The effect of immigration on the unemployment rate in the
United Kingdom
Bachelor Thesis June 2016
Faculty: Faculty of Economics and Business Student Name: Gabriele Kusaite
Student Number: 10630392
Specialization: Economics and Finance Supervisor: dr. E.W.M.T. Westerhout JEL Classification: C30, E24, F22, J20, J61
2
Statement of Originality
This document is written by Gabriele Kusaite who declares to take full
responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and
that no sources other than those mentioned in the text and its references have
been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of
completion of the work, not for the contents.
3 Abstract:
This paper analyses the relationship between immigration and the unemployment rate in the United Kingdom for the period of 1972-2015. The theoretical section of this paper analyses the trends of immigration in the UK over the period researched and discusses the theories of possible effects of immigration on the host country’s labour market. The empirical research done in the paper includes regressions for the full period of 1972-2015, as well as for the two split periods of 1972-2003 and 2004-2015. It is found that depending on the variables included in the regression immigration is predicted to have either no effect or a negative effect on the unemployment rate. That is, an increasing number of immigrants is expected to not affect unemployment or decrease unemployment in the UK. In addition to this, it is found that the effect of immigration on the unemployment rate has not experienced any significant changes over time.
4 CONTENTS:
1. Introduction 5
2. Theoretical framework 6
2.1 Immigration to the UK 6
2.1.1. The level of immigration, emigration and net migration 6
2.1.2. Composition of immigrants in the UK 7
2.2 Literature review 9
2.3 The effect of immigrants on the labour market of the host country. 11 2.3.1 Neoclassical economic theory: standard model of labour. 11
2.3.2 Theory of productive endowments 13
2.3.3 Self-selection of immigrants 13
2.4 Relationship between GDP and unemployment: Okun’s law 14
3. Methodology 15
3.1. The Model 15
3.2. The Data 16
3.2.1. Unemployment 17
3.2.2. Immigration 17
3.2.3. Gross Domestic Product (GDP) 18
3.2.4. Emigration 18
4. Results 18
5. Conclusion 20
5 1. Introduction
On June 23rd this year the United Kingdom will hold a referendum on whether or not to stay in the EU.
The decision made by the voters will have a tremendous effect on the future of the UK and its economy. As with any other topic, staying in the EU has its advantages and disadvantages and arguments of both sides have been a widely discussed topic over the past couple of months.
One of the main statements of the “Vote Leave” campaign is that leaving the EU would significantly decrease the levels of immigration, since the UK could set its own rules regarding immigrants from the EU, who now have rights to free movement. In addition to that, some supporters of the campaign believe that restricting immigration would help to increase the levels of employment. According to them, the high number of immigrants is a concern, because it is thought to increase the unemployment rate of the locals, especially those who are low-skilled. However, these are just opinions of the supporters’ of leaving the EU campaign and they are representing their political views as well, so the validity of this statement is uncertain. Therefore, it is very important to separate politics and assess arguments from a purely economic perspective. Thus, the research question to be answered in this paper is to what extent does the level of immigration affect the unemployment rate in the UK?
The annual number of immigrants coming to the UK has increased significantly over the past few decades. In fact, in the last 20 years it doubled. The biggest part of the increase in the immigration is due to the immigrants from the European Union. Because of the free labour market movements among the EU members, the UK cannot control the immigration from these countries. Therefore, they do not need to meet any education or skill requirements, as they do not need any visa in order to enter the country. Because of that a significant part of those immigrants is likely to not be high skilled and not have high levels of education. This may cause higher competition among low skilled workers in the country, and as the foreign labour force is more likely to accept worse working conditions and lower wages, this might lead to negative outcomes for the UK labour market, in particular higher unemployment.
The sections of this paper are as following. Section 2 discusses changes in immigration in the UK, the motivation for the thesis, previous literature on the topic and the relevant theoretical models. Section 3 describes the model and the data used in order to answer the research question. Section 4 gives the results and the possible reasons behind the results. Finally, Section 5 concludes the paper and discusses the possible limitations of the research.
6 -200 -100 0 100 200 300 400 500 600 700 1970 1975 1980 1985 1990 1995 2000 2005 2010
Migration 1970-2014
Net migration Immigration Emmigration 2. Theoretical framework
2.1 Immigration to the UK
2.1.1. The level of immigration, emigration and net migration
The graph below shows the annual figures of long term international migrants from 1970 to 2014 in the UK. More specifically, it includes the levels of immigration and emigration, as well as the level of net migration, which is the level of immigration less the level of emigration. All numbers are in thousands.
Side note: due to the revision of net migration estimates for the period from 2001 to 2011, the net migration for those years is not equal to the level of immigration less the level of emigration (the levels of immigration and emigration have not been revised).
As it can be seen from the graph, the number of immigrants and emigrants has stayed around the same level from the beginning of the period until 1994. In the year 1994 the level of immigration exceeded the level of emigration and has remained higher since then. From 1994 to 2004 the number of immigrants almost doubled reaching the level of nearly 600 thousand people per year. It remained at that level after that, with the exception in 2012 and 2013. In 2014 the yearly number of immigrants exceeded 600 thousand making it the all time high. Therefore, it can be concluded that immigration has increased sharply over the past 20 years and it could be the case that such a significant number of immigrants does indeed have an effect on the unemployment of the natives. However, in order to make any further conclusions about the immigration and to explain these changes, first, the composition of the immigrant population is needed to be analysed.
7 -150 -100 -50 0 50 100 150 200 250 300
Net Migration by Nationality
EU citizens Non-EU citizens British Citizens 2.1.2. Composition of immigrants in the UK
The graph below shows the long term international net migration by nationality from 1975 to 2014. All numbers are in thousands.
As it is clear from the graph, the net migration from non-EU countries has risen significantly from 1998 to 2000. In that period it increased from 129 thousand in 1998 to 214 thousand in 2000. Since then it has stayed at around the level of 200 thousand, except for the drop in 2012 and 2013. However, since the immigration from non-EU countries is being controlled by the rules of the British government and those immigrants need visas and working permits in order to be able to work in the UK, it is not so relevant to this study. The more relevant issue for this paper is immigration from the EU countries, since immigrants from the EU have a full access to the labour market of the UK, and one of the reasons of the leaving the EU is the idea to be able to control their free movement.
The level of net migration of the EU citizens has stayed around the level of 0 until 2004. That means the immigration was equal to emigration. Therefore, the level of EU immigrants in the UK has not been significant during the period up to 2004. In 2004 the net migration increased by almost 100 thousand. As it can be seen from the previous graph, 2004 saw almost 100 thousand people increase in total immigration as well. Therefore, it can be concluded that the main reason behind this increase was the immigration from the EU. Such an increase of EU immigrants can be explained by the fact that ten countries joined the European Union that year, namely, Cyprus, the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Malta, Poland, Slovakia, and Slovenia. Since then the level of net migration from the EU has been increasing moderately until the year 2007. Then, there was a drop in net migration numbers that lasted until 2012.
After that, it again increased sharply from 82 thousand in 2012 to 174 thousand in 2014. One of the reasons behind this might be the immigration from Romania and Bulgaria. Even though these countries joined the EU in 2007, it was only on 1 January 2014 that Romanians and Bulgarians gained
8 0 20 40 60 80 100 120 140 YE Jun 06 YE Jun 07 YE Jun 08 YE Jun 09 YE Mar 10 YE Sep 10 YE Mar 11 YE Sep 11 YE Mar 12 YE Sep 12 YE Mar 13 YE Sep 13 YE Mar 14 YE Sep 14 YE Mar 15 YE Sep 15
EU immigration
EU 15 citizens EU 8 citizens EU 2 citizensthe full access to the UK’s labour market, as it was the end of transitional employment restrictions. According to the data from the UK’s Office on National Statistics (ONS) the total number of immigrants from Romania and Bulgaria has reached more than 200 thousand in the end of 2014. In 2014 migrants from Bulgaria and Romania represented 6% of the total migration to the UK.
It is also important to note, that the level of immigration of the EU citizens and non-EU citizens is becoming closer. According to the provisional immigration numbers from 2015 from ONS, the level of immigration from the EU came close to the level of immigration from non-EU countries (it was around the level of 270-280 thousand, with the level of immigration from non-EU countries still slightly above).
All of the statements from above about EU immigration can also be confirmed by the graph below. The graph shows the immigration of the EU citizens from June 2006 to December 2015. Figures for the year 2015 are provisional. EU 15 includes Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain, Sweden and United Kingdom. EU 8 includes Malta, Cyprus, Estonia, Latvia, Lithuania, Poland, Czech Republic, Slovakia, Slovenia and Hungary. EU 2 is Bulgaria and Romania.
As it can be seen from the graph, immigration from the EU2 did increase sharply, since the end of transitional employment restrictions. Immigration from EU8 has stayed around the level of 60-80 thousand. However, the trend of immigration from EU15 has been upwards since the beginning of the period, especially in the last three years. The possible reason behind this could be the financial crisis that started in 2008. Since countries experienced significant long-lasting negative effects on their economies during the years of the crisis, residents of the countries that took longer to recover could have migrated to the UK in search of a better economic situation.
In fact, immigration from the “old” EU states (EU15), such as Greece and Italy, increased significantly, reaching more than 1 million working age people in the end of 2014. In addition to that, the latest Labour Force Survey data from the employment statistics shows that the estimate of the
9
employment level of the EU nationals (not including British) was almost 2.15 million in March 2016, representing nearly 7% of total working population, while the employment level of non-EU nationals was only 1.2 million, representing nearly 4% of total working population.
In conclusion, it can be seen that the composition and nationalities of immigrants have changed significantly over time. There has been a sharp increase in the number of immigrants from the EU countries since 2004 and yearly levels of immigrating EU-citizens have almost reached the levels of non-EU immigration. Therefore, the results of this study could be different from the results of the previous research done about the effect of immigration on the unemployment level in the host country.
2.2 Literature review
This paper will focus on the relationship between immigration and unemployment in the UK. There has been a significant number of papers analyzing the relationship between immigration and various labour market variables for different countries. Most of the results concluded that immigration has no significant long term effect on country’s economy (unemployment, wages, etc.). The most relevant studies and their results will be reviewed below.
Sebastien Jean and Miguel Jimenez (2011) conducted a study on the unemployment effect of immigration in OECD countries on both skill and aggregate levels. For the skill-level each country’s labour market was divided into 18 segments with a corresponding level of education and experience, whereas for the aggregate level the entire labour market was researched as one entity. The results for the skill-level analysis would show how different the outcomes are for different types of workers, while the aggregate level analysis would show the total average effects. The data for the research only included males for eighteen OECD countries, including the UK, and covered the period of 1984-2003 for aggregate analysis and the period of 1992-2003 for disaggregate analysis.
The findings were that immigration does not have any permanent effect on unemployment of the natives. Moreover, it was found that if the share of immigrants in the labour force remains constant, there is no short-term effect on unemployment as well. However, when the share of immigrants increased, the findings were different. On the skill-level it was found that natives, who possess similar skills to immigrants, are affected with a temporary and limited rise in unemployment compared to other natives. Natives from the categories that felt a higher increase in the share of immigrants than in the previous years experienced temporary and insignificant rise in unemployment as well. On the aggregate level the short-term effect could have been larger than on the skill-level, but it was difficult to evaluate and it was concluded insignificant after adjusting for economic shocks.
Research done more on an individual country basis, includes a paper by Boubtane et al (2013). They performed a Granger causality analysis on immigration and host countries’ economic conditions (unemployment and GDP per capita). The data was collected for a period of 1980-2005 on 22 OECD countries. Their findings revealed that only in one out of twenty two countries, namely Portugal, unemployment had a negative causality on immigration inflow, while immigration did not cause unemployment in any country. Furthermore, they found that immigration inflow did not cause per
10
capita economic growth in any country, while GDP per capita growth caused immigration in four countries, namely France, Iceland, Norway and the United Kingdom.
There are papers studying the effects of immigration for other countries as well. For example, Chassamboulli and Palivos (2013) did a research on Greece for the period of 2000-2007. They found that skilled natives experienced increase in both wages and employment if they were complements to immigrants, while the effect on low-skilled workers depended on the size of the minimum wage (or if it is present at all) and government policies on it. After that Latif (2015) did a research on the relationship between immigration and unemployment in Canada for 1983-2010. The results showed that in the long run immigration had no significant effect on unemployment rate.
A different approach on the matter was taken by E. Moreno-Galbis and A. Tritah (2015) in one of the more recent papers. They studied the effects of immigration on labour markets in 13 Western European countries in the period from 1998 to 2004. The researchers introduced the differences in host country specific assets between the locals and immigrants into their model. In other words, they believed that immigrants lack country specific labour market knowledge and assets, thus they have worse bargaining positions with their employers. That would mean, that immigrants would accept lower wages, and give firms more profit, making them able to hire more employees. The same would hold for more productive and less productive immigrants, as well as for immigrants with the same productivity level as natives. Therefore, they predicted that immigrants could have a positive effect on natives’ employment prospects. Findings matched the expectations, as it was proven that immigrants have a positive short term effect on employment rate of natives in job sectors with the highest immigration rates.
The first study that was concentrated only on the UK was performed by Dustmann et al. (2015). They conducted research on the impact of immigration on the British labour market. The study included the period 1983-2000. They researched labour market variables including employment, unemployment, participation and wages. After looking at the composition of the labour force of the natives and immigrants, the authors found that the skill and education distributions of both were similar. Therefore, they believed that immigration could have different effects on labour market than in other countries, such as USA. The results of the study concluded that there is no significant evidence of immigration having overall adverse effect on aggregate unemployment and other variables. In fact, the impact of immigration on wages was positive, yet, statistically insignificant. However, in some groups there was evidence of negative effects on employment, and it was those groups with intermediate education. The negative effect in those groups was offset by the positive effect on employment of people with higher qualifications.
Few of the previous articles, namely Sebastien Jean and Miguel Jimenez (2011) and E. Moreno-Galbis and A. Tritah (2015) did include the UK in their research. However, the results were not reported on the individual country basis. This makes it unclear if the results could be applied to the UK. The situation of immigration in the UK is likely to be different from other countries’ situations, because of factors, such as language, free healthcare and strong economic state. Therefore, it is worth studying the individual case of the UK.
11
In addition to that, the newest data used in all of the previously discussed studies that did research effects of immigration in the UK was from 2005. However, this paper will include the data up until the end of 2015. The results from the previous studies may be different from the results of this paper because the nature of immigration has changed from being controlled by the laws of the UK, to being controlled by the laws of the EU. Thus, as it is already mentioned in the previous section of this paper, the nationalities of immigrants have significantly changed over the past 10 years, as the population of EU immigrants has increased. Therefore, it is likely that the immigrants from the EU, who have the rights to free movement, have different skill and education levels than the immigrants from the earlier periods.
As it can be concluded from the previous research, immigration should have no significant negative effect on unemployment. Therefore, this is the expectation of the results of this study. However, since this paper will include more recent data, it is possible that the results will differ.
2.3 The effect of immigrants on the labour market of the host country. 2.3.1 Neoclassical economic theory: standard model of labour.
This model is based on the paper from Borjas (1989). Firstly, some assumptions are needed to be made in order to simplify the model. One of the restrictions is that the production function is assumed to have constant returns to scale. In other words, if the amount of all inputs is doubled, the output will double as well. Furthermore, the higher level of any input while holding other input levels constant will have diminishing returns. That is, the output will increase at a diminishing rate. It is also worth noting, that in the competitive market the firms that profit-maximise will hire the factors of production as long as their price does not exceed the value of the marginal productivity. Also, the quantities employed of all the inputs will determine the wage of every input. Using this standard model, it is possible to analyse what will happen to the employment and wages of various inputs when the supply of immigrants is shifted (there is a change in the supply of the immigrants in the market of the host country).
For example, when the supply of immigrants increases (Si shifts to Si’), the wage of the immigrants will decrease (from ri to ri’) and the employment of the immigrants will increase (from xi to xi’). The decrease in the wage is caused by the diminishing returns (the more immigrants there are in the market, the smaller their marginal productivity is). It can also be explained as an increase in competition in the labour market of immigrants. This situation is presented in Figure 1.
The impact of immigrants on the employment and wages of the natives depends on their relationship. Both of the possible outcomes are exhibited in Figure 2. If the natives and immigrants are substitutes, meaning they can substitute each other in the process of the production, the demand function of the native labour will shift to the left (from Dn to Dn’) decreasing the wages and employment of the natives (from rn and xn to rn’ and xn’). According to Borjas (1989, p. 480), this happens due to the increased competition between natives and immigrants. In this situation increased supply of the immigrants has a negative effect on the productivity of the natives.
12
On the other hand, if they are complements, the demand function of the native labour will shift to the right (from Dn to Dn’’) making wages and employment of the natives increase (from rn and xn to rn’’ and xn’’). This implies that natives gain from the additional supply of the immigrants, since immigrants increase the productivity of the natives.
This model, therefore, shows that it is impossible to tell how exactly immigrants will affect the native labour market, without analysing the situation empirically. That is, in order to be able to tell if the demand of natives will move to the right or to the left it is essential to find a degree to which natives will be complemented or substituted by the additional supply of the immigrants. However, the immigrants and the natives will not always have the same technological relationship. Thus, in some cases, immigrants will be complements with certain groups of natives, while with the other groups they will
Figure 1
13
exhibit the relationship of substitutability. Most importantly, it is necessary to realise that immigrants cannot be treated as a single input into the labour market.
2.3.2 Theory of productive endowments
The similar thoughts were presented in one more paper by Borjas (1999). The analysis concluded that if the natives of the host country have different productive endowments, which can also be explained as different skill levels, than those of the immigrants, they will benefit from the inflows of immigrants. Moreover, the greater the differences in productive endowments are, the more benefits natives can get. However, the benefits are not equally divided to the population of the natives. The natives who have productive endowments which compete with the endowments of the immigrants lose, while those who have complementary productive endowments to those of the immigrants experience gains.
2.3.3 Self-selection of immigrants
Borjas (1994) talked about the model of self-selection that was first described by Roy (1951). This model suggests that there are three types of selection that describe immigrant flows: positive and negative selection and refugee sorting. However, only the first two types are relevant for this study.
Positive selection occurs when immigrants have above average abilities in their home (source) country, as well as the host country. Therefore, their earnings in both of the countries are also above average. If the host country has a more dispersed income than the country of origin and if the earnings across those two countries are highly correlated, those immigrants will be from the upper tail of their home country income distribution and they will outperform demographically similar natives. The reason behind this is that the home country is “taxing” high ability workers and “insuring” low ability workers. Therefore, the income of the host country is more unequally distributed and high income workers that are migrating from the source country experience higher benefits than low income workers. This situation is called “brain drain”, as the host country with higher inequality becomes more attractive for people who are more successful in the labour market.
Negative selection occurs when people who earn below average in their home country and perform poorly in the host labour market migrate from the source country to the host country. The necessary condition for it to happen is the source country having more inequality in income distribution that the host country. The earnings between two countries have to have a positive correlation as well. In other words, this situation happens when the host country has higher “taxes” on high income workers and insures low income workers better than the source country. This leads to migration of low skilled workers to the host country, as they can improve their living situation, while high income workers have no high incentives to leave, as they have relatively good opportunities in their home country already.
There are a number of important conclusions to be made from all of these theories. Firstly, the effect of immigration on the economy of the host country mostly depends on how comparable the skills of the immigrants are to the skills of the native work force. If most of the immigrants have similar skills to the natives, the effect to the economy will be negative, while if the skills of the immigrants
14 -6 -4 -2 0 2 4 6 8 10 12 14 19 71 19 73 19 75 19 77 19 79 19 81 19 83 19 85 19 87 19 89 19 91 19 93 19 95 19 97 19 99 20 01 20 03 20 05 20 07 20 09 20 11 20 13 20 15
Unemployment and GDP growth
Unemployment Real GDP growth
significantly differ from the skills of the native population, immigration will have a positive effect on the employment and wages of the host country. However, the effect will not be equally distributed throughout the economy, groups with different skills will gain and groups with similar skills will lose. In addition to that, the skills of the immigrant flows are highly dependent on the differences of the income distribution between host and the source country. If the high income workers are taxed relatively more in their home country than in the host country, they will migrate and immigrants in the host country will be high skilled, while if there is less income inequality in the host country than in the source country, low income workers will migrate and that will lead to low skilled immigrants in the host country.
2.4 Relationship between GDP and unemployment: Okun’s law
One of the main factors affecting unemployment rate is GDP. The relationship between unemployment and GDP is negative and can be described by Okun’s law. According to Prachowny (1993), the main idea of the Okun’s law is that a one percentage point reduction in unemployment would increase the output by around 3 %. However, these numbers may vary among different countries and different time periods. The output is measured by the level of the GDP, while the change in the output is measured by the growth rate of GDP. In the paper by Lee (2000), the “gap” model equation of Okun’s law is as following:
𝑦𝑡− 𝑦𝑡∗= −𝛽1∗ (𝑢𝑡− 𝑢𝑡∗)
Where the 𝑦𝑡 is the level of output, 𝑦𝑡∗ is the potential level of output, 𝑢𝑡 is the rate of unemployment
and 𝑢𝑡∗ is the natural rate of unemployment. Therefore, if unemployment goes above the natural rate of
unemployment, GDP falls below its potential level and vice versa.
The graph below shows the relationship between the unemployment and real GDP growth in the UK for the period of 1970 to 2015.
15
As it can be seen from the graph, the relationship between unemployment and GDP is negative. In the periods of negative GDP growth unemployment increased, and in the periods of high GDP growth unemployment decreased. Thus those two variables are negatively correlated. However, the correlation is far from -1. Therefore, it can be concluded that real GDP growth is an important factor in determining the unemployment rate, and it should be added into the model of this paper.
3. Methodology 3.1. The Model
This study aims to analyse the effects of immigration on unemployment in the UK. In other words, it will be tested if the flows of immigrants have any effect on the unemployment rate in the country. Therefore, the empirical model used in this study will include the unemployment rate as a dependent variable, and immigration as an independent variable. The hypotheses will be as following:
𝐻0: 𝛽𝐼𝑀 = 0; 𝐻1: 𝛽𝐼𝑀 ≠ 0
That is, the null hypothesis will test if the coefficient of migration is zero, which would mean that migration does not have a significant effect on unemployment in the UK. The alternative hypothesis would indicate a coefficient different from zero, which would mean that migration does have an effect on unemployment. The positive coefficient would indicate immigration having a positive effect on unemployment and, therefore, a negative effect on the UK’s economy, while a negative coefficient would indicate that immigration decreases unemployment in the country.
In total three time periods will be researched: the full period of the available data from 1972-2015 and two split periods, including the period from 1972 to 2003 and the period from 2004 to 2015. The division point is the year 2004, since, as it is mentioned in the previous section, it was the year of significantly increased number of immigrants from the EU, who are likely to have a different effect on the UK’s economy, than immigrants from earlier years.
The models used in this research employ OLS multiple regression framework and include variables discussed in previous sections. Two different equations will be used for the regressions: with and without emigration. However, in order to have a larger number of data points in the regressions for the two shorter time periods, the coefficients for those will be separated and both periods will be put into one regression. That is, the regression will include two immigration variables, one of them being zero after the year 2003 and the other being zero up to the year 2004. Then, the first variable will result in the coefficient for immigration for the period of 1972-2003 (period A) and the second variable will result in the coefficient for the period of 2004-2015 (period B). For both the full and separated time periods two regressions will be performed and in total that will result in four regressions. The equations are as following:
16
1) ∆𝑢𝑡 = 𝛼 + 𝛽𝑔∗ 𝑔𝑡+𝛽𝑖𝑚∗ 𝑖𝑚𝑡+ 𝜀𝑡
2) ∆𝑢𝑡= 𝛼 + 𝛽𝑔∗ 𝑔𝑡+ 𝛽𝑖𝑚∗ 𝑖𝑚𝑡+𝛽𝑒𝑚∗ 𝑒𝑚𝑡+ 𝜀𝑡
3) ∆𝑢𝑡 = 𝛼 + 𝛽𝑔∗ 𝑔𝑡+𝛽𝑖𝑚1∗ 𝑖𝑚𝑡1+ 𝛽𝑖𝑚2∗ 𝑖𝑚𝑡2+ 𝜀𝑡
4) ∆𝑢𝑡 = 𝛼 + 𝛽𝑔∗ 𝑔𝑡+𝛽𝑖𝑚1∗ 𝑖𝑚𝑡1+ 𝛽𝑖𝑚2∗ 𝑖𝑚𝑡2+𝛽𝑒𝑚∗ 𝑒𝑚𝑡+ 𝜀𝑡
Where ∆𝑢𝑡 is an absolute change in the unemployment rate at time t (𝑢𝑡− 𝑢𝑡−1), 𝑔𝑡 is the real GDP
growth at time t, 𝑖𝑚𝑡 is the percentage of the immigrants of the economically active people at time t
(𝑖𝑚𝑡1 refers to the period A, 𝑖𝑚𝑡2 refers to the period B), 𝑒𝑚𝑡 is the percentage of the emigrants of the
economically active people at time t and 𝜀𝑡 is the error term.
This model is derived using previous articles on the topic. Firstly, Latif (2015) states that unemployment is a function of the immigration rate and per capita real GDP. However, since this paper is looking at change in unemployment, the more accurate measure is the real growth rate of GDP. This model is also similar to the model researched by Asif (2013), except that it excludes population growth. However, Asif researched Pakistan, India and China for the period of 1987 to 2009, which all had a significant average yearly growth in population for that period that could affect unemployment (2.43%, 1.82% and 0.96% respectively), whereas the UK had a relatively stable population (average yearly growth of 0.34% for the period researched in this paper). Other variables included in that model were the exchange rate and inflation, but the real growth rate of GDP already takes these into account. Emigration is added into second equation, since its effect might be as important as the effect of immigration.
In order to perform the regressions some assumptions need to be made. According to Stock and Watson (2011) the assumptions are as following. Firstly, exogeneity will be assumed throughout the study. Secondly, the following assumptions have to be made as well:
1. 𝑢𝑖 is a random variable with 𝐸(𝑢𝑖|𝑋𝑖) = 0
2. Large outliers are unlikely
In addition to testing the coefficients being different from zero, the coefficients of the regression with two periods will be compared. Since the period B starting in 2004 had much higher levels of immigration from the EU than the previous years, the coefficients for it will be expected to be higher than the coefficients for the period A.
3.2. The Data
The data for annual unemployment rates, annual levels of immigration and emigration, and annual real GDP growth for the researched period is available from the UK’s Office of National Statistics (ONS). All data from ONS was extracted from their official website, thus, it is reliable. Regressions for all of the periods include 44 data points.
17 3.2.1. Unemployment
The dependent variable used in this paper is the change in unemployment rate, which is calculated as the difference between unemployment rate in the current period and the unemployment rate in the previous period. Therefore, the data needed included yearly unemployment rates from 1971 to 2015 (45 observations).
As mentioned previously, the data on yearly unemployment rates is from the ONS. Unemployment rates are measured by the Labour Force Survey, which is a survey of the employment circumstances of the UK population. Their definition of unemployed people is the definition that is specified by the International Labour Organisation (ILO). Unemployed people are those without a job who have been actively seeking for a job in the previous four weeks and are available to start work in the next two weeks and those who are out of work but have accepted a job that they are waiting to start in the next two weeks. The rate of unemployment is then a percentage of economically active people who are unemployed. Economically active people include those who are working or unemployed.
3.2.2. Immigration
The independent variable used in the research is immigration. The data for long-term international migration is from The Migration Statistics Quarterly Report from ONS. This paper uses the data from the latest report from 26 May 2016. The report uses statistics on migration published by the Office for National Statistics (ONS), the Home Office and the Department for Work and Pensions (DWP). The definition of a long-term international migrant used in the statistics is the recommended definition of the UN: “A person who moves to a country other than that of his or her usual residence for a period of at least a year (12 months), so that the country of destination effectively becomes his or her new country of usual residence” (United Nations Department of Economic and Social Affairs: Statistics Division, 1998). Estimates of the Long-Term International Migration are based mostly on data from the International Passenger Survey, with some adjustments made for certain groups, including, asylum seekers, non-asylum enforced removals, people resettled in the UK under the various resettlement schemes, visitor and migrant switchers and flows to and from Northern Ireland.
The data of the long-term international migration includes the number of immigrants, emigrants and net migration (immigration minus emigration). Data for 2015 is only provisional. The yearly total number of immigrants will be used for the estimate of yearly immigration, since it was used in most of the other research, such as Jean, S., & Jimenez, M. (2011); Latif, E. (2015) and Moreno-Galbis, E., & Tritah, A. (2015). However, in order to make the coefficient easier to interpret, the number of immigrants will be divided by the total number of economically active people for the each year. The number of economically active people is calculated by using the unemployment rate and the level of unemployed. This percentage of immigration will be included as a variable in both regressions. According to the OECD statistics the average duration of unemployment over the years 1990-2015 in the EU was 14 months. Because of the definition of the long term international migrant being someone who
18
has lived in a different country for more than a year, the regression does not need to include any lags for those variables.
3.2.3. Gross Domestic Product (GDP)
GDP growth will be included as an independent explanatory variable. The annual data for the period researched (1990-2015) on GDP growth is available from the ONS. In the ONS dataset, GDP growth is calculated by using the method of chained volume measures (CVM), which means that the growth is already in real terms.
From previous research on Okun’s law (Sögner, L. (2001); Lee, J. (2000); Moosa, I. A. (1997)), it can be concluded that the data on unemployment and GDP can be taken from the same time period. Therefore, it will be assumed that GDP growth does not have a delayed effect on unemployment (both variables will be taken from the same year).
3.2.4. Emigration
The level of emigrants will be included as an additional explanatory variable in some of the regressions. The data for emigration is taken from the same data set as immigration. That is, it is one of the variables in the data set on long-term international migration from The Migration Statistics Quarterly Report from the ONS. However, because of the reason mentioned previously in the immigration section, the number of emigrants will be divided by the total number of economically active population as well, to get a percentage variable.
4. Results
Before looking at the results of the regressions it is useful to check the correlations between the change in the unemployment rate and the rate of immigration as a percentage of the economically active people for the periods researched. The correlations are given in Table 1.
Period 1972-2015 1972-2003 (A) 2004-2015 (B)
Correlation -0,1576 -0,2818 -0,1034
From Table 1 it can be seen that the correlation is negative for the full period as well as for period A and period B. In addition to this, the correlation is the most negative for period A, whereas it is the least negative for period B. Therefore, immigration is likely to have a negative coefficient in the regressions, meaning it would decrease the unemployment rate rather than increase the unemployment rate, and this coefficient is likely to be more negative for the period A than for the period B.
19
Four OLS regressions were performed in order to find the effect of immigration on unemployment in the UK. Two of them included only the rate of immigration and real GDP growth as explanatory variables. The other two included the rate of emigration as well. The first and the second regressions were for estimating the coefficient for the full period (1972-2015). The third and the fourth regressions were for estimating the coefficients for period A and period B, that is, the coefficient for the period of 1972-2003 and the coefficient for the period of 2004-2015 respectively. The results are given in Table 2.
Variable Full period Split periods
Immigration (1972-2015) -0,3901 -0,8316 - - [-1,82] [-2,44]* - - Immigration (1972-2003) - - -0,5178 -1,1121 - - [-1,46] [-2,31]* Immigration (2004-2015) - - -0,4218 -0,9371 - - [-1,85] [-2,57]* Emigration - 1,5866 - 1,7582 - [1,65] - [1,78] Real GDP growth -0,2777 -0,2474 -0,2727 -0,2350 [-5,88]** [-4,97]** [-5,57]** [-4,50]** Constant 0,0113 0,0010 0,0122 0,0016 [3,68] [0,14] [3,26] [0,23] R-squared 0,4708 0,5045 0,4734 0,5130 N 44 44 44 44
(The first number is the coefficient, the second number in the square brackets is the t-value, * indicates the coefficient being significant at 5% significance level, ** indicates the coefficient being significant at 0.1% significance level. The coefficients and t-values of the main explanatory variables are in bold.)
From the results, it can be seen that immigration does not have a significant positive effect on the unemployment rate (it does not increase the unemployment rate) in any of the regressions. In fact, all the coefficients of immigration are negative, meaning immigration is predicted to have a negative effect on unemployment.
The results for the full period show that the coefficient of immigration is only significant in the regression that includes emigration as well and, as mentioned previously, that coefficient is negative. The coefficient of emigration is not significant and the coefficient for real GDP growth is significant and negative in both regressions. This fits the theory of the negative relationship between GDP growth and unemployment.
The results for the shorter periods show that the coefficients of immigration for the regression without emigration are negative but insignificant for both period A and period B. The regression including emigration shows that immigration has a significantly negative effect on unemployment for both periods. Again, the coefficient of immigration is insignificant and the coefficients of the real GDP growth are significantly negative in both cases.
20
From the R-squares it can be claimed, than for both full and split-up periods, the regressions which include emigration fit the data better. Therefore, the results of the regressions including emigration are likely to be more accurate than the results of regressions without emigration. Furthermore, the regressions with split coefficients fit the data better than the regressions with one coefficient for the full period.
When comparing the coefficients of immigration of period A and period B, it is noticeable that the coefficients of period B (2003-2015) are less negative than the coefficients of period A (1972-2015). Therefore, the effect of immigration on unemployment may have become less negative in recent years. In order to see whether the difference between coefficients is statistically significant, an F test is performed. However, due to the insignificance of the coefficients of the regression with emigration, an F test will only be performed for the coefficients of the regression including emigration. The test gave an F-value of 0.68 and a p-value of 0.4131, meaning that the coefficients for period A and period B are not significantly different from each other. Therefore, the effect of immigration can be concluded to have remained constant over time.
The overall effect of immigration on the unemployment rate in the UK can be concluded to be negative in all the regressions including emigration and to be not significantly different from zero in all the regressions not including emigration. In other words, an increasing rate of immigrants as a percentage of economically active people is predicted to either decrease or have no effect on the unemployment rate, depending on which variables are included in the regressions. However, as the regressions with emigration fit the data better, the negative effect is likely to be more accurate. This effect can be explained by the theories discussed previously and it allows for the drawing of a few conclusions about the population of the immigrants in the UK. Firstly, most of the immigrants are likely to be complements to the natives, as they increase the employment prospects in the country. This means, that the skill and education levels of the immigrants are different to the ones of the natives and that is beneficial to the economy. Secondly, from the positive selection theory, the larger part of the immigrants can be concluded to be high skilled and originate from the countries that have lower inequality than the host country. Finally, as the coefficients of immigration have not changed significantly over time, the composition of the immigrant skill and education levels can be concluded to have stayed similar as well. Therefore, even though the number of immigrants from the EU has increased significantly over the recent years, the majority of immigrants are likely to be high-skilled and have a positive effect on the employment in the UK.
5. Conclusion
This paper analyses the effects of immigration on the unemployment in the UK. In the theoretical part of the paper, it is found that the composition of immigrant population have changed significantly over the past couple of decades. The immigration from EU countries has increased by a significant amount, thus, changing the nationalities of immigrants. The reason behind this is a changed nature of immigration: instead of the UK setting its own rules, the rules of the EU currently apply to immigrants from the European Union. The free movement of the EU immigrants is likely to have changed the skill and education levels of the immigrant population as well.
21
The previous research done in this field concluded that there was no negative effect of immigration on the unemployment rate and this was the expectation of the study. However, as the population of immigrants has changed over the recent years and more recent data was used in this research, this paper could have reached different conclusions.
The results were as expected, as it was found that immigration had no positive effect on unemployment in the UK in any of the regressions. In fact, it was found that immigration had a negative effect on the unemployment rate in the UK in some of the regressions and that this effect has not changed significantly over time. According to the theories analysed, this effect is negative because immigrant population is likely to be comprised of more complements than substitutes for the natives and those immigrants are likely to be high skilled. Furthermore, as the effect has remained around the same throughout the research period, it can be concluded that the skill and education levels of the immigrants has not experienced any drastic changes.
However, this study has some limitations as well. The immigrants were treated as one entity in this research. Therefore, the total impact on the economy was researched. A more accurate way to study the impact would have been to study the effect of different educational and skill groups of immigrants on the corresponding groups of natives. That could be a relevant topic for the further research in the area. One more limitation could be possibly omitted variables in regressions. In that case the estimates of the coefficients would be biased. In addition to that, exogeneity might be an issue. That is, the changing rate of unemployment may cause immigration. Further research is needed in order to deal with this and other potential issues mentioned above.
22 6. Reference list:
Asif, K. (2013). Factors Effecting Unemployment: A Cross Country Analysis. International Journal of
Academic Research in Business and Social Sciences,3(1), 219.
Borjas, G. J. (1989). Economic theory and international migration. International migration review, 457-485.
Borjas, G. J. (1994). The economics of immigration. Journal of economic literature, 32(4), 1667-1717. Borjas, G. J. (1999). The economic analysis of immigration. Handbook of labor economics, 3, 1697-1760. Boubtane, E., Coulibaly, D., & Rault, C. (2013). Immigration, unemployment and GDP in the host country:
Bootstrap panel Granger causality analysis on OECD countries. Economic Modelling, 33, 261-269. Chassamboulli, A., & Palivos, T. (2013). The impact of immigration on the employment and wages of
native workers. Journal of Macroeconomics, 38, 19-34.
Dustmann, C., Fabbri, F., & Preston, I. (2005). The Impact of Immigration on the British Labour Market*. The Economic Journal, 115(507), F324-F341.
Jean, S., & Jimenez, M. (2011). The unemployment impact of immigration in OECD countries. European
Journal of Political Economy, 27(2), 241-256.
Latif, E. (2015). The relationship between immigration and unemployment: Panel data evidence from Canada. Economic Modelling, 50, 162-167.
Lee, J. (2000). The robustness of Okun's law: Evidence from OECD countries. Journal of
macroeconomics,22(2), 331-356.
Moosa, I. A. (1997). A cross-country comparison of Okun's coefficient. Journal of comparative economics,24(3), 335-356.
Moreno-Galbis, E., & Tritah, A. (2015). The effects of immigration in frictional labor markets: Theory and empirical evidence from EU countries. European Economic Review.
Prachowny, M. F. (1993). Okun's law: Theoretical foundations and revised estimates. The review of
Economics and Statistics, 331-336.
Roy, A. D. (1951). Some thoughts on the distribution of earnings. Oxford economic papers, 3(2), 135-146. Sögner, L. (2001). Okun's Law Does the Austrian unemployment–GDP relationship exhibit structural
breaks?. Empirical Economics,26(3), 553-564.
Stock, J. H., & Watson, M. W. (2007). Introduction to econometrics.
United Nations Department of Economic and Social Affairs: Statistics Division (1998). Definition of “international migrant” for the purpose of measuring flows”. Recommendations on Statistics of
International Migration (p.10). Retrieved from
