Comparing US states by their unauthorised immigrant population

California is a common rhetorical example used to illustrate the harms of immigration (unauthorised or otherwise) in the US. People point to California’s runaway government debt, poor public school system, and rising rates of social disorder/crime as the inevitable consequences of more liberal immigration policies. I thought it might be worth pulling together a 50-state view (plus the District of Columbia) to see what we can generalise from a ranking of jurisdictions by their unauthorised immigrant populations.

The Pew Research Center has provided some estimates of the unauthorised immigrant population by state in 1990 and 2010¬†(see tables A3 and A4), and I combined these with US Census Data of the population by state to calculate the share of unauthorised immigrants in each state’s population in 1990 and 2010. It was then a simple step to calculate how much the unauthorised immigrant population has grown or shrunk over the intervening 2 decades.

In 1990, these were the top 10 states (and DC) by share of unauthorised immigrants in their population:

  1. California
  2. Texas
  3. District of Columbia
  4. Arizona
  5. Nevada
  6. New York
  7. Florida
  8. Illinois
  9. New Mexico
  10. New Jersey

In 1990 the bottom 10 were:

  1. Tennessee
  2. Wisconsin
  3. Missouri
  4. Mississippi
  5. Indiana
  6. Iowa
  7. South Carolina
  8. Kentucky
  9. Alabama
  10. Ohio

And as of 2010, here are the top 10 states by share of population:

  1. Nevada
  2. California
  3. Texas
  4. New Jersey
  5. Arizona
  6. Maryland
  7. District of Columbia
  8. Florida
  9. Georgia
  10. New Mexico

The bottom 10 are:

  1. South Carolina
  2. Alaska
  3. South Dakota
  4. Missouri
  5. Ohio
  6. Vermont
  7. North Dakota
  8. Montana
  9. Maine
  10. West Virginia

While I’m not sure what life in these United States was like in 1990, I do know that in 2010 I would much prefer to live in any of the top 10 states ranked by the proportion of unauthorised immigrants in their population than I would prefer to live in the bottom 10. (In fact, I almost live in the District of Columbia: it’s literally walking distance from my current home, though to be fair, the parts of DC that are most accessible to me are also the swankiest. I am quite sure I would not have wanted to live in the District in 1990, however.)

Another way to rank the states would be how much their unauthorised immigrant population has grown. Here are the top 10 states ranked according to the absolute percentage point change in their unauthorised immigrant population:

  1. Nevada
  2. New Jersey
  3. Texas
  4. Maryland
  5. Georgia
  6. Arizona
  7. Oregon
  8. North Carolina
  9. New Mexico
  10. Utah

Meanwhile the bottom 10 (the bottom 3 are actually negative, i.e. the proportion of unauthorised immigrants fell over these 20 years):

  1. New Hampshire
  2. Missouri
  3. Wyoming
  4. South Dakota
  5. West Virginia
  6. Maine
  7. Alaska
  8. Montana
  9. North Dakota
  10. Vermont

Again, I would much prefer to live in most any of the top 10 states than I would in the bottom 10. I lived in New Hampshire/Vermont for the first 4 years of my time in the US (I lived actually on the border of those two states) and as beautiful as they are in the autumn, I can’t say they have much to offer otherwise, especially in the depths of winter (though it would also depend on how much you love skiing or other winter sports). Obviously there is cause and effect here: nice states attract more immigrants. But it does seem clear that if unauthorised immigrants “kill the goose that lays the golden egg” by laying waste to the land of these attractive states, it isn’t terribly apparent from these rankings.

There is one way to slice the data that might be more favourable to restrictionist conclusions, though: we can rank states by the percentage change in their unauthorised immigrant population. So Alabama, with 0.12% of its population unlawfully present in 1990 versus 2.5% in 2010 would have a (2.5 Р0.12) / 0.12 = 1920.2% increase. The low base effect means that these rankings are somewhat suspect, but for your benefit, here they are (along with all the other data I used to construct the rankings above):

State/District 1990 % of pop 2010 % of pop %age point change over 20 years % growth over 20 years
Alabama 0.12% 2.50% 2.38% 1920.19%
Iowa 0.18% 2.50% 2.32% 1288.42%
Kentucky 0.14% 1.80% 1.66% 1227.28%
Tennessee 0.21% 2.20% 1.99% 972.98%
Indiana 0.18% 1.80% 1.62% 897.95%
Ohio 0.09% 0.90% 0.81% 876.24%
North Carolina 0.38% 3.50% 3.12% 828.54%
Wisconsin 0.20% 1.80% 1.60% 780.52%
Arkansas 0.21% 1.80% 1.59% 746.22%
South Carolina 0.14% 1.20% 1.06% 736.71%
Mississippi 0.19% 1.60% 1.41% 724.15%
Georgia 0.54% 4.40% 3.86% 714.40%
Nebraska 0.32% 2.40% 2.08% 657.64%
Hawaii 0.45% 3.10% 2.65% 587.10%
Maryland 0.73% 4.60% 3.87% 528.33%
Pennsylvania 0.21% 1.30% 1.09% 517.91%
Connecticut 0.61% 3.40% 2.79% 458.81%
Michigan 0.27% 1.50% 1.23% 457.72%
New Jersey 1.23% 6.20% 4.97% 404.50%
Oregon 0.88% 4.30% 3.42% 388.88%
Minnesota 0.34% 1.60% 1.26% 366.74%
Missouri 0.20% 0.90% 0.70% 360.52%
Utah 0.87% 3.80% 2.93% 336.46%
Oklahoma 0.48% 2.00% 1.52% 319.41%
Washington 0.82% 3.40% 2.58% 313.67%
Delaware 0.75% 3.00% 2.25% 299.70%
Kansas 0.61% 2.40% 1.79% 296.41%
Colorado 0.91% 3.60% 2.69% 295.34%
Louisiana 0.36% 1.40% 1.04% 293.88%
Nevada 2.08% 7.20% 5.12% 246.08%
Virginia 0.81% 2.70% 1.89% 234.22%
New Mexico 1.32% 4.30% 2.98% 225.74%
Rhode Island 1.00% 3.00% 2.00% 201.04%
New Hampshire 0.45% 1.20% 0.75% 166.22%
Massachusetts 0.91% 2.40% 1.49% 162.53%
Texas 2.65% 6.70% 4.05% 152.91%
Arizona 2.46% 6.00% 3.54% 144.36%
Florida 1.85% 4.50% 2.65% 142.59%
Illinois 1.75% 4.10% 2.35% 134.33%
Idaho 0.99% 2.20% 1.21% 121.48%
District of Columbia 2.47% 4.50% 2.03% 82.07%
West Virginia 0.28% 0.50% 0.22% 79.35%
New York 1.95% 3.20% 1.25% 64.49%
South Dakota 0.72% 1.00% 0.28% 39.20%
Wyoming 1.10% 1.50% 0.40% 36.08%
California 5.04% 6.80% 1.76% 34.90%
Maine 0.41% 0.50% 0.09% 22.79%
Alaska 0.91% 1.00% 0.09% 10.01%
Montana 0.63% 0.50% -0.13% -20.09%
North Dakota 0.78% 0.50% -0.28% -36.12%
Vermont 0.89% 0.50% -0.39% -43.72%

One interesting reaction to all these numbers might be that two decades is too little time to truly assess the long-run impact of unauthorised immigration on a state’s economy and society. So we should be looking for states like Nevada, Texas, New Jersey, Maryland, etc. to become “wastelands” like California over the next decade or two (or at least see some pernicious effects such as bankrupt local governments or increasingly horrid public schools). Then again, many of these states were already in the top 10 in 1990, so it’s not all that clear that we shouldn’t be seeing these supposed effects already.

If you have any thoughts or reactions, feel free to share in the comments. I’ve also uploaded the same numbers in Excel spreadsheet format for ease of use. Hopefully these figures can drive some interesting conversation going forward; it’s quite plausible that I or another Open Borders blogger may return to them in the future.

John Lee is an administrator of the Open Borders website. Liberal immigration laws are a personal passion for him. See all blog posts by John.

Twitter 

4 thoughts on “Comparing US states by their unauthorised immigrant population”

  1. Many pro-immigration arguments, and some restrictionist ones based on state figures need more attention to dose-dependence (immigration proponents are more likely to make this error because it is convenient to claim that if small doses do not have large negative effects, then large doses won’t either).

    http://en.wikipedia.org/wiki/Dose_(biochemistry)#Effects_are_dose-dependent

    Is New York going to collapse over the course of 20 years if it goes from a 2% to a 3% illegal immigrant share, while those immigrants have not had a chance to gain the vote, access most welfare benefits, have children (who receive benefits, vote, and drive the effects of immigration on crime)? Not to mention that much of the law governing New York is not set locally but by the federal government?

    No, that would be almost mathematically impossible. But it would not make sense to expect that such a result generalizes to 50% or 75% population shares, to time horizons long enough for naturalization/amnesty and child-rearing, to populations that have had systematically lower performance in rich countries and in their home countries, and to the long-run equilibrium effects on institutions and savings.

    Take the hypothesis that the long-run effect of migration from third world countries on national income and welfare is a convergence of log income to that typical of countries with the average IQ or human capital of the post-immigration population, using the work of the economist Garrett Jones and others:

    http://mason.gmu.edu/~gjonesb/

    If that hypothesis, on which open borders would in the long run reduce world GDP rather than double it, were true, then the effects of the state-level changes being discussed here would be so small as to be lost almost completely in the noise of the data. If the data are consistent with the predictions of restrictionist hypotheses, that should not be taken as evidence against those same hypotheses.

    Moving away from immigration, here’s an analogy: California and New York both have destructive rent control regimes covering a modest share of the housing stock, yet still have high incomes relative to other states. Economists think that these have distortionary effects, but wouldn’t predict that the controls would make the states poorer than others: the rent control regimes have limited coverage within the states, are not as destructive in their terms as they could possibly be, are offset by many other advantages, and housing is only one part of the economy.

    This does not mean that universal rent controls setting prices below the cost of landlords would not have very bad effects on the housing market, let alone that systematic price controls setting all prices below the cost of production would turn out well. And rent control advocates should not be arguing that the aggregate success of California or New York show their policies are harmless. Instead, they should look at more appropriate scales, like the housing market in the areas where rent control actually operates, the effects when controlled rents are actually much below market rates (so that they have an effect), and over time horizons long enough for the negative effects of rent control on the construction of new housing stock to be visible.

Leave a Reply