Tuesday 24 March 2015

Why restricting immigrants' access to benefits might not save the Treasury any money

It has been argued that immigrants in the UK ought not to be able to claim unemployment benefits for some period of time after their arrival, or perhaps indefinitely. Some people make this argument by appealing to a notion of fairness: they contend that immigrants should not be able to receive such benefits until they have first paid into the transfer system through their taxes. Others make the argument with reference to projected savings for the Treasury: they contend that immigrants will be deterred from coming to the UK if they know they will not be able to claim such benefits upon arrival. 

Regarding the latter argument, government figures indicate that if all immigrants who are currently receiving unemployment benefits in the UK stopped receiving them tomorrow, and nothing else changed, the amount of money saved by the Treasury would be not be particularly large. Yet even then, it is unlikely that nothing else would change under such circumstances; in fact, changes would likely occur that would render the amount of money saved by the Treasury smaller than the total sum currently being paid to immigrants. 

Many immigrants to the UK originate in countries where wages for low-skilled workers and average standards of living are much lower than in the UK. It stands to reason that even if they were not able to claim unemployment benefits, low-skilled immigrants from such countries would still find it desirable to come to the UK, albeit somewhat less desirable than at present. More importantly however, reducing immigrants' access to unemployment benefits would presumably make them willing to accept jobs at lower wages. Indeed, it might make them willing to accept jobs at wages that some native workers would find less attractive than unemployment benefits.

In other words, one effect of reducing immigrants' access to unemployment benefits could be to bid down wages to the point where some native workers would shift onto unemployment benefits, thereby curtailing any projected savings for the Treasury. 

Monday 16 March 2015

Historical racism and the black/white incarceration rate ratio

A couple of days ago, I came across a fascinating table of data indicating that states with a longer history of racism tend to have lower black/white incarceration rate ratios. As the person who posted the table noted, "The more stereotypically 'racist' a state is, the *lower* the ratio of black to white imprisonment". It is fairly obvious that this is true just from eyeballing the data, but I thought I'd conduct a slightly more detailed analysis anyway.

The original table can be found at the website for the Bureau of Justice Statistics. (If you click on the link that says "PDF", it is Table 14 on page 11.) Here I examine two alternative measures of historical racism: first, the number of years with anti-miscegenation laws (those that prohibit interracial marriage); and second, school segregation prior to Brown v. Board of Education. The black/white incarceration rate ratio is unavailable for New Mexico and Wyoming. School segregation is not reported for Alaska or Hawaii since both of these states joined the union after Brown v. Board of Education.

The first chart below plots the black/white incarceration rate ratio in 2005 against the number of years for which the state had anti-miscegenation laws. The relationship is negative and significant, namely r = –.53 (p < 0.001, n = 48). It becomes even stronger if Alaska and Hawaii, which joined the union after many states had already repealed their anti-miscegenation laws, are excluded: r = –.65 (p < 0.001, n = 46). 


Years with anti-miscegenation laws is negatively related to the black incarceration rate at r = –.29 (p = 0.042, n = 48), and positively related to the white incarceration rate at r = .28 (p = 0.055, n = 48). Again, both of these relationships become stronger if Alaska and Hawaii are excluded, namely r = –.40 (p = 0.006, n = 46) and r = .31 (p = 0.035, n = 46), respectively. 

The second chart plots the mean black/white incarceration rate ratio by school segregation prior to Brown v. Board of Education. Consistent with the results above, it is monotonically decreasing in the stringency of school segregation laws. States that prohibited school segregation have the highest mean black/white incarceration rate ratio, while states that required school segregation have the lowest mean black/white incarceration rate ratio. 


The mean white incarceration rate is significantly higher in states that required school segregation than in states that prohibited school segregation (difference = 167, p < 0.001). However, the mean black incarceration rate is not significantly lower in states that required school segregation (difference = –305, p = 0.2). In fact, the mean black incarceration rate is highest in states that permitted school segregation (yet there are only two states in this group: Arizona and Kansas). 

It is not clear why black/white incarceration rate ratios tend to be lower in states with a longer history of racism. One possible explanation for the finding that white incarceration rates tend to be higher in such states is simply that whites in these states commit more crime. And one possible explanation for the finding that black incarceration rates tend to be the same or lower in such states is that these states have larger black populations, meaning that there is more likely to be at least one black person on the jury when a black defendant is being tried.

An important caveat is that white and black incarceration rates correspond to inmates in state prisons and local jails, but not those in federal prisons. So it is possible that some or all of the negative association between historical racism and black/white incarceration rate ratio is attributable to a tendency for states with a longer history of racism to send relatively more of their black convicts to federal prisons. 

Saturday 14 March 2015

Which Americans think it is wrong to watch porn?

YouGov just published an interesting poll on attitudes toward pornography in the United States. One of the questions they posed was, "Do you personally consider watching porn to be morally acceptable or morally wrong?" Of those who agreed to answer the question, 39% said "morally acceptable", 38% said "morally wrong", and 23% said they were unsure. Overall then, about as many people think it is acceptable to watch porn as think it is wrong to do so. 

Yet more interesting than the overall percentages were the differences between various socio-economic groups. Unsurprisingly perhaps, as the first chart below indicates, women were more likely than men to say that watching porn is morally wrong (by 14 percentage points, to be specific). In addition, as the second chart below indicates, Republicans were much more likely than Democrats to say that watching porn is morally wrong (by a full 27 percentage points).



Looking at these differences, it occurred to me that, historically, the fiercest opposition to pornography has come from two groups with almost diametrically opposed views on the proper roles for men and women within society: the anti-pornography feminists, and the socially conservative right. Incidentally, this is not to say that all social conservatives or all feminists are opposed to pornography. Indeed, many of the so-called sex-positive feminists take a contrary position on the issue. 

Wednesday 4 March 2015

How spatially concentrated are JSA claimants?

Following on from my last post, the present one examines the spatial concentration of JSA (Jobseeker's allowance) claimants, using data from the ONS. The grey line in the chart below, which corresponds to the general population, indicates how much spatial concentration emerges purely from differences in the populations of local authorities. It has approximately the same integral as the line for White British in the chart from my last post. The line for JSA claimants is slightly further out, indicating that JSA claimants are more spatially concentrated than would be expected purely from differences in the populations of local authorities. 


Here, the top 5% of local authorities (for the relevant group) contain 15% of the general population, and 23% of JSA claimants. The top 10% of local authorities contain 25% of the general population and 35% of JSA claimants.

Monday 2 March 2015

Which ethnic group in England and Wales is the most spatially concentrated?

This post examines the spatial concentration of different ethnic groups in England and Wales, using data from the 2011 census. The eight largest ethnic groups in 2011 were (from largest to smallest): White British, Indian, Pakistani, African, Caribbean, Irish, Bangladeshi, and Chinese. Here, ethnicity is defined by the respondent herself, who ticks whichever box on the census form (e.g., 'Chinese British') corresponds most closely to her own identity. (For the sake of clarity, I omit 'British' from all ethnic group names other than 'White British'.) 

The ONS reports the number of each ethnic group in 348 local authorities in England and Wales. To quantify spatial concentration: I first rank local authorities by percentage of the total population of a particular ethnic group contained within a particular local authority, separately for each ethnic group; and I then plot cumulative percentage of the total population against cumulative percentage of local authorities from highest to lowest, separately for each ethnic group.  


In the legend on the chart above, ethnic groups are ordered from least spatially concentrated to most spatially concentrated: White British is the least spatially concentrated, while Pakistani is the most spatially concentrated. The top 5% of local authorities (for the relevant ethnic group) contain: 15% of White British, 29% of Irish, 32% of Chinese, 52% of Indians, 50% of Africans, 58% of Caribbeans, 62% of Bangladeshis, and 57% of Pakistanis. The top 10% of local authorities contain: 25% of White British, 44% of Irish, 47% of Chinese, 65% of Indians, 69% of Africans, 75% of Caribbeans, 75% of Bangladeshis, and 75% of Pakistanis.