long-term unemployed workers?
A case study of an innovative Dutch ALMP provider that matches unemployed
workers with temporary assignments, takes away administrative burdens and
makes work pay
Jelbert Jan van Ommeren, s1457772
Supervisor: Prof.dr. P.W.C. (Pierre) Koning
Second Reader: Prof.dr. M.G. (Marike) Knoef
Master Thesis ‘Innovative Active Labour Market Policies’
Leiden University, Public Administration: Economics & Governance
3
Abstract
Flexibilisation of the labour market poses both risks and opportunities for the actors that are involved in the reintegration process of long-term unemployed workers. Flextensie Nederland B.V., a Dutch social enterprise, aims to activate and to reintegrate welfare beneficiaries to the labour market by giving them the opportunity to accept temporary and flexible assignments and by making this work pay. The stepping-stone effect of these ‘Flextensie assignments’ (treatment) from unemployment to employment will be investigated in this thesis. A model of duration analysis, which compares the unemployment durations of treated candidates with the unemployment durations of non-treated candidates, will be used to estimate the size and significance of the stepping-stone effect of Flextensie assignments. The results show that participation in ‘Flextensie assignments’ does significantly and substantially decrease the unemployment durations of candidates within the first six months after such participation. Furthermore employer interviews are used to explore trends and selection effects regarding the demand side of the Flextensie-instrument. The interviews largely support earlier notions that employers are more willing to engage in welfare-to-work policies when this furthers their corporate social responsibility agendas, financial risks are limited, candidates are motivated and labour market shortages are experienced. We also find that candidates are selected for treatment based on their motivation. This implies that motivated candidates are overrepresented in the treatment group, indicating that the estimations suffer from upward bias. The direction and the significance of the estimated stepping-stone effect of participation in Flextensie assignments are however unambiguous and clearly indicate that temporary employment increases the reemployment opportunities of welfare beneficiaries in the short term.
4
Management Summary (Dutch)
Deze Nederlandstalige Management Summary vat de belangrijkste bevindingen en conclusies van de master thesis in ‘Innovative Active Labour Market Policies’ samen. In deze thesis wordt onderzocht of het door Flextensie Nederland B.V.1 ontwikkelde arbeidsmarkt instrument er in slaagt om de activatie en de reïntegratie van langdurig werklozen naar de arbeidsmarkt te bevorderen. Alle informatie die in deze samenvatting is gepresenteerd is terug te vinden in de thesis. De samenvatting vangt met de probleemstelling van het onderzoek en met enige noodzakelijke achtergrond informatie aan en presenteert achtereenvolgens de algemene statistieken van het Flextensie-instrument, de belangrijkste resultaten van de duuranalyse, de bevindingen van de interviews met betrokken werkgevers en de conclusies en discussiepunten van het onderzoek.
1. Probleemstelling en achtergrond
De flexibilisering van de arbeidsmarkt en de lagere budgetten voor arbeidsmarktbeleid genereren nieuwe uitdagingen voor gemeenten om de werklozen in hun bijstandsbestand te activeren en te reïntegreren. De alsmaar toenemende uitstroom van werklozen naar tijdelijk flexibel werk verhoogt de werklast van case-managers en leidt eveneens tot financiele onzekerheid bij de bijstandsgerechtigden zelf. De reïntegratie van met name de werklozen met een grotere afstand tot de arbeidsmarkt lijdt hieronder, aangezien het accepteren van flexibel werk voor hen vaak niet aantrekkelijk is door de administratieve druk en de financiele onzekerheid die dit met zich meebrengt.
Flextensie Nederland B.V. (Flextensie), een private onderneming die onderdeel uitmaakt van de Nederlandse Vereniging voor Sociale Innovatie, heeft een arbeidsmarktinstrument ontwikkeld dat flexibel werk aangrijpt als een kans voor langdurig werklozen om hun afstand tot de arbeidsmarkt te verkleinen op een financieel en administratief aantrekkelijke wijze voor de bijstandsgerechtigden, de werkgevers en de gemeenten. In de gemeenten die gebruik maken van het Flextensie-instrument kunnen werklozen zich vrijwillig inschrijven als Flextensie kandidaat. Lokale werkgevers kunnen vervolgens de lokale Flextensie-intercedent benaderen om geschikte kandidaten voor tijdelijke opdrachten te vinden. Wanneer er een match is stelt de gemeente het uurtarief, doorgaans gelijk aan het wettelijk minimum loon, voor de werkgever vast. De kandidaat behoudt tijdens de ‘Flextensie opdracht’ zijn of haar volledige uitkering en ontvangt daar bovenop een uurpremie van circa twee euro. Na aftrek van de door Flextensie gemaakte kosten vloeit het overgebleven deel van het uurtarief terug naar het reïntegratiebudget van de gemeente. Peels en Engelen (2017) hebben berekend dat het Flextensie-instrument kostendekkend is in de meeste gemeenten. Daarnaast
5 beargumenteren Flextensie, Peels en Engelen (2017) en de staatssecretaris van het Ministerie van Sociale Zaken en Werkgelegenheid dat de arbeidsmarkt-verdringing door het instrument beperkt en acceptabel is vanwege het tijdelijke karakter van Flextensie opdrachten en het arbeidsmarktconforme uurtarief.
Het doel van deze scriptie is om te onderzoeken of de participatie van werklozen in Flextensie opdrachten voor hen daadwerkelijk een opstap naar de arbeidsmarkt kan bieden. Flextensie en een aantal deelnemende gemeenten hebben hiervoor data over de kandidaten, de aangenomen opdrachten en de uitstroom van kandidaten naar arbeidsmarktcontracten ter beschikking gesteld. Het effect van participatie in Flextensie opdrachten kan geschat worden door middel van een duurmodel dat de werkloosheidsduren van de kandidaten die hebben deelgenomen aan een Flextensie opdracht (treatment groep) vergelijkt met de werkloosheidsduren van de kandidaten die geen Flextensie-opdracht hebben doorlopen (controle groep).
Ter aanvulling zijn er vier interviews gedaan met werkgevers die het Flextensie-instrument hebben gebruikt om tijdelijk personeel in te huren. Deze interviews bieden verdere diepgang in de motivaties van werkgevers en de selectie van kandidaten voor de Flextensie opdrachten en eventueel vaste(re) contracten. Achtereenvolgens zullen de belangrijkste statistieken van het Flextensie-instrument, de resultaten van de duuranalyse, de trends in de interviews en de conclusie van de scriptie samengevat worden.
2. Statistisch overzicht
Uit Tabel 1 kan worden afgelezen dat een derde (0,33) van alle Flextensie kandidaten die zich in 2016 of 2017 hebben ingeschreven heeft deelgenomen in een Flextensie opdracht. Van bijna 1.000 kandidaten is bekend of zij wel of niet zijn uitgestroomd naar een arbeidsmarktcontract, dit zijn kandidaten die zijn ingeschreven in Almere, Utrecht, Zaanstad of een gemeente in het samenwerkingsverband ‘WerkSaam Westfriesland’. In totaal ligt de uitstroom van kandidaten naar een contract op zo’n 16 procent in 2016-2017. In Almere, Utrecht en Zaanstad heeft ruim een derde (0,35) van de kandidaten die hebben deelgenomen aan een Flextensie opdracht een arbeidsmarktcontract getekend tegenover ongeveer zes procent van de controle groep.
De beschikbare gegevens tonen aan dat ruim de helft van de geplaatste kandidaten (0,54 totaal en 0,72 in Almere, Utrecht en Zaanstad) een contract tekent bij de werkgever waarvoor zij eerder een Flextensie opdracht hebben gedaan. Het grootste deel van de kandidaten (0,89) tekent een contract voor meer dan twaalf uur per week. Gemiddeld verbleven kandidaten ruim twee jaar in de bijstand voordat zij zich inschreven bij Flextensie en hebben zij de middelbare school of een MBO-opleiding voltooid.
6
Tabel 1 – Beschrijvende statistieken van de kandidaten die ingeschreven zijn in 2016 of 2017
Kandidaten 2016-2017 Treatment Controle Totaal
Almere, Utrecht en Zaanstad Alle gemeenten
Gem. Obs. Gem. Obs. Gem. Obs.
Dagen in bijstand tot inschrijfdatum 781,82 143 818,62 396 835,54 1.035
Leeftijd 41,82 215 - - 42,08 678
Man 0,52 220 - - 0,56 601
Voltooide opleiding 2,74 105 - - 2,74 228
Uitstroom naar contract 0,35 230 0,06 615 0,16 986
Uitstroom naar zelfde werkgever 0,72 64 - - 0,54 123
Uitstroom >12 uur / week 0,88 72 0,96 28 0,89 105
Uitkeringsgemeente Almere 0,39 230 0,31 615 0,13 2.181
Uitkeringsgemeente Utrecht 0,32 230 0,36 615 0,14 2.181
Uitkeringsgemeente Zaanstad 0,29 230 0,33 615 0,13 2.181
Deelname aan Flextensie opdracht 1,00 230 0,00 615 0,33 2.175
Aantal Flextensie opdrachten 1,27 223 - - 1,25 689
Aantal kandidaten 230 615 2.181
De inhoud van ruim 1.000 Flextensie opdrachten is beschreven in Tabel 2. Ongeveer een derde (337 van de 1.122) van de opdrachten zijn voortijdig beëindigd en Flextensie opdrachten hebben een gemiddelde looptijd van tien weken en 24 uur per week. De standaarddeviatie voor deze twee variabelen is groot wat betekend dat er veel variatie is. Het uurtarief dat werkgevers betalen ligt rond de twaalf euro en de kandidaten ontvangen zo’n twee euro premie per gewerkt uur.
Tabel 2 – Statistieken van Flextensie opdrachten
Gemiddelde Std. Dev. Obs.
Duur Flextensie opdracht (in dagen) 72 64,94 1.122
Voortijdig eind van Flextensie opdracht - - 337
Uren / week Flextensie opdracht 24 11,98 808
Uurtarief voor werkgevers (in Euro’s) 11,87 0,57 1.082
Premie / uur voor kandidaten (in Euro’s) 1,94 1,64 681
3. Resultaten van het duurmodel
Het duurmodel maakt het mogelijk om het effect van participatie in Flextensie opdrachten op de werkloosheidsduur van kandidaten te schatten. Tabel 3 presenteert de bevindingen van de duuranalyse en schat daarbij de langere termijn effecten van participatie in Flextensie opdrachten op de werkloosheidsduur van kandidaten. Kandidaten die geplaatst zijn op een Flextensie opdracht worden gevolgd tot 91, 182, 274 en 365 dagen na het begin van hun plaatsing.
Uit Tabel 3 blijkt dat participatie in een Flextensie opdracht de werkloosheidsduur van kandidaten substantieel en significant verlaagd tot zes maanden na participatie. Het treatment effect wordt insignificant en verdwijnt wanneer een kandidaat nog niet is uitgestroomd binnen negen maanden na participatie. Verder blijkt dat de werkloosheidsduur in Utrecht significant lager en in Zaanstad significant hoger is, alhoewel robustheidscontroles het negatieve effect voor Zaanstad lager
7 inschatten. Er worden geen significante effecten op de werkloosheidsduur van kandidaten gevonden voor hun leeftijd, geslacht, voltooide opleidingsniveau en uitkeringsduur. Wel laten de data een trend zien die lijkt te impliceren dat vrouwelijke kandidaten een kortere werkloosheidsduur hebben dan mannen. Dit is in lijn met diverse andere studies, bijvoorbeeld Martin en Grubbs (2001) en Jongen et. al (2015), die vinden dat arbeidsmarktinstrumenten de arbeidsmarktparticipatie van vrouwen doorgaans meer bevorderen dan de arbeidsmarktparticipatie van mannen.
Tabel 3 – Effect van participatie in Flextensie opdrachten over tijd
De afhankelijke variabele ‘duur-tot-uitstroom’ (werkloosheidsduur) is gemeten als de duur tot Flextensie kandidaten uitstromen naar een arbeidsmarktcontract
Model 1 91 dagen Model 2 182 dagen Model 3 274 dagen Model 4 365 dagen
Coeff. Haz. Coeff. Haz. Coeff. Haz. Coeff. Haz.
Treatment effect 1.22*** (0.36) 3.40 1.73*** (0.23) 5.65 1.91*** (0.21) 6.75 1.80*** (0.20) 6.06 Langere termijn 1.80*** (0.22) 6.02 1.53*** (0.28) 4.61 0.22 (0.54) 1.24 -0.56 (1.03) 0.57 Gemeente Referentie: 1. Almere 2. Utrecht 0.72** (0.23) 2.04 0.72** (0.23) 2.04 0.73** (0.23) 2.07 0.72** (0.23) 2.05 3. Zaanstad -1.48*** (0.31) 0.23 -1.49*** (0.31) 0.23 -1.46*** (0.31) 0.23 -1.45*** (0.31) 0.23 Leeftijd 0.02 (0.01) 1.02 0.02 (0.01) 1.02 0.01 (0.01) 1.01 0.01 (0.01) 1.01 Man -0.43 (0.26) 0.65 -0.43 (0.26) 0.65 -0.42 (0.26) 0.66 -0.43 (0.26) 0.65 Opleiding 0.04 (0.18) 1.05 0.03 (0.18) 1.04 0.05 (0.18) 1.05 0.03 (0.18) 1.03 Uitkeringsduur -0.00 (0.00) 1.00 -0.00 (0.00) 1.00 -0.00 (0.00) 1.00 -0.00 (0.00) 1.00 Observaties 2,439 Kandidaten 1,256 Uitstroom 106
8 Het opsplitsen van het effect van Flextensie opdrachten op de werkloosheidsduur van kandidaten per gemeente levert eveneens interessante resultaten op, welke zijn gepresenteerd in Tabel 4. Opvallend is dat het effect van participatie in Flextensie opdrachten op de werkloosheidsduur, hoewel nog steeds sterk aanwezig en significant, met afstand het kleinst is in de gemeente Utrecht. Voor een groot deel wordt dit verklaard doordat de Flextensie database in de gemeente Utrecht vaker wordt gebruikt om uitkeringsgerechtigden te koppelen aan reguliere vacatures. Hierdoor stroomt een groter deel van de controle groep uit naar werk dan in de andere gemeenten het geval is, wat leidt tot een relatief lager treatment effect van Flextensie opdrachten op de werkloosheidsduur van kandidaten in Utrecht. Dit laat zien dat inschrijving bij Flextensie op zichzelf de reïntegratiekansen van bijstandsgerechtigden al kan verhogen. De coefficient voor mannelijke kandidaten in Almere is tenslotte sterk negatief en licht significant, wat impliceert dat mannelijke kandidaten doorgaans een langere werkloosheidsduur hebben dan vrouwelijke kandidaten.
Tabel 4 – Effect van participatie in Flextensie opdrachten per gemeente
De afhankelijke variabele ‘duur-tot-uitstroom’ (werkloosheidsduur) is gemeten als de duur tot Flextensie kandidaten uitstromen naar een arbeidsmarktcontract
Model 1 Model 2 Model 3
Almere Utrecht Zaanstad
Coeff. Hazard Coeff. Hazard Coeff. Hazard
Treatment effect 2.49*** (0.47) 12.05 1.05*** (0.28) 2.86 2.07*** (0.53) 7.92 Leeftijd 0.03 (0.02) 1.03 0.03 (0.02) 1.03 0.00 (0.03) 1.00 Man -0.91* (0.44) 0.40 -0.01 (0.41) 0.99 -0.77 (0.63) 0.46 Opleiding 0.22 (0.35) 1.24 0.10 (0.27) 1.11 0.00 (0.33) 1.00 Uitkeringsduur -0.00 (0.00) 1.00 0.00 (0.00) 1.00 -0.00 (0.00) 1.00 Observaties 555 548 1,336 Kandidaten 282 306 668 Uitstroom 29 58 19
9
4. Werkgever interviews
Er zijn vier interviews afgenomen met werkgevers die gebruikmaken van het Flextensie-instrument. Het nemen van sociale verantwoordelijkheid in combinatie met de betaalbaarheid (i.e. lage financiële risico’s) van het instrument zijn de belangrijkste drijfveren voor werkgevers om Flextensie kandidaten in te huren voor tijdelijke klussen. Daarnaast vinden zij dat de kandidaten vaak zeer gemotiveerd zijn en is het instrument handig om eventuele tekorten op de arbeidsmarkt op te vangen. Deze bevindingen zijn in lijn met de bevindingen van Peels en Engelen (2017).
De meeste werkgevers bieden laagdrempelige arbeid aan waardoor zij kandidaten niet selecteren op basis van hun werkervaring en opleidingsniveau, maar voornamelijk naar de motivatie van de kandidaat kijken. De meeste werkgevers geven dan ook aan dat zij bereid zijn om kandidaten een vaster contract aan te bieden wanneer de kandidaat gemotiveerd is en goed werk heeft geleverd tijdens duur van een Flextensie opdracht.
Deze selectiemethode impliceert echter wel dat gemotiveerde bijstandsgerechtigden overgerepresenteerd zijn in de treatment groep. Dit veroorzaakt naar alle waarschijnlijkheid een overschatting van de treatment effecten. Het is echter aantoonbaar dat selectie niet de enige drijfveer is van de geschatte treatment effecten, aangezien de ‘gemotiveerde treatment groep’ dan ten allen tijde een substantiële en significante voorsprong zou moeten hebben op de ‘minder gemotiveerde controle groep’. Tabel 3 toont echter aan dat dit niet het geval is, wat impliceert dat tijdelijk werk daadwerkelijk een tijdelijke opstap biedt naar regulier werk voor bijstandsgerechtigden.
5. Conclusie en discussie
De resultaten van de duuranalyse zijn robuust en tonen aan dat participatie in Flextensie opdrachten de werkloosheidsduur van kandidaten sterk en significant verminderd. Een kanttekening daarbij is dat kandidaten die een half jaar na hun plaatsing nog niet zijn uitgestroomd naar de arbeidsmarkt opnieuw geplaatst moeten worden of een andere vorm van begeleiding nodig hebben met betrekking tot hun zoektocht naar een baan, aangezien hun opgedane voordeel op de arbeidsmarkt daarna verwatert. Deze bevindingen zijn in overeenstemming met De Graaf-Zijl et. al (2011) die eveneens vinden dat tijdelijk werk de werkloosheidsduur van bijstandsgerechtigden verkort, maar niet hun baankansen vergroot in de lange termijn. Dit korte termijn effect wordt verklaart doordat bijstandsgerechtigden aan hun sociale netwerk en menselijk kapitaal bouwen tijdens dergelijke periodes van tijdelijk werk, maar niet tijdens de daaropvolgende periodes van werkloosheid. Samengevat kan het Flextensie-instrument dus een grote stap van werkloosheid naar de arbeidsmarkt voor bijstandsgerechtigden betekenen, terwijl de kansen op reïntegratie voor deze groep werklozen doorgaans zeer beperkt zijn.
10 Leeftijd, geslacht, opleiding en uitkeringsduur hebben geen consistent significante invloed op de werkloosheidsduur van kandidaten, alhoewel er wel een trend lijkt te zijn die onderschrijft dat vrouwelijke kandidaten meer profiteren van hun deelname in het Flextensie-instrument dan mannelijke kandidaten. Uit de interviews met werkgevers blijkt dat met name de motivatie van kandidaten belangrijk is om geselecteerd te worden voor een Flextensie opdracht en om eventueel uit te stromen naar een contract. Dit veroorzaakt wel een ‘upward bias’ van de resultaten, doordat gemotiveerde kandidaten overgerepresenteerd zijn in de treatment groep, maar het is aantoonbaar gebleken dat dit de kern van de bevindingen niet onderuit haalt.
De duurzaamheid van de onderzochte werkloosheid-naar-werk transities is niet aan bod geweest in deze scriptie aangezien er nog weinig informatie aanwezig is over soorten en duren van contracten die zijn aangegaan door Flextensie kandidaten, wat daarmee een interessant onderwerp voor vervolgonderzoek blijft. Daarnaast kan vervolgonderzoek naar vergelijkbare arbeidsmarktinstrumenten bijdragen aan de generaliseerbaarheid van de bevindingen met betrekking tot de inzet van tijdelijk werk als arbeidsmarktinstrument. Het onderzoek naar de effectiviteit van actief arbeidsmarktbeleid, zoals flexibel werk, op de werkloosheidsduur van bijstandsgerechtigden zou daarbij bevordert worden door een centraal databeheer van arbeidsmarkttransities en de inzet van arbeidsmarktinstrumenten op gemeentelijk en landelijk niveau.
11
Table of Contents
Abstract ... 3
Management Summary (Dutch) ... 4
1. Introduction ... 13
2. Literature Review ... 15
2.1 Introduction to ALMPs ... 15
2.2 ALMP categories: Studying ALMP effectiveness ... 16
2.2.1 Category 1: Employment Services ... 17
2.2.2 Category 2: Labour market training ... 18
2.2.3 Category 3: Subsidised employment, job creation and temporary employment ... 18
2.2.4 Category 4: Labour market activation measures and financial incentives ... 19
2.3 The supply side of ALMPs: Public authorities and quasi-markets ... 20
2.4 The demand side of ALMPs: Employer engagement ... 22
3. The Flextensie Instrument ... 23
3.1 The functioning of the Flextensie-instrument ... 24
3.2 Lack of opportunities for long-term unemployed workers ... 25
3.3 Financial insecurity ... 26
3.4 Administrative burdens ... 28
3.5 Adverse side effects and gaming activities ... 28
3.6 Employer involvement ... 30
4. Research Methods ... 30
4.1 Introduction to the duration analysis research design ... 30
4.2 Employer interviews ... 32
5. Description of the Data ... 32
5.1 Description of the datasets ... 32
5.2 Summary statistics... 34
5.3 Operationalization of the data ... 37
6. Research Model ... 39
6.1 Duration analysis model ... 39
6.2 Assumptions of the Cox Proportional Hazards model ... 41
6.3 Null-hypothesis ... 43
7. Duration Analysis ... 44
7.1 Descriptive analysis ... 44
12
7.3 Robustness checks ... 51
7.4 Alphen aan den Rijn and Leeuwarden ... 54
8. Employer Interviews ... 55
8.1 Motives for contracting Flextensie candidates ... 55
8.2 Candidate selection ... 57
9. Conclusion ... 59
References ... 61
Appendix I – Flextensie population statistics ... 65
Appendix II – Graphic PH-assumption tests ... 68
Appendix III – Violations of the PH-assumption ... 71
Appendix IV – Longer run effects, by municipality ... 74
Appendix V – Alphen aan den Rijn and Leeuwarden ... 77
Appendix VI – Transcripts of employer interviews (Dutch) ... 81
Transcript respondent 1 ... 81
Transcript respondent 2 ... 83
Transcript respondent 3 ... 84
13
1. Introduction
In 2013 the Dutch government announced to update the ‘Employment and Assistance Act’ (Wet werk en bijstand; WWB) with a set of new measures aiming to increase the pressure on welfare beneficiaries to reintegrate to the labour market quickly2. At the same time, a period of retrenchment has caused the Dutch government to cut down almost half of its expenses on ‘active labour market policies’ (ALMPs) from 2002-2013 (CPB, 2016: 163). Divosa, the association of public caseworkers for welfare recipients, therefore voiced concerns that these new regulations would increase the bureaucratic workload of municipality case-managers and decrease their space for a more customised approach3.
Next to the new measures and the lower reintegration budgets a continuing process of labour market flexibilisation has further increased the workload of municipality case-managers. A labour market that grows more flexible implies that the outflow of welfare beneficiaries to temporary and part-time jobs increases which in turn causes their labour market incomes to fluctuate. Case-managers accordingly have to invest an increasing amount of time in recalculating the level of a clients’ benefits eligibility. For welfare beneficiaries in turn this generates financial insecurity as additionally earned market income results in a decrease or even in a cancellation of their welfare benefit payments. Such financial insecurity makes unemployed workers more reluctant to accept flexible jobs. It is calculated that the share of flexible and temporary labour in the Netherlands has increased from 14 per cent in 2003 to 22 per cent in 2016, which is one of the highest rates in all of Europe (SZW, 2017: 5-6)4.
This research will focus on Flextensie Nederland B.V., which is a private actor in the Dutch market of welfare-to-work services and has developed an innovative active labour market instrument that aims to support the labour market reintegration of welfare beneficiaries. The Flextensie-instrument matches welfare beneficiaries with temporary flexible jobs (Flextensie assignments) and makes work pay as candidates remain fully eligibly to welfare benefit payments during their period of temporary employment and receive an additional hourly fee. The idea is that candidates are no longer reluctant to accept flexible jobs and can thus start to acquire (recent) work experience, to increase their human capital and to decrease their distance to the labour market in general.
2 Overheid.nl (2013), ’33 801: Wijziging van de Wet werk en bijstand en enkele andere sociale
zekerheidswetten (Wet maatregelen Wet werk en bijstand en enkele andere wetten)’,
https://zoek.officielebekendmakingen.nl/kst-33801-3.html (visited at 26-10-2017).
3 Divosa (2013), ‘Divosa: WWB-maatregelen belemmeren gemeenten bij uitvoering’,
https://www.divosa.nl/nieuws/divosa-wwb-maatregelen-belemmeren-gemeenten-bij-uitvoering (visited at 26-10-2017).
4
CBS (2015), ‘Sterke toename flexwerk’, https://www.cbs.nl/nl-nl/nieuws/2015/23/sterke-toename-flexwerk
14 Flextensie assignments may therefore act as stepping stones for the activation and the reintegration of long-term unemployed workers to the labour market.
Flextensie Nederland B.V. has entered the market for ‘welfare-to-work services’ in late 2014 and they provided their services to over forty Dutch municipalities in late 2016 already. Their own estimates suggest that ± 25 per cent of their candidates, having received benefits for three years on average, re-enters the labour market after having participated in a Flextensie assignment. This percentage is remarkably high as normally the odds of reintegrating to the labour market after having been unemployed for such a long time are about 8 per cent5.
The research question of this thesis is thus phrased as followed:
To what extent is the Flextensie-instrument able to provide a stepping stone towards the activation and the reintegration of long-term unemployed workers?
Flextensie’s quick expansion did result in a media campaign where critics argued that they have developed an instrument that promotes cheap forced labour at the expense of the weakest members in society6. This wave of criticism was picked up by several national parliamentarians and the ‘Ministry of Social Affairs and Employment’ (SZW) responded to the criticism by commissioning an investigation of the instrument. This policy research has been executed by Peels and Engelen (2017), part of research agency ‘De Beleidsonderzoekers’, and has presented a theoretical and statistical overview of the instrument, its participants and its effects.
Flextensie has provided this research with the same data as ‘De Beleidsonderzoekers’ on their candidates and placements on ‘Flextensie assignments’. Furthermore Flextensie has provided up-to-date outflow data of candidates for several municipalities. With their data, Peels and Engelen (2017: 12, 56) find that 33 per cent of all candidates that have been placed on a Flextensie assignment make the transition from unemployment to the labour market, although it remains unclear how structural this job finding rate is as long-term effects are still unknown and data on employment and contract types is limited.
The aim of this research is to increase the scientific understanding of the effectiveness of temporary jobs as a stepping stone for the activation and the reintegration of long-term unemployed workers. This will be done by conducting a literature review and by analysing the data made available by Flextensie through an econometric model. As an experimental research design is out of reach,
5
Flextensie Nederland B.V., ‘Welkom bij Flextensie’, https://flextensie.nl/ (visited at 27-08-2017).
6
Doorbraak.eu (2017), ‘Campagne tegen Flextensie komt op stoom’, www.doorbraak.eu/campagne-flextensie-komt-op-stoom/ (visited at 27-08-2017).
15 data on the timing of Flextensie enrolments, Flextensie placements and Flextensie outflow will be analysed in a ‘duration analysis model’.
It is furthermore interesting to investigate the employers’ side in the instrument as the success of ALMPs much depends on the willingness of private employers to participate (Ingold & Stuart, 2015: 443-444). According to the interviews conducted by Peels and Engelen (2017: 28-32) employers are satisfied with the Flextensie-instrument for various reasons, ranging from corporate social responsibility motives to the limited financial risks related to the instrument. Additional employer interviews will be conducted in this thesis in order to learn more about the motives of employers and the selection processes of candidates.
The results of the literature review are discussed in the next chapter. The framework of this analysis is inspired by The Economics of Imperfect Labour Markets by Tito Boeri and Jan van Ours (2013). Chapter 3 will then discuss the Flextensie-instrument more extensively. The research methods, a description of the data and the research model will be discussed in Chapters 4, 5 and 6 respectively. The results of the duration analysis and the employer interviews are presented and discussed in Chapters 7 and 8 and are followed by the conclusion and discussion in Chapter 9.
2. Literature Review
Boeri and van Ours (2013) describe four categories of ALMPs that will be used in this literature review, namely public employment services, labour market training programs, subsidised
employment measures and labour market activation measures. The Flextensie-instrument doesn’t
exactly fit into one of these types, but instead contains elements that can be attributed to all four types. The literature review will start with a general introduction to ALMPs, followed by discussions of the four ALMP-categories, their applicability to the Flextensie-instrument and their effectiveness. The last two paragraphs of this literature review will then focus on the supply and demand sides of ALMPs respectively.
2.1 Introduction to ALMPs
ALMPs are mainly implemented in order to improve the functioning of the labour market. Such an improvement can be achieved when ALMPs enhance labour market mobility and adjustment, provide ways to invest in the human capital of unemployed workers and lower unemployment and welfare benefit expenses by reintegrating unemployed workers to the labour market. Besides that, ALMPs are also meant to overcome the moral hazard problem that is caused by ‘Unemployment Insurance’ (UI). Modern welfare states with more generous UI systems tend to experience that a larger share of unemployed workers receive disincentives to quickly reintegrate to the labour
16 market. These UI systems either run through unemployment benefit (UB) or through welfare benefit (WB) schemes. Individuals are eligible for UB schemes when they have worked for a specific period of time and when their former employer has contributed into the state’s UB fund. WB schemes, on the other hand, are tax funded and are meant to support individuals that have fallen below certain specified thresholds for income and assets (Boeri & van Ours, 2013: 307-308, 351 and 369; CPB, 2016: 21).
The disincentives to find a job can be further explained with three traps that may be experienced by unemployed workers. The (1) unemployment trap discourages job search efforts when benefit levels for unemployed workers and their families are relatively high compared to net market income levels. In this situation individuals may experience more utility from staying in benefit schemes rather than finding a job. The (2) inactivity trap concerns individuals that are able to work, but have fallen under benefit schemes different from the UB scheme, for instance WB or disability schemes. Low net gains from work and financial insecurity provide disincentives to find a job in this situation as benefits tend to gradually decline or disappear as soon as these individuals start to receive a labour market income. Lastly the (3) poverty/low-wage trap discourages low-paid workers to work more or to move to higher wage scales because, due to disappearing income-tested benefits, increases in gross earnings do not result in a higher net income (Boeri & van Ours, 2013: 369).
2.2 ALMP categories: Studying ALMP effectiveness
This paragraph will discuss the difficulties in the study of ALMP effectiveness in general as well as the adverse side effects that can influence the effectiveness of ALMPs. The following four paragraphs will then continue to discuss the four categories of ALMPs, how they can be related to the Flextensie-instrument and their effectiveness.
ALMP effectiveness found in micro studies tends to differ from the effects found in macro studies, which can largely be attributed to the so-called micro-macro paradox in ALMP studies. Micro studies have the advantage of having many observations, but suffer from a selection bias and are only able to provide estimates of partial-equilibrium effects. Macro studies, on the other hand, use aggregate data to measure the effectiveness of ALMPs on aggregate employment levels, but are limited by a few number of observations, have to put a multitude of differing ALMPs into a few ‘simplified’ categories and have to deal with simultaneity bias (Boone & van Ours, 2007). Such a simultaneity bias occurs when the explanatory variable X causes the dependent variable Y, but Y causes X as well (Toshkov, 2016: 206), which often leads to an overestimation of ALMP effectiveness. For example, an estimated change in aggregate unemployment levels may be caused by both the reaction of a government to unemployment and the response of unemployment to ALMPs (Calmfors, 1994).
17 Important to keep in mind is that the effects of ALMPs on job finding rates that are found in macro as well as micro studies tend to be small in general. In part this is explained by the adverse side effects that are related to ALMPs. Locking-in (1) occurs when the participation of unemployed workers in ALMPs decreases their job search intensity, whereas crowding out (2) occurs when ALMPs reintegrate the one unemployed worker more quickly at the expense of another. In this case ALMPs may be effective for individual unemployed workers, but turn out to be ineffective on aggregate levels of unemployment (Boone & van Ours, 2004; Calmfors, 1994).
Closely related to the crowding out effect are the displacement effect (3), where jobs created by an ALMP program come at the expense of regular jobs, and the substitution effect (4), where jobs created for a certain category of workers replace jobs for other categories due to changes in relative wages. Furthermore there could be dead weight effects (5) when ALMPs subsidise hiring that would have occurred anyways and taxation effects (6) when the financing of ALMPs affect everyone’s behaviour in society (Boone & van Ours, 2004; Calmfors, 1994).
2.2.1 Category 1: Employment Services
Employment services are provided to all workers that enter the UB scheme and are aimed to assist them in finding a new job. Examples of employment services provided to unemployed workers are services that support them with UB related administrative issues, job counselling and job search courses, and provide them with information on relevant vacancies, training programs and other available ALMPs (Boeri & van Ours, 2013: 351). The services provided by the Flextensie-instrument can partly be considered as employment services, since the instrument aims to support Flextensie candidates in reconnecting them to the labour market by matching them with temporary vacancies.
By conducting a meta-analysis of 199 ALMPs evaluated in 97 micro studies between 1995 and 2007, Card et. al (2010) find that employment services are relatively effective in both the short and the longer run. By analysing country level ALMP data of twenty OECD countries over 1985-1999, Boone and van Ours (2007) also find that employment services have some positive effect on bringing down unemployment. In their analysis of literature and analytical studies of ALMPs, Martin and Grubb (2001) similarly find that employment services work for most unemployed workers and tend to be somewhat more effective for women and single parents.
Key to the good functioning of employment services are good management in the form of active placement efforts, stimulation of participants’ motivation to reintegrate and strong monitoring and control mechanisms (Martin & Grubb, 2001). Furthermore, good targeting and timing is important as individuals have different needs and resources are limited and therefore need to be allocated as optimally as possible (Boeri & van Ours, 2013: 351, 367; CPB, 2016: 161).
18
2.2.2 Category 2: Labour market training
Labour market training programs aim to increase the attractiveness of their participants for employers by updating and extending their skillsets. Training programs can be made available for unemployed workers as well as for workers at risk of losing their job and even employed workers (Boeri & van Ours, 2013: 351). Flextensie assignments can be regarded as a form of labour market training as they aim to increase the human capital and labour market experience of the participants. In some municipalities, the Flextensie-instrument is specifically used as a type of trial placement for candidates where they can gain specific on-the-job experience and potentially enter into a more sustainable relationship with the employer (Peels & Engelsen, 2017: 10).
Card et. al (2010) find that the effectiveness of labour market training programs increases over time and starts showing relatively positive effects after a period of two years. The findings of Boone and van Ours (2007) and Martin and Grubb (2001) confirm this, although Martin and Grubb (2001) more specifically find that training tends to work best for women and single mothers. It is important however that training programs meet labour market needs, have strong on-the-job components and result in valuable certifications.
2.2.3 Category 3: Subsidised employment, job creation and temporary employment
Subsidised employment measures are used to create or to stimulate employment for targeted and usually disadvantaged groups (Bouri & van Ours, 2013: 351). As will be discussed more extensively in Chapter 3, Flextensie assignments are not so much a form of subsidised employment. The Flextensie-instrument is however targeted on the disadvantaged group of long-term unemployed workers and does, in some cases, stimulate the creation of new flexible jobs (Peels & Engelsen, 2017: 42-43).
Card et. al (2010) and Boone and van Ours (2007) find that subsidised jobs are relatively ineffective over both the short and the longer run. Martin and Grubb (2001) similarly find that subsidised employment and direct job creation are ineffective for most unemployed workers, but also find that subsidised employment can be effective for long-term unemployed workers. In his evaluations of a natural experiment in Slovakia’s labour market, van Ours also finds that subsidised employment can be effective in increasing job finding rates and decreasing job separation rates. The effectiveness of these temporary subsidised jobs did however significantly decrease when the Slovakian authorities decided to increase the durations of these jobs. This drop in effectiveness can be attributed to the occurrence of locking-in effects as the increased duration of the subsidised employment programs likely reduced the job search intensity of the participants (van Ours, 2000; van Ours, 2002).
The literature that specifies on temporary rather than subsidised employment measures also finds contradicting effects of such ALMPs. De Graaf-Zijl et. al (2011) find a positive effect of
19 temporary jobs as a stepping stone for regular work as they find that temporary jobs significantly reduce unemployment durations. However they do not find this effect to remain significantly positive in the longer run. Contrary to the findings of De Graaf-Zijl et. al (2011), Babos’ (2014) concludes in his macro study, using data of Central Eastern European countries over 2005-2010, that temporary jobs are more often a trap than a step. However, Babos’ (2014) findings further show that there are large differences between countries and that his results tend to be more positive when they are narrowed down to unemployed workers. Atkinson and Williams (2003) argue that a stepping-stone effect of temporary work to regular employment may appear to be limited from the currently available research, but that the participants of such programs tend to belong to the more disadvantaged groups of jobseekers and that a non-existence of such stepping stones would significantly damage “their ability to progress through the labour market”.
According to Martin and Grubb (2001), the effectiveness of subsidised employment depends on tight targeting, short durations and close monitoring of employer behaviour. There is however a risk that, if tight targeting is closely tied to “disadvantage” characteristics, participating unemployed workers become stigmatised and don’t really benefit from these programs. In line with subsidised employment in general, the use of directly created jobs should also be short in duration in order to avoid the locking-in of participants and they should be closely targeted to the most disadvantaged groups of unemployed workers.
2.2.4 Category 4: Labour market activation measures and financial incentives
Labour market activation measures aim to provide ‘stick’ incentives to unemployed workers in order to increase their job search intensity. Examples of such measures are mandatory participation for beneficiaries in training programs, monthly job application requirements and benefit sanctions when beneficiaries fail to meet certain obligations (Boeri & van Ours, 2013: 351-352). Opposite to such ‘stick’ incentives, the Flextensie-instrument aims to activate unemployed workers voluntarily by using ‘carrot’ incentives; namely by providing them with financial security and an additional fee per hour worked.
The studies of van den Berg et. al (2004) and Abbring et. al (2005) measured the reaction of Dutch welfare beneficiaries on benefit sanctions and found that the imposed benefit sanctions significantly and largely increased the transition rates of unemployed individuals to work. Therefore, changes in benefit levels appear to have large impacts on the job search behaviour of welfare recipients, which especially applies to benefit sanctions as it tends to be difficult to smooth consumption for welfare recipients. The Dutch Central Planning Agency acknowledges that benefit sanctions and lower benefit levels increase labour market participation. However, when making decisions regarding benefit levels and sanctions, authorities also have to take into consideration that
20 lower benefit levels and benefit sanctions increase the risks of poverty and backlashes in income levels and also might lower the average productivity (CPB, 2015: 12-13, 76).
By studying the effectiveness of ‘fiscal participation policy’ in the Netherlands, Jongen et. al (2015) find that financial incentives are most effective for women with young children. They further find that policies that increase the income gap between ‘working’ and ‘not working’ are more effective in increasing labour market activation than financial incentives that work through changes in the taxation system or a guaranteed standard income. Another interesting note made by Jongen et. al (2015) is that the effect of financial incentives on labour market participation in the Netherlands has decreased over the last couple of decades. This is explained by the increase of the structural level of labour market participation in the same period, thus making further decreases in unemployment more difficult and financial incentives less effective.
Evaluations of re-employment bonus programs for UI recipients that managed to find a job quickly in Japan and Korea found that this financial incentive significantly reduced unemployment durations (Martin & Grubb, 2001). Studies of the Canadian ‘Self-Sufficiency Project’ (SSP) also found that financial incentives are able to increase both the level of earnings and the stable (full-time) employment rates of participating welfare beneficiaries. The SSP project itself turned out to run break-even as the costs of the financial incentives were levelled by the decreased income assistance payments and the increased payroll tax revenues (Lei & Michalopoulos, 2001; Lin et. al, 1998; Michalopoulos et. al, 1999; Michalopoulos & Hoy, 2001).
The Canadian SSP-program was designed in a way that ensured that welfare beneficiaries would, from a financial point of view, be better off when they would participate in the program rather than to remain in their welfare benefit schemes. In other words, the program was designed to overcome the unemployment, inactivity and poverty traps discussed earlier in this chapter. Evaluations found that four years after the program had started the share of unemployed workers that had been in a full-time job for at least one year was nine percentage points higher for SSP participants than for beneficiaries in the control group. The evaluation results for beneficiaries that participated in a SSP Plus program, that also offered additional pre- and post-employment services, turned out to be even more positive (Lei & Michalopoulos, 2001; Lin et. al, 1998; Michalopoulos et. al, 1999; Michalopoulos & Hoy, 2001).
2.3 The supply side of ALMPs: Public authorities and quasi-markets
The provision of ALMPs to unemployed workers is traditionally organised by (local) authorities. However, in a number of countries so-called quasi-markets –where private providers take over the provision of ALMPs and public authorities act as ‘purchasers’, have been created. The introduction of such quasi-markets is expected to increase the efficiency and effectiveness of welfare-to-work
21 services as competition incentivises all producers to innovate and improve their products (Boeri & van Ours, 2013: 351, 374-367; Gerdes, 2015: 10-11; van Berkel & van der Aa, 2005: 331-334).
The Dutch model of quasi-markets is an interesting topic in itself as the provision of welfare-to-work services used to be left to private providers entirely in the years prior to 2007. However disappointing results of this privatisation in combination with retrenchment measures made the Dutch government decide to partly reverse this model in 2007. The activation of long-term unemployed workers in particular remained unsuccessful. Amongst others, this implies that additional social services, next to activation measures, appear to be necessary in order to successfully activate the long-term unemployed. Nowadays municipalities have taken over most of the welfare-to-work services themselves again and tend to limit the contracting of private service providers to the activation of unemployed workers with poorer job prospects only (Gerdes, 2015: 10-11; Koning & Heinrich, 2013: 465; van Berkel & van der Aa, 2005: 331-334; van Berkel, 2010: 30).
Quasi-markets are still largely controlled by public authorities as they decide which private welfare-to-work instruments to use for whom and when (CPB, 2016: 21). Municipalities are thus responsible for determining which individuals are eligible for what benefit schemes and services and have to adequately refer those individuals to the contracted service providers. Inadequate referral processes generate inefficient matches between clients and employment service providers and can eventually lead to the failure of private providers to meet their contractual obligations. Their role as a as purchaser of welfare-to-work services further enables public authorities to measure the effectiveness and efficiency of the contracted private providers and to filter out weak labour market instruments (Boeri & van Ours, 2013: 367; van Berkel & van der Aa, 2005: 334).
The introduction of welfare-to-work quasi-markets and performance related incentives has, however, raised some concerns about the possibility of unintended gaming activities of private ALMP providers. Cream-skimming, or cherry picking, occurs when providers focus their resources on welfare beneficiaries with a relatively shorter distance to the labour market. This implies that those unemployed workers with a larger distance to the labour market can be ‘forgotten’ as they receive less support to reintegrate to the labour market. In the most extreme cases this could lead to
parking, where hard-to-place beneficiaries aren’t provided with any activation and reintegration
services at all (Kok et. al, 2017: 24; Koning & Heinrich, 2013: 462-463).
Koning and Heinrich (2013) did find some evidence of gaming activities, such as cream-skimming and parking, amongst welfare-to-work providers in the Netherlands. They found that gaming activities are more likely to occur when providers are at a greater risk of failing to meet their contractual performance standards. Taken their findings in mind, Koning and Heinrich (2013) suggest that gaming activities can likely be reduced if the risks for private service providers are reduced or if
22 municipalities decide to outsource the reintegration of easier-to-place beneficiaries rather than the reintegration of hard-to-place beneficiaries.
A somewhat different perspective to the ALMP debate is presented by Bonvin (2008), who suggests that the activation of UI recipients should be “capability-friendly” and – rather than leaving the implementation and goal setting of ALMPs to public authorities and local actors alone, beneficiaries should be invited to take part in the definition and implementation of ALMPs. Such a capability-friendly approach requires beneficiaries to have adequate means and freedom of choice on the labour market. For this benefit recipients should be incentivised by making work pay programs that overcome the earlier discussed traps and public authorities and private providers should act as intermediaries towards a quick and long-term reintegration. This also implies beneficiaries’ freedom of choice with regard to their participation in ALMPs as Bonvin (2008) argues that consumers’ freedom of choice will result in a higher quality of services provided by private employment agencies.
By analysing the Swedish quasi-market for job coaching, Gerdes’ (2015) findings call for a more nuanced view regarding the freedom of choice for beneficiaries. Although freedom of choice and available performance information of private providers can result in efficiency gains, since better performing providers were found to be able to increase their market shares, Gerdes also finds that consumers’ freedom of choice leads to an efficiency-inequality trade-off. This efficiency-inequality trade-off occurs because the more vulnerable groups of beneficiaries, for instance those with lower education and non-European immigrants, tend to be less capable of using the available quality ratings of private providers to their advantage.
2.4 The demand side of ALMPs: Employer engagement
It is lastly important to discuss literature on the demand side of welfare-to-work policies. The engagement and interest of employers in welfare-to-work policies is often assumed, but research of the ‘Work Programme’ in the United Kingdom reveals that the level of employers’ engagement and involvement in the ‘Work Programme’ is disappointing. This points out that, in order to increase the effectiveness of ALMPs, it is crucial to understand what motivates or demotivates employers to engage in welfare-to-work programs (Ingold & Stuart, 2015: 443-445).
Research shows that employers often view unemployed workers as high risks accompanied with low quality compared to regular jobseekers. There are concerns about a possible lack of skills, experience, qualifications and other social issues. Especially during economic recessions, when the supply of labour is at its largest, employers’ demand for unemployed workers jobless tends to be low (Ingold & Stuart, 2015: 446; McCollum, 2012: 534).
23 Advantages of contracting unemployed workers through welfare-to-work providers are that there are no recruitment costs involved and that these workers are usually immediately available, have low wage costs and could enhance possible corporate social responsibility agendas. Incentives, such as wage subsidies, trial periods, the screening of applicants, tailored training and ‘interventions’ offered by the welfare-to-work providers could further decrease the risks and costs for employers (Ingold & Stuart, 2015: 446-447; McCollum, 2012: 533).
The effectiveness of financial incentives for employers is debatable however, due to the discussed adverse side effects and strong stigma’s associated with subsidised workers. Besides that service providers in Scotland criticised the use of extensive financial incentives and state support for employers in the recruitment and training of unemployed workers as they found that employers contributed little in return (Atkinson and Williams, 2003: 48-49; McCollum, 2012: 536).
Providers play an important role in filling information gaps, persuading employers to engage in welfare-to-work programs and in making a turn towards a more sustainable reintegration of unemployed workers. Good relationships between providers and employers are vital for the success of ALMPs as this makes it possible to engage employers more actively and to solve issues before contracts end. Building sustainable relationships is difficult when multiple providers are fighting over the attention of the same employers, thus a central ‘employer-recruitment’ approach and local partnerships between providers are recommendable. In-work support and regular direct contact between service providers and ‘placed unemployed workers’ could also further the ‘sustainable employment’ agenda, but should always be in agreement with the employer in order to avoid that they experience such support as unnecessary and inconvenient (Ingold & Stuart, 2015; McGollum, 2012; Schrijver & Vogels, 2013).
3. The Flextensie Instrument
The founders of Flextensie, all having a ‘public background’, experienced that the reintegration process for unemployed workers with a larger distance to the labour market was rather inefficient and could be improved. The instrument that Flextensie has developed therefore aims to take the flexibilisation of the labour market as an opportunity in reintegrating long-term unemployed workers, whilst at the same time taking away administrative and financial burdens for all involved parties7’8.
The instrument basically aims to solve three issues, namely: (1) the lack of opportunities regarding work resumption for unemployed workers with a larger distance to the labour market, (2)
7
Chapter 3 has been sent to Flextensie Nederland B.V. for review; no major changes of the content have been suggested or made.
24 financial concerns and (3) administrative burdens that appear when activating or attempting to activate these unemployed workers.
The functioning of the Flextensie-instrument will be discussed in the first paragraph. The three following paragraphs will further address the three issues that the instrument aims to solve, after which the possible adverse side effects and gaming activities related to the instrument are discussed. The last paragraph of the chapter will discuss the findings of Peels and Engelen (2017) on the demand-side of the instrument.
3.1 The functioning of the Flextensie-instrument
This paragraph will discuss how the instrument functions in a municipality context and is based on information obtained from Flextensie and the Dutch Central Planning Agency’s (CPB) report on “promising labour market policy” (CPB, 2016). First of all, Flextensie has similar contracts with all participating municipalities and these contracts do not include any financial incentives, for instance bonus payments for placed or reintegrated candidates. The contracts do include an implementation period and a post-implementation period. The implementation period covers the training of the case-manager that has been appointed as the Flextensie-intermediary as well as giving advice regarding the setup of a system that matches unemployed workers with Flextensie assignments. In the post-implementation period Flextensie mostly just covers all administrative duties related to the use of the instrument.
The type of contract between Flextensie and participating municipalities implies that most income generated by Flextensie comes from the wages paid by employers that hire Flextensie candidates and some additional income is earned with the advice salaries during the implementation phase. Flextensie uses the wages paid by the participating employers to pay out an hourly premium to the candidate, to cover its own costs and to transfer the remainder to the municipality. Overall, data from the Dutch ‘Chamber of Commerce’ shows that Flextensie performed break-even over 2014-20159.
Figure 3.1 shows the flows of labour between the pools of unemployed workers in a municipality and the local labour market. As municipalities tend to focus their efforts on reintegrating the easier-to-place unemployment beneficiaries (see for instance Koning & Heinrich, 2013), most participants that enrol in Flextensie are harder-to-place welfare beneficiaries.
Enrolment in Flextensie can theoretically boost reintegration in two ways; either (1) by getting placed on a temporary Flextensie assignment by the instrument or (2) by getting ‘back on the
9 Kamerstukken II, 2016-2017, 2526.
25 Unemployment beneficiaries Guidance, mediation, control Welfare beneficiaries Training, school, social work, trial jobs
1 1 2 2 3 3 1
radar’ for reintegration support provided by the municipality. When and what ALMPs are deployed will be discussed in the next paragraph.
Figure 3.1 further shows that the municipality is in charge of deploying the Flextensie-instrument. Flextensie itself operates more in the background as its involvement is limited to the implementation of the instrument and, later on, to taking care of the administrative duties related to the instrument. The flow of workers from the labour market (back) to unemployment is not included in order to keep the figure simple.
3.2 Lack of opportunities for long-term unemployed workers
The first issue encountered in the reintegration process of long-term unemployed workers is the lack of reintegration opportunities for this group of welfare beneficiaries. This is related to financial incentives that push the focus of municipalities towards reintegrating unemployed workers with a shorter distance to the labour market as this is faster, more efficient and more beneficial in the shorter run. Workers that enter unemployment benefits have access to the ‘basic-reintegration-package’ offered by municipalities, which usually consists of (job search) guidance and mediation. Case-managers further control whether the beneficiaries meet their requirements; for instance a
Flextensie
Case-managers Flextensie-intermediary
Municipality department of Employment Services
Labour Market Flextensie assignments
26 monthly amount of job applications. Case-managers have the possibility to deploy other available labour market instruments when it turns out that a beneficiary isn’t easy-to-place (CPB, 2016: 176-180, 203).
ALMPs available for welfare beneficiaries with a larger distance to the labour market are (theoretically) focused on investing in human capital rather than on quick labour market reintegration. Examples of such ALMPs are training and schooling programs, social workplaces, subsidised labour and also the Flextensie-instrument (CPB, 2016: 21-22, 177). Although this approach is evidence based, it can also be regarded as a process of cherry picking or cream-skimming by municipalities that causes those unemployed workers with a larger distance to the labour market to become ‘forgotten’ with regard to their quest for a job. It is difficult for this group of unemployed workers to become activated as they get less job search guidance and mediation of the municipality and are more likely to get locked-in ALMPs that cost a lot of time and might turn out to be ineffective as well.
Unemployed workers in welfare benefit schemes can voluntarily apply to become a Flextensie candidate. Case-managers in municipalities that work with the instrument inform their clients about this possibility. As has been mentioned in the previous paragraph and is visualised in Figure 3.1, becoming a Flextensie candidate can boost the reintegration process of unemployed workers in two ways. The more regular path is that candidates increase their working experience and their human capital during their placement on a Flextensie assignment, but just entering the Flextensie database can already boost their reintegration process as the municipality now knows that these candidates are motivated to work10. The Flextensie assignments have a maximum duration of six months when they are on a part-time basis (less than 28 hours per week) and three months when they are on a full-time basis. As has been discussed in the literature review, such temporary jobs might be a stepping stone for unemployed workers to more regular work and at least keep them in touch with the labour market.
3.3 Financial insecurity
The second issue that the Flextensie-instrument aims to solve applies to all involved actors, being the candidates, the potential employers and the municipalities. As has been discussed in the literature review, welfare beneficiaries often experience financial insecurity when they manage to find a job (the inactivity trap). This applies even more in a time of labour market flexibilisation as jobs offered to welfare beneficiaries more often tend to be on a temporary basis and for flexible hours. Generated labour market income normally decreases the level of benefits or even takes away a
27 beneficiaries’ eligibility for benefits entirely. This potentially causes welfare beneficiaries to get trapped in their current situation of unemployment, inactivity or poverty and thus decreases or takes away their job search efforts.
To tackle this problem, placement on a Flextensie assignment does not cause a decrease in the level of welfare benefits for candidates and furthermore aims to make work pay. This way the financial insecurity related to accepting flexible jobs for welfare beneficiaries is taken away. Participants in the instrument are further incentivised as they receive an additional hourly premium of roughly €2, -. Article 31, paragraph 2 under j. of the Dutch Participation Act allows welfare beneficiaries to earn a maximum yearly ‘premium for work resumption’ of € 2.416,-11. The maximum set duration of Flextensie assignments ensure that the candidates stay within these legal boundaries. Combined with their welfare benefits, participants earn about as much as the statutory hourly minimum wage. The monthly income level of Flextensie participants is higher than it would have been when they would have accepted a similar flexible job without Flextensie-instrument12. As has been discussed in the literature review, these financial incentives are important in successfully activating welfare beneficiaries.
Hiring personnel with a larger distance to the labour market generates risks for employers as well. As has been discussed in the literature review, employers often expect unemployed workers to be of relatively low quality. Peels and Engelen (2017: 8) find some evidence of this as well as the employers that they interviewed mentioned that Flextensie participants are in need of additional guidance and tend to have been alienated from the labour market. The instrument takes away most of these risks as it doesn’t include any recruitment costs that employers would normally incur, hourly costs are usually established at a minimum wage level and the duration of the instrument is set beforehand. There are no additional financial incentives for employers that participate in the Flextensie-instrument as municipalities do not pay wage subsidies to employers during Flextensie placements.
Activating welfare beneficiaries with a larger distance to the labour market is costly and risky for municipalities as well, especially during a time of retrenchment and smaller available ‘reintegration budgets’. Municipalities hope to save expenses by outsourcing the reintegration of hard-to-place beneficiaries to private service providers. Although there are some implementation costs, the Flextensie-instrument is also designed to take away financial burdens from municipalities. Municipalities set the hourly wage-level for employers that hire personnel through the Flextensie-instrument and employers transfer the agreed upon amount to Flextensie, who then subtracts their own costs, pays the hourly fees to the participants and transfers the remainder to the municipality.
11
Overheid.nl, ‘Participatiewet’, http://wetten.overheid.nl/BWBR0015703/2018-01-01 (visited at 03-01-2018).
28 These savings are then reinvested in the municipality’s budget for labour market reintegration services. Peels and Engelen (2017: 57-62) calculated that most municipalities do save money by using the Flextensie-instrument.
3.4 Administrative burdens
The third issue of administrative burdens is somewhat related to the issue of financial concerns. In a situation without Flextensie, matching welfare beneficiaries with flexible vacancies implies a process of bureaucratic hassle. There are two possible extreme scenarios to take into account. In the first scenario a municipality decides to stop paying out welfare benefits as soon as a welfare beneficiary starts working on a flexible job. In this scenario the beneficiary has to apply for welfare benefits again when the job is over and possibly gets assigned to a new case-manager. Furthermore, whilst being on the flexible job, the welfare beneficiary won’t be active in any mediation processes towards a more constructive labour market reintegration.
The second scenario is that the municipality continues to pay out welfare benefits, but calculates the income of the beneficiary and subtracts this income from the level of benefits afterwards. The result is that accepting a flexible job will be unattractive from both a financial and an administrative point of view. Flextensie solves a large part of the administrative hassle as municipalities continue to pay out full welfare benefits to all Flextensie candidates and Flextensie takes care of all of the financial issues regarding the instrument13.
3.5 Adverse side effects and gaming activities
As has been discussed in the literature review, the instrument might not be very effective if the adverse side effects related to the instrument are strong. In this case there is no strong locking-in effect as Flextensie candidates are still available for participation in other labour market instruments, Flextensie assignments are short in duration and –even during placement on a Flextensie assignment, candidates are free to accept regular labour market contracts at any time. Placement on a Flextensie assignment may even overrule lock-ins from other labour market instruments as, for example, candidates that get matched with a Flextensie assignment and are also active in a compulsory training program can stop training and start working, which in other situations is not always possible.
There have been some concerns regarding crowding out and job displacement effects however. The before mentioned media campaign did result in parliamentary questions that led the state secretary of the ‘Ministry of Social Affairs and Employment’ (SZW) in the Rutte II cabinet, Jetta
