Bryan Caplan and I clearly have a disagreement about this question. I argue that anti-foreign bias, identified by Bryan in his book The Myth of the Rational Voter, ensures a significantly high enough dislike of immigrants that changes in our welfare programs don’t significantly influence public demand for stricter immigration policies. Bryan, in his latest post in the ongoing debate about liberaltarians and immigration, argues that this is not the case:

Anti-foreign bias is indeed strong and durable.  But this hardly implies that it is invariant to circumstances.  Immigration really was much more free before the welfare state arose.  Welfare state abuse really is one of the most popular arguments against immigration.  And there is good evidence that opposition to the welfare state heavily depends on the perception that out-groups disproportionately benefit from it.  It’s no stretch to flip this evidence and say that support for the welfare state heavily depends on the perception that out-groups don’t benefit from it.

To his point about immigration being freer in the pre-welfare era, I would point out that this was true only for some people, some of the time. For instance there was the Chinese Exclusion Act in 1882 which severely limited Chinese immigration to this country until it was overturned in 1943. There was also the Immigration Act of 1917 and the National Origins Act of 1924, the latter of which completely banned immigration from any Asians. Notice these laws were passed prior to any kind of welfare program and removed the decade after this country’s first major welfare programs were past in the New Deal.

So while overall immigration may have been freer, there has been a lot of really ugly and, depending on who you are, more restrictive immigration laws that existed prior to any sort of welfare state in this country.

But as Karl has recently pointed out, Bryan is a betting man. So I’d like to propose an empirical test to this. If welfare can drive opinions on immigration, then the welfare reform passed in 1996 should have motivated a significant increase in tolerance for immigrants.  The GSS asks whether respondents whether immigration should be increased or decreased, and I think it also has a “neither” option. I propose taking data for the full timeperiod of the GSS and we will not see a positive spike in 1996 or 1997 more than two standard deviations from the mean for those supporting more immigration, or a negative spike for the percent against more immigration. Thats four tests.  I’ll call Bryan the winner if any of them are in his favor, and I’ll bet him $50. I promise I haven’t peeked at the data, in fact I wouldn’t know how since I’ve never worked with GSS data before. This is why I also want to request that, if he accepts the bet, Bryan crunch the numbers since I know he has used the data many times before.

UPDATE: Eli Dourado points out in the comments the series is too short to have a meaningful standard deviation, so let me alter the bet offer to this: support for immigration increases by less than 10% from 1995 to 1997.

UPDATE 2: To put my point above in perspective, think about what the percent of the world that lived in all of Asia from around 1880 to 1943 was. I won’t venture a guess at what that number is, but lets say a large percent. For this part of the world immigration was harder in a pre-welfare state period than it is today.