The Economics of Integration

(1)

The Economics of Integration

Does social capital play a role?

JEL Classification: Z13, J2, J31, J6, J7

Master thesis for Economics

Abstract:

Becker (1957) stated that there is no economic reason to discriminate. He pointed at undiscovered, immeasurable factors that make, for instance, ethnic minorities less attractive to hire. This thesis hereby tries to find the factors that determine income. I hereby included social capital as the undiscovered factor. We measured social capital on different aspects namely size of

the network, bridging and bonding social capital and the acceptance of western values. From the results we found that social capital indeed matters for the achieved economic prosperity of immigrants in the Netherlands. Hereby this research did not only discover

immeasurable factors but also confirmed the importance of social capital.

(2)

The Economics of Integration

Does social capital play a role?

PART I: Introduction

-1. Introduction

Lately the topic of the integration of minority groups in the Dutch society has become the topic of a debate, involving both the public and the media. The idea that especially Muslims do not integrate enough in Dutch society leads to increasing discrimination against them. Many native Dutch citizens see immigrants as lazy, and believe them to be profiting of the Dutch Social Security System. Fear for ghettos of immigrants in the big cities is increasing.

Before 9-11, the aim of policy makers was to create a multicultural society. In such a multicultural society, all cultures are allowed to exist next to each other and participants of the society are tolerant towards each other. Nevertheless, after the increasing threat of terrorism and the growing resistance against foreigners1_{the public climate seems changed forever. All political}

parties now state opinions regarding minorities that previously were never expressed because of fear of discrimination. In addition, the answer of policy makers to this public debate is often “integration policy”. Minorities should integrate into the Dutch society and behave like the Dutch.

This problem also can be seen in many other European countries. They are faced with the problem of how to treat immigrants, and what to expect from them. In all these international discussions, there is no consideration to the economic aspects. Therefore an interesting question is: what is, in economic terms, the best for the immigrants themselves? Is it beneficial for an immigrant to integrate? Perhaps contacts with the native Dutch majority can help immigrants to obtain better jobs. They are more accepted by native Dutch and more familiar with Dutch ways of doing business.

Immigrants can also add much value to the society because of their special characteristics. We see that many ethnic minorities establish their own businesses. By making use of their cultural backgrounds, they are able to exploit a niche in the market. And in today’s increasing globalizing world immigrants can even become of increasing value.

There exists an extensive amount of research on the reasons for the lower economic performance of minorities. While these ethnic minorities are characterized by substantially lower levels of human capital, there still exists a large unexplainable disadvantage of minorities. This

(6)

unexplainable part is often attributed to discrimination, but strong evidence for this does not exist. Gary S. Becker has become famous for his work on discrimination economics, in which he stated that discrimination does not exist. This paper will adopt his statements, and investigate whether the concept of social capital plays a ‘missing link’ role.

This paper has two main goals. First of all, I will try to find factors determining the income of minorities. I will hypothesize that these factors include integration and social capital, and test their effect on income. I will do this in order to investigate whether social capital can (partly) explain the large income differences existing between the native Dutch and ethnic minorities. In this sense, it could play a ‘missing link’ role. My second goal will be to provide further evidence for the social capital theory. Because the social capital theory is broadly debated, more support for the theory and the effects of social capital would be valuable.

I will focus on differences between the immigrants instead of comparing native Dutch with ethnic minorities. This implies that I compare the successful immigrants with the less successful ones. I use this method because it is unfeasible to compare integration of minorities to integration of native Dutch. When measuring integration, the characteristics of native Dutch will be used as benchmark. Since I measure integration by the number of contacts with native Dutch and the acceptance of Dutch values, these are typically aspects only relevant for immigrants. Furthermore, since I want to investigate what determines income of minorities, I should look at the position of the minorities themselves, and find the reasons why some immigrants perform better than others.

The structure of the paper is as follows: First, I will give an extensively review of the economic situation of the minorities and describe their characteristics. Here I also will give the reasons for the lower economic performance of these minorities. By focusing on the social situation of the minorities, I want to prove that there are reasons to believe that social capital may cause this lower performance. The conclusions drawn in this first part are used for sketching the assumptions regarding the determinants of income.

(7)

In the third part of this thesis, from section six onwards, I will start to investigate the situation of the minorities, by sketching a model incorporating the effect of social capital. Furthermore this section will introduce the other factors influencing income and describe my approach towards social capital. Based on these conclusions I will define hypotheses that will be tested in the following parts of this thesis. Section seven will explain how I measure the diverse components of the model. Special attention is given to this, since, as can be read from the fourth section, measurement of social capital is very problematic and procedures are contentious.

From part four onwards I will present the empirical evidence. Part four mainly focuses on the direct determinants of income. In section eight I present the results of a regression involving the diverse determinants of income. Section nine will apply the same regression to some of the different subgroups. I do this in order to increase knowledge on how social capital works. Section ten investigates to what extent religion and race matter for income. Finally, the most important conclusions from part three are summarized in section eleven.

I will further investigate the effect of social capital on the labour market in part five. An often stated hypothesis is that higher proficiency in Dutch helps immigrants to succeed in their educational career. Furthermore, it is frequently stated that social capital aids in the creation of human capital.2_{Since human capital is one of the most important determinants of income, I will}

investigate whether these relations hold in section twelve. Section thirteen explores if empirical evidence can be found for some, as deducted from the mathematical model of section 6, expected effects of social capital. In section fourteen I investigate whether the determinants of income, including social capital, also influence the achieved position of jobs of immigrants.

In part six, section fifteen gives the final conclusions of this study. This section will present the factors determining income of immigrants and answer the question whether social capital plays a role here. Also, I will provide some useful policy suggestions to decrease the income gap between ethnic minorities and native Dutch. Finally, I will elaborate on the implications of my results for the general discussion regarding the social capital theory.

(8)

2. The situation of minorities in the Netherlands

Most research regarding non-western minorities in the Netherlands focuses on the four largest groups of minorities: Turks, Moroccans, Antilleans and Surinamers. Since the position of minorities is often discussed in the Netherlands, there is an extensive amount of statistics available on this topic. From these it follows that minorities are faced with a lower socio-economic position in the society, compared to both native Dutch and western immigrants. However, there are large differences between the different ethnic groups. This section will describe the situation of the four largest groups of non-western minorities.

Figure 1 Standardized3_{, average family income of native Dutch and the four largest, non-western immigrant groups.}

Source: CBS Statline.

Starting with the income differences, figure 1 shows the standardized disposable income of the native Dutch families and the four largest groups of immigrants. It follows that the income of immigrants is 25 to 35 percent lower than the income of the native Dutch. There are also large differences in income between the four immigrant groups. Surinamers and Antilleans achieve higher income than Turks and Moroccans. From table 1 we see that there have been hardly any improvements during these recent years. On the contrary, the income gap between native Dutch and the ethnic groups even increased.4

3_{Standardized average income here implies that the CBS adjusted the income figures of families to hold for a}

standard family (two parents, two children).

4_{From figure one we see that the Antillean family income follows another pattern than that of the other ethnic}

minorities. In contrast to the other groups, income rose for the Antilleans after 2002. Later on, I treat the differences between these ethnic groups in more detail.

0 5 10 15 20 25

Dutch Surinamers Antilleans Turks Moroccans

(9)

Dutch Surinamers Antilleans Turks Moroccans 2000 100.00 100.00 100.00 100.00 100.00 2001 108.94 109.86 104.76 109.60 107.56 2002 112.85 112.68 109.52 113.60 113.45 2003 113.41 111.27 111.11 110.40 112.61 2004 114.53 111.27 114.29 108.80 110.92

Table 1: Index figures of the average income of the four minorities and of the native Dutch with 2000 as base year.

Source: CSB Statline.

Looking at the average wages in table 2, there also seems to be a large gap between the ethnic minorities and the native Dutch. However, the differences between the Dutch native and the ethnic minorities are decreasing over these years. This is quite remarkable, since differences in total income, as can be seen in table 1, increased.

Dutch Surinamese Antilleans Turks Moroccans

1999 100.00 100.00 100.00 100.00 100.00

2000 104.11 105.59 104.22 106.08 106.62

2001 111.87 115.64 113.86 116.22 118.38

2002 115.07 120.11 118.67 120.95 122.06

2003 118.26 123.46 122.89 125.00 127.94

Table 2: Index figure of the average wages of the four minorities and the native Dutch with 1999 as base year.

The most plausible explanation is that unemployment increased dramatically among the ethnic minorities. In recent years, unemployment increased much for most of the ethnic groups. Table 3 shows the developments in unemployment. When there are more unemployed, this can decrease the average income of ethnic minorities even when their wages increase. A closer look at unemployment rates confirms our conclusions: minorities have a large disadvantaged position and are much more likely to be unemployed.

There are several reasons for the lower performances of the immigrants. The following section will look at possible explanations. Furthermore, I will examine research aimed at discovering reasons for the disadvantaged position of minorities.

Dutch Surinamers Antilleans Turks Moroccans

1999 2.4 % 6.7 % 9.3 % 13.8 % 11.1 % 2000 2 % 5.4 % 7.4 % 10.8 % 10.3 % 2001 1.7 % 3.3 % 4.5 % 7.7 % 5 % 2002 1.9 % 3.9 % 5.2 % 7.4 % 5.3 % 2003 2.9 % 6.1 % 8.2 % 10 % 7 % 2004 3.7 % 6.5 % 12.6 % 10.7 % 8.9 %

(10)

3. Reasons for the disadvantaged position of ethnic minorities on the labour market

Many different explanations exist for the lacking position of immigrant groups in the labour market. Of course, human capital theory plays a major role here. However, as will follow from this section, different levels of human capital cannot explain all of the differences between native born and immigrants. This section, therefore, will give some other factors that play a role in explaining the lower economic performance of minorities. Besides some intuitions and theories, it will also present past research on this topic. I divided the arguments into two categories: demand side characteristics and supply side characteristics. I am looking at the labour market which implies that characteristics of the employers determine the demand side, and the characteristics of the employees constitute the supply side of the labour market.

I will also look at some social capital indicators for the ethnic groups to investigate whether social capital can play a role by determining economic achievements.

3.1 Demand side characteristics and discrimination

The labour demand side includes factors leading to unwillingness of employers to hire an immigrant worker. Some influential theories will be treated here.

An often stated explanation is that minority labour is ‘at the end of an imaginary hiring queue’. They are the last ones that will be appointed a job (Thurow 1976). Empirical evidence can for instance be derived from the fact that in periods of economic growth, unemployment of minorities decreases more than proportional. The opposite is observed in periods characterized by lower economic growth (see for instance Veenman 1998, SCP 2003 and SCP 2005a).

Another explanation is that minorities are constantly ‘in the wrong place on the labour market’. They are attached to certain industries in which demand for labour is decreasing (Veenman 1998). This is especially due to the fact that immigrants are focused on the lower end of the labour market, but these jobs are increasingly relocated to low-wage countries.

These theories have in common that they do not give any explanation why the immigrants will ‘be last in queue’ and ‘in the wrong markets’. It is not unimaginable that discrimination makes minorities the least attractive workforce and provides them no opportunities to enter other markets than the ones in which they are currently working. In section five I will outline the different theories of discrimination in more detail.

(11)

hard to find. This fact combined with the economic literature which states that discrimination does not exist, leads to doubts regarding the influence of discrimination5. A method often used to estimate discrimination effects is making a regression involving many determinants of wages. The unexplainable part of a regression that tries to explain the differences in wages of native Dutch and minorities is assigned to discrimination (van Ours et al. 2002, Schwieren 2003). Similar methods are used to explain differences in unemployment between native and ethnic groups. However, it is doubtful whether these differences are exclusively due to discrimination. First of all, regressions from different authors hardly show similar results (Schwieren 2003). Secondly other factors can be involved, like contacts that provide information on better paying jobs6_.

Another approach often used is ‘correspondence testing’, where equally qualified pairs of candidates, varying only in race, apply to a vacancy. This kind of research often concludes that there indeed is discrimination (Schwieren 2003). In about 20 percent of the cases, the native Dutch received a job at the expense of the ethnic minority applicant. The authors stated that, since the candidates were equal, the only reason for choosing the native Dutch applicant should be discrimination (Veenman 2003). From other publications (as for instance Onderwijsinspectie 2006) we also see that ethnic minorities were faced with discrimination when trying to obtaining a trainee post.

The major disadvantage of this method is that it is very difficult to find two persons equal in everything except ethnic origin. It is imaginable that the researcher believes that the two candidates are identical whereas an employer does see some differences. For instance, an employer might think that an immigrant worker is not representative enough, or less able to maintain important contacts for the firm. Thereby Veenman (2003) stated that the required equality of the ethnic and native candidates in these kinds of researches could be disputed.

From surveys it follows that about 40 percent of the Turks, Moroccans, Surinamese and Antilleans stated that they were personally faced with discrimination (SCP 2005a). Furthermore, data from different Dutch institutions that register discrimination complaints show that 27% of all complaints are labour market related. Also, more complaints of this nature were registered from 2002 onwards (LBR 2004). It is remarkable that reported discrimination is much larger in sectors where unemployment is higher (Veenman 2003). The author states that this can be due to frustration, which leads immigrants to report discrimination more easily. This argument makes it even more doubtful whether registration by victims is a good indicator of discrimination.

(12)

From these researches, we can only conclude that there exist some reasons to believe that there could be discrimination at the labour market. However, clear evidence for this statement does not exist.

3.2 Supply side characteristics

The characteristics of immigrants include several factors that could explain their lower economic performance. An especially important factor is the quality of the workforce, as measured by education and proficiency. This section will therefore look at the proficiency of Dutch and the level of human capital. Other factors I will take into account are the amount of time the immigrant spent in the Netherlands, and the participation of women on the labour market.

3.2.1 Education of ethnic minorities

It is common knowledge that ethnic minorities are not performing well in the Dutch educational system, and therefore most of the ethnic minorities end their educational career with lower levels of education. We see that the dropout rate of ethnic minorities is much higher7_{, there}

are hardly ethnic minorities receiving higher education, immigrants have lower graduating rates and spent more time to finish a particular type of education (SCP 2005a). Finally, the number of immigrants without an educational certificate is much higher, compared to native Dutch. 8

Figure 2 illustrates the lower performance of ethnic minorities in education by showing the distribution of the levels of achieved secondary education. As can be seen in the figure, the native Dutch get much higher levels of education, while the Turks and Moroccans have the lowest level of education. A reason for the slightly higher performance of Surinamese and Antilleans could be their advantage in language, which will be treated in Section 3.2.2.

This evidence confirms the often stated impression that ethnic minorities score much lower on education and are less equipped with human capital. The large differences in income can therefore be partly attributed to human capital. From research it follows that the reason for high unemployment under minorities is their lower level of education. However, research gives us different conclusions regarding the influence of human capital. The estimations vary widely, where education can determine from about 30 to 80 percent of the differences in income

7_{For non-western, ethnic minorities the drop-out rate in secondary education is 5-6 percent, compared to 1-2}

percent for the native Dutch. Source: CBS Statline.

8_{The percentage of non-western, ethnic minorities without any “starting qualification” in 2005 was 47 percent,}

(13)

inequality (Ours and Veenman 1999, Ours et al. 2002). These studies conclude that the education followed in the Netherlands is especially of importance, while education attained in the country of origin has hardly any influence at all (Ours and Veenman 1999, Ours et al. 2002).

Figure 2: Distribution of secondary education under the different types of education for the ethnic groups and the native Dutch. Numbers given are based on the third grade of secondary education in 2004 / 2005.

3.2.2 Proficiency in Dutch of immigrants

We saw that Antilleans and Surinamers perform better during their educational career, compared to the other ethnic groups. This higher performance could be explained by the history of these countries and their resulting advantages in the use of the Dutch language. From research (see table 4) it follows that the proficiency in Dutch of Surinamers and Antilleans is much higher than that of Moroccans, while Turks have even more problems speaking the Dutch language.

Surinamers Antilleans Turks Moroccans No problems with speaking Dutch 85 72 28 47

No problems with reading Dutch 50 49 8 12

Table 4: Proficiency in the Dutch language among ethnic groups, in percentages. Source: SCP 2003 (table V.1 page 401).

From SCP (2005) it also becomes clear that, when looking at the diverse skills trained during education, the differences in language skills between ethnic minorities and the native Dutch are especially large. Scores on the CITO examinations show that discrepancy in calculating skills between the ethnic minorities and the native Dutch are much smaller compared to the differences in language related skills.

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Nat ive Dut ch Turk s Mor occa ns Sur inam ese Ant illean s

Vwo : scientific level

Havo : higher level

(14)

However, the higher proficiency in Dutch among in particular Antilleans does not lead to higher results during their education. Furthermore, the figures do not always show a relative high proficiency in Dutch for the Antilleans. For instance, figures of the CITO9_{show that results of}

Antilleans in tests of Dutch proficiency are comparable to the results of Turks and Moroccans10_.

Furthermore, from these scores on the CITO examination, it follows that Antilleans scores lower than other ethnic minorities11_{. I thereby conclude that the results with respect to education of}

minorities are somewhat mixed. Nevertheless, an important part of the educational fall behind is rooted into proficiency.

3.2.3 Time spent in the Netherlands

The time an immigrant has spent in the Netherlands influences his attained economic status. Many studies have shown that the generation to which immigrants belongs plays an important role. For instance, Ours et al. (2002) conclude that for the second generation income differences are much smaller than for the first generation. This can be explained by the higher level of integration into the Dutch society of the immigrants of a higher generation combined with their higher proficiency. Still, obvious proof for the relation between integration and economic performance cannot be found in the literature. Therefore, I will investigate whether I can find such a relation between achieved economic status and the integration into the society. Section 3.3.2 will describe some characteristics of integration in more detail.

3.2.4 Ethnic women

The discrepancy in labour market outcomes between the ethnic groups could be explained by the effect of gender. It could, for instance, partly be due to participation of women. If immigrant women do not work, while native women do, this will increase the differences in family income between immigrant and native families. In table 5 we see that there are large differences in participation rates among the diverse ethnic groups. The participation of Turkish and Moroccan women is low, while participation of Surinamese women is even higher than the participation of Dutch women (SCP 2003). The most reasonable explanation seems to be religion. Another explanation is that for ethnic minority females it is required to work in order to earn enough family income.

native Dutch Surinamers Antilleans Turks Moroccans

Man 81 78 73 67 65

9_{The CITO examination is a centralized examination at the end of primary education.} 10_{See SCP 2003 (figure V.3, page 403 and figure V.4 page 405).}

(15)

Woman 59 64 59 32 30

Table 5: Participation on the labour market of some ethnic minorities in 2002, in percentage.

Source: SCP (2003, table 7.1 page 207).

On the other hand, it can be argued that especially ethnic minority women earn much less in the labour market compared to both ethnic minority men and native Dutch women. There exists a ‘pecking order’: native men earn the most, then native women, ethnic men and finally ethnic women. The question remains why ethnic women earn less. Are there structural factors involved that make ethnic women even less productive and as a result less attractive to hire? In other words, are ethnic women equipped with much less human capital and other important endogenous factors? Or do these women exogenously earn less?

This analysis will be conducted at the level of the individual, and will highlight the personal characteristics determining income. By including a dummy variable determining gender of the person in question I can investigate whether, regardless of other personal characteristics, women earn less.

3.3 Social capital as an explanation for the disadvantaged position of minorities

In the previous subsection I described several probable causes for the lower income of immigrants. These included a lower proficiency in Dutch, lower human capital, and less participation of women.

The unexplainable part is often attributed to discrimination. However, as noted earlier, there are many alternative factors that may play a role. Among these are integration in the Dutch society and the fact that immigrants often do not have a large network that can help them to find a job or gain additional sources to start their own business. These alternative explanations take social capital into account. This section will to find some evidence from which we can argue that social capital can play a “missing link” in explaining immigrant’s lower economic performance. To do this I will focus on three aspects related to social capital: how do immigrants obtain a job, the effect of integration and the characteristics of self-employment. An extensive review of the social capital theory will be given in section four.

3.3.1 How ethnic minorities obtain their job

In section two, we saw that unemployment is much higher under minorities. According to Granovetter, especially “weak ties” contacts increase the chances of employment12. Reports on the situation of minorities (SCP 2003) also concluded that people mostly find a job by

(16)

information obtained from friends or contacts. This is especially true for Turks and to a lesser extent also for Moroccans, as can be seen in table 6.

Turks Moroccans Antilleans Surinamers

Family / friends / contacts 43 32 16 16

Employment agency 6 7 8 10

Temporary employment agency 20 18 26 26

Open application 11 18 13 13

School /stage 6 6 5 5

Advertisement 8 11 16 16

Asked by employer 3 4 7 7

Other 4 5 8 8

Table 6: Way of obtaining a job, for some ethnic minorities in 2002. Source: SCP 2003 (table 8.2 page 236).

Hooft et al. (2004) did a comparative study on searching for and finding a job. They asked Turks and a representative native Dutch group about their plans regarding their search for a job, and after four months they asked for the results. Some remarkable conclusions were drawn. First of all, they concluded that Turks are much more active on the labour market but their chances of success are lower compared to the representative Dutch group. When searching a job, the representative group made much more use of formal methods of searching a job. They also discovered that Turks, who stated that their culture is similar to the Dutch culture, had a higher chance of obtaining a job. A very surprising result is the fact that Turks make more often use of informal contacts when searching for a job. The authors conclude that Turks should rely more often on formal ways of searching jobs, because especially for the Turkish minority, informal contacts often lead to jobs of lower quality. It follows that from the results of Hooft et al the importance of weak ties clearly emerges.

Recently a study on ethnic minorities with higher education became available (RWI 2006) 13_.

This research concludes that employers and ethnic employees are often not ‘member of the same networks’. An employer stated: “We would gladly hire immigrants with higher education. We believe that there is much talent to discover there […] they have additional value because of their knowledge of other cultures […] however, we do not reach those groups by our traditional recruitment methods. Recently we used an ethnic recruitment agency […] which led to five immigrant candidates. […] we do not want to see or treat ethnic minorities different from the autochthones. Nevertheless, in practice this is required. (RWI 2006, page 60, translated).

This publication also stated that cultural differences play a role during job interviews. Especially cultural differences in style of communicating are important. Immigrant candidates sell

13_{The Raad voor Werk en Inkomen, (translated as council for employment and income), is a commission consisting}

(17)

themselves less and take a humble position towards the employer, because of their respect for a higher authority. However, many employers interpret this attitude as a lack of motivation and interest for the vacancy. In addition, the letter of application written by an immigrant candidate often does not match the expectations of the employer (RWI 2006, page 66-67).

This research confirms the importance of networks. Besides supplying vacancies, a network can also provide support when applying for a job. Members of a network can give the candidate some practical tips regarding the Dutch way of applying for a job. This is confirmed by a project that linked immigrants with higher education that had problems obtaining a job to an native ‘buddy’. This buddy acted as a coach; he gave the immigrant practical advice and used his own network to help the immigrant. The method proved to be very successful (RWI 2006). The media also gave attention to this success (for example NRC October 2006).

3.3.2 Self-employment under ethnic groups

Especially in the United States, immigrants’ response to their failure in getting a job is to start their own business. Light (1972) describes that this is specially the case for immigrants with strong ties to their own ethnic group and less contact with the native community. Especially the successful businesses try to use their ethnical characteristics to exploit a niche in the market. Sometimes the entrepreneurs only provide services for their own ethnic group, but others are able to increase their clientele to the native majority (like Chinese restaurants that have a broad base of customers).

In the Netherlands, especially Turks establish their own businesses. Still, from research it follows that ethnic minorities are less self-employed compared to native Dutch. While of the Dutch population about 7.5 percent has his own business, for Turks this is 6 percent, for Moroccans 3 percent, for Surinamers 4 percent and for Antilleans 3 percent. The Chinese minority can be seen as an exception with 15 percent of entrepreneurs (SCP 2005b).

(18)

3.3.3 Integration in the Dutch society

One can imagine that better integrated immigrants are more able to improve their economic position. From the second section it followed that income differences between the four different groups of minorities are large. We also see that levels of integration into the Dutch society differ between these groups. The choice between trying to become part of the Dutch society, or instead forming an own subgroup within the Dutch society is largely connected to the bridging / bonding theory of social capital14_{. Therefore, testing whether the level of integration}

into the Dutch society influences the achieved economic performance is a good way to test the bridging / bonding theory.

From this section it followed that there are large differences between the four ethnic groups. Thereby there can also be large differences into the level of integration between these diverse groups. A possible consequence could be that strong ties with the own ethnic minority are more valuable for less integrated groups than for better integrated groups. This is another hypothesis I will try to test later on.

3.4 Conclusions.

From this section, it followed that in particular human capital is important, especially the level of education and where it had been attained (in a foreign country or the Netherlands). A second factor (related to education) is proficiency in Dutch. A final determinant is the amount of time the immigrant has spent in the Netherlands.

These conclusions are confirmed by research of the ‘Raad voor Werk en Inkomen’. This study investigated the position of ethnic minorities with higher education on the labour market (RWI 2006). It states that unemployment of ethnic minorities is relatively high and the positions occupied by them are relatively low. The authors claim that this is partly due to education, age15_,

and race. However, these factors could not explain the differences between the autochthones and their ethnic counterparts. The authors therefore emphasize the importance of social capital.

We saw that there exist many theories explaining the lower economic performance of the ethnic minorities but consensus can hardly to be found. And since there still exist a large unexplainable part, this is often stated to be due to discrimination. This is not a very satisfying way of doing research, especially since there are strong reasons to believe that other factors cause

14_{See section 4.4 for more information on the bridging / bonding theory.}

15_{The authors concluded that on average ethnic minorities take more time to graduate, which implies that they are}

(19)

these differences. This section attested that social capital can be one of these “missing factors involved”. The conclusions here drawn provide a point of departure for further research on the effect of social capital on income of immigrants.

(20)

PART II: Review of relevant literature

4. Social capital

Råda, a local village in Sweden with about 500 inhabitants. Life is not too easy in this part of Sweden (Värmland), characterized by cold, long winters and many small villages with almost no facilities. However, Råda, as many other villages in the midlands and north of Sweden, is characterized by local cooperation. Besides many local activities, the inhabitants themselves invest collectively in a central heating system, to decrease the cost of energy, especially in wintertime. An secondary water system was also created, to use water from the local lake for purposes that do not require clean water.

However, lately especially young inhabitants move away towards the south of Sweden, since there are hardly any jobs available in the region of Värmland. The houses that are left empty are bought by foreign tourists, mostly as a holiday cottage, but also for permanent occupation. This has had consequences for collective initiatives. It is becoming harder to control the collective facilities, left alone to establish new initiatives, since the new inhabitants hardly speak Swedish, are not used to this kind of local cooperation and are less involved in the local community.

Another example of social cooperation is the “Rotating Credit Association”, as described in Light (1972). At the beginning of the 20th_{century, many ethnic minorities in the United States}

were faced with increasing discrimination, leading to high unemployment. Getting self-employed after they were discharged was difficult, since discrimination also led to difficulties obtaining finance from official institutions16_{. Some immigrants solved this problem by starting Rotating}

Credit Associations, consisting of a fund contributed by participants from a local ethnic minority. From that fund, participants were able to borrow money for their own initiatives. These organizations were especially established by immigrants from China, Japan, West Africa and West India, since in their home country these kinds of initiatives were also often used. Light states that the American born ethnic minorities (like the Afro-Americans) were unable to start these kinds of initiatives, since they had lost the tradition of cooperating in this way.

16_{Note that the financial institutions were discriminating directly by refusing to give credit. Furthermore they saw}

(21)

As can be seen in the previous examples, cooperation between actors embedded in a society results in production. This is the basic idea underpinning the theory of social capital. In the first example, we saw that because of the arrival of new ethnic groups and their failure to invest in social capital, some productive initiatives became neglected. The new inhabitants were not used to these kinds of initiatives. After their arrival, these initiatives therefore gradually ceased to exist. Living together, interacting with others and making use of shared resources enhances local cooperative initiatives. This concept is what the term social in social capital refers to.

From the second example we saw that an ethic minority responded to a threat in the society (an inability to borrow) by using an ‘own ethnic tradition’. The fact that the American born were not able to establish these projects leads us to believe that this kind of cooperation will be lost forever if not well maintained, and it needs to be build up over time to be of use. From the first example, we see that social capital can depreciate when there is no reinvestment. One can create and destroy collaboration over time, and invest (or disinvest) in social capital. This explains the word capital in the term social capital. While the conceptualization is simple, economic- and social scientists disagree with regard to the definition of social capital and its effects on the economy.

This section will therefore give an overview of the social capital theory. An important distinction I want to draw here is between social capital as a collective good (as described by for instance Coleman 1994 and Putnam 1995, 2000, 2001) and as an individual good (see for instance Granovetter 1973 and Fukuyama 1999). When treating social capital as a collective good17, social capital emerges from group participation and involvement. Here, social capital develops at a macro level. At the individual level, social capital can be seen as an additional pool of resources, embedded in social networks. Individuals make use of these resources in combination with other personal resources to achieve their goals.

4.1 Conceptualizations of social capital

The largest drawback of the social capital theory is the absence of a widely accepted definition. Knack and Keefer (1997, p. 1283) therefore stated: “the elasticity of the term social

17_{Note that collective good implies that the good is “non exclusive” and “non rivalry”. These terms imply that no}

(22)

capital has made these discussions difficult”. Because there is no agreement on what social capital exactly is, we also do not have a commonly accepted form of measurement (Portes 1998 and Sobel 2002).

The way social capital should be defined and conceptualized is broadly debated. As identified by Durlauf (2002), one can identify three major types of definitions of social capital. The conceptualizations arising from these definitions provide the main ingredients for the social capital debate. These conceptualizations, which are “trust”, “informal associations” and “shared norms and values”, are treated in the following paragraphs.

4.1.1. Social capital and trust

Arrow remarked (1972, p. 357) that “Virtually every commercial transaction has within itself an element of trust, certainly any transaction conducted over a period of time.” Higher levels of trust lead to more transactions, since trust decreases transaction costs, mainly by minimizing the risks involved in doing business. As stated by Arrow, especially exchanges where time elapses before the transaction is completed are influenced by trust. Transactions can be secured by mutual trust or by the legal system. Relying on law involves much higher transaction costs compared to mutual trust. Besides this direct effect, we can also expect that in a society characterized by higher mutual trust, fewer resources are necessary to protect private and public property. For instance, while in the Netherlands the lock of a bicycle is often even more expensive than the bicycle itself, locking a bicycle in Sweden is not necessary.

Trust is also very important for government policy. If individuals see their government as trustworthy, then government’s promises and policy announcements are more likely seen as credible. This leads to more reliance on these promises by the actors in the economy, and will increase their investments and other economic activities. Knack and Keefer (1997) did prove that there is a positive relation between trust and economic performance.

4.1.2. Social capital and informal organizations

(23)

By using factor analysis and combining many indicators of social capital18_{, Putnam ranked}

the different states of the US according to their level of social capital. He discovered that in states that had higher levels of social capital, schools and children performed better, there was less violent crime, people were less pugnacious, health was better, tax evasion was lower, there was higher income equality, higher civic equality and people were more tolerant. We see that Putnam’s (2000) definition of social capital especially refers to the macro level:

“Social capital refers to connections among individuals – social networks and the norms of reciprocity and trustworthiness that arise from them … A society of virtuous but isolated individuals is not necessarily rich in social capital. (p. 19)

Putnam’s conceptualization of social capital as the amount of participation in organisations is broadly debated. Research shows that these kinds of organizations create trust among their members, but it is doubtful whether this trust also extends to outsiders (Claibourn and Martin, 2000). Knack and Keefer19_{(1997) proved that the membership of organizations as}

defined by Putnam is unrelated to trust and civic norms at a macro level.

Knack (2003) updated the research of Knack and Keefer (1997) by including more countries and a larger period (1980 – 1999 instead of 1992). Furthermore, Knack questioned the comparability of the used measurements of membership rates of the different countries included in the sample. Hence, he adjusted the measurements to make them more comparable. After these adjustments, Knack found a positive effect of membership on trust, but the effect on economic performance was negligible.

It seems that we cannot draw definite conclusions from research on the influence of both formal and informal organization membership on trust rates and economic performance. The research intended however only searched to macro effects on the society, due to higher membership rates. Thereby it is also possible that a third variable related to membership causes the positive macro effects. An example of such a variable could be the quality of organizations. First of all, better functioning organizations will probably receive more support from the public, and thereby they will have more members. Moreover, these organizations can act as valuable institutions in the economy leading to higher productivity, as described by Hall and Jones (1999).

18_{Measured as participation in different kinds of organizations and collaborations.}

19_{Knack and Keefer included membership rates into a linear regression, where they hoped that this number of}

(24)

A second variable affecting the results could be the size of the network. Someone who is a member of multiple organizations has a larger network, and more contacts could lead to both higher trust and a higher economic performance.20

4.1.3. Social capital and shared norms and values

Fukuyama sees social capital as the amount of shared norms and values in a society. He defines (1999) social capital as:

“Social capital can be defined simply as a set of informal values or norms shared among members of a group that permits cooperation among them. If members of the group come to expect that others will behave reliably and honestly, then they will come to trust one another. Trust is like a lubricant that makes the running of any group or organization more efficient.” (Fukuyama 1999, p. 16)

Fukuyama stated that the network is social capital. He described it in the following terms:

“A network is a group of individual agents who share informal norms or values beyond those necessary for ordinary market transactions.” (Fukuyama 1999, p. 199)

According to Fukuyama (1999) a network arises spontaneously, resulting from the initiative of local actors. He distinguishes a network from a formal organisation by the fact that “economic exchange will be conducted on a different basis from economic transactions in a market” (p. 200). Following his definition, all sorts of “communities” can be seen as networks: both informal groups of friends, as well as the organizations mentioned by Putnam and the rotary credit associations described previously. The critical element is that the network has its own norms and values. We can use this definition both at the micro as well as at the macro level. An individual can profit from being a member of a network, but the network can also provide a society with some collective facilities. Different ethnic groups however each have their own distinct norms and values. Therefore, social capital cannot be homogenous once a society includes multiple different cultures.

(25)

4.2 Strength of weak ties

In 1973, Granovetter published the theory of strong and weak ties. In his influential paper, he explained the importance of strong networks. He presented an imaginary individual, called Ego. The inner circle of close friends around Ego forms a strong network in which information is shared quickly. However, it is unlikely that new information and knowledge can be acquired from this network, because the actors in the network are already member of the same clique. Ego can gain new information and knowledge from weak ties, his connections with actors within other networks. These weak ties, which act like a bridge between networks, allow information to flow between them. In addition, a person can reach more people by making use of these ties.

This theory is often applied in the analysis of the search of jobs (Granovetter 1995). Jobs are often obtained by informal contacts. When someone hears from a close friend of a vacancy, the chance that other friends will compete for the job is high. This happens because a network of friends is often composed of similar people with comparable education and interest. Furthermore, since the actors know each other well, information spreads quickly and all friends know about the job vacancy. Information obtained by weak ties is quicker, more exclusive and more relevant for Ego. Therefore Granovetter concluded that Ego will obtain a job easier (with less transaction costs) by using weak ties, and jobs obtained in this way are better in terms of income and prestige compared to jobs obtained by strong ties.

Flap and Boxman (1999) tested these hypotheses of Granovetter by investigating how managers obtained their job, their position and their social capital for a large sample of Dutch managers. They found that many (about 74%) of the managers relied on informal contacts. Furthermore, managers having more work related contacts, attend many receptions and other social gatherings and are member of associations, are endowed with more information regarding the labour market and relevant vacancies. Furthermore, these managers judged their chances at the labour market higher. Remarkable was the observation that for these managers, weak ties did not bring better jobs compared to strong ties. However, the authors stated that weak ties still seem to be of great importance, since they are a valuable source of labour market information.

These conclusions are in line with research on the unemployed in Sweden (Korpi 2001). This research also found that ties did matter, but difference between strong and weak ties could hardly be found.

(26)

Concluding, research on weak ties confirms the importance of ties in general, although it was found to be difficult to make a clear distinction between strong and weak ties. This might be the reason that not every research was able to confirm the higher value of weak ties compared to strong ties. Furthermore, a connection can be made between the “strength of weak ties” described here and the method of finding a job described in section 3.3.1. In that paragraph, we saw that ethnic minorities too often rely on contacts within their own ethnic group (strong ties) to obtain jobs. Ethnic minorities lack weak ties with native networks, as can also be concluded from the study of the RWI described in section 3.3.1.

4.3 Bridging, bonding and linking social capital

An interesting extension of the social capital theory is the bridging / bonding concept from Putnam. He stated (2000) that ‘negative social capital’ can be created, which he called bonding (or excluding) social capital. Bonding social capital brings people together who already know each other. Examples are certain religious networks. Members of these groups distinguish themselves from non-believers or people of another faith. Thereby bonding social capital creates boundaries between groups, since it focuses on the differences between them. This might even lead to giving preferential treatment to the own group.

The opposite of bonding social capital is bridging (or inclusive) social capital. Bridging social capital brings people from different networks together and creates shared values between them. It can be measured as an increase in tolerance and trust in others. One can see an analogue with the theory of weak ties, as identified by Lin (2001). The weak ties act as a bridge towards other networks and thereby form bridging social capital.

Bridging social capital is also important for immigrants when looking at the effect of culture on economic performance. From diverse publications, it follows that culture does not only influence the method of doing business (see for instance Greif 1994), but also assumptions regarding the motives of an actor. Henrich et al (2004) state that we can no longer assume that an actor acts rational to increase his own welfare (homo economicus). As game theory concludes, social preferences and fairness are also of major influence on behaviour. However, it is not possible to develop a new universal theory of behaviour, since social preferences and perceived fairness differ greatly between cultures.

(27)

Note that Sabatini (2006b) created a third concept of social capital: linking social capital. This term refers to memberships of formal and informal organizations as also identified by Putnam (2000). It thereby focuses more on concrete memberships, instead of group characteristics, like culture and norms and values. Sabatini argues that these associations can increase face-to-face interaction and trust and facilitate the learning of cooperative attitudes and behaviour.

4.4 Social capital and its economic effects

Much research has been published regarding the economic effects of social capital. However, it seems to be extremely difficult to test the effect of social capital, since there is no consensus how to measure it. Some authors have focused on ‘trust’, some on institutions and social organizations, and others on a broader notion of the concept (relations, kinships, lobbying organizations and hierarchical organizations). In this section, I will summarize some influential work on this field. I will distinguish between studies that focus on the micro-economic level and those focussing on the macro-economic effects of social capital.

4.4.1. Micro-economic effects

As we saw earlier, social capital can be seen as a network of relations providing the actor with valuable information. Especially for finding a job, relations proved to be important. Research did not only show that 35 - 55 percent of the jobs are obtained by informal contacts (Völker and Flap 1999), but also that a larger network leads to higher chances of obtaining a job (Korpi 2001). Research also confirms that employers use social networks to obtain workers. The percentages are about 35 to 50 percent, similar to those of employees using informal contacts (Fontaine 2004).

By creating some models of trade between communities, Routledge and Amsberg (2003) found that social capital decreases transaction costs. However, when labour has to be more mobile, the authors stated that social capital had a negative effect on transaction cost. If employees have lower social capital, they have fewer objections to moving for their job. On the contrary, if employees have higher levels of social capital, these employees will have more objections to moving, which will increase transaction costs.

(28)

time, the demand for workers with altruistic behaviour increased, and this development could explain the increasing wages of women.

Sabatini summarized several studies on social capital and especially the three different kinds of social capital: bridging, bonding and linking (see section 4.3). In one article (2006b) he concluded that companies increasingly invest in team building activities and the establishing of external contacts. He therefore investigated whether bridging, bonding or linking social capital increase labour productivity in small and medium sized enterprises. Bonding social capital was measured by family closeness and (to a lesser extent) by closeness to friends. Bridging social capital was measured by social engagement and both strong and weak ties. He measured linking social capital by memberships of voluntary associations and amount of contacts between members of these associations. He found that bonding social capital had a large negative influence on productivity. The effect of bridging social capital was also negative, but to a lesser extent. Linking social capital, however, had a positive effect on productivity.

4.4.2. Macro-economic effects

Stephen Knack and Philip Keefer (1997) wrote one of the most influential works on social capital and its economic effects. They measured social capital as trust and civic cooperation and found that these variables had a “significant impact on aggregate economic activity” (p. 1283). The authors emphasize the role of secondary education in creating trust, while they conclude that the formation of and participation in organizations can sometimes decrease trust by creating more segmentation into the economy.

This conclusion was confirmed by others, like Whitely (2000), who incorporated social capital (defined as trust) into a neo-classic growth model (Solow-Swan model). The author stated that “social capital defined as interpersonal trust, is an important factor in explaining cross national variations in economic growth” (p. 460). However, Beugelsdijk, De Groot and Van Scheijk (2002) dispute the conclusions of Knack and Keefer. They state that the significance, direction and size of the effect is largely influenced by the conditional set of variables that are included in the regression, and thereby questioned the robustness of Knack and Keefer’s regression

(29)

Dasgupta (2005) described the economic effects of trust and social capital on the society. His emphasis was on the role of trust for the purpose of secure transactions. He stated that social capital can best be seen as a “system of interpersonal networks”. While his paper focuses on the collective effects, from his definition it follows that we have to focus on the personal networks. He stated that trust is a product of networks and thereby of social capital.

Glaeser et al (2000) made a regression containing many geographic, religious and other collective variables. They used different definitions of social capital and found that maximally 10 percent of individual variation in social capital can be explained by these collective variables21_{. By}

conducting experimental games on a group of Harvard students, Gleaser et al. (1999) also showed that social capital is mainly made up of individual aspects that determine the outcomes of the performed games.

We can conclude that the discussion regarding social capital and its effect on the aggregate economy has not ended yet. While there is mixed evidence regarding the effect of social capital on a macro level, it is obvious that it matters at an individual level. This follows especially from the research of Dasgupta (2005) and Gleaser et al. (1999, 2000) from which they concluded that social capital is largely individual driven. Therefore, it seems more justified to test the effect of social capital on the micro level.

4.5 Social capital and human capital

Social capital is strongly related to human capital, and this is sometimes even described as the most robust empirical relationship (Glaser et al. 2000). Research on this topic was influenced mainly by the work of Coleman (1988). He stated that social capital can stimulate children to perform better at school.22_{If a child has many connections with relatives or neighbours provided}

with high levels of social capital, the child may receive more support. His evidence mainly consists of lower dropout ratios among children that receive more support from their parents or visit Catholic schools.

These results were confirmed in other studies. Putnam (2001) found that in states with higher levels of social capital schools performed better. By investigating a Mexican minority in the San Francisco – San José region, Stanton–Salazar and Dornbusch (1995) found that students

21_{To derive these results, Glaeser et al used factor analysis, utilizing many different definitions of social capital. The}

definitions leading to the best result implied that only ten percent of individual variation in social capital was explained by collective variables. Hereby the authors concluded that social capital is mostly individual determined.

(30)

with higher grades were characterized by significantly higher levels of social capital23_{. They}

emphasize the role of weak ties, especially ties with non-Mexican and non-kin persons. When Goyette and Conchas (2002) compared the grades of Vietnamese and Mexican American students, they also found that more weak ties led to higher results for Vietnamese students.

Dika and Singh (2002) reviewed thirteen studies that discussed the relationship between social capital and educational attainment, and the achieved level of education. They concluded that many of the relationships were significant and in the expected direction. Dika and Singh also mentioned fourteen other studies that examined the relationship between social capital and grades received by students. Here the relationships were again significant and in the expected direction. While all of the relationships regarding social capital and education were found to be positive, the authors state that theoretical and empirical support “could be stronger”. The authors point at problems due to the gaps in conceptualization and measurement of social capital. They state that clarification of this discussion could improve evidence on the effect of social capital.

It follows that there are many reasons to believe that social capital does not only have a direct effect on economic performance, but also an indirect effect by supporting the creation of human capital.

4.6 Conclusions

This section tried to present a brief overview of the literature on social capital. Furthermore, some empirical results of the effect of social capital were presented. We saw that many different approaches to social capital exist. I will discuss the approach that I will use throughout the rest of this paper later in this paragraph.

A distinction between social capital at a macro- and micro level was made. From the work of Putnam arises the idea that social capital can act as a collective good that brings, besides many other positive effects, economic prosperity. However, from section 4.4.2 it did not only follow that this effect is doubtful, but it was also concluded that the determinants of social capital are largely individual driven. Gleaser et al. (1999) stated: “Social capital is a meaningful, individual-level variable that can be studied with the tool of price theory, once it is understood that aggregation will be much more difficult than for other forms of capital” (p. 33).

The aim of this research is to find the effect of social capital on the economic outcomes of minorities. This implies that social capital should be treated as an individual aspect. Fukuyama emphasized the role of networks in social capital. They provide actors with information and,

(31)

because of shared norms and values, make cooperation easier. These networks can be seen as an asset that improves the ability of an actor to achieve a higher position in society. For building a network, an actor needs skills and the opportunity to meet others. One opportunity to meet others and to include new members in a network can be found in informal organizations. By participating in these kinds of organizations an actor can increase his network, and therefore these organizations play an important role.

(32)

5. The economics of discrimination.

In 1957, Gary S. Becker was the first to describe the problem of discrimination from an economic point of view. Before his publication, discrimination was only treated from a sociological or psychological perspective. His publication still plays a leading role in the economic theory on discrimination. This section will briefly describe the theoretical background of discrimination.

Becker distinguished between two practices through which discrimination takes place and affects the labour market: wage discrimination and employment segregation. Wage discrimination refers to different payments for workers who have equal qualifications but differ in their characteristics (ethnic background, race etcetera). Segregation practices involve hiring employees based on their characteristics (ethnic background, race etcetera).

Becker stated that these practices imply that people have a “taste for discrimination”, which means that they prefer an employees with particular characteristics. For example, employers might prefer native Dutch employees. These tastes are based on expectations regarding the quality of the employee. In Chiswick (1995) an example is treated, based on preferences of the public to go to a Chinese restaurant employing Chinese waitresses. It assumed that this leads to a higher demand of, and therefore a higher salary for, Chinese waitresses. Waitresses of another ethnic background hereby will get lower wages and when someone starts a restaurant with non-Chinese waitresses, it has to offer lower prices to compensate for the absence of non-Chinese waitresses. When there is at least a small group of costumers who believe that food served by a non-Chinese waitress will taste the same, they will visit the cheaper restaurant. Therefore, when a significant group finds out that the quality of a non-Chinese waitress is comparable to a Chinese one, the public will chose the cheaper restaurant until the price differences are vanished. The ‘preference’ for Chinese waitresses proved to be due to discrimination, we could infer this because of the price differences that arose.

The Economics of Integration

The Economics of Integration

Does social capital play a role?

Master thesis for Economics

The Economics of Integration

Does social capital play a role?

Contents:

PART I: Introduction

PART II: Review of relevant literature