Backlash Against Globalization: The
effects of salience and demographic
changes on immigration attitudes in the
United Kingdom from 2006-2014
Laurens Sloot
Student number: s3055140
Email:
l.n.sloot@student.rug.nl
University of Groningen, Faculty of Economics and Business
M.Sc. International Economics and Business
Supervisor
Dr. Dimitrios SOUDIS
Faculty of Economics and Business
University of Groningen
1
Abstract
Previous research often focuses on either individual or national characteristics to determine attitudes towards immigration. This paper looks at how individual, regional and national variables interact to politicize people into developing an anti-immigration point of view. Using NUTS 1 data for the United Kingdom (UK) to look at social attitudes and immigrant statistics, and using national data to determine the national salience of immigration, this paper finds that a stable immigrant share of the population fosters acceptance of immigrants, while large and rapid increases to immigrant shares lead to negative attitudes towards immigration. Furthermore, this paper hypothesises that the national salience on immigration has a negative effect on immigration attitudes, which becomes greater when immigration shares are rapidly growing. However, the effects of salience and the interaction between it and immigration change are insignificant, possibly due to limitations of the data. Further research at a NUTS2 or 3 level is recommended.
2
Contents
1. Introduction
3
2. Literature Review
4
3. Methodology and Data
11
4. Results
21
5. Conclusion
24
References
26
3
1. Introduction
The world is currently experiencing a severe backlash against globalization.
Especially attitudes towards immigrants seem to be at an all-time low in the Western world. This is despite that the 21st century is often referred to as “the age of migration” (Castles and Miller, 2009). In recent years the anti-immigration argument has been gaining grounds. In Europe, following large inflows of Syrian, African and other refugees, many European countries, such as Bulgaria and Austria, have simply closed their borders to them. Although the thousands of drowning refugees in the Mediterranean Sea are undoubtedly heart-breaking stories, public stances towards immigrants and refugees have not improved. To make matters worse, a continued string of terrorist attacks across Europe, but specifically targeting France, started in 2015 and has been continuing. At the same time, all around Europe right-wing populist parties have been gaining more votes in political elections (for example, see recent elections in the Netherlands, Austria, and France). In the USA they elected Donald Trump for president who ran on an aggressive anti-immigration platform. This anti-immigration stance is not a new sentiment and has been a “European” social identity ever since the post-war period (Likata and Klein, 2002). This can be noticed in the United Kingdom’s (UK) immigration policies. After WWII the UK adopted policies with a goal to obtain “zero net immigration” (Givens and Luedtke, 2005). Following the break-up of the Soviet Union and the war in Yugoslavia huge amounts of asylum seekers from Eastern Europe came to many European countries, including the UK. This lead to a further tightening of immigration policy (Somerville et al., 2009). At the same time right-wing populist parties have been growing over the years (Betz, 1994) as has anti-immigration sentiment and this is reflected in the results of the referendum concerning Brexit. According to Lord Ashcroft’s polls , 1/3rd of
people that voted to leave the EU, did so to regain control over immigration (Lord Ashcroft, 2016).
This begs the question: what determines attitudes towards immigration in the UK? This is an important question as immigration attitudes directly influence immigration policy through policymakers and indirectly through elections (Rodrik 1995). Much research has been done to determine the reasons for anti-immigration attitudes, yet scholars remain researching for answers as many results seem ambiguous or highly dependent on the models and variables used or unique local factors. The first obvious reason is that immigrants are seen as an economic threat, by reducing real wages, employment levels and being a burden on tax revenues. But research has shown that immigrants have a small negligible impact on wages, unemployment, GDP and welfare costs (Dustmann, Fabbri, and Preston, 2005). Other research has pointed to education, and that it fosters tolerance towards different ethnicities and multi-culturalism (Hainmueller and Hiscox, 2007). An extensive amount of research has looked at the perceptions that natives have of the immigrant’s impact on the economy, society, culture and politics.
More recent research also focuses on the salience of immigration and how for
4 2005 look at restrictiveness of immigration policy rather than individual attitudes). Despite the extensive amount of research, much of it focusses on cross-country data and often in a single year. There also seems to be a specific focus on the US and Europe as a whole in the literature. I analyse the politicized places hypothesis, developed by Hopkins (2010) to look at how regional shares of immigrants and changes to them interact with national salience to influence immigration attitudes in the UK from 2006 to 2014. Using NUTS1 (see Appendix A for list of regions) data on immigration attitudes obtained from the Europeans Social Surveys, national salience data from the Migration Observatory (Allen, 2016), and the population statistics from the Office for National Statistics (2016a) for the UK, I find that close contact with immigrants improves immigration attitudes, but that fast immigrant influxes lead to anti-immigration. Finally this paper finds that the effect of salience and its interaction with immigrant flows is insignificant, possibly due to data limitation.
This paper is structured as follows: in section 2 I discuss relevant previous literature concerning economic and non-economic determinants of immigration attitudes as well as the effects of salience on attitudes. Section 3 outlines the methodology and data used in the regressions. In section 4 the results are presented and examined. Finally, section 5 concludes this thesis and provides implications for further research.
2. Literature Review
Economic threat
Immigrants are often perceived to have a negative economic impact. Specifically, the perceived negative impact on the labour market has been often used to justify
5 expansion will simply absorb any additional labour. If capital is not completely elastic then owners of capital will also see their wages rise, which may overall decrease average wages and not only low-skilled wages. The model dealing with skill composition is also called the “factor proportions” analysis (Borjas 1999b).
Obviously this model hardly reflects economic reality. In an open economy model with more than 1 sector (Heckscher-Ohlin), an influx of low-skilled labour into a country will cause a shift in that country’s industrial structure and lead to a boost in the production of low-skill intensive products, while the production of other products decreases (Rybczynski theorem). If the country in question is a small open economy, with prices fixed to the world market, low-skilled immigration will first push down low-skilled wages. This will attract firms intensively using low-skilled workers, thus increasing demand, and in turn, low-skilled wages in the long-run. This process can be very quick however if firms foresee the change in skill composition or if capital is easily available. This is known as the hypothesis of factor price insensitivity (Leamer and Levinsohn, 1995). The flexibility of the industrial structure is crucial in this model and is determined by how large the product mix is. If it is small enough, i.e. has less products than factors of production, then the industrial structure cannot adjust (Card, 2001). Another adjustment mechanism is technology. Firms that see a rise in the skilled labour force may opt to adjust their production technology to one that uses more low-skilled labour. Empirical papers that analysed which adjustment mechanism, i.e. either changing production mix or production technology, was more important found that for
around two-thirds of extra labour absorption happens through changes in technology (Hanson and Slaughter 2002).
If the economy is large enough to affect world prices, a boost in the country’s low-skill intensive products output can decrease world prices and subsequently decrease real wages of low-skilled labour. If, on the other hand, immigration inflows are large enough that it causes a change in the number of tradable products of a country, low-skilled labour may experience a fall in real wages. In the case of Europe, which has many small open countries and has not experienced massive immigration inflows, multiple scholars have shown that immigration has little to no effect on long-run wages and unemployment (See Friedberg and Hunt 1995 for a general review; for Europe see Zimmerman 1995; Winkelman and
Zimmermann (1993); Mühleisen and Zimmermann (1994); Hunt 1992; DeNew and
Zimmerman 1994; Hartog and Zorlu 2005; Dustmann, Fabbri, Preston and Wadsworth 2005; Pischke and Velling 1997; Winter-Ebmer and Zweimuller 1999; for the US see Lalonde and Topel 1991; Altonji and Card 1991; Card 2001).
6 move to areas which are doing well economically, which would lead to upward biased
estimates. Furthermore, assumptions on the mobility of the local labour market heavily influence results. Hatton and Tani (2003) find some evidence that in the UK natives crowd out to other areas following large enough immigration inflows, although there results are statistically weak. These problems result in a wide variety of results. Borjas, Freeman, and Katz (1997), for example, find a negative correlation between immigration and native employment in 1960s US, a positive one for the 1970s, and then a negative one again for the 1980s. Schoeni (1997) found similar alternating results for the 1970s and 1980s. In the case of the UK, Gregg et al. (2004) demonstrate that labour mobility for the low-skilled/educated is low, which theoretically would result in lower real wages and employment levels for low-skilled/educated natives.
However Dustmann, Fabbri and Preston (2005), using 2000 British Labour Force Survey data, found that for Britain, most immigrants actually have a very similar skill
composition to the native labour force. This means that immigration should not affect wages, which is in line with their findings, although they do mention a weakness in their data. Other research reinforce their results however. Dustmann, Frattini and Preston (2013) looking at the UK between 1997-2005, concluded that should the share of immigrants to the native
working-age population increase by 1%, average wages would increase by 0.1 to 0.3%. They also found that for the same increase in immigrant share wages for the 5% lowest paid workers would decrease by 0.6%, while higher paid wages increased. Reed and Latorre (2009) looked at the UK for the period 2000-2007 and found that a 1% increase led to 0.3% reduction in average wage. Nickel and Salaheen (2008 and 2015), also looking at the UK, found that a 1% migrant share increase led to 0.5% reduction in the average wage of the unskilled and semi-skilled between 1992-2006 and a 0.2% decrease in 1992-2014.
Borjas (1999a, 2003), however found that immigration negatively affects average real wages of natives in the US. Although looking at average wages is somewhat misleading, since most immigrants are skilled and thus as theory suggests, we would expect low-skilled wages to be more negatively affected, while high-skill wages in fact rise. Negative immigration impacts are also found in models that assume all goods are traded. These results, however do not hold when nontraded products are included in the model (Borjas, 1999b). If immigrants enter a sector of nontraded goods and increase supply faster than demand and thus reduce prices, then the impact of immigration on real wages is largely dependent on consumer tastes.
7
Contact with immigrants: threat and acceptance
The economic impact of immigrants appears to be negligible and it seems that anti-immigration attitudes largely stem from people’s perception of the impact rather than any actual impact on the economy, society, culture and political system. Such perceptions are influenced by group psychology and whether groups feel challenged by other groups
depending on the external circumstances. One of the most prominent theories used to model these attitudes is the racial threat theory, which is part of the realistic group conflict theory (Levis, 2014). The key idea is that groups perceive they are in a competition for scarce resources and thus feel threatened by other groups if these are large enough. This suggests that contact with immigrants would lead to anti-immigration sentiment amongst the native population if natives feel threatened. However, other theories suggest that close contact with different cultures fosters acceptance. Evidence of this however is very varied. In some papers close contact with immigrants reduces the level of perceived threat (eg McLaren 2003). In others contact with immigrants increases threat (Taylor, 1998), where again others find threat levels to remain the same (Escandell and Ceobanu, 2008). This paper matches the line of argument of Hopkins (2010) and Karreth et al. (2015) in distinguishing between levels of immigrants and net inflows of immigrants. When high immigration shares or levels of the population are something that has existed for a long period of time, repeated exposure to immigrants leads to pro-immigration attitudes.
H1: Higher immigration shares of populations positively affect immigration attitudes.
On the other hand, Karreth et al. (2015) argue that such an acceptance process occurs over a long period of time and thus cannot take effect when there are rapid influxes of
immigrants. According to the politicized places hypothesis perceptions about immigration are not so much influenced by immigration levels, but by sudden changes to those levels
8 Figure 1.
(The Economist, 2016).
There are also a number of US case studies that reveal how local politics can change dramatically when faced with rapid demographic changes (e.g. Hopkins, 2009). Overall studies point to the fact that given enough time, high levels of immigration will eventually lead to acceptance, but in the short-run, tensions will arise as natives feel threatened. This argument leads to this paper’s second hypothesis.
H2: Recent influxes of immigrants negatively affect immigration attitudes.
Politicized places
However, the observation of change does not automatically lead to anti-immigration sentiment. For this to happen salient frames need to exist that link immigration to politics and economic issues. Individual ideologies, beliefs, and political alignment are some common frames. Chandler and Tsai (2001) show that people with more conservative and right-wing political ideologies and those with a strong national pride, are more likely to be
anti-9 immigration is evident, the causality between the two is less so. Hajnal and Rivera (2012) argue that white people with more anti-immigration stances or more negative views of Latinos are less likely to favour the Democratic party. Whereas Hainmueller and Hopkins (2010) show that those who align themselves on the right-wing of the political spectrum are more likely to be anti-immigration. No matter the direction of causality, there is little doubt that right-wing parties can frame people’s immigration attitudes. A political alignment variable is therefore also included in the regression. Betz’s (1994) review of the rise of right-wing populist parties in Europe argues that the low-skilled and educated, the young, and the unemployed are the ones more likely to live in areas on the periphery of cities where
immigration growth tends to be high. Feeling increasingly encircled by immigrants and abandoned by the other central and left-wing parties they vote for right-wing populist parties. Right-wing populist parties not only influence the public’s opinion, but also that of colleague politicians, even if right-wing populist parties are not part of the ruling government. Van Spanje’s (2010) work proves empirically that the presence of anti-immigration parties has a contagious effect on immigration policy of other political parties.
Another major frame and of particular interest in this paper, is the media, which plays a large role in framing people’s attitudes and can often mislead people into believing
immigrants pose a larger economic threat than they do in reality. Nelson et al. (1997) suggest framing changes the importance of an issue in people’s eyes. In other words the reported issue becomes more important than others (Chong and Druckman, 2007). Others suggest that framing effects also influence belief content (de Vreese, 2010; Slothuus, 2008). This implies that the media not only influences people’s perception of the importance of an issue, but also whether it is good or bad. This means that the media also affects people’s existing beliefs and not only their importance. Shah et al. (2004) shows how people may change their beliefs when confronted with new information contexts. de Vreese et al. (2011) using their own survey did two studies (one with 304 respondents and the other with 1,632 respondents) to research media framing effects on individual opinions in the context of negotiations about the Turkish membership of the EU. They found that news frames increased the perceived
importance of factors highlighted in them, including economic, cultural and security factors. This in turn influenced the support and valence of attitudes. However they only look at a very small time frame (4 weeks). de Vreese et al. (2011) argue that for framing effects to be useful the content needs to be analysed. In line with other work (see Schneider et., 2001; Shah et al., 2004; van Klingeren, 2014; Soroka, 2006), they showed that negative framing has a much greater impact than positive framing. They also found that positive frames were more
influential on “high political sophisticates” who are better able to understand new information (see also Druckman and Nelson, 2003). Unfortunately, due to a lack of access to data I am unable to measure content of the media articles and thus I only use media attention in my analysis. Nonetheless this paper assumes that the majority of content in immigration articles are negative in nature and thus proposes the following hypothesis.
10 However, the magnitude of influence the media has on attitudes depends on whether immigration is perceived as a problem or threat, meaning when there are large immigrant inflows. Hopkins (2010) finds a significant negative correlation between the interaction of immigration inflows and salience and its effect on immigration attitudes. Hopkins simulated shifting immigration growth levels from the 5th to 95th percentiles at a time when salience was low and found that fast changing counties are only 0.1% more likely to be
anti-immigration. When salience was high, the same analysis meant individuals were 9.9% more likely to be anti-immigration in fast-changing counties. The effect of media on immigration attitudes is a fairly new research topic and not many papers specifically focus on its
interaction with demographics. Givens and Luedtke (2008) find that the immigration salience influences immigration policy in France, Germany and the UK. In regards to immigration attitudes in the Netherlands, van Klingeren et al. (2015) found that immigration statistics had little effect. Also interestingly negative framing of the news had no negative effect, while a positive framing did have a positive effect. They conclude with caution that the issue has been salient for such a long-time, it’s effects are numbed. The numbness of UK citizens to the media is unknown and therefore this paper hypothesises that salience will increase the
negative effect of recent immigrant inflows.
H4: The interaction between national salience and regional change in immigration populations negatively affects immigration attitudes.
The perception that natives are in competition against immigrants is an important factor in determining attitudes. Scheve and Slaughter (2001), Kessler (2001), and Mayda (2006) all find that low-skilled natives perceive immigrants to have a negative impact on the economy. Dancygier and Donnelly (2013), using ESS data look at how growth in industrial sectors influences immigration attitudes and conclude that individuals employed in growing sectors are more likely to be pro-immigration compared to those in shrinking sectors. Dustmann and Preston (2007) prove that British people are more concerned about the immigration impact on welfare than the labour market. Welfare concerns are also deemed important in Hanson, Scheve and Slaughter (2005) and Facchini and Mayda (2006), although they used a different approach. Hainmueller and Hopkins (2014) find that immigration attitudes are shaped by sociotropic perceptions about its cultural impact and to a lesser extent its economic impact. Attitudes are clearly influenced by the perceived economic impact of immigrants. These perceptions are in turn determined by the economic conditions. If economic conditions are poor immigrants become a larger threat. Thus a regional
unemployment rate is included as a control variable to determine the effects of the economic conditions on attitudes.
One crucial independent variable in many papers is educational attainment or skill level. The prevalent finding is that more educated people are less racist and more appreciative of cultural diversity (Hainmueller and Hiscox, 2007; Mayda, 2006) Education also makes students more socially tolerant by increasing knowledge about foreign knowledge,
11 education and pro-immigration attitudes (Hainmueller and Hopkins, 2014), because despite rising average education levels, anti-immigration sentiment is also rising. It may be that those who opt or are selected for more education already possess pro-immigration feelings. Overall most researchers agree that education captures the differences in immigration attitudes extremely well (Chandler & Tsai 2001, Hainmueller and Hiscox 2007, Card et al. 2011). Thus education is also included as a control variable in the regressions.
3. Methodology and Data
The previous section showed why it is sensible to analyse the effects of national salience and changes to immigration populations on immigration attitudes at a regional level. Overall there is a vast amount of literature discussing determinants of immigration attitudes but many weaknesses exist in their models. Literature often looks only a single time period (see for example: Card, Dustmann and Preston, 2005; Hainmueller and Hiscox,2007) and fails to incorporate whether attitudes have changed over time. Those that do, often focus on cross-country comparisons and individual characteristics. While it is no doubt interesting to observe the differences across countries, it makes it hard to analyse determinants of
immigration attitudes. The variation in characteristics between countries and even between regions within a country, means further research should be done at a regional level. This paper attempts to do just so. Unfortunately access to longitudinal data on immigration attitudes is also hard to find at a NUTS2 level and therefore I only look at NUTS1 level data.
Model
The regression is described as follows:
(1) attitudei,j = f(β0j + immigration changeij * β1 + saliencej * β2 + saliencej * immigration changeij * β3 +immigration sharetij * β4j + Xij * β5j
where i is the individual respondent, j corresponds to the wave of the European Social Survey (ESS), and t is time.
The dependent variable, attitude, is a binary variable that is equal to 0 if respondents are anti-immigration and is equal to 1 if they are pro-immigration. Immigration share represents the share of the regional immigrant population at time t (2005). Immigration
change indicates the change in immigrant share of the population from time t to the time
when survey j was held. Salience measures the salience, or the amount of newspapers related to immigration and/or migration at the time of survey j. Xij x β5j incorporates any other covariates, which may influence the regression. These include educational attainment, age, citizenship, and left/right political alignment.
As the dependent variable is binary I employ a logistic regression. Moreover, since I
12
Immigration attitudes
The European Social Survey (ESS) is a cross-country survey that is held every two years since 2002. It measures attitudes, beliefs and behaviours of people in more than 30 European countries, although it only has consistent data for about 22. For this paper I will only consider data for the UK from 2006 to 2014, which consists of waves, 3, 4, 5, 6 and 7. Together these surveys consist of 11,718 respondents in the UK who answered an hour-long questionnaire. Within each wave respondents are linked to the NUTS1 region in which they live. This paper’s empirical tests on immigration attitudes involve answers to the following question:
• To what extent do you think [respondent’s country] should allow people from the poorer countries outside Europe to come and live here?
o Allow many to come and live here (1) o Allow some (2)
o Allow a few (3) o Allow none (4) o (Don’t know) (8)
In this paper, responses 1 and 2 are said to be pro-immigration and responses 3 and 4 are anti-immigration and are given the values of 1 and 0 in the dependent variable atti. In many other social surveys regarding immigration, respondents are asked whether or to tighten or relax immigration policy, but Card, Dustman and Preston (2005) argue that this phrasing is inappropriate because it can be difficult to judge the stance of respondents on immigration. Moreover, the definition and connotation of immigrants differs on an individual level, which is why it is not used in the question. Figure 3 shows a breakdown of immigration attitudes per region for 2006-2014 and figure 4 shows those for attitudes for each survey year.
Figure 3. 0 100 200 300 400 500 600 700
UKC UKD UKE UKF UKG UKH UKI UKJ UKK UKL UKM UKN
Immigration attitudes per region 2006-2014
13 Figure 4.
What becomes immediately clear is that there is in fact quite a strong
anti-immigration sentiment present. While most responses are between “allow some” and “allow a few”, in almost all regions and in all years, there are many more “allow none” than “allow many” responses, with the only exception being London (UKI). There also appears to be a small trend over time, with a slight growth in “allow none” responses and a decrease in all other responses, although that could simply due to the fact there are nearly 100 less
respondents in the years 2012 and 2014 than in previous years.
In the rotating immigration module of wave 5 two other versions of the same question but with different source countries have also been asked.
• Poor countries inside Europe (also in wave 7) • Rich countries outside Europe
• Rich countries inside Europe
Unfortunately these questions have been excluded from all other waves despite a willingness to include them in this analysis. It must be noted that respondents connotate immigrants from poorer countries as being low-skilled (Hainmueller and Hiscox 2007). And as Card, Dustman and Preston (2005) have shown, people are more willing to accept
immigrants from rich countries. Hainmueller and Hiscox (2007) show similar results for the UK. Natives favour immigrants from rich European countries (0.56), poor European
countries (0.53), rich non-European countries (0.51) and poor non-European countries (0.49). Dustman, Fabbri, and Preston (2005), using data from the 2000 British Labour Force Survey, computed the education levels and skill composition of natives, immigrants, and recent immigrants. Surprisingly, immigrants have a very similar education level and skill
composition and are actually slightly more skilled. Thus while this question may not reflect the type of immigrants actually entering the UK, it seems unlikely that most UK citizens know the actual skill composition of immigrants and thus it is more likely to capture any anti-immigration sentiment as it somewhat reflects people’s perceptions about immigrants.
0 100 200 300 400 500 600 700 800 900 1000 2006 2008 2010 2012 2014
Immigration attitudes per year
14 Table 1 shows the summary statistics for UK responses about how many immigrants should be allowed in from poorer countries outside Europe (see also figures 3 and 4) and the perceived effect of immigrants on the overall country, economy, taxes, jobs, culture, and crime. Excluding the first question, a value of 0 indicates anti-immigration sentiment and a value of 10 indicates favouring immigrants. On average immigrants are seen to be a negative influence on all of these problems, especially crime. When asked about what qualifications immigrants should have, respondents of ESS-7 gave a score of 10 for extremely important and 0 for not important. They placed much importance on immigrants being able to speak English, being committed to the way of local life, and to having skills that are needed in the country. Being Christian and white were seen as unimportant (Table 2).
Table 1.
(1) (2) (3) (4) (5)
VARIABLES N mean sd min max
Allow many/few immigrants from poorer countries outside Europe
11,470 2.705 0.883 1 4
Immigration bad or good for country's economy
11,474 4.558 2.471 0 10
Country's cultural life undermined or enriched by immigrants
11,422 4.880 2.607 0 10
Immigrants make country worse or better place to live
11,496 4.557 2.479 0 10
Taxes and services:
immigrants take out more than they put in or less
2,186 4.553 2.409 0 10
Immigrants take jobs away in country or create new jobs
2,217 4.683 2.346 0 10
Immigrants make country's crime problems worse or better
2,179 3.868 2.048 0 10
15 Table 2.
(1) (2) (3) (4) (5)
VARIABLES N mean sd min max
Qualification for immigration: Christian background
2,239 2.913 2.954 0 10
Qualification for immigration: committed to way of life in country
2,247 7.939 2.183 0 10
Qualification for immigration: good educational qualifications
2,241 6.960 2.266 0 10
Qualification for immigration: speak country's official language
2,249 8.080 2.113 0 10
Qualification for immigration: be white
2,239 1.741 2.476 0 10
Qualification for immigration: work skills needed in country
2,242 7.557 2.291 0 10
Number of groups 12 12 12 12 12
Salience
To analyse the influence of salience on attitudes I have obtained immigration salience data published through The Migration Observatory, part of the University of Oxford (Allen, 2016). Allen (2016) in turn used two archiving services, Nexis UK and Factiva, which incorporates many full-text articles of many newspapers and periodicals, including all 19 national UK publications during 2006-2015. Allen (2016) looked at 11 of the most popular publications (figure 5) to do a frequency analysis of how many articles had migration-related words in their titles or content (see Allen, 2016 for full details on data collection and
methods).
Figure 5: Publications considered in the salience variable.
Tabloids Midmarkets Broadsheets
Daily Mirror (Sunday Mirror) Daily Mail (Mail on Sunday) The Daily Telegraph (Sunday Telegraph)
Daily Star (Daily Star Sunday) The Express (Sunday Express) Financial Times
The People The Guardian (The Observer)
The Sun The Independent (Independent on Sunday)
The Times (Sunday Times)
16 “immigration” or “migration” was slowly decreasing from 2006 until 2011 (see figure 6). The small spike in 2010 is accredited to the General Election, which introduced a new
Conservative-led coalition, which had a strong goal to reduce net migration. Policies to achieve this were introduced as of 2011 and onwards. As of 2012 the number of immigration articles increased again, with 2014 having more than double the amount of articles as 2011. The correlation between immigration inflows and salience is only 0.148, but appears higher when looking at 2014. Figures 6 and 7 show the fluctuations in immigration related articles and net immigration and both rapidly increased in 2014.
Figure 6.
Figure 7.
(Office for National Statistics, 2016b)
Due to the fact that quite some fluctuation exists in the number of articles, I
normalized the values of the variable by subtracting the mean from the value and dividing it by the standard deviation. Summary statistics can be found in table 3. There are, however,
0 50 100 150 200 250 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 M igr a tion ( th o u s a n d s )
17 some severe limitations to the data. Firstly, this paper fails to incorporate data on web or television salience, which probably have significant influences on framing. There is also the fact that this paper does not distinguish between articles that have pro-immigration content and those that have anti-immigration content. This makes it harder to draw concrete
conclusions on the regression results, specifically whether British people are more influenced by negative news or positive news. These are possibilities for further research.
Local demographics
To measure the share of immigrants of the regional population and its growth, I have used local area population data, obtained from the Office of National Statistics (ONS, 2016a). The ONS uses the same definition for long-term international migrants (LTIM) that the United Nations recommends, which is “a person who changes their country of usual
residence for a period of at least a year, so that the country of destination effectively becomes the country of usual residence” (United Nations, 1998). Their main source of data is the International Passenger Survey (ISP), which is a continuous voluntary survey held at all main air and sea routes. It provides estimates of both inflows and outflows, but not exact counts, nor does it cover all migration types. The ONS uses the Labour Force Survey (LFS) to calibrate the geographical distribution of immigrants. To measure the immigrant share of the regional population, I looked at estimates of the total-and the Non-British population,
obtained from the ONS and derived from the Annual Population Survey (APS) for 2005. I chose not the use estimates of Non-UK born population, since being born in the UK does not automatically give British citizenship. I then calculated the change of immigration shares between 2005 and subsequent years. However it is likely that the data will vary depending on what definition of migrants you use.
18 Figure 8.
Control variables
Several control variables are also included in the regressions. These variables are almost all found in the ESS waves and describe characteristics of the respondents. Education is added as a control variable, since literature suggests it is one of the strongest predictors of immigrant attitudes (Hainmueller and Hopkins, 2007). The variable is called eduyrs and measures the amount of years spent in full-time education. Age (agea) is included as studies report it has a strong negative correlation with attitudes (Dustmann and Preston, 2001). British citizenship is also included to control for the fact that some respondents are not UK citizens and may in fact be immigrants themselves. It seems logical that non-UK citizens are more likely to favour immigrants. ctzcntr is the binary variable that equals 1 when the respondent is a British citizen and 2 if not. lrscale measures whether respondents politically identify themselves with either the right (10) or the left (0). This is included as several studies prove that right-wing people are more probable to dislike immigrants. Finally the regional unemployment rates (variable unemp) is added to check whether higher rates have an adverse effect on attitudes. Figure 10 shows the unemployment rate over time in each region.
0 1 2 3 4 5 6 7 8 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
Immigrant share of population
Scotland Wales
Northern Ireland North East North West
Yorkshire and The Humber East Midlands
West Midlands East
19 Figure 10.
(NOMIS 2016)
Table 3 shows the summary statistics and table 4 shows the correlation matrix. There are several interesting correlations to note. First of all there is the strong correlation between immigrant shares and immigrant change (0. 6839). This goes to show that immigrants do indeed largely go to places where there are already large immigrant populations. This is also shown in London which has both the largest immigrant share and inflow. As mentioned before immigrant change and salience have a small positive correlation (0.148). Furthermore there is a negative correlation between salience and unemployment (-0. 354). This suggests that salience on immigration is in fact lower when economic conditions are poor, which is somewhat surprising, although due to the limited timeline, this hardly constitutes as proof. Peculiar is the positive correlation between unemployment and immigrant change (0.4775). one might expect these to be negatively correlated, because immigrants would migrate to areas with low unemployment. Instead this correlation seems to suggest that immigration lead to higher unemployment levels, even though theories suggest and studies have shown this not the case. More likely it is a mixture of causes. One likely cause is the after-effects of the 2008 financial crisis, which caused many European countries to experience higher unemployment rates. Higher unemployment rates may therefore have not greatly influenced immigrants’ decisions to move to the UK, since many European countries were worse off (see Appendix C). Furthermore, some of the political and economic circumstances immigrants and refugees are leaving are so poor in comparison that almost any country is likely to have better
prospects. 0.0 2.0 4.0 6.0 8.0 10.0 12.0 2006 2007 2008 2009 2010 2011 2012 2013 2014
Regional unemployment rates
UKC UKD UKE UKF UKG UKH
20 Table 3. Summary statistics
VARIABLES Abbreviations Obs Mean Std. Dev Min Max
Years of full-time education completed eduyrs 11,233 13.36776 3.792868 0 56
Age of respondent agea 11,279 50.577 18.8762 15 123
UK citizen ctzcntr 11,356 1.043061 0.2030033 1 2
Political left/right scale lrscale 9,909 5.047835 1.888637 0 10
Regional unemployment rate unemp 11,358 6.641838 1.588468 3.8 10.8
Regional immigration share imsh 11,358 4.869644 4.485946 2.088028 18.57958 Change in immigration share imch 11,358 1.517146 1.350413 0.1031919 7.361795
Salience sal 11,358 .0021823 1.000362 -1.163318 1.806244
Table 4. Correlation Matrix
imsh imch salnorm salnximch eduyrs agea ctzcntr lrscale unemp
21
4. Results
This section contains the regression results. The first column in table 5 shows the results of the regression with only the immigrant share as an independent variable, while allowing for random effects at a regional level. Immigration share is statistically significant at a 95% level and has a positive effect on immigration attitudes, which corroborates the first hypothesis. In the second column the only independent variable is immigrant change and this has a negative and significant effect on immigration attitudes which is in line with the second hypothesis. In the third column only national salience is considered. Here the results are not corroborated with the third hypothesis. Salience is not only insignificant, but has a positive influence on immigration attitudes. In column 4 the three previous variables are included together in the regression, which largens the magnitude of the coefficients but has no effect on their sign or significance. Column 5 includes the interaction variable between immigrant change and national salience, which is insignificant. Hypothesis 4 therefore also seems to be unsupported. The subsequent columns include control variables including education, age, citizenship, political alignment, and the regional unemployment rate. These, with the exception of the unemployment rate, are all significant and have the expected signs. Their results show in column 6 that more time spent in education has a positive influence on immigration attitudes. Column 7 shows that being older has a negative effect on attitudes. Column 8 reveals that a non-UK citizen is more likely to favour immigration than UK citizens. In column 9 people who are more right-wing politically are more likely to be anti-immigration. Finally in column 10, the coefficient for regional unemployment rates is negative, which was predicted, but is also insignificant.
22 Table 5.
23 Both salience and the interaction variable are insignificant in all regressions and thus do not support hypotheses 3 and 4. There are several possible explanations for the
insignificant results of the salience and interaction variables in the regressions. It may be due to the absence of data on online and television media. It could also be that the observations need to be made over a longer time-period and perhaps that data at a NUTS 2 or even 3 level needs to be analysed to better capture changes in immigrant shares and local media attention. The attitudes of larger cities, where there tend to be more immigrants and pro-immigration university students, may overshadow the attitudes in smaller towns within a single region and the effects that the media has on them. Likewise, on a national level, it could be that London simply dwarfs the other regions, since it has the largest share and change in immigrant
populations and also has the most positive attitudes towards them. It is also quite possible that the UK people are not really affected by the printed media. To further analyse this would require a breakdown of the content of salience data to see the difference in effects of positive and negative media. Although I do not have access to this type of data for the period of 2006-2014, Grierson (2017) summarizes the results of a trust barometer survey by PR firm
Edelman. In an article for The Guardian, he writes that UK people’s trust in the media fell from 36% in 2016 to 24% in 2017. This would mean that people are numbed by the news and do not let it affect them as much. Subsequently this implies the possibility of other frames that have a larger influence on attitudes. These frames could include different forms of online media (e.g. social media) and television. In effect the results indicate that this paper studies a too narrow selection of frames.
For further analysis consider the average predicted probability of 0.43 of the last regression in column 10 in table 6. This value means that the average probability that any respondent’s attitude is equal to 1, or in other words is pro-immigration, is 43%. This corresponds with the mean value of the dependent variable attitude of 0.42, which means slightly more respondents are anti-immigration. To visualise the effects of salience and immigration change I plot the predicted margins by changing the values of the predictors. The figures in Appendix D show the probability that attitude =1 (i.e. pro-immigration): when change of immigrant share varies from 0 to 15%, and when salience ranges from -2 to 5. As expected both immigrant change and salience have negative slopes. The predicted
24 Figure 11
Figure 12.
Overall hypothesis 1 is completely supported by the evidence and hypothesis 2 is mostly supported, whilst hypotheses 3 and 4 are not supported at all. The results indicate that education and citizenship capture the largest effect on immigration attitudes, after which come immigrant share and change. This is clear evidence towards the importance of the distribution of immigrants within a country, and immigration policy should take this into consideration. While salience was shown to be insignificant to attitudes, this is likely due to limitations on the data. Overall this paper shows that immigration share and change are important determinants of immigration attitudes.
5. Conclusion
In this paper I have empirically investigated the correlation between immigration attitudes and immigrant shares, influxes and salience. Contrary to most other research I used panel data to look at regional and national variables over time. Furthermore the interaction of salience and immigration change has not yet been widely researched, which I suspect is due to a lack of data. I chose to use a mixed effects logistic model to appropriately capture the
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 P red ictiv e p ro b ab ilit y th at a tti =1 Change in % immigrant
Interactive effect - change in % immigrant
Sal min Sal max 0 0.1 0.2 0.3 0.4 0.5 -2 -1 0 1 2 3 4 5 P red ictiv e p ro b ab ilit y th at a tti = 1 Salience
Interactive effect - salience
25 multilevel structure of my data and because my dependent variable was binary. My results coincide to some extent with previous literature. The first hypothesis is accepted as the correlation between attitudes and immigrant share is positive, because prolonged contact with immigrants breeds understanding and acceptance. However, when areas experience rapid increases to immigrant levels this acceptance process cannot catch up, thus making natives feel threatened by rising immigrant populations and fostering negative sentiment. Hypothesis 2 is thus also confirmed. The effect of salience is hypothesised to be negative on immigration attitudes but is insignificant in this paper. The interaction term between salience and
immigrant change is negative, but again insignificant. The null hypothesis for hypotheses 3 and 4 can therefore not be rejected. However, to conclude that salience and the interaction term does not have any effect on immigration attitudes would be putting the cart before the horse as there are several possible explanations for this weak statistical result. The main explanation is that printed media has become less influential and thus other forms of media need to be analysed, for example online media. Overall only half of the key results of the regressions are significant, yet this paper suggests that salience and its interaction with local demographic changes could still be a determinant of social attitudes towards immigration. My research has found significant results for education, age, citizenship, and political alignment that are consistent with previous research.
Further research is required to better analyse the correlation between attitudes, salience and local demographics. However continued research is likely to be difficult for the UK and also possibly for other countries due to a lack of available data. Continued research at a NUTS 2 or even NUTS3 level is recommended, but unfortunately there are not many social attitudes surveys that provide information at that level, instead often pooling it together at a NUTS 1 or national level. The British Social Values Survey does have such local data but does this type of immigration survey inconsistently. Access to older reports are also not available online. There is also a lack of publicly available data on salience of the media. All research papers concerning media salience referenced in this paper have used the LexisNexis archive, including, indirectly, this paper. Access to this database requires paid membership, which unfortunately I did not have. Luckily, Allen (2016) published some of their
26 only highly interrelated but able to deeply influence individuals around the entire world. While this paper has expressed a need to focus on the local level, it would be a mistake to ignore international and global matters when analysing individual attitudes.
References
Altonji, J.G. and Card, D. (1991). The Effects of Immigration on the Labor Market Outcomes of Less-skilled Natives, in Immigration , Trade and Labor, edited by J.M. Abowd and R. B. Freeman, University of Chicago Press.
Allen, W.L. (2016). A Decade of Immigration in the British Press. The Migration Observatory report, COMPAS, University of Oxford, UK.
Betz, H.-G. (1993). The New Politics of Resentment: Radical Right-Wing Populist Parties in Western Europe. Comparative Politics, 25(4): 413-427.
Borjas, G.J. (2003). The Labor Demand Curve Is Downward Sloping: Reexamining the Impact of Immigration on the Labor Market. Quarterly Journal of Economics, 118: 1335–1374. Borjas, G.J. (1999a). Heaven’s Door: Immigration Policy and the American Economy. Princeton,
N.J.: Princeton University Press.
Borjas, G.J. (1999b). The Economic Analysis of Immigration Chapter 28, Handbook of Labor Economics, 3: 1697-1760
Borjas, G. J., Freeman, R.B., and Katz, L.F. (1997). How much do immigration and trade affect labor market outcomes. Brookings Papers on Economic Activity, 1: 1-67
Card, D. (2001). Immigrant Inflows, Native Outflows, and the Local Labor Market Impacts of Higher Immigration, Journal of Labor Economics, 19(1): 22-63.
Card D. Dustmann C. Preston I. (2011). Immigration, Wages, and Compositional Amenities. J. Eur. Econ. Assoc. 10:78-119.
Castles, Stephan and Miller, Mark J. (2009). The Age of Migration: International Population Movements in the Modern World (4th edition). Basingstoke: Palgrave MacMillan, Chandler, C.R., and Tsai, Y.M. (2001). Social Factors Influencing Immigration Attitudes: An
Analysis of Data from the General Social Survey. Social Science Journal 38(2). pp. 177–188 Chong, D., and Druckman, J. N. (2007). A theory of framing and opinion formation in competitive
elite environments. Journal of Communication, 57(1): 99-118.
Dancygier, R.M., Donnelly, M.J. (2013). Sectoral economies, economic contexts, and attitudes toward immigration. The Journal of Politics. 75:17–35
De New, J. P. and Zimmermann, K. F. (1994): Native Wage Impacts of Foreign Labor: A Random Effects Panel Analysis Journal of Population Economics, 7: 177-192.
27 de Vreese, C. H., Boomgaarden, H. G., Semetko, H. A. (2011). (In)direct framing effects: The
effects of news media framing on public support for Turkish membership in the European Union. Communication Research, 38: 179-205.
Druckman, J.N., and Nelson, K.R. (2003). Framing and deliberation: How citizens’ conversation limit elite influence. American Journal of Political Science, 47: 729-745.
Dustmann, C., and Preston, I.(2001). Racial and Economic Factors in Attitudes to Immigration. IZA Discussion Paper 190. Bonn, Germany: IZA (Institute for the Study of Labor)
Dustmann, C., Fabbri, F. and Preston, I. (2005). The Impact of Immigration on the British Labour Market”, Economic Journal, 115(507): 324-341.
Dustmann, C., Fabbri, F. Preston, I. and Wadsworth, J. (2005). The Local Labour Market Effects of Immigration in the UK. Unpublished manuscript, United Kingdom Home Office
Dustmann, C., Frattini, T., and Preston, P. (2013). The Effect of Immigration along the Distribution of Wages. Review of Economic Studies, 80(1): 145-173.
Dustmann C., Glitz A., and Frattini T. (2008). The labour market impacts of immigration, Oxford Review of Economic Policy, 24(3): 477-494.
Escandell, X. and Ceobanu, A. (2009). When Contact With Immigrants Matters: Threat, 21 Interethnic Attitudes and Foreigner Exclusionism in Spain's Comunidades Autonomas. Ethnic and Racial Studies, 32: 44-69.
European Social Survey data:
ESS Round 3: European Social Survey Round 3 Data (2006). Data file edition 3.6. NSD - Norwegian Centre for Research Data, Norway – Data Archive and distributor of ESS data for ESS ERIC ESS Round 4: European Social Survey Round 4 Data (2008). Data file edition 4.4. NSD - Norwegian
Centre for Research Data, Norway – Data Archive and distributor of ESS data for ESS ERIC. ESS Round 5: European Social Survey Round 5 Data (2010). Data file edition 3.3. NSD - Norwegian Centre for Research Data, Norway – Data Archive and distributor of ESS data for ESS ERIC. ESS Round 6: European Social Survey Round 6 Data (2012). Data file edition 2.3. NSD - Norwegian Centre for Research Data, Norway – Data Archive and distributor of ESS data for ESS ERIC. ESS Round 7: European Social Survey Round 7 Data (2014). Data file edition 2.1. NSD - Norwegian Centre for Research Data, Norway – Data Archive and distributor of ESS data for ESS ERIC. Eurostat (2017). Unemployment statistics. Available at:
http://ec.europa.eu/eurostat/statistics-explained/index.php/Unemployment_statistics.
Facchini, G. and Mayda, A.M. (2006). Individual Attitudes Towards Immigrants: Welfare-State Determinants Across Countries. CReAM Discussion Paper 04/06, University College London Economics Department.
28 Gang, I.N., Rivera-Batiz, F.L., and Yun, M.S. (2002). Economic Strain, Ethnic Concentration and
Attitudes Towards Foreigners in the European Union. IZA Discussion Paper 578. Bonn, Germany: IZA (Institute for the Study of Labor)
Givens, T. and Luedtke, A. (2005). European Immigration Policies in Comparative Perspective: Issue Salience, Partisanship and Immigrant Rights. Comparative European Politics, 3(1): 1-22
Gregg, P., Machin, S. and Manning, A. (2004). Mobility and joblessness, in (R. Blundell, D. Card, R. Freeman, eds.), Seeking a Premier League Economy, pp. 371–410, Chicago: University of Chicago Press.
Grierson, J. (2017). Britons' trust in government, media and business falls sharply. The Guardian. Available at: https://www.theguardian.com/politics/2017/jan/16/britons-trust-in-government-media-business-falls-sharply. (Accessed 15 May 2017)
Hainmueller, J. and Hiscox, M. J.. (2007). Educated Preferences: Explaining Attitudes toward Immigration in Europe. International Organization 61 (2): 399–442.
Hainmueller, J. and Hopkins, D.J. (2014). Public Attitudes toward Immigration. Annual Review of Political Science 17: 225–249
Hajnal, Z. and Rivera, M. (2012). Immigration, Latinos, and white partisan politics: the new Democratic defection. Working Paper, Dep. Polit. Sci., Univ. Calif. San Diego
Hanson, G. (2005). Challenges for Immigration Policy. in C. Fred Bergsten, ed., The United States and the World Economy: Foreign Economic Policy for the Next Decade, Washington, D.C.: Institute for International Economics, pp. 343–372.
Hanson, G. H. and Slaughter, M. J. (2002). Labor-market Adjustment in Open Economies: Evidence from US States. Journal of International Economics, 57 (1): 3-29.
Hanson, G.H., K. Scheve, and. Slaughter, M.J. (2005). Individual Preferences over High-Skilled Immigration in the United States, mimeo.
Hartog, J. and Zorlu, A. (2005). The Effect of Immigration on Wages in Three European Countries+ Journal of Population Economics 18 (1):113–51
Hatton, T. J. and Tani, M. (2005). Immigration and Inter-Regional Mobility in the UK, 1982- 2000. Economic Journal 115 (507): 342-358.
Hopkins, D.J. (2009). The Diversity Discount: When Increasing Ethnic and Racial Diversity Prevents Tax Increases. Journal of Politics 71 (01): 160–77.
Hopkins, D.J. (2010). Politicized places: explaining where and when immigrants provoke opposition. American Political Science Review, 104:40–60
Hunt, J. (1992). The Impact of the 1962 Repatriates from Algeria on the French Labor Market, Industrial and Labor Relations Review, 45(3): 556-72.
29 Kessler, Alan (2001). Immigration, Economic Insecurity, and the “Ambivalent” American Public.
Working Paper 41. University of California, San Diego: Center for Comparative Immigration Studies
Krugman, P. and Obstfeld, M. (2000). International Economics: Theory and Policy. 5th ed. Reading, Mass: Addison-Wesley, pp.72-73
Lalonde, R.J. and Topel, R.H. (1991). Labor Market Adjustments to Increased Immigration, in Immigration , Trade and Labor, edited by J.M. Abowd and R. B. Freeman, University of Chicago Press
Leamer, E. and Levinsohn, J. (1999). International Trade Theory: The Evidence, in Handbook of International Economics, Volume 3, edited by G. M. Grossman and K. Rogofff, North Holland Press, pp. 1339-1394.
Levis, Laura (2014). Unraveling "Racial Threat". Harvard Magazine. Available at:
http://harvardmagazine.com/2014/09/unraveling-racial-threat. (Accessed 20 May 2017). Lord Ashcroft (2016). How the United Kingdom voted on Thursday… and why. Lord Ashcroft
Polls. Available at: http://lordashcroftpolls.com/2016/06/how-the-united-kingdom-voted-and-why/. (Accessed at 08/04/2017).
Mayda, Anna Maria (2006). Who Is Against Immigration? A Cross-Country Investigation of Attitudes Towards Immigrants. Review of Economics and Statistics 88 (3): 510–530 McLaren, L. (2003.) Anti-immigration Prejudice in Europe: Contact, Threat Perception, and
Preferences for the Exclusion of Migrants. Social Forces 81 (3): 909–36. Mühleisen, M. and Zimmermann, K.F. (1994). A Panel Analysis of Job Changes and
Unemployment. European Economic Review. 35(3/4): 793–801.
Nelson, T. E., Clawson, R. A., and Oxley, Z. M. (1997). Media framing of a civil liberty conflict and its effects on tolerance. American Political Science Review, 91, 567-583
Nickell, S. and Salaheen, J. (2008). The Impact of Immigration on Occupational Wages: Evidence from Britain. Working Paper No. 08-6, Federal Reserve Bank of Boston, Boston.
Nickell, S. and Salaheen, J. (2015). The Impact of Immigration on Occupational Wages: Evidence from Britain. Staff Working Paper No. 574, Bank of England, London.
NOMIS (2016). Annual Population Survey. Available at:
https://www.nomisweb.co.uk/query/construct/summary.asp?mode=construct&version=0&dat aset=17.
Office for National Statistics (2016a). Local Area Migration Indicators, UK. Available at:
https://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/internation almigration/datasets/migrationstatisticsquarterlyreportprovisionallongterminternationalmigrat ionltimestimates.
Office for National Statistics (2016b). Long-term international migration 2.06, area of destination or origin within the UK. Available at:
30 almigration/datasets/longterminternationalmigrationareaofdestinationororiginwithintheuktabl e206
Parker, S. (2015). ‘Unwanted invaders’: The representation of refugees and asylum seekers in the UK and Australian print media. Myth and Nation, 23. Available at:
http://orca.cf.ac.uk/79108/.
Pischke, J. and Velling, J. (1997). Employment Effects of Immigration to Germany: An Analysis Based on Local Labor Markets, The Review of Economics and Statistics.
Razin, A., and Sadka, E. (1995). Resisting Migration: Wage Rigidity and Income Distribution+ American Economic Review, 85(2): 312–316
Reed, H. andLatorre, M. (2009). The Economic Impacts of Migration on the UK Labour Market. IPPR Economics of Migration Working Paper 3, Institute for Public Policy Research, London.
Rodrik, D. (1995). Political economy of trade policy. In Grossman, G. and Rogoff, K., editors, Handbook of International Economics, Vol.3, chapter 28, pages 1457—1494. North-Holland Elsevier Science, The Netherlands.
Scheve, K.F. and Slaughter, M.J. (2001). Labor Market Competition and Individual Preferences over Migration Policy, Review of Economics and Statistics, 83(1): 133- 45
Schneider, T.R., Salovey, P., Apanovitch, A. M., Pizarro, J., McCarthy, D., Zullo, J., et al. (2001). The effects of message framing and ethnic targeting on mammography use among low-income women. Health Psychology, 20: 256-266.
Schoeni, R. F. (1997). The effect of immigrants on the employment and wages of native workers: evidence from the 1970s and 1980s. Unpublished paper (RAND, Santa Monica, CA). Shah, D., Kwak, N., Schmierbach, M., and Zubric, J. (2004). The interplay of news frames on
cognitive complexity. Human Communication Research, 30(1): 102-120
Sjaastad, Larry A. (1962). The costs and returns of human migration. Journal of Political Economy 70: 80-93
Slothuus, R. (2008). More than weighting cognitive importance: A dual-process model of issue framing effects. Political Psychology, 29: 1-28.
Somerville, W., Sriskandarajah, D., and Latorre, M. (2009). United Kingdom: A Reluctant Country of Immigration. Migration Policy Institute. Available at:
http://www.migrationpolicy.org/article/united-kingdom-reluctant-country-immigration. (Accessed 8 May 2017)
Soroka, S. N. (2006). Good news and bad news: Asymmetric responses to economic information. Journal of Politics, 68: 372-385.
31 The Economist (2016). Britain’s Immigration Paradox. The Economist Group Limited. Available at:
https://www.economist.com/news/britain/21701950-areas-lots-migrants-voted-mainly-remain-or-did-they-britains-immigration-paradox. (Accessed at: 20 May 2017).
United Nations (1998). Recommendations on Statistics of International Migration, Revision 1. Available at: https://unstats.un.org/unsd/demographic/sconcerns/migration/migrmethods.htm. (Accessed 8 April 2017).
van Klingeren, M. (2014). “Welcome” to Europe. How media and immigration affect increasing euroscepticism. Doctoral dissertation, University of Amsterdam. Available at
http://hdl.handle.net/11245/1.416265
van Klingeren, M., Boomgaarden, H.G., Vliegenthart, R., and de Vreese, C.H. (2015). ‘Real World Is Not Enough: The Media as an Additional Source of Negative Attitudes Toward
Immigration, Comparing Denmark and the Netherlands’. European Sociological Review 31 (3): 268–83.
Van Spanje, J. (2010). Contagious parties: Anti-immigration parties and their impact on other parties’ stances on immigration in contemporary Western Europe. Party Politics 16(5): 563– 586
Winkelmann, R. and. Zimmermann, K.F. (1993). Aging, Migration and Labour Mobility. In Johnson, Paul, and Klaus F. Zimmermann, eds., Labour Markets in an Ageing Europe. Cambridge: Cambridge University Press, 1993, pp. 255–83.
Winter-Ebmer, R. and Zweimuller, J. (1999). Do Immigrants Displace Young Native Workers: The Austrian Experience, Journal of Population Economics, 12(2): 327- 340.
32
Appendix A - NUTS 1 regions UK
UKC = North East UKD = North West
UKE = Yorkshire and the Humber UKF = East Midlands
UKG = West Midlands UKH = East of England UKI = London
33
Appendix B – Annual immigrant net change
Appendix C – EU unemployment rates
(Eurostat, 2017) -0.20 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
Annual change in the immigrant share of population
Scotland WalesNorthern Ireland North East North West
Yorkshire and The Humber East Midlands
West Midlands East
34
Appendix D – Predicted Probabilities
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 P red icted p ro b ab ilit y th at a tti =1 change % immigrant
