar, and was possibly inspired by what I’d said, but is structurally quite different in that it involves shifting between very different styles of managing immigration. It may be a more realistic account of how the US, or another similarly situated country, might actually move towards open borders.

  2. “auctions, then tariffs, then taxes”

    Tariffs can be paired with maximum quotas (which you set high enough that you expect them not to be met at your tariff price, but prevent a surprise surge), which can be increased over time. That separates the risk of a sudden rush of immigrants from price uncertainty for the buyers.

    Basically, this taxonomy is missing the “sell n visas for price p” model..

    ” And the government can easily afford to wait and garnish wages rather than getting a lump sum up front.”

    Governments could extract more total money out of people by exploiting their self-destructive discount rates (experimentally, people are suckers for/terrified of lump-sum payments, depending on who’s paying). However, up-front payments would be more tempting to governments as a way of alleviating budget crises and freeing up money to provide immediate goodies and compensation to voters.

    Also, up-front payments would induce more self-selection: the people who would come would be more productive, more patient, and more self-disciplined. Earlier immigration waves involved large up-front payments for transportation and settlement, which helped to get above-average immigrants relative to the sender countries. The more you make the immigration process convenient for shortsighted people, the more of them you’ll get.

  3. “Basically, this taxonomy is missing the “sell n visas for price p” model.”

    True, good point. The advantage of an auction is that you could get more revenue, but if you want to make the price predictable without opening yourself up to an unlimited rush, this would be a way to do it.

  4. >The advantage of an auction is that you could get more revenue

    You could probably get higher revenue by price discriminating with different tariff rates based on observable hard-to-fake features, like country of origin, ability to pass examinations (general ability or in particular fields), age, etc.

    1. Do you mean that a discriminatory tariff would raise more money compared to a flat tariff (uncontroversial claim) or compared to an open auction (controversial claim)?

      1. I was thinking of an auction where people place bids for n visas, and the top n bidders get visas and are obligated to pay the bid of the nth+1 bidder.

        Those with high reservation prices wind up making out like bandits in consumer surplus if n is large enough that it will have to include marginal migrants who can’t bid much.

Leave a Reply