Tag Archives: attitudes to immigration

Collected comments on the World Values Survey data

Co-blogger Nathan Smith recently had a post titled who favors open borders? that looked at some data from the World Values Survey on attitudes to immigration. Nathan’s post was mentioned by Bryan Caplan at EconLog here and by Steve Sailer on his own blog here and here. Commenters at all places have raised a number of interesting points. This post is meant to expound a bit on my own interpretation and mention issues raised by commenters across all these posts.

Interesting theories for the general patterns

This is an expansion and restructuring of some stuff I already mentioned in a comment on Nathan’s post. I’ll first offer the individual theories, then the synthesis.

  1. Countries that people generally want to leave (emigrate from) tend to have a larger proportion of people supporting “let anyone come” and in general seem to have a more pro-open borders position. If this holds up empirically, one simple explanation may be a sort of intuitive Golden Rule: people who want to migrate to other countries take the right to migrate more seriously for immigrants to their country as well.
  2. Countries that generally see a higher proportion of immigrants generally tend to be more restrictionist, while countries that have a low proportion of immigrants (and a low proportion of the foreign-born in general) tend to be more pro-open borders. If this holds up empirically, then the simplest explanation might be that high levels of immigration lead to a nativist backlash by making the native-immigrant distinction more salient. One confounding factor here is that countries with a high proportion of immigrants also tend to have a high proportion of people who are more pro-open borders on account of being immigrants or related to immigrants. My suspicion is that the relation between high immigration levels and low support for open borders would be even stronger if we restricted attention to natives who are native-born and do not have a foreign-born spouse, sibling, parent, or child.
  3. Slightly related to (2), but different: countries that have higher proportions of immigrants tend to be less likely to favor extreme solutions. In other words, in addition to leaning more restrictionist, they’re generally less likely to have lots of people at the extremes of “Let Anyone Come” and “Prohibit” whereas countries like India that have a very low share of the foreign born have large proportions of people at both extremes. Of course, it’s possible that India is an outlier in this regard. I’m less sure of this pattern than the others. The simple explanation for this pattern, if it holds up, is that countries with a large resident foreign born population (whether immigrants or guest workers/students) is more tuned to the practical constraints and “arguments on both sides” and hence would be more likely to support middle-of-the-road solutions.

My overall guess, based on looking at the table, is that the very high “Let Anyone Come” countries are mostly explained by (1). Take a look at the top five countries: Vietnam, Burkina Faso, Rwanda, Ethiopia, and Mali. With the exception of Vietnam, they all have GDP (PPP) per capita per year less than $2000 (they’re in the $1000-1500 range by a number of estimates — see here). Vietnam, with a GDP (PPP) of about $3000-3500 per year, is somewhat of an outlier on the GDP front, but still poor enough to go with the general thesis here. The only other country that I can see on the list with a notably low GDP (PPP) is Ghana, but the estimates for Ghana (#11 on the “Let Anyone Come” ranking) vary quite a bit between $1800 and $3100.

From #6 onward on the list (percentages 28% or lower for “Let Anyone Come”) we start seeing middle-income countries and “upper low-income countries” which are generally not places that people want to desperately leave, but also aren’t attractive destinations for immigrants in general (though they see some traffic from their bordering countries). Countries 6 to 10 are Morocco, Romania, Uruguay, Peru, and India. Of these, Morocco and India are low-income but not the extreme low-income levels of sub-Saharan Africa, whereas Romania, Uruguay, and Peru are solidly middle-income countries — GDP (PPP) between $8000 and $16000 for all of these. The case of India is a little confusing because of its huge size — there are parts of India that have income levels comparable to the extreme low-income sub-Saharan African countries, and other parts that almost make it to middle-income status. One reason for the unusual response percentages in India may be this considerable diversity in the income levels between regions.

For these countries, then, I think the main operative factors are (2) and (3) — they tend to generally be more pro-open borders but also have high numbers of people at both extremes. However, unlike the extreme low-income countries case where almost all these countries are strongly pro-open borders, middle-income countries overall are all over the map. Malaysia, the most restrictionist country by the “Let Anyone Come” metric, is also middle-income. So a more careful statistical analysis would be needed in order to decipher the patterns here.

Patterns for specific countries

Some outlier countries have been pointed out in various comments:

  • The causes of Vietnam’s top position are unclear. Eric speculates a bit about this here (comparing Vietnam to Indonesia), and some commenters on Steve Sailer’s blog post also offer their thoughts.
  • India has received a lot of attention for having unusually high percentages of people in both the “Let Anyone Come” and Prohibit” categories. Regional variation within India may be part of the story. I offer some thoughts on India at this comment.
  • Among developed (high-income) countries, Sweden is a bit of an outlier with respect to its “Let Anyone Come” percentage. Its percentage, 18%, is much higher than the 8% and lower values for other developed countries. Steve Sailer and his commenters offer some theories about Sweden here.
  • Malaysia’s unusually low “Let Anyone Come” number has sparked the interest of my co-blogger John Lee, who offered some preliminary thoughts in this comment.
  • In this comment, Brian Moore points out that Canada and the USA have very similar views on immigration but very different immigration policies.

Who favors open borders?

The World Values Survey records quite a bit of information about public opinion related to immigration. I’d like to do in-depth analysis of it at some point. Here are a few things I’ve noted so far (no rich statistical analysis yet though):

  • Young people worldwide are more favorable to open borders, but the effect is very slight. There is no sign– yet– that generational change will tilt the world towards open borders.
  • Children of immigrants are somewhat more favorable to immigration.
  • There seems to be NO correlation worldwide between attitudes towards immigration policy and self-positioning on the left-right spectrum. (This surprised me.)
  • There seems to be no correlation between social class and attitudes towards immigration policy, unless it’s that the middle classes are a bit more favorable.
  • Correlations with life satisfaction are weak; however, the most strongly restrictionist attitudes seem to be more common among people leaning towards dissatisfaction with their lives.
  • People who trust foreigners “completely” are more favorable to a welcoming immigration policy (well, duh), yet 13% of those who don’t trust foreigners at all still say “let anyone come.”
  • People who don’t want immigrants as neighbors are more likely to favor strict limits on or prohibition of immigration (58%, to 42% of those who don’t mind immigrant neighbors) but some of these, too, favor “letting anyone come.”
  • No difference between men and women.

There are large differences across countries in attitudes towards immigration policy. Only 48 countries seem to be covered by the survey, but among those, two-thirds have public opinion more favorable to immigration than the United States, as measured by the share saying “let anyone come.” In particular, Mexican attitudes towards immigration policy are more liberal than Americans’. Some commenters at this site have suggested Asia as an example of a more restrictionist society that nativist Americans might desire to emulate. The WVS data suggest that this is true at the level of public opinion: Japan, South Korea, Taiwan, and Malaysia have some of the lowest shares of open borders supporters in the world, though in the terms of the number favoring “strict limits” or more, South Koreans are more liberal on immigration than Americans are.

What I find most interesting in the international data is that some developing countries have far more favorable attitudes towards immigration than any rich country. In Vietnam, Burkina Faso, Rwanda, and Ethiopia, nearly half the population favors letting anyone come. India has an unusually large number of open borders supporters as well, though it is also tied for highest in terms of the number of people supporting complete prohibition of immigration. Several countries in Eastern Europe and Latin America also seem to have more open borders supporters than any of the rich countries, except Sweden, which appears to be an outlier, with a far more pro-open borders populace of any rich country included in the survey.

Country Let anyone come As long as jobs available Strict limits Prohibit

1

Vietnam

49%

27%

22%

1%

2

Burkina Faso

43%

45%

10%

1%

3

Rwanda

41%

48%

8%

2%

4

Ethiopia

40%

28%

27%

5%

5

Mali

34%

46%

16%

4%

6

Morocco

28%

41%

20%

11%

7

Romania

23%

42%

23%

11%

8

Uruguay

23%

56%

17%

3%

9

Peru

23%

50%

21%

6%

10

India

23%

22%

25%

30%

11

Ukraine

21%

53%

19%

7%

12

China

20%

51%

21%

8%

13

Ghana

18%

39%

36%

6%

14

Sweden

18%

54%

27%

1%

15

Guatemala

17%

55%

21%

7%

16

Argentina

15%

45%

34%

6%

17

Serbia

14%

26%

46%

14%

18

Bulgaria

13%

55%

24%

8%

19

Moldova

13%

50%

26%

11%

20

Poland

12%

35%

46%

6%

21

Mexico

12%

45%

25%

17%

22

Zambia

11%

30%

44%

15%

23

Brazil

9%

47%

33%

11%

24

Georgia

9%

19%

56%

16%

25

Finland

9%

40%

48%

3%

26

Turkey

9%

43%

27%

21%

27

Italy

8%

49%

37%

6%

28

Canada

8%

51%

39%

2%

29

Spain

8%

48%

42%

3%

30

Slovenia

7%

56%

29%

8%

31

Germany

7%

43%

45%

5%

32

USA

7%

37%

49%

8%

33

Chile

6%

50%

35%

9%

34

Cyprus

6%

36%

51%

7%

35

S Africa

6%

16%

48%

30%

36

Switzerland

6%

67%

26%

1%

37

Indonesia

6%

15%

72%

8%

38

Andorra

5%

72%

22%

1%

39

Egypt

5%

25%

43%

26%

40

Thailand

5%

16%

65%

14%

41

Norway

4%

53%

42%

1%

42

Trinidad And Tobago

4%

32%

55%

10%

43

Australia

3%

54%

41%

2%

44

S Korea

3%

56%

36%

5%

45

Japan

3%

42%

50%

5%

46

Taiwan

3%

30%

58%

9%

47

Jordan

2%

28%

46%

25%

48

Malaysia

2%

8%

72%

18%

 

Another very interesting pattern emerged when I dug down into the data involving religion. When asked “How important is God in your life?” on a scale of 1 to 10, about half the respondents answered “10” and half answered something less.  I was distressed to discover that those for whom God was very important in their lives seemed to have less favorable attitudes towards immigration. But when I broke it down by religious demonination, I found something different. While Muslims who regard God as very important in their lives tend to be more restrictionist, Christians of each denomination are more likely to support open borders if they are strongly in touch with God, as shown in the table below (which includes all denominations for which there were over 500 observations in the WVS dataset):

 

How important is God in your life? (scale: 1-10)
Religious Denomination <10 10
Roman Catholic 9% 15%
Protestant 7% 15%
Evangelical 7% 11%
Orthodox 13% 19%
Church of Sweden 16% 19%
Muslim 19% 13%
Buddhist 7% 9%
Ancestor worship 44% 57%
Hindu 12% 15%

 

The percentage in each cell represents the share of respondents saying “Let anyone come.” Note that it is not the case that Christians are more supportive of open borders in general. Many factors affect support for open borders, and it seems that public opinion in rich countries is often less favorable to open borders. And of course most rich countries are nominally/historically Christian. So Muslims are actually more likely than most Christian denominations to favor open borders. But within each Christians denomination, there is a statistically significant (though fairly small) positive correlation between rating God’s importance in one’s life “10” and advocating “let anyone come.”

Continue reading Who favors open borders?

Statistical discrimination: a perverse incentive for high-skilled immigrants to oppose low-skilled immigration from their homeland

There are many reasons for high-skilled immigrants to support immigration (high-skilled and low-skilled) from their homelands. Loyalty to their fellow homelanders, and a greater interest in the welfare of these, might be a good start. However, there are plenty of selfish reasons as well. More immigration from their homeland would likely mean more people who share the same language and culture, making it possible to gossip in the home language or participate in cultural traditions. It increases the likelihood of the availability of homeland-specific cuisine, and lowers the price of such cuisine. Homeland-specific connections might make it easier to get introductions on jobs or find people to bond with in new and unfamiliar situations.

Armed with knowledge of statistical discrimination (see here and here for starters, or here for a more sophisticated discussion) I came to see that there is a perverse incentive in the opposite direction. Namely, immigrants, like many other people, are often judged statistically based on the average qualities of their “group” which in many cases is all people from their homeland. What this means is that if the average skill level of immigrants from their homeland drops, the immigrant is less likely to be judged as skilled, particularly by strangers. Thus, high-skilled immigrants have a reason to oppose low-skilled immigration from their homeland for the simple reason that it would pull the average down. It’s unclear to me whether this disincentive to support low-skilled immigration would cancel out the pluses.

To make the statistical discrimination analogy clear, consider immigration from Mexico and India to the United States. On average, Mexicans are more, not less, skilled than Indians. However, the stereotype in the US for immigrants from Mexico is that of lower skilled immigrants, whereas Indians are stereotyped as brainy high skilled workers (see here for an anecdotal answer comparing a person’s experience of Indians in the United States versus Indians in India). Why? The reason is that immigration from India is far more selective, with very little illegal immigration, and a much narrower fraction being selected from a considerably larger population. In the case of Mexico, considerable illegal immigration, much of it low-skilled, skews the average skill level downward.

Based on my limited knowledge, I am not aware of any empirical evidence that supports high-skilled immigrants specifically opposing low-skilled immigration from their homelands (as opposed to opposing low-skilled immigration in general). I’m not even sure if this issue has been researched. If, as an empirical matter, they do not, then that probably means that the positive effects far outweigh any worries regarding statistical discrimination.