• No results found

The impact of immigration influx on the unemployment rate in the Netherlands

N/A
N/A
Protected

Academic year: 2021

Share "The impact of immigration influx on the unemployment rate in the Netherlands"

Copied!
28
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

The impact of immigration influx on

the unemployment rate

in the Netherlands

Bachelor Thesis

June 2015

Faculty: Faculty of Economics and Business

Student Name: Asti Ramdhani

Student Number: 10621598

Specialization: Economics

Supervisor: dr. Audrey Xianhua Hu

(2)

STATEMENT OF ORIGINALITY

This is to certify that to the best of my knowledge, the content of this thesis is my own work. This thesis has not been submitted for any degree or other purposes. I certify that the intellectual content of this thesis is the product of my own work and that all the assistance received in preparing this thesis and sources have been acknowledged.

(3)

TABLE OF CONTENT

1. INTRODUCTION 4

2. THEORETICAL BACKGROUND

1. Literature Review 6

2. Standard Competitive Model 7

3. Heckscher-Ohlin Model: An Open Economy 9

4. Reverse Causality 10

5. Other factors affecting unemployment rate 11

6. Immigration in the Netherlands 12

7. Labor market performance of the immigrants in the Netherlands 15

8. Unemployment and GDP in the Netherlands 17

3. DATA AND METHODOLOGY

3.1. The Model 19 3.2. The Data 21 4. RESULTS 22 5. LIMITATIONS 25 6. CONCLUSION 26 7. BIBLIOGRAPHY 26

(4)

1. INTRODUCTION

Why do some countries restrict immigration, while others fully support open borders? After the 1970’s oil crisis, European countries experienced some economic turbulence and have therefore stopped recruiting actively for the guest worker program, the program designed to recruit migrant workers as a solution to labor shortages after the World War II. Some EU countries such as Denmark and Finland have chosen restrictive policies with regards to immigration, in particular, for immigrants with family reunification motives. On the other hand, Germany, the most populous country in the European Union has almost 10 million immigrate population in 2013. This represents a total of 20.5% of the population. Naturally, one would predict that a high immigration rate would harm the economy and employment opportunities of the receiving country. However, as Germany’s birth rate is one of the lowest in the world, this large migration influx does not harm its unemployment level. There has also been a growing concern that the high standard of living in Europe would not be maintained if they do not receive as many immigrants as they do now, likewise, they may increase productivity and income of other factors of production, and are also the consumers of domestic goods and services.

Even though immigration has its drawbacks; the most prominent being that a surplus of labor supply would depress the wages in its host country, immigration also has its own advantages. Immigration is needed in countries with demographic problems such as a declining and aging population. As the EU member states compete with other countries in attracting highly skilled migrants, it is important to have a careful cost and benefit analysis before deciding on restricting one country's immigration.

Borjas (2003) mentioned three substantive questions that modern literature on unemployment mainly focuses on: Firstly, how do immigrants perform in the host country’s economy? Secondly, what impact do immigrants have on the employment opportunities of the native workers? Finally, which immigration policy most benefits the host country? These three questions are important in the shaping policies with respect to immigration.

Since the common policies for immigration and asylum in the EU were initiated in the Amsterdam Treaty of 1997, the topic on immigration policy and its economic impacts has gained much attention. Moreover, international immigration has increased significantly in the last few decades as technology and communication has become more advanced. The

(5)

immigration in Europe began to rise drastically in the 1980s. It was not until 50 years ago that European countries became net immigration countries; previously they have been countries of emigration for more than two hundred years. In 2012, there were an estimated 1.7 million immigrants from non-EU countries and some 1.7 million people migrated from one EU member state to another (Eurostat, 2014). Since 2005, all Western Europe countries (EU’s first 15 Member States, Norway and Switzerland) have had a positive net migration rate including the Netherlands. In 2014, almost 3,6 million population in the Netherlands are of a foreign background, corresponding to a large portion of 21,36% (CBS, 2015). Each year it receives more than 100,000 immigrants of whom the majority come from Turkey, Morocco, Netherlands Antilles and Indonesia.

This paper studies the impact of recent migration on the unemployment landscape in the Dutch labor market. It tries to observe the extent to which immigration depress the labor market opportunities of natives, if such effect exists. Empirical observation will be conducted on the relationship between immigration and unemployment rate using national level data on a quarterly-basis from 2003 to 2014. During that period the immigration rate experienced a hike following a global economic slowdown in 2003, and after a short recovery the unemployment increased again as an effect of the 2008 global financial crisis. The linear OLS regression will be used to observe the relationship, by regressing immigration on unemployment rate while controlling for several covariances including GDP, wage rates and job vacancy. This paper contributes to the growing literature and research on the labor market outcomes in Europe and aims to provide understanding of the economic mechanisms and effects brought about by immigration. According to the Netherlands statistical office CBS,(Centraal Bureau voor de Statistiek) the unemployment rate in the Netherlands has been low and quite stable, ranging from 3.5% to 6.5% in 2001 to 2012.

Sections of this thesis following the introduction are structured as follows: Section 2 summarizes existing literature that studies the labor market consequence of immigration. Subsequently, the definition of immigrants, immigration and unemployment history in the Netherlands will be explained. Section 3 describes the data and the model used for empirical analysis. Section 4 elaborates the results of the economic analysis. Finally, the conclusion will be drawn in Section 5 followed by the limitations of the study in Section 6.

(6)

2. THEORETICAL FRAMEWORK

2.1 LITERATURE REVIEW

I'm the last of five sons. If my father divides his farm between us, no one will have enough land to feed our families. If he gives it all to my oldest brother, I get nothing. If I'm drafted into the prince's army, I may not live long enough to worry about farming! There is nothing for me here. I need to move away from all this confusion and find a peaceful place to farm.

--Karl, Prussia, 1840

The debate on whether immigration effects a country in an adverse way has increased significantly in the last decades. Researchers and experts have produced a growing collection of scholarly articles on this matter, especially in countries with high level of immigration such as the US, Australia and Germany. Most of the studies try to observe the negative labor market consequences that are predicted by standard textbook models, but researches supporting the prediction are scarce. In general, existing studies on various countries and time periods do not agree on one conclusion and often find no or insignificant effect, be it positive or negative.

One of the most important and compelling studies on the effect of immigration on labor market opportunities is conducted by David Card (1990) where he exploited a 7% increase in Miami labor force through the 1980 Mariel Boatlift. Yet he found no effects on wages and unemployment rates of the less-skilled non-Cuban workers, although the immigration influx was mostly of low skilled workers. So no harmful effect on employment of natives was observed.

The central finding of Card’s study has been replicated by Pischke and Velling (1997). They analyzed the employment effect of increased immigration in Germany in late 1980s by studying the change in amount of foreigners while controlling for variables such as share of employment in 12 industries, share of highly skilled-workers, share of female workers and population density logarithm. The predicted adverse effect on employment outcomes of natives, considering the rigid wage structure in Germany and how labor unions play an important role in wage setting was not found on the study. The period studied coincides with an economic boom, during which unemployment rates were naturally falling and the economy absorbs immigration easily. Studies during different economic condition suggest

(7)

otherwise; Winkelman & Zimmermann (1993) studied the 1982 recession and concluded a detrimental effect on wages and unemployment especially in foreign intensive industries. The study by Borjas (1994) also contributes to the predicted adverse effect on wages of natives. Findings from other researchers conclude different results. A study by Docquier, et al. (2011) carried out on the OECD countries during the 1990s surprisingly concluded a positive effect on average wages of less-educated native workers in all countries, to a higher extent on countries that strictly selects highly educated workers. The effect on employment level was also positive. On the other hand, the effect of emigration on less educated wages is negative, to a higher extent on countries with more educated emigrants. These impacts are driven by the fact that migration flows in the 1990s consist of workers that are more “college-intensive” than the domestic labor force as they are much more mobile. Thus, they concluded that the skill composition of migrants holds an important role in determining the labor market effects. Building upon the conclusions of existing studies, which are country and time dependent, we cannot use these to predict the impact of immigration on employment opportunities in the Netherlands. There has not been many studies established on the effects of immigration in the Netherlands. However, the paper by Galloway and Josefowicz (2008) studied the regional unemployment effect of immigration in the Netherlands from 1996 to 2003. They used the OLS regression by controlling for variables such as occupational shares, educational attainment, and population density, among others. The results show that an increase in the share of foreign workers significantly increases the change in regional unemployment rates, thus a harmful unemployment effect was observed. Moreover, the study emphasizes the importance of controlling for educational attainment because it increases the overall goodness of fit and has a significant impact in unemployment rates.

2.2 Standard competitive Model

In a standard labor market, the labor supply has an upward sloping curve because higher wages generate a higher supply of labor, and the labor demand curve is downward-sloping because firms hire more workers when labor is inexpensive, and will refrain from hiring when labor is costly, as explained by Borjas (2003).

Friedberg and Hunt (1995) stated that in a closed economy immigration influx is predicted to reduce the price factor of production with which they are perfect substitutes, raises the price

(8)

factor of production with which they are complementary, and has an ambiguous effect on those which are imperfect substitutes.

Therefore, an increase in low-skilled immigrants will lower the wages of existing low-skilled workers and will have an ambiguous effect on other factors of production. The effect of lower wages will only take place if the immigrants are willing to work for less than natives, as in the case of illegal immigrants. Subsequently, this fall in low-skilled wages makes it more attractive for producers to use this factor of production so they substitute away from capital and high-skilled workers. The large supply in unskilled-workers suggests that optimal output is now higher, and as firms increase their production the scale effect takes place and they use more of all inputs. On the other hand, if the immigrants are skilled workers, the increased supply of skilled workers will lower skilled wages and raises unskilled wages.

Figure 1.a Figure 1.b Figure 1.c The figures above are taken from the book Macroeconomics by Mankiw, N. G. (2009)

Figure 1.a shows the short-term impact of immigration when they are perfect substitutes with natives. As the two groups compete with each other in the same market, the supply curve shifts out, lowering wages and increasing total employment. Notice that although the total amount of employment increased, the total number of native workers employed decreased from N0to N1. In the long-term, firms expand production and take advantage of cheap labor and the demand curve shifts outwards (Figure (b)). Wages adjust to an initial level and employment increases without harming the employment of natives. Figure (c) shows the short-term impact when immigrants and natives are complements of each other. The two labor groups do not compete on the same labor market, thus making natives more productive, shifting out the demand curve. As a result, wages are higher and more workers are hired, while capital is fixed.

(9)

In practice, it is essential for the immigrants to have skills that are lacking or required in the host country in order to prevent higher unemployment. If the immigrants have the same skills as the native population, this could lead to an exile of native employees. However, immigrants should not be of low ability, those who lack skills that are required by employers cannot adapt easily to the host country and may eventually be a burden due to a higher expenditure in social welfare programs. On the other hand, highly skilled immigrants can contribute to a country’s economic growth (Borjas, 1994).

In his paper, Borjas (2003) observes that an increase in labor supply in a certain skill-group affects the earning and opportunities of that particular group. He groups the data according to educational attainment and work experience, and finds that immigration substantially worsens labor market opportunities. Moreover, immigration influx reduced the wage of average native worker by 3.2 percent.

Furthermore, Jean Baldwin (1982) argues that the fear of economic replacement may come from the uneven distribution of immigrants. Therefore the dispersion of new labor across the country is an important element in formulating immigration liberalization policy, so as to minimize the detrimental economic effects on native workers.

2.3 Heckscher-Ohlin Model: An Open Economy

In a country with a large amount of immigrants, it is natural for workers who have been residing in an area to fear of displacement riske; the risk of being replaced by the new immigrants in the labor market. The view that immigration influx will have a negative effect on employment is based on the assumption of perfect competition in the labor market. In his paper, Jean Baldwin (1982) argued that the degree of substitutability of natives for immigrants is crucial. This is evident in existing studies on disaggregated labor substitutability, one of which concluded that different categories of labor in example according to age or educational attainment are all substitutes for one another, and that capital is a substitute to all categories. The study by Docquier et al. (2011) mentioned earlier contributes to this theory, as they concluded a positive effect on wages and unemployment level in the receiving countries. This positive effect is due to the college educated net migrants are complementary to low-skilled natives in the receiving country.

In an open economy, the Heckscher-Ohlin model is more relevant according to Friedberg & Hunt (1995). The model states that between two exactly identical countries besides their

(10)

resource endowments, trade will be determined by the unique factor endowments that are abundant in each country. Each country will produce and specialize in producing a particular good that is abundant and this will then be traded with the other country’s good. As a consequence of the production specialization, factor price equalization does not obtain but it will allow for a difference in price of labor and determine the geographic pattern of specialization, as stated by Dornbusch et al. (1980).

The resulting difference in wages will attract immigrants to go to the country that offers higher wages and is capital-abundant, as mentioned by Friedberg & Hunt (1995). The effect of this migration on the wage rates will depend upon the size of the influx; a large wave of immigration can deteriorate wage levels and encourage the country to be labor-intensive in its production. So a large migration can create wage differentials between two countries. On the contrary, a small migration inflow will hardly have any effect on wages and increases the production of labor-intensive goods without decreasing wage levels. In this way, more goods will be sold in the world market and factor price equalization will be achieved.

Moreover, Friedberg & Hunt (1995) stated that the effect of immigration also depends on the size of the receiving country. A large enough receiving country would increase the world production of the labor-intensive goods and results in an excess supply, reducing the world price as well as the wages.

The effect of immigration on unemployment takes place after the wage changes. Immigration influx of unskilled workers will reduce the wages in the receiving country, even though the employment level of unskilled workers increases. The resulting wage decrease will cause some native workers to leave the country in search for a higher wage. Therefore the employment rate decreases. On the other hand, the impact on the group of workers whose wage increased is a higher employment level.

2.4 Reverse Causality

As mentioned earlier, existing research does not draw a consistent conclusion on the effect of immigration on unemployment. One possible reason is reverse causality. According to Gelman and Imbes (2013), a reverse causal question generally does not have a well-defined answer, even when all the possible data are made available.

(11)

In this paper, the author studies the effect of immigration on unemployment rates. We can avoid having reverse causality by studying the behavior of the two variables within a period of time when an exogenous immigration wave occurs and the change in unemployment rate is not triggered by attractive unemployment benefits or changes in wage rates. The author noted an increase in unemployment rates in the Netherlands during the cyclical economic slowdown in 2003 and after a slight decrease the unemployment rate increased again following the 2008 global financial crisis, which will be discussed thoroughly in section 2.8.

2.5 Other factors affecting unemployment rate

Many factors besides immigration rates are important in determining the changes in the unemployment rate. To name a few are a country’s GDP, job vacancy and wage rates. As immigrants leave their home country in search of a more developed place both in terms of economic and social welfare, a country’s GDP is an important factor.

GDP: Okun’s Law

The negative correlation between unemployment and real GDP growth can be illustrated by the ‘gap version’ of Okun’s Law. An increase in unemployment rate is associated with a lower real GDP growth compared to its natural level. The law takes the form as follows: (  𝑌 − 𝑌* ) =   −β. (  𝑈 − 𝑈*)

where  𝑌  is real GDP and 𝑌* is the natural level of output. 𝑈 is actual unemployment rate and 𝑈* is the natural level of unemployment. β  is the Okun coefficient, which is predicted to have a value of around 3%. This means that a one-percentage point reduction in unemployment rate would produce approximately 3% more output, as explained by Prachowny (1993). The economic influences that are considered in making geographic move is the assessment of current and future net benefits according to Greenwood (1985) and choosing the location with maximum present discounted value. One of the important variables is the calculation of expected income from employment, which can be directly measured by wages and the available job opportunities in the potential country destination.

(12)

Consumer’s confidence adjusts quickly to changes in economic conditions; in an economic downturn the consumer’s confidence declines and so does aggregate demand. In such situations firms cut production costs by decreasing wages or laying off workers to cope with the decreasing demand. The latter is usually more feasible, because workers are reluctant to accept wage cuts and firms fear that it would reduce their productivity. Wage rigidity or the failure of wages to adjust downwards to the equilibrium level when labor supply equals labor demand is one of the reasons of unemployment. This is what the equilibrium model of the labor market explains as mentioned in Mankiw’s book (2009). The research by Verbeek (2014) found that there is 39% downward nominal wage rigidity in the Netherlands, which is about average compared to other countries.

When the real wage is set above the equilibrium wage there is a surplus of labor supply relative to the labor demanded by firms. In other words more people are willing to take a job at the prevailing wage rates than actually needed. When firms cannot hire all these job seekers job rationing occurs, leaving some portion of the labor force unemployed.

The wage curve explained in David Card’s paper (1995) illustrates more clearly how unemployment and wages are negatively correlated; the curve is negatively sloped as individuals are willing to work more when wages increase. However, in the case of a tight market at which the wages are already high and unemployment is low, it is hard to attract more workers at the prevailing wage so firms have to increase their wage offers. But in a labor market with high unemployment rate, the wage of individuals would be lower.

The Beveridge Curve: Job vacancies and unemployment

The Beveridge curve illustrates a negative relationship between unemployment rate and job vacancies. The curve shows how aggregate demand fluctuations generate movements in the total number of job vacancies. The logic behind the relation is that an economy suffering from a contraction has fewer vacancies therefore a portion of the labor force are left unemployed, while during economic expansions the higher demand creates more vacancies, therefore the unemployment rate is kept low.

2.6 Immigration in the Netherlands

There must be some strong reasons for people to relocate to another country, described as ‘push and pull’ factors by sociologists. Some leave their home countries due to the ‘push

(13)

factors’ that are usually negative such as lack of job opportunities, population growth and political turmoil, while others base their immigration decisions as they are attracted to its ‘pull factors’ like freedom and economic opportunity. Generally, people are drawn to countries that are more developed, industrialized and have higher standard of living. Moreover, countries with growing economy are attractive, as they demand more labor to fulfill the increasing demand.

The Netherlands receives immigration of more than 150,000 people yearly with a total population of 16,8 million in 2014. According to the Netherlands statistics (CBS) an immigrant, termed as allochtoon (plural allochtonen), is a person with at least one parent born in a foreign country. The immigrant is a first generation if born in a foreign country, and is a second generation if born in the Netherlands, while the third generation from foreign-born parents is not considered as immigrants. The first and second-generation immigrants are roughly equally distributed. In this paper the word immigrant includes both first and second-generation immigrants.

The immigrants according to the CBS are categorized into immigrants coming from ‘Western countries’ – Europe (excluding Turkey), North America, Oceania (including Australia and New Zealand), Japan and Indonesia, and immigrants coming from ‘non-Western countries’. Japan is included in the socio-economic grounds, while Indonesia is categorized into the ‘Western countries’ because many Dutch descents were born there during the colonial times.

Figure 2.a Immigration in the Netherlands Figure 2.b Emigration in the Netherlands Both Figure 2.a and 2.b are taken from a CBS publication (2003)

Since 1865 the figures on international migration have become available. The two figures above show immigration and emigration flows from 1860s to 2000 taken from a CBS publication (2003). It shows steady increase in both migration flows, indicating zero net

(14)

immigration until around 1910, approaching World War I. The volatile figures during postwar periods are caused by the World War I refugees and deportations from World War II. In 1945 the significant immigration inflow is due to Indonesian decolonization where 350,000 Dutch descents repatriated and 13,000 Moluccan came. The influx and economic uncertainty caused some 200,000 Dutch populations to emigrate to Australia and Canada. Some countries employ supply-driven system such as in the US, Canada, or Australia. In the Netherlands, the immigration policy is demand-driven, where residency and employment request is granted only if suitable vacancies exists; employers who searching for employees can expand their choices to job seekers from outside European Economic Area. When the Dutch economy was booming in the 1960s, firms hired low-skilled workers from Italy, Spain, Greece, Turkey and Morocco, therefore the upward trend in migration.

Even though the initial intention of the guest workers program was for temporary recruitment, after it had stopped in 1973 most of the Turkish and Moroccans stayed permanently in the Netherlands. As immigrants who have settled and found better economic opportunities usually brought their family with them, family migration grew and in 1990s the ethnic group accounts for half a million population. Moreover, immigrants from EU and Western countries are counterbalanced by emigration to those countries, but it is not the case for non-Western immigrants. Therefore net migration comes from non-Western countries in particular Turkey, Morocco and Suriname, together representing 44.6%. Family reunification and formation are the main reasons for migration into the Netherlands, accounting for 41%. Other motives include asylum (26.3%), labor (19.5%), school (7.8%) and others.

Figure 3 Immigration by country of origin in 2010. (Source: CBS, 2015)

Turkey has the largest share of 29%, followed by Morocco by 18,8%, Suriname by 18,4%, the former Netherlands Antilles and Aruba by 7,4% which includes Aruba, Caribbean Netherlands, Curaçao, St. Marten and the old Netherlands Antilles, China by 2,9%, Iraq by 2,8% and other countries together accounting for 29.

21%   19%   18%   7%   3%   3%   29%  

Immigration in the Netherlands by country of origin

Turkey   Morocco   Suriname  

(former)  Netherlands   AnElles  and  Aruba   China  

Iraq   Others  

(15)

2.7 Labor market performance of the immigrants in the Netherlands

How do immigrants perform in the host country’s economy? To answer this question it is necessary to study both educational attainment as well as employment rate of immigrants. The data available from the CBS is the education disaggregation of the labor force while the data on population disaggregation of educational attainment is not available.

Influx of migration in the Netherlands has been quiet volatile variances between each month of a given year. Immigration reaches its peak in September when the academic calendar starts. It increased again in January followed by a slight decrease and raised again approaching the month of September. Noticing the intensity of the increase in the month of September, putting their children to school is an important consideration for immigrants in deciding when to move with their family to the Netherlands.

Figure 4.a and 4.b. Monthly immigration inflow from January 2003 to April 2015 0   5000   10000   15000   20000   25000  

Immigration in the Netherlands 2003-2008 (Monthly)

0   5000   10000   15000   20000   25000   30000  

(16)

We can observe the education attainment of immigrants to study how well they perform in the labor market. According to Statistics Netherlands, the eight levels of education in the Netherlands are primary education, Vmbo, Mbo1 & Avo onderbouw (pre-secondary and lower secondary education), Havo & Vwo (higher secondary education), Mbo 2 & Mbo 3 (higher secondary vocational education), Mbo 4 (completion of Mbo), Havo & wvo (higher secondary education and university or professional preparation), Hbo (higher professional education) & Bachelor’s degree, master’s & doctorate and unknown.

The table belowpresents the labor force participation by ethnic background shown by the columns and by the highest level of education by rows. The distributions of educational attainment between native Dutch and Western background are roughly similar, but the education of non-Western are much more concentrated in below Havo, as can be seen in Turks (39.08%), Moroccans (61.11%), Surinamese (49.2%) and Arubans (42.71%). Moreover, there are more proportions of Dutch natives with Mbo & Bachelor’s degree, but more Western background proportions have attained Masters and PhD in the Dutch labor force. We can conclude that there are more educated immigrant workers coming from Western countries than those from non-Western countries.

Table 2.1 Labor force by the level of education and ethnic background in 2010 per thousand people

Below Havo

Havo, vwo,

mbo Mbo 4 Havo, vwo

Hbo, wo bachelor WO masters, doctor Total Total 531 (4.8%) 4440 (40.3%) 1896 (17.2%) 1084 (9.8%) 1986 (18%) 1080 (9.8%) 11017 Native Dutch 140 (1.61%) 3618 (41.6%) 1630 (18.7%) 826 (9.5%) 1665 (19%) 822 (9.4%) 8701 Western foreign 66 (6.32%) 394 (37.7%) 129 (12.4%) 116 (11.1%) 179 (17.1%) 160 (15.3%) 1044 Non-western foreign 309 (24.7%) 426 (34%) 138 (11%) 140 (11.2%) 141 (11.3%) 98 (7.83%) 1252 Turks 102 (19.1%) 86 (33%) 28 (10.7%) 28 (10.7%) 17 (6.5%) . (0%) 261 Moroccans 132 (61.1%) 63 (29.2%) 24 (11.1%) 17 (7.9%) 20 (9.3%) . (0%) 216 Surinamese 123 (49.2%) 96 (38.4%) 39 (15.6%) 27 (10.8%) 35 (14%) . (0%) 250 Antilleans 41(42.7%) 40 (41.7%) 15 (15.6%) . (0%) . (0%) . (0%) 96 Other 88 (20.5%) 140 (32.6%) 31 (7.2%) 57 (13.3%) 55 (12.8%) 59 (13.7%) 430 Unknown 15 (71.4%) 2 (9.5%) 1 (4.8%) 2 (9.5%) 1 (4.76%) . (0%) 21

The economic performance of immigrants is positively linked to labor market participation. Figure 5 below shows employment and unemployment rate of the labor force by country of origin in 2003 and 2012.(the data for 2014 was not available) The employment rates show similar trend for the native Dutch and immigrants from Western countries. But because non-Western immigrants have obtained lower levels of education the employment rate is much lower, to a higher extent on the Turks and Moroccan. Even though in 2012 the Turks and

(17)

Moroccans employment rate increased, unemployment rate increased to between 15% and 20%. Moreover, it is evident that unemployment rate of non-Western immigrants were much higher than those of Dutch and Western immigrants.

Not only is the unemployment rate of Turks and Moroccans high, the employment level is also relatively low. Bovenkerk, et al.(1995) mentioned that the reasons for low employment level include discriminatory behavior of employers and language barriers, as more than half have problems with speaking and reading Dutch. Moreover, Rodenburg et al. (2003) argued that the generous social transfer system in the Netherlands may demotivate immigrants to work, but this should not be much of a concern as the non-EU immigrants employment rate in the Netherlands are not as low compared to other European countries. In this paper, the reasons behind the changes in unemployment rate will be analyzed which is closely related to the size of immigrants arriving in the Netherlands.

Figure 5. Employment and unemployment rate of population age 15-64 by origin (source: CBS, 2015)

2.8 Unemployment and GDP in the Netherlands

Since 1995, the economy of the Netherlands has increased steadily with some fluctuations throughout the year. The unemployment rate in the Netherlands has been relatively low compared to other European countries. During the economic slowdown in 2003 to 2004, the unemployment rate increased steadily, reaching a peak of 6.5 percent in February 2005. Unemployment was steadily decreasing as the economy grew, but it started increasing again

0   10   20   30   40   50   60   70   80  

Employment and unemployment rate by origin (2003)

Unemployment  rate  

Net  labour  parEcipaEon  (percentage   employed)   0   10   20   30   40   50   60   70   80  

Employment and unemployment rate by origin (2012)

Unemployment  rate   Net  labour  parEcipaEon   (percentage  employed)  

(18)

as the economy was hit by the 2008 global financial crisis. The falling stock and house prices reduced consumer’s wealth and therefore, the willingness to spend. As the aggregate demand was pushed down, firms decreased production and the unemployment rate increased.

The international orientation in the Dutch financial sector and the large export sector formed speculations that the Netherlands would be vulnerable in effect of the financial crisis. But there was a quick world trade recovery and the flexible Dutch labor market quickly adjusts to the changing economic conditions, they were able to supply the high vacancy rate. Therefore the unemployment rate was decreasing throughout the first second 2008 and reached a low level of 3.2 percent in August. It increased again until early 2010 but remained below 5%. After an economic recovery in 2011, it suffered a recession and the GDP dropped in mid-2013. This increased the unemployment rate, and reached an 8.2% peak in February 2014 and since then it has moderately decreased and has been steady on the range of 6.8 to 7.4 percent.

Figure2.7 below clearly shows the relationship between GDP (green line) and unemployment rate (blue line). The economy was growing between 2003 and 2007, consumer’s confidence was high and therefore demanded more goods and services. As firms produce more output to fulfill the increasing demand, they employed more workers and thus, unemployment fell.

Figure 6 Unemployment rate (left axis) and immigration and real GDP (right axis) (source: CBS, 2015)

As mentioned before, the expected effect of increased immigration is a decline in employment rates of resident workers, as their competitors in the labor market have grown. The effect on each group of low-skilled and high-skilled workers depends largely on how the immigrants perform in the labor market in relative to the currently residing workers. If

0   20000   40000   60000   80000   100000   120000   140000   160000   180000   0.00%   1.00%   2.00%   3.00%   4.00%   5.00%   6.00%   7.00%   8.00%   9.00%  

Jan-­‐03   Sep-­‐05   Jun-­‐08   Mar-­‐11   Dec-­‐13  

G D P an d I mmi gr ati on U n emp loyme n t r ate

Unemployment and GDP in the Netherlands (2003-2014)

Unemployment   Rate  

ImmigraEon   (quarterly)   GDP  (2010)  

(19)

immigrants are substitutes, a detrimental effect is expected by having a downward pressure on wages or depressing the employment opportunities if wages are rigid and labor union are not powerful in the wage setting. But if they own complementary skills, immigration will help the economy and expand production.

3. Data and Methodology

3.1 The Model

As stated earlier, this paper aims to study if there is any negative effect of immigration on unemployment. The model to be carried out will investigate the following hypothesis: “Immigration influx has a negative effect on the labor market opportunities” that is, immigration has a negative effect on employment level in the receiving country. An Ordinary Least Square (OLS) regression analysis will be conducted. The null hypothesis that will be tested is as follows,  H!: β!" = 0, which refers to immigration having zero effect on

unemployment. The null hypothesis will be tested against the alternative hypothesis  H!: β!"≠ 0.

There are some assumptions that need to be made in order to make reliable conclusions from the data. Firstly exogeneity is assumed in the OLS approach and that all regressors are linearly related to the dependent variable. Moreover, these assumptions are made:

1. The conditional distribution of error term 𝑢! given 𝑥!!, 𝑥!!, and 𝑥!! has a mean of zero

2. (𝑥!!, 𝑥!!, 𝑥!!, 𝑦!), 𝑖=1, ..., 48) are identically and independently distributed (i.i.d.)

3. Large outliers are unlikely

The OLS population multiple regression model with 𝑛 regressors has the following form, as mentioned in Stock & Watson (2012):

   Y! =    β!+ β!  𝑥!! +  β!𝑥!!+ … + β!  𝑥!"+ 𝑢!, 𝑖 = 1,2, … , 𝑛.

where Y is the dependent variable, 𝑥! are the independent variables, and 𝑢 is the error term.

The model used in this study adapts the OLS multiple regression equation by controlling for the variables discussed in previous sections. We employ the differenced equation of the OLS

(20)

regression following Pischke and Velling (1997) and we have two models: one is a regression will all the regressors, and another regression using only Immigration and GDP as the regressors, following the research done by Fernandez (2014). The two models are as follows:

  1    ∆𝑈𝑛 = α + β!"  ∗ ∆  IM! +  β!"#  ∗ ∆  GDP! +  β!∗ ∆  W!+  β!∗ ∆  V!+  ε!

  2    ∆𝑈𝑛 = α + β!"  ∗ ∆  IM! +  β!"#  ∗ ∆  GDP! +  u!

where ∆𝑈𝑛 is the dependent variable, which is the change in unemployment rate in the Netherlands. ∆  IM! is the percentage change in immigration. As stated by Pischke and Velling (1997), we can therefore assume that it is the inflow of immigration, not the levels, that yield an unemployment rate above its steady state. The model controls for several variables: GDP for real GDP, W for wages, and V for job vacancies. The term ∆  means a

percentage change in the variable from one quarter to the previous quarter. The error term

is represented by ε and u, while 𝑡 indicates the quarter observed.

We include more independent variables that are correlated with the included regressor (immigration) and also determine the dependent variable in order to refrain from omitted variable bias, as suggested by Stock and Watson (2012). Moreover, as immigrants have certain location preferences we control for the three control variables so as to avoid the possible endogeneity problem mentioned by Pischke and Velling (1997). The potential endogeneity bias we hope to eliminate is based on the fact that Immigrants entering the Netherlands mostly aim to reside in urban areas such as in Amsterdam and Rotterdam, which is expected to create an upward bias in the resulting estimates.

Several regressions will be run on different time frames. The first one will be on the whole period from 2003 to 2014, and in order to see if immigration necessarily increases unemployment, we also observe separately the period when the unemployment increased (1st

quarter 2003 to 4th

quarter 2004), and a drastic decrease in unemployment (1st

quarter 2005 to 2nd

quarter 2008) and again, a significant increase (3rd

quarter 2008 to 4th

quarter 2014).. Therefore, four time frames are studied and on each period the two models will be regressed, resulting in 8 regression outcomes.

(21)

3.2 The Data

The dependent variable is the change in unemployment rates and the independent variable is the percentage change in the number of immigrants arriving into the Netherlands. The control variables are the country’s real GDP, average wages, and number of job vacancies. All data is obtained from the Statistics Netherlands, Centraal Bureau voor de Statistiek (CBS). The data used are quarterly and corresponds to a national level data from 2003 to 2014, with a total of 48 observations. However, there are 2 data missing for the variable wages.

The data on monthly unemployment rate, monthly wages and monthly job vacancies were obtained from the website, while the data on monthly immigration and quarterly real GDP were requested to the Info Service of the CBS. By using the data from the CBS, the data is confirmed to be valid and reliable, and will not lead to any bias.

Dependent variable: Unemployment rate

Unemployment rate is the ratio of unemployed people to the total labor force, which is the number of economically active people. The definitions of an unemployed person in the Netherlands is a person without work who is actively searching for paid work and are directly available. The labor force comprise of the population of 15-65 years old.

The unemployment rate refers to all residents, both Dutch and immigrants. The figures on unemployment rate from the CBS are based on the Dutch Labor Force Survey (LFS), which uses a sample size of 64,000 respondents and uses time series model to make the monthly figures. The unemployment rate on a monthly-basis was first published in 2010 providing data from 2003, therefore the period studied in this paper starts from January 2003. The unemployment rate published by the CBS are measured according to 12-hours threshold monthly labor participation and according to the international definition which comprise of people who are neither paid in employment nor self-employed (ILO).

Independent variable: Immigration

The dependent variable is the percentage change in immigration, measured as the change in immigration population from one quarter to the previous quarter, divided by the immigration population in the base quarter. The unit is in percentage change to associate a 1% change in immigration with the change in unemployment rate, which is already in a percentage form.

(22)

Control variables: wages, GDP and job vacancies.

The variable wage is the monthly gross wage for full-time working termed by the CBS as CAO wages, a Dutch acronym for collective labor agreement of the three sectors: government, subsidized corporations and private companies. The data available are the quarterly average CAO wages relative to those in the base year 2000.

The regressor GDP is quarterly real GDP measuring the value of economic output adjusted to current prices, thus inflation is already taken into account. The data on quarterly real GDP from CBS is calculated using the production approach, which sums up the gross added of various institutional sectors with taxes, minus subsidies on products. The value of GDP is expressed in the 2010 value in millions of euro.

The data on job vacancies from the CBS is quarterly data expressed in thousands of job vacancies available during the quarter, it include jobs that are immediately available in a company or institution and can be filled by internal or external recruitment. The data available are the quarterly average job vacancies relative to those in the base year 2000, and the data used in the model are the percentage change from one quarter to the previous quarter.

4. Results

This paper employs OLS regressions to study the effect on immigration on unemployment. Eight regressions on the unemployment rate were performed; four of which immigration and the three control variables were used as regressors, four others that only used immigration and GDP.

As mentioned in the previous section, the regressions carried out in this research studies the whole period from 2003 to 2014 as well as three different economic periods that took place during the period studies. The following results are presented; Table 2 presents equation (1) and (2) which studied the whole period and equation (3) and (4) is focused on the 2003-2004 economic slowdown. Table 3 presents equation (5) and (6) which are focused on the economic expansion from 2004 to 2008, lastly equation (7) and (8) studied the period from 2008 to 2014, which follows the global financial crisis.

In any set of periods, the first equation is regressed using Model (1) from page 19, and the second equation uses Model (2) with the restricted regressors of only Immigration and GDP.

(23)

Table 2 (1) (2) (3) (4) UN UN UN UN IM -0.807** -0.803*** -0.000424 -1.601* (.232) (-3.89) (-0.00) (-3.24) GDP -7.417** -6.762** 1.566 -10.02 (2.67) (-3.28) (0.13) (-2.4) W -10.02 114.2 (23.78) (0.88) V -0.365 3.296 (0.68) (1.09) Constant 0.200 0.160** -0.244 0.303* (0.12) (2.72) (0.42) (2.97) N 46 48 8 8 R-squared 0.255 0.254 0.769 0.678 Prob > Chi2 0.015 0.0014 0.238 0.0586 (t statistics in parentheses) * p<0.05, ** p<0.01, *** p<0.001

The table above presents the coefficients with its standard deviation in parentheses below each coefficient for the variables IM, GDP, W, V, and the constant term. The number of observation, N, the R-squared and significance of the model, Prob>Chi2

are also presented. Regression (1) and (2) are done on the whole period from 2003 to 2014. The key coefficient of interest is β!", the coefficient of immigration. Both regressions produce negative and

significant coefficient for Immigration and GDP. The significance of equation (1) is quiet low, as the p-value higher than 0.01. When we regress without the control variables Wage and Vacancy in equation (2), the equation has a low p-value of 0.0014. The null hypothesis  H!: β!" = 0 can therefore be rejected in favor of the alternative hypothesis at the 5% level.

(24)

The number of observations in equation (2) is two amounts lower than equation (1) because it does not have the variable Wages, which has two missing variables.

Standard model predicts that increased immigration would have an adverse effect on labor market outcomes, which on the unemployment rate would be an increase in unemployment in the host country. The coefficient on Immigration of -0.8% in equation (1) and (2) can be interpreted as: a 1% increase in immigration population leads to 0.8% lower unemployment rate, which actually improves the labor market. Moreover, the negative predicted effect of GDP is observed. Equation (1) shows a 7.47% decrease in real GDP as the immigration population increased by 1%, and equation (2) predicts the negative effect to a lower extent, of 6.72%. However, this effect is very large and needs further investigation.

In equation (3) and (4) we observe the period during which the Dutch economy underwent a slower GDP growth and unemployment increased as an effect of the economic slowdown in 2003. The results show no significance on any variables, and the model’s significance is also very low. Restricting the regressors only on Immigration and GDP increased the significance of Immigration and the significance of the model at the 10% level.

Table 3 below presents the results of equation (5) and (6), which focus on the economic growth period. Equation (5) generates positive coefficients on both Immigration and GDP, the positive GDP coefficient is contrary to Okun’s Law, which predicts that a growth in GDP would decrease unemployment, however the model has very low level of significance and all variables are insignificant as well. Regressing the model on the Immigration and GDP gives us both negative coefficients but still insignificant. Moreover, the R-square is very low. Equation (7) and (8) studies the period from 3rd

quarter 2008 to the 4th

quarter 2014, which is when the economy was affected by the global financial crisis. All variables in equation (7) were insignificant, and Immigration had a negative sign. The restricted equation on equation (8) produces significant coefficients on all variables. The coefficient on Immigration is negative, like what was found in the previous equations. Moreover, the p-value is quiet low and is significant at the 10% significance level.

(25)

Table 3 (5) (6) (7) (8) UN UN UN UN IM 0.375 -0.399 -0.425 -0.545* (0.47) (-0.60) (-1.50) (-2.34) GDP 10.65 -1.464 -2.729 -5.847* (1.09) (-0.23) (-0.65) (-2.33) W 78.99 50.32 (1.63) (0.90) V 0.903 0.144 (0.47) (0.13) Constant -0.859 -0.153 0.101 0.270** (-1.77) (-0.86) (0.52) (3.73) N 14 14 22 22 R-squared 0.3048 0.083 0.2868 0.2487 Prob> F 0.462 0.622 0.1944 0.0661 (t statistics in parentheses) * p<0.05, ** p<0.01, *** p<0.001 5. Limitations

An important limitation to this study is the limited data availability, especially of the disaggregation of educational attainment on the labor force in the Netherlands on a quarterly basis. The data was available on a yearly basis, and using yearly would generate a very small number of observations. This data was needed to control for share of high-skilled and low-skilled workers as the model by Pischke & Velling (1997) and Galloway & Josefowicz (2008) did in their paper. Other possible control variables such as the share of workers over the age of 55, the share of part-time workers, and the share of female workers were not available on a quarterly or monthly basis.

(26)

6. Conclusion

This paper studies the effect of change in immigration to the change in unemployment rates in the Netherlands. Using quarterly data of average unemployment rate and the amount of immigration population in a given quarter, an OLS regression analysis was conduction while using several variables as control variables. They are the country’s percentage growth in real GDP, the quarterly average wage in all sectors, and the quarterly job vacancies. The results do not confirm the standard prediction of a negative effect on unemployment rate, rather it concludes that an increase in immigration population improves the labor market by decreasing unemployment rates.

7. Bibliography

Baldwin, J. (1982). The Substitutability of Natives and Immigrants in Production. The Review

of Economics and Statistics, 64(4), 596-603.

Borjas, G. (2003). Labor Economics. McGraw Hill Irwin. Sixth Edition.

Borjas, G. (1994). The Economics of Immigration. Journal of Economic Literature, 32(4), 1667-1717.

Borjas, G. (2003). The Labor Demand Curve is Downward Sloping: Reexamining the Impact of Immigration on the Labor Market. NBER Working Paper, 9755.

Card, David. (1990). The Impact of the Mariel Boatlift on the Miami Labor Market.

Industrial and Labor Relations Review, 43(2), 245-257. doi: 10.3386/w3069.

Card, D. (1995). The Wage Curve: A Review Review. Journal of Economic Literature, 33(2), 785-799.

Docquier, F., Özden, C., and Peri, G. (2011). The Labor Market Effects of Immigration and Emigration in OECD Countries. IZA Discussion Paper, 6258.

Dornbusch, R, Fischer S, and Samuelson P. (1980). Heckscher-Ohlin Trade Theory with a Continuum of Goods. The Quarterly Journal of Economics, 95(2), 203-224.

(27)

Eurostat Statistics Explained. 2015. Job vacancy and unemployment rates - Beveridge curve

<http://ec.europa.eu/eurostat/statistics-explained/extensions/EurostatPDFGenerator/getfile.php?file=145.18.110.152_1435478215_3 5.pdf>

Fernandez, P. (2014). The Impact of Immigration on natives’ unemployment: Some evidence for Spain. University of Amsterdam, Amsterdam.

Friedberg, R. M. & Hunt, J. (1995). The Impact of Immigrants on Host Country Wages, Employment and Growth. The Journal of Economic Perspectives, 9(2), 23-44.

Gelman, A. & Imbens, G. (2013). Why ask why? Forward causal inference and reverse causal questions. NBER Working Paper, 19614. doi: 10.3386/w19614.

Greenwood, M. J. (1985). Human Migration: Theory, Models and Empirical Studies. Journal

of Regional Science, Vol.25(4) pp. 521-544. doi: 10.1111/j.1467-9787

Jozefowicz, J.; Galloway, R. M. (2008). The Effects of Immigration on Regional Unemployment Rates in The Netherlands. International Advances of Economic Research. doi: 10.1007/s11294-008-9157-8.

Mankiw, N. G. (2009). Macroeconomics. Worth Publishers; 7th

Edition.

Masselink, M & Noord, P. (2009). The Global Financial Crisis and Its Effect on the Netherlands. European Commission Publication, 6(10).

Pischke, J. & Velling, J. (1997). Wage and employment effects of immigration to Germany: An analysis based on local labor markets. Review of Economics and Statistics, 79(4), 594-604. doi: 10.1162/003465397557178.

Peri, G. & Ottaviano, G.I. P. (2008). Immigration and national Wages: Clarifying the Theory and the Empirics. NBER Working Paper, 14188. doi: 10.3386/w14188.

Prachowny, M. F. J. (1993). Okun's Law: Theoretical Foundations and Revised Estimates.

(28)

Stock, J & Watson, M. (2012). Introduction to Econometrics. Pearson Education Limited. 3rd

Edition.

Roodenburg, H., Euwals, R., and ter Rele, H. (2003). Immigration and the Dutch economy.

CPB Netherlands Bureau for Economic Policy and Analysis, 1–121.

Tregub, I. V.& Kabanova, E. (2012). Okun's Law Testing Using Modern Statistical Data.

Forum for Research in International Trade Working Paper, 488.

Verbeek, W. P. (2014). Downward Nominal and Real Wage Rigidity in the Netherlands.

Erasmus University Rotterdam, Rotterdam.

Winkelmann, R & Klauz. F. Z. (1993). Ageing, Migration and Labour Mobility. Cambridge

University Press, 255-283.

Referenties

GERELATEERDE DOCUMENTEN

De onderzochte musea, met name het Rijksmuseum, het Stedelijk Museum en het Mauritshuis, hebben zich ten tijde van de verbouwing ontwikkeld van collectiegerichte instellingen

The European Commission has supported the cross- industry social partners through various mechanisms, while the European Social Dialogue succeeded in negotiating

- Customizability: The degree to which repositories themselves support customization of components to provide a tailored learning experience (cf. Kellogg, et al. ) or to support

consistency in brand advertisements and customer-based brand equity, brand attitude and brand image and whether this relationship is moderated by brand familiarity, a 3 (level

Finally our hypothesis was that an interaction between type of composite and duration of target presentation would be found: based on Frowd et al., (2005a) and his suggestion

Internal selection of ideas by marketing department &amp; MT C Mode of co-creation Elite circle Elite circle Elite circle Innovation mall External participants

Daane, Beirne and Lathan (2000:253) found that teachers who had been involved in inclusive education for two years did not support the view that learners

Meetplan voor de monstercampagne in week 31-32 2014 voor de Noordzee kust van Ameland en Schiermonnikoog ten behoeve van het meerjarige onderzoek naar de effecten van kustsuppleties