The impact of the current economic crisis on youth unemployment in the Netherlands

(1)

The impact of the current economic crisis on

youth unemployment in the Netherlands

Linde de Visser

10113290

(2)

Abstract

This is a study on the effect of the 2008 global financial crisis

on Dutch Youth Unemployment. The EU currently is trying to

find a solution to the growing European Youth Unemployment

Rate. This thesis is focussed on the effect of the 2008 global

financial crisis on the Dutch Youth Unemployment Rate in the

Netherlands, beside this a comparison to other European

countries will be made. The method in this thesis is a literature

research and a detrending analysis over the period 2004-2008.

During the 2008 global financial crisis the Dutch Youth

Employment Rate increased faster than the Total

Unemployment Rate. It cannot be concluded that this effect is

due to the crisis, because this result can only be found when

taking into account that, due to Employment Protection

Legislation, the Unemployment Rates started to increase later

because employers are more reluctant to fire employees due to

high costs. When researching other countries with similar

Employment Protection Legislation levels, this effect was not

found. Therefore it cannot be stated that there is a causal effect

of the crisis on the faster increasing Youth Unemployment

Rate.

(3)

Lists of abbreviations

(EPL) Employment Protection Legislation (EU) European Union

(TUR) Total Unemployment Rate (YAR) Youth/Adult Ratio

(YGS) Youth Guarantee Scheme (YUR) Youth Unemployment Rate

(5)

List of graphs

Graph 1: European Youth Unemployment Rates Graph 2: European Youth/Adults Ratios

Graph 3: YUR per country Graph 4: TUR per country

(6)

List of tables

Table 1: impact of crisis on youth unemployment, five variables Table 2: impact of crisis on youth unemployment, seven variables

(7)

1. Introduction

‘Why is it so hard to find a job?’ Is a question I hear almost daily during my work as a guest lecturer for FNV Bondgenoten, the largest Dutch union. I travel trough the country to teach at different schools. I rarely see the same student twice. Many of these students are struggling to find and keep a job, even though they try hard, This made me wonder whether these youngsters are hit harder by the crisis than the average Dutch person.

1.1 General framework

Youth unemployment is a growing problem. The European Union is trying to find a solution to this problem. Trough the articles below, a part of the European ideas are expounded.

According to the Guardian (2012), an English newspaper, European Youth Unemployment has increased from 15% in 2008 to 22.4%. Moreover, the Youth Unemployment Rate has increased faster than the overall unemployment rate in Europe (Traynor, 2012). Leading to a total of 5.5 Million unemployed youngsters in 2012.

On the 28th Of June, 2013, Rijksoverheid (the Dutch central government) published an article on youth unemployment. Discussing that throughout the EU, 7.5 million youngsters (under 25) are unemployed (Rijksoverheid, 2013) - the difference between the Guardian’s estimation and that of the Rijksoverheid could be explained by a different description of unemployment (Dietrich, 2013). A stimulation package of 16 million euro aimed at diminishing youth unemployment was reserved for

implementing a Youth Guarantee Scheme (YGS).

The European parliament published an article on their website on the 27th of February 2014 on the YGS (Euro parliament, 2014). The goal of the YGS is to ensure all people younger than 25 get a ‘good-quality, concrete offer’ within 4 months after leaving school or after they became unemployed. ‘The good quality offer should be for a job, apprenticeship, traineeship or continued education and be adapted to each individuals need and situation’ (Euro parliament, 2014). Twenty countries within the EU were asked to submit a plan on how they would implement the YGS in their countries.

(8)

Graph 1, European Youth Unemployment Rates. Source: European parliament (2014)

Notice in graph 1 that the Dutch Youth Unemployment Rate (YUR) is less that of many other European countries. Still, Dutch newspapers publish articles on the subject. NRC Handelsblad (2013), a Dutch newspaper published an article on the rising youth employment rate in the Netherlands (Spijkerman, 2013, UWV, 2014). According to the article, the number of unemployed people under 27 increased by more than 50%, herewith the article uncovers a growing problem.

These articles are all written after the start of the current economic crisis. What is the cause of the increasing youth unemployment? And is the youth unemployment really reacting differently than the Total Unemployment Rate? This leads to the following question: is the impact of the economic crisis on the Youth Unemployment Rate in the Netherlands different than the impact on the Total Unemployment Rate?

1.2 Earlier research

There has been research on this subject before. This earlier research is focused more on countries where youth unemployment is higher than in the Netherlands. Important empirical research on the subject was for example done by Hans Dietrich

(9)

(2012), the most important conclusion for this thesis is the youth/adult ratio, which shows that, in most countries, youth unemployment is about twice as high as adult unemployment. Also theoretical research was done on this subject, for example by Verick (2009), whose study contains a theoretical explanation for why, during and after an economic crisis, youth unemployment increases faster than total

unemployment due to an economical crisis.

1.3 Central question and design

The central question in this research is; is the impact of the 2008 global financial

crisis on youth employment rates in the Netherlands different than the impact on total employment rates? This question will be answered by doing a detrending analysis

where a timeline for unemployment will be researched to see if the interaction variable, Youth*Crisis, is significant. This interaction variable makes it possible to compare YUR to TUR. The dataset that is used is for the time period of 2004-2013 and collected by Eurostat. The prior Eurostat goal is to collect European Union (EU) data that is comparable between EU countries. Because in this thesis different European countries are compared, this is an interesting source for data. There is not much prior research done on this subject for the Netherlands. This thesis, therefore, is focussed on the Netherlands.

1.4 Structure

Prior research will be discussed more extendedly in the second part of this thesis. A distinction will be made between theoretical and empirical findings. The empirical findings will be on the Netherlands, Germany, France and the EU. Then the method will be discussed followed by the results and the conclusion of this thesis.

In this part of the thesis different theoretical arguments will be discussed on why youth unemployment increases more than total unemployment due to a crisis, followed by an exposition on earlier empirical evidence.

(10)

2. Youth unemployment in theory and practice

In this part of the thesis, earlier empirical as well as theoretical research will be

discussed. The theoretical part will be a general explanation for youth unemployment, in the empirical part, the Netherlands, France and Germany will be discussed per country.

2.1 Theoretical arguments on youth unemployment

Due to financial crises, unemployment rates often increase. As Verick (2009) stated, ‘unemployment rates across the world will continue to rise and stay stubbornly high for some time to come’. Youth unemployment is affected more during, and after, a crisis than total unemployment and young men are affected more than young women. Verick (2009) gives an explanation for this effect.

According to Verick (2009), the first reason for this occurrence is the decrease of aggregate demand. When aggregate demand decreases, employment decreases as well, depending on the unemployment elasticity in the particular country and/or sector. Lefresne (2012) argues that youngsters are more affected because the YUR overreacts to changes in economic welfare. When overall employment increases, youth employment increases faster and vice versa. The reason that young people are more affected by an economical crisis, is because of the sector they are working in. The building sector is an example, because there are mostly young employees. Before the recent financial crisis, the house prices increased and there was a lot of activity in the sector. Due to the financial crisis, this sector has collapsed leading to many unemployed young people, especially men (Verick, 2009).

Another explanation for the increase of the unemployment rate is temporary employment, especially young people have a temporary contract (Verick 2009). In the EU, on average 41% of young men are temporarily employed, 11% of employees between 26 and 65 were temporarily employed (Eurostat, 2014). In the Netherlands, 40% of the youngsters are temporarily employed, 11% of the employees between 26 and 65 were temporarily employed. (CBS, 2013). These employees are easier to fire than employees that have a permanent contract.

The last argument Verick (2009) has for the increasing unemployment rate is the Employment Protection Legislation (EPL). The more rigid EPL in a country is, the less the unemployment rate will increase in that particular country (Verick, 2009).

(11)

Thus, the less EPL in a country is, the more quickly the unemployment rate will rise. This, in combination with the two arguments above, will hit young men hardest due to the fact that young people often have a temporary contract (more often than older people) thus they are more easy to let go. And aggregate demand decreased more within the sectors where young people (especially men) are active.

2.2 Earlier empirical findings

Dietrich (2013) found a relation between youth- and adult unemployment. According to him, youth unemployment is about two and a half times as high as the adult unemployment rate. He calls this the youth/adult ratio (YAR). The European YAR increased between 2001 and 2010 and reached its maximum in 2008 (Dietrich, 2013). In the graph below the YAR’s of several countries are depicted.

(12)

2.2.1 Germany

In Germany, the YUR is traditionally low, about the same as the adult unemployment. German YAR increased during the recession from around 1 to around 1.5 (Parker et al, 2014). Since 2004 German YAR was higher than 1.5 twice (Eurostat, 2014). During the crisis, German YUR declined to 8% despite of the increasing YAR. Even though the German YAR decreased, it is still one of the lowest European YAR’s. According to Verick’s theory, if there are many temporarily employed youngsters, YUR is expected to be high, German YUR should be high since the percentage of temporarily employed youngsters is over 50%, there must be taken in account that most German youngsters are able to find a permanent job. 6.7% of the youngsters were unable to find job with a non temporary contract during parker’s (2014)

research. There is one group of youngsters in Germany that does have a hard time finding a job, which does correspond with Verick’s theory. Lower educated

youngsters are not easily employed. Therefore, youngsters that are without a job, often attended lower education or dropped out of school and are likely to be unemployed for a longer period of time (Parker et al, 2014).

In Germany EPL is extensive, employees can only be fired if the employer can proof the employees behaviour is bad (Jahn, 2009).

2.2.2 France

In France, the economic crisis had a great impact on the French YUR, it increased to 26.2%. In France the reason for the high YUR, besides the reasons discussed in paragraph 2.1, might be the lack of experience French youngsters have (Balaram et al, 2014). The lack of experience within these youngsters in France come from the French vocational education system, which is, other than many other countries, more general education. This means the student don’t have to choose their future

profession at a young age. Also, students do not have much work experience when they become part of the working population, only 15% of French youngsters that are still in the education system also have a (small) job (Balaram et al, 2014), due to the lack of experience and the increased availability of more experienced jobseekers (whom have lost their job due to the crisis and thus are available again). The French YAR did stay intact during the crisis and still lies between the usual 2 and 2.5

(13)

Many researchers think the high unemployment rate in France is due to the extended EPL, however this was not statistically supported (Balaram et al, 2014)

2.2.3 The Netherlands

The Netherlands used to have a very high Youth Unemployment Rate, therefore the Dutch government implemented a series of policies to reduce youth unemployment. Partly due to the implemented policies, Dutch YUR started decreasing in the early 90’s (Parker et al, 2014, Eurostat, 2014). In 2008, Dutch YUR reached it’s lowest point of 6%. Between 2008 and 2010 Dutch YUR increased to 8.7%, which is relatively low compared to other European countries.

In the Netherlands many youngsters are temporarily employed this was 52.2% in 2012 (Parker et al, 2014).

In 2012 Dutch YUR started to increase more rapidly and was at it’s top in 2013 at a level of 11.2% (Eurostat). Parker (2014) thinks Dutch YUR started to increase in 2012 instead of immediately after the start of the crisis is probably due to the strong EPL in the Netherlands. Dutch YAR has been relatively constant between 1.5 and 2 between 2004-2013.

2.2.4 Employment Protection Legislation compared

In the three countries discussed before, EPL is almost at the same level. Zientara (2006) made a scale to measure EPL. The scale is from zero up to three where zero is no protection at all, level one is ‘protection for the regular worker against dismissal’, level two stands for ‘specific requirements for collective dismissal’ and level three is ‘regulation of temporary forms of employment’ (Zientara, 2006). The Netherlands score 2.1, Germany 2.5 and France scores 2.9 points on Zientara’s scale. These countries, thus, are quite close to each other in EPL perspective. Zientara (2006) found in his research that there is a positive correlation between unemployment and a higher EPL. This could explain why the French unemployment rate is higher than the German and Dutch unemployment rate, however, it can not explain why the Dutch unemployment rate is higher than the German unemployment rate. It also can not explain why the German unemployment rate is low in general because the German EPL is quite high compared to other European countries.

(14)

3. Methodology

In this chapter, the research method is discussed. Starting with why certain data was used. After this, the actual method is discussed.

3.1 Data

The data used in this research is collected from Eurostat. Eurostat is an

establishment that collects European data. Typically, each country collects their own data, under their own definition. For unemployment, these definitions can vary. The Netherlands has three restrictions for listing a person as unemployed., the first restriction is that they are not employed for more than twelve hours, so some

employed people are listed as unemployed. Another Dutch restriction is; the person has to be aged between 15 and 67. The last Dutch restriction is that the person has to actively search for employment. These Dutch restrictions will not be used, because Eurostat restrictions are used. Eurostat uses the following restrictions; The age has to be between 15 and 74 (in this thesis ages used are from 15 to 64), the person has been actively searching for employment in the four weeks before the reference week, the person must be unemployed during the reference week, the person must be available for employment within two weeks from the reference week (Eurostat, 2014). Every country within the EU is treated this way by Eurostat so the data are

comparable between countries. This is important for this thesis since a set of European countries is used.

3.2 Statistical model

This research is based on a detrending analysis. The timetrend will be researched and, after the defined moment for crisis, a series of variables will measure whether or not the 2008 global financial crisis had an impact on the timetrend for both YUR and TUR. The timeline analysis will be researched by the following regression models:

t tY t t CY t Y t t C t C t Y C Y t Y UR =β0 +β ⋅ +β ⋅ +β ⋅ +β ⋅ ⋅ +β ⋅ ⋅

(15)

And: t t CdelayY t Cdelay t tY t t CY t Y t t C t C t Y C Y t Y Cdelay Cdelay Y UR =β₀ +β ⋅ +β ⋅ +β ⋅ +β ⋅ ⋅ +β ⋅ ⋅ +β ⋅ +β ⋅ ⋅

The first regression model describes the impact of the crisis on TUR by variable C, Crisis. The impact on YUR is measured by variable C*Y, Crisis*Youth. In the second model, Cdelay is added. This variable is added because of Parker’s (2014) findings, he argued that, due to high EPL, the crisis had a delayed impact on Dutch

unemployment rates. Cdelay measures the effect on TUR and Cdelay*Y measures eventually extended effect on YUR.

In the following subchapters, the variables will be discussed more extendedly.

3.2.1 Crisis (C)

According to the article by Choudry et al (2012) there are two kinds of crises, the systematic- and the non systematic banking crisis. The systematic crisis indicates a crisis with a large number of financial and corporate defaults. If there is a small number of corporate and financial defaults, it is called a non systematic banking crisis. In this case, the crisis started with a mortgage crisis in 2007 (Peicuti, 2013), followed by the Lehman Brothers bankruptcy in 2008 in the USA. After this occurred, the crisis spread to Europe where many banks were saved by the government. In the Netherlands DSB bank went bankrupt in October 2009 (DSB Bank, 2014). Because this is a wide spread crisis and the first bank failure was in September 2008 (Lehman Brothers), the fourth quartile of 2008 will be used as the start of the financial crisis.

3.2.2 Time (t)

The time span used in this thesis is the first quartile of 2004 till the last quartile of 2013. The time variable counts from 1 to 40 and is in chronological order, this means that t=1 is the first quartile of 2004, t=40 is the last quartile of 2013. The period is before and during the 2008 global financial crisis, this is to measure whether YUR was more affected by the crisis, or YUR was higher to begin with and was not hit harder.

(16)

3.2.3 Youth (Y)

The variable youth is a dummy variable; it can only have the values 0 and 1. 0 if the person is no youngster and 1 if it is a youngster. In this research youngster are the youngest people that fit the Eurostat description of being unemployed and are younger than 25.

3.2.4 Delayed effect on unemployment rates (Cdelay)

Parker’s (2014) theory explains that the Dutch TUR and YUR started to increase in 2012 due to the high level of EPL. Therefore the second regression includes another variable, Cdelay. This variable has two possible values, 0 or 1, 0 if it was before 2012, 1 for 2012 and 2013. This variable is used on itself and in combination with youth. It is not used in combination with the variable crisis because it is irrelevant to include C*Cdelay*Y since variable Cdelay only has a value of 1 in times of crisis. This variable thus measures the delayed effect on unemployment, which Parker describes is due to EPL. According to Parker’s findings, this effect started in 2012. Therefore, 2012 and 2013 the value of variable Cdelay will be 1 and 0 before 2012.This effect will be measured fur TUR and YUR.

3.2.5 Unemployment rate (UR)

Is the unemployment rate, or the probability that someone with the filled in characteristics is unemployed. The small c stands for country.

3.3 Statistical method

In this thesis two timeline analysis were used. The first one with five beta’s, the second one with 7 beta’s. In the first regression Crisis*Youth is the most important coefficient, because it measures whether the crisis actually has a larger impact on youth unemployment compared to total unemployment. If this coefficient is

significant, the 2008 global financial crisis has an influence on the youth unemployment.

The second regression has two added variables, EPL and youth*EPL. With these extra variables, it is possible to measure the delayed effect of the 2008 global financial crisis on unemployment. According to Parker et al (2014), the Netherlands experienced a delayed effect on unemployment, due to EPL. The real effect started,

(17)

according to him, in 2012. Therefore, for the second regression EPL*youth also measures the (delayed) effect from the crisis on youth unemployment.

The regressions are done with Stata, which also performs a t-test for testing significance of the coefficients. Exact Stata outcomes are found in attachment 1.

(18)

4. Results

In this chapter results from the two regressions will be discussed per country.

4.1 Regression results

Doing the two regression led to the following results.

Table 1: Impact crisis on youth unemployment, five variables. Significance levels 10%>5% *, 5%>1% **, 1%>0.1% ***, 0.1%> ****

Table 2: Impact crisis on youth unemployment, seven variables. Significance levels 10%>5% *, 5%>1% **, 1%>0.1% ***, 0.1%> **** β NL EU DE FR Crisis 0.22 1.03 0.25 0,81 (0.81) (0.80) (0.54) (0.76) time 0.05 0.05 -.19**** 0.00 (0.04) (0.04) (0.03) (0.03) youth 3.49**** 9.05**** 4.33**** 12.91**** (-0.65) (0.48) (0.50) (0.61) crisis*youth 0,72 2.01* -0.01 2.53** (-1.15) (1.13) (0.77) (1.07) time*youth 0.01 0.06 -0.04 -0.01 (-0.05) (0.05) (0.04) (0.05) constant 2.84**** 6.06**** 11.41**** 7.09**** (-0.46) (0.59) (0.35) (0.43) R-squared 0.75 0.96 0.97 0.92 n 80 66 80 74 β NL EU DE FR Crisis 0.70 1.89** 0.67 1.34* (0.64) (0.73) (0.57) (0.74) time -.042 -0.04 -.23**** -.05 (0.03) (0.04) (.03) (0.04) youth 4.03**** 9.85**** 4.38**** 13.25**** (0.57) (0.90) (0.61) (0.66) crisis*youth 1.27 2.64** .03 2.88*** (0.91) (1.03) (0.81) (1.04) time*youth -.045 -0.00 -0.04 -.04 (0.05) (0.06) (0.05) (0.06) Cdelay 2.31**** 1.82*** 0.92* 1.35** (0.56) (0.66) (0.51) (0.64) Cdelay*youth 1.39* 1.32 0.09 0.89 (0.79) (0.93) (0.72) (0.91) Constant 3.73**** 7.17**** 11.88**** 7.62**** (0.40) (0.63) (0.43) (0.46) R-squared 0.87 0.98 0.98 0.92 n 80 66 80 74

(19)

4.2 Timeline analysis discussed per country

In this part the outcomes of the regressions will be discussed per country.

5 10 15 20 25 Y UR % 0 1 0 s tar t c ris is 3 0 4 0 2004Q1 _2006Q1 _2008Q1 _2010Q1 2012Q1 time eu Germany France Netherlands

Youth Unemployment Rates

Graph 3: YUR per country, first vertical line is for start C, second vertical line is for start Cdelay

2 4 6 8 10 12 T UR% 2004Q1 4 0 3 0 2 0 1 0 2006Q1 2008Q1 2010Q1 _2012Q1 time eu Germany France Netherlands

Total Unemployment Rates

(20)

4.2.1 Germany

As described in chapter two, German YUR decreased during the crisis. According to table one, Crisis had no significant effect on unemployment, nor was there a

difference between YUR and TUR. A negative trend was found for the variable time, this means unemployment rates decrease over time.

In the second table, where EPL is added, EPL has an effect on unemployment which is slightly significant (between 5 and 10%). This effect is much lower than in the other countries, even though EPL in the other tested countries is also at a high level on the scale from Jahn (2006). In Germany, only 6.7% of the youngsters are unable to find a permanent job (Parker et al, 2014).The 2008 global financial crisis had a remarkably low impact on German unemployment. Also, there was no delayed effect found. These findings do support Verick’s theory (2013) where it is claimed that a crisis will have less effect on the unemployment rates, if youngsters are able to find a permanent job. This is also consistent with the relatively stable YAR, because total unemployment does not increase much, youth unemployment increases faster, but also not much.

4.2.2 France

In the first table, it is visible that the crisis had a significant impact on French youth unemployment, crisis*youth is significant and positive with 2.5%, meaning that the French YUR increased faster than French TUR due to the 2008 global financial crisis by approximately 2.5% per year. According to Balaram et al (2014), France was hit this hard due to the low level of work experience under French youngsters.

Youngsters that are in school, often do not have a job leading to a disadvantage on the labour market, compared to other, older, applicants.

Table two is not very different from the first one, the delayed effect due to EPL on the total unemployment has a significant effect, meaning that there is also a delayed effect on total unemployment, this is not larger for youngsters. This implies there was a direct effect on YUR specifically and a delayed effect on TUR , which includes YUR, due to the 2008 global financial crisis. This might be explained by a high percentage of temporarily employed youngsters. These youngsters have less

(21)

EPL because they are temporarily employed and thus the crisis has a higher direct effect on YUR, this should be further researched.

4.2.3 The Netherlands

Table one shows significant coefficients for the constant and for the variable youth, which means youth unemployment significantly differs from total unemployment. For every other variable, no significant variables were found. This means time and crisis is not proven to have any effect on unemployment.

When looking at the second table, EPL*youth as well as EPL have significant beta coefficients. Which supports the hypothesis of Parker et al (2014), due to the high level of EPL in the Netherlands, the Netherlands experienced a delayed impact of the crisis on (youth) unemployment.

4.2.4 Europe

According to the first table, the 2008 global financial crisis has an impact of small significance on youth unemployment. The YUR on itself is much higher than the TUR, this coefficient is very significant. After adding EPL to the equation, the effect of the 2008 global financial crisis on YUR becomes more significant and is quite big, also EPL has a positive effect on TUR but not on YUR specifically. This could be explained by the fact that many European countries score high on the scale Zientara (2006) made, this scale measures whether EPL is high or low. All countries taken into account in this research score a high EPL.

4.3 Shortcomings

The results in this research in table 2 show that there was a different effect of the crisis on Dutch YUR compared to Dutch TUR, but only when the delayed effect due to the EPL is taken into account. For further research it is recommended to renew Zientara’s (2006) scale of EPL and compare countries with different levels of EPL. In this research it cannot be stated the delayed effect on Dutch YUR is really because of EPL because there is no comparison material. Also it should be tested why the results are so different per country even though the EPL levels are all high and

(22)

taking the delayed effect is taken into account, it is not possible to explain why these three countries differ so much, even though the EPL levels are comparable.

(23)

5. Conclusion

By doing a detrending analysis, the conclusion was drawn that in the Netherlands, the impact of the crisis on YUR was higher than on TUR, at least after from 2012.

There are theoretical reasons for this to happen, youngsters often have a temporary job. This makes it more easy to fire youngsters, thus youngsters are more sensitive to losing their job. Another reason is aggregate demand, during a crisis aggregate demand decreases first in the sector youngsters work. The last reason is EPL, if EPL is low, people are more likely to loose their jobs because it is more easy for the employer to fire them (Verick, 2009).

By doing a time series analysis, it was tried to find a direct impact of the crisis on Dutch YUR, this was not found. Only when the later effect of the crisis, which could be due to the high EPL rate, was taken in account in the model, an effect could be found. This effect starts from the first quartile of 2012. Further research would be recommendable, to find out if this impact is really delayed by EPL.

(24)

Bibliography

DIETRICH, H. 2013, Youth unemployment in the period 2001-2010 and the European crisis – looking at empirical evidence, Transfer, vol. 19, pp. 306-324.

DOLPHIN, T. AYRES, S. BALARAM, B. & PARKER, I. 2014, States of Transition: Youth Unemployment, Education, and Labour Market Policy in Europe and the US. 3rd, Institute for Puplic Policy Resource, London.

DSB Bank, [cited june 26th 2014], available from: www.dsbbank.nl

Euro parliament, [cited 26 may 2014], available from:

http://m.europarl.europa.eu/EPMobile/en/news/product.htm;jsessionid=ABD0134416 925FFD03B68664649C607D?reference=20140226STO37008

Eurostat, [cited july 12th], available from:

http://epp.eurostat.ec.europa.eu/cache/ITY_OFFPUB/KS-SF-11-008/EN/KS-SF-11-008-EN.PDF

JAHN, E. J. 2009, Do firms obey the law when they fire workers? Social criteria and severance payments in Germany, Economic Affairs, vol. 26(4), pp. 46-53.

LEFRESNE, F. Youth Unemployment and Youth Employment Policy: Lessons from France. 1st ed. Friedrich Ebertstichtung, 2012. 19 p.

PEICUTI, C. 2013, Securization and the subprime mortgage crisis, Journal of Post Keneysian Economics, vol. 35(3), pp. 443-456

Rijksoverheid, [cited 26 may 2014], available from:

http://www.rijksoverheid.nl/nieuws/2013/06/28/eu-8-miljard-voor-aanpak-jeugdwerkloosheid.html

(25)

SPIJKERMAN, Catrien. 2013, Aantal werkloze jongeren 50 procent gestegen in jaar tijd. NRC handelsblad, section business, 27 june 2013, [cited 26 may 2014] available from: http://www.nrc.nl/carriere/2013/06/27/aantal-werkloze-jongeren-in-jaar-tijd-met-50-procent-gestegen/

TRAYNOR, Ian. 2012, Unemployment of Europe’s young people soars by 50%. The Guardian, section business, 16 april 2012, [cited 26 may 2014] available from: http://www.theguardian.com/business/2012/apr/16/european-youth-unemployment-soars

UWV, [cited july 12th, 2014], available from:

https://www.werk.nl/werk_nl/arbeidsmarktinformatie/publicaties/thematische-publicaties/rapportage_jongeren

VERICK, S, 2009, Who is hardest hit during a financial crisis? The vulnerability of young men and women to unemployment in an economic downturn, EZA discussion papers, nr. 4359, pp.3 - 35.

ZIENTARA, P. (2006), Employment protection legislation and the growth of the service castor in the European Union

(26)

Appendix

Supplement 1: data Data used for regressions

quartile YUR EU YUR Germany YUR France YUR Netherlands TUR EU TUR Germany TUR France TUR Netherlands 2004Q1 20,1 8,5 8,0 4,1 2004Q2 18,7 13,0 18,8 8,0 8,1 10,6 7,2 4,0 2004Q3 19,7 7,2 7,3 3,6 2004Q4 21,5 8,1 7,8 3,9 2005Q1 15,8 20,3 9,3 11,1 7,7 4,6 2005Q2 18,6 15,5 18,9 8,6 7,7 10,9 7,2 4,1 2005Q3 16,8 20,7 7,7 10,4 7,3 3,8 2005Q4 18,4 14,1 22,4 7,3 7,7 10,5 7,9 3,9 2006Q1 14,2 22,6 7,7 11,0 8,0 4,0 2006Q2 17,2 13,3 20,8 6,0 7,1 9,9 7,2 3,4 2006Q3 14,9 20,8 6,3 9,2 7,2 3,1 2006Q4 16,8 12,6 22,5 6,2 6,8 9,5 7,1 3,1 2007Q1 16,3 12,1 21,0 7,1 6,8 9,3 7,4 3,2 2007Q2 15,4 12,1 18,6 6,1 6,1 8,2 6,6 2,6 2007Q3 15,5 12,6 17,7 5,5 5,9 8,0 6,6 2,4 2007Q4 15,0 10,9 19,3 5,2 6,0 7,9 6,4 2,4 2008Q1 15,0 10,7 17,6 5,7 6,2 8,0 6,3 2,5 2008Q2 15,2 10,9 17,3 5,6 5,8 7,5 5,8 2,3 2008Q3 15,7 11,1 18,1 4,9 5,7 6,7 5,9 1,9 2008Q4 16,5 9,6 21,3 5,0 6,3 6,7 6,4 2,1 2009Q1 19,0 11,1 23,0 6,5 7,5 7,8 7,3 2,5 2009Q2 19,7 11,6 22,7 6,3 7,5 7,4 7,2 2,7 2009Q3 20,5 12,2 22,6 6,6 7,6 7,4 7,4 2,8 2009Q4 20,4 10,1 24,8 7,0 8,0 7,0 8,1 3,2 2010Q1 21,6 11,2 23,4 9,6 8,9 7,7 8,3 4,0 2010Q2 20,9 9,6 22,2 8,9 8,3 6,8 7,5 3,7 2010Q3 20,6 10,3 22,7 8,3 8,0 6,4 7,5 3,5 2010Q4 20,8 8,4 23,2 8,1 8,3 6,4 8,0 3,5 2011Q1 21,5 9,3 23,3 8,1 8,6 6,5 8,0 3,9 2011Q2 20,8 8,6 21,2 6,9 8,1 5,6 7,4 3,7 2011Q3 21,3 9,1 20,7 7,4 8,1 5,4 7,7 3,6 2011Q4 21,9 7,4 23,4 8,3 8,7 5,2 8,2 4,1 2012Q1 23,0 8,0 23,1 9,9 9,4 5,8 8,6 4,4 2012Q2 22,5 8,1 22,0 9,2 9,0 5,1 8,1 4,3 2012Q3 22,7 9,1 23,4 9,2 8,9 5,0 8,2 4,4 2012Q4 23,4 7,4 27,1 9,6 9,4 5,0 8,7 4,8 2013Q1 24,1 7,8 25,4 11,1 10,1 5,7 8,9 5,7 2013Q2 23,1 7,7 23,1 10,6 9,6 5,1 8,3 5,9 2013Q3 23,0 8,7 23,1 11,2 9,2 4,8 7,9 6,0 2013Q4 23,0 7,4 24,1 11,2 9,4 4,8 8,7 6,1

(27)

Supplement 2: Regression outcomes _cons 3.7337 .4046474 9.23 0.000 2.92705 4.54035 YE 1.392656 .7905241 1.76 0.082 -.1832246 2.968538 EPL 2.313563 .558985 4.14 0.000 1.199247 3.427879 Yt -.0450882 .0480584 -0.94 0.351 -.1408909 .0507146 YC 1.267969 .9104082 1.39 0.168 -.5468962 3.082835 Y 4.035092 .5722578 7.05 0.000 2.894317 5.175867 t -.0423174 .0339825 -1.25 0.217 -.1100602 .0254254 C .697321 .6437558 1.08 0.282 -.5859827 1.980625 nl Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 489.9995 79 6.20252532 Root MSE = .95756 Adj R-squared = 0.8522 Residual 66.0180214 72 .916916964 R-squared = 0.8653 Model 423.981479 7 60.5687827 Prob > F = 0.0000 F( 7, 72) = 66.06 Source SS df MS Number of obs = 80 . regress nl C t Y YC Yt EPL YE _cons 2.835899 .4581062 6.19 0.000 1.923103 3.748696 Yt .0089552 .0496119 0.18 0.857 -.0898986 .107809 YC .7176374 1.146813 0.63 0.533 -1.567438 3.002712 Y 3.494658 .6478601 5.39 0.000 2.203769 4.785548 t .0474627 .0350809 1.35 0.180 -.0224375 .1173629 C -.2169229 .8109194 -0.27 0.790 -1.832715 1.398869 nl Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 489.9995 79 6.20252532 Root MSE = 1.2842 Adj R-squared = 0.7341 Residual 122.033071 74 1.64909556 R-squared = 0.7510 Model 367.966429 5 73.5932857 Prob > F = 0.0000 F( 5, 74) = 44.63 Source SS df MS Number of obs = 80 . regress nl C t Y YC Yt

(28)

_cons 7.174533 .6332942 11.33 0.000 5.906856 8.442209 YE 1.317587 .9293732 1.42 0.162 -.542756 3.177929 EPL 1.8163 .6571661 2.76 0.008 .5008388 3.131761 Yt -.0022244 .0634352 -0.04 0.972 -.1292038 .124755 YC 2.635798 1.032851 2.55 0.013 .5683231 4.703273 Y 9.852805 .8956133 11.00 0.000 8.06004 11.64557 t -.0412959 .0448555 -0.92 0.361 -.1310839 .0484921 C 1.89147 .7303357 2.59 0.012 .4295438 3.353395 eu Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2617.69955 65 40.2723007 Root MSE = 1.0199 Adj R-squared = 0.9742 Residual 60.3322658 58 1.04021148 R-squared = 0.9770 Model 2557.36728 7 365.338183 Prob > F = 0.0000 F( 7, 58) = 351.22 Source SS df MS Number of obs = 66 . regress eu C t Y YC Yt EPL YE _cons 6.063907 .5940612 10.21 0.000 4.875608 7.252206 Yt .0622295 .0536922 1.16 0.251 -.0451708 .1696299 YC 2.009793 1.133213 1.77 0.081 -.2569709 4.276556 Y 9.047131 .8401294 10.77 0.000 7.366622 10.72764 t .0475541 .0379661 1.25 0.215 -.0283894 .1234976 C 1.028518 .8013026 1.28 0.204 -.5743261 2.631361 eu Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2617.69955 65 40.2723007 Root MSE = 1.2378 Adj R-squared = 0.9620 Residual 91.9340337 60 1.5322339 R-squared = 0.9649 Model 2525.76551 5 505.153102 Prob > F = 0.0000 F( 5, 60) = 329.68 Source SS df MS Number of obs = 66 . regress eu C t Y YC Yt

(29)

_cons 11.88393 .4302387 27.62 0.000 11.02493 12.74293 YE .0864528 .7160611 0.12 0.904 -1.343209 1.516115 EPL .9149732 .5063316 1.81 0.075 -.0959505 1.925897 Yt -.0447721 .0471185 -0.95 0.345 -.1388472 .0493031 YC .0321953 .8123411 0.04 0.969 -1.589696 1.654087 Y 4.378032 .6084494 7.20 0.000 3.163224 5.592841 t -.2277081 .0333178 -6.83 0.000 -.2942292 -.1611869 C .6749421 .5744119 1.18 0.244 -.4719083 1.821792 de Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 603.360541 73 8.26521288 Root MSE = .81458 Adj R-squared = 0.9197 Residual 43.7942006 66 .663548493 R-squared = 0.9274 Model 559.56634 7 79.9380486 Prob > F = 0.0000 F( 7, 66) = 120.47 Source SS df MS Number of obs = 74 . regress de C t Y YC Yt EPL YE _cons 11.41074 .3541336 32.22 0.000 10.70408 12.11741 Yt -.0408415 .0353361 -1.16 0.252 -.1113537 .0296707 YC -.0080765 .7683799 -0.01 0.992 -1.541355 1.525202 Y 4.333322 .5008206 8.65 0.000 3.333951 5.332694 t -.1861093 .0249864 -7.45 0.000 -.2359689 -.1362496 C .248726 .5433266 0.46 0.649 -.8354654 1.332917 de Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 603.360541 73 8.26521288 Root MSE = .84503 Adj R-squared = 0.9136 Residual 48.5566178 68 .714067909 R-squared = 0.9195 Model 554.803923 5 110.960785 Prob > F = 0.0000 F( 5, 68) = 155.39 Source SS df MS Number of obs = 74 . regress de C t Y YC Yt

(30)

_cons 7.616373 .4640327 16.41 0.000 6.691341 8.541405 YE .8927097 .90654 0.98 0.328 -.9144447 2.699864 EPL 1.351807 .6410206 2.11 0.038 .0739557 2.629658 Yt -.0422544 .0551114 -0.77 0.446 -.152117 .0676082 YC 2.882953 1.044018 2.76 0.007 .8017413 4.964165 Y 13.25412 .6562413 20.20 0.000 11.94593 14.56232 t -.0516373 .0389697 -1.33 0.189 -.1293219 .0260473 C 1.341581 .7382322 1.82 0.073 -.1300579 2.81322 fr Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 4180.9795 79 52.9237911 Root MSE = 1.0981 Adj R-squared = 0.9772 Residual 86.8172796 72 1.20579555 R-squared = 0.9792 Model 4094.16222 7 584.880317 Prob > F = 0.0000 F( 7, 72) = 485.06 Source SS df MS Number of obs = 80 . regress fr C t Y YC Yt EPL YE _cons 7.091791 .4288886 16.54 0.000 6.237212 7.94637 Yt -.0076119 .0464477 -0.16 0.870 -.1001609 .084937 YC 2.530184 1.07367 2.36 0.021 .3908491 4.669518 Y 12.9077 .60654 21.28 0.000 11.69914 14.11626 t .0008209 .0328434 0.02 0.980 -.0646211 .0662629 C .8073916 .7591995 1.06 0.291 -.7053464 2.32013 fr Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 4180.9795 79 52.9237911 Root MSE = 1.2023 Adj R-squared = 0.9727 Residual 106.963113 74 1.44544747 R-squared = 0.9744 Model 4074.01639 5 814.803277 Prob > F = 0.0000 F( 5, 74) = 563.70 Source SS df MS Number of obs = 80 . regress fr C t Y YC Yt

The impact of the current economic crisis on youth unemployment in the Netherlands