1 Introduction

Turkey created an unemployment insurance programme relatively late, in 1999, with the first benefits only being paid out in 2002. Despite its short lifespan, the programme has already been transformed through successive reforms. While key statutes to protect employees in the case of job loss have remained intact, new policy instruments have been added to the programme. Collectively, these new instruments have changed the logic of the programme from passive labour market policies to active labour market policies and employment generation.

In this chapter, the causal dynamics behind the introduction and the transformation of Turkey’s unemployment insurance programme up to 2019 are traced (for an overview of developments in 2020 see Öktem 2021). I identify four causal mechanisms in the process: double benefit, business-led reform, transnational cooperation and outcompeting (note that this list may not exhaust all relevant mechanisms).

First, the double benefit of unemployment insurance as both a social policy programme and a financing tool is crucial for understanding its development (double benefit mechanism). In the context of repeated failures to prudently design and manage social security schemes and against the backdrop of a deep economic crisis and the prospect of IMF intervention, unemployment insurance was designed so that benefit payments would in all likelihood never exceed contributions. This led to a structural surplus and ever-increasing fund reserves, possibly beyond what policymakers had envisioned when designing the policy. As a result, unemployment insurance had a “double benefit” for policymakers: it served as a social policy tool, but also as a device for financing the budget deficit (during a period of strict IMF supervision).

This structural surplus and the fund’s ever-growing reserves, over time, generated pressures for reform. Business and labour unions demanded a lowering of contribution rates to a level that would be sufficient to pay benefits or increasing the quality of benefits. Yet, these demands were largely ignored. Instead, successive governments established a number of new instruments in the unemployment insurance programme. These new instruments were mainly active labour market policies (ALMP), such as vocational training courses, which had long been demanded by business. Therefore, these changes can be described as business-led reforms (business-led reform mechanism). ALMP also broadly fit into the agenda of a Europeanisation of labour market policies and mirrored developments in European countries, where an increasing emphasis on ALMP had been visible in the previous decades (Kenworthy 2010). In this sense, transnational cooperation with the European Union also shaped the direction of policy change (transnational cooperation mechanism). Furthermore, these instruments were often launched and expanded with a view on “outcompeting” the opposition. Governments aimed to maximise their prospects in elections by flexibly deploying tangible benefits to participants of these policies (outcompeting mechanism).

The findings add to the literature on causal mechanisms in the development of social insurance schemes and also resonate with the literature on the fiscal politics of the welfare state (Koreh 2017a, 2017b). This literature argues that the expansion of social insurance may be driven by fiscal concerns and that fiscal arrangements may influence the subsequent development of social insurance. Thus, social insurance programmes may have a double benefit, as social policies and as state-financing devices. Usually, this double benefit is most important for pension insurance (see also Chap. 3). However, the chapter shows that it can also apply to unemployment insurance. Yet, it is important to emphasise that the double benefit mechanism, by itself, does not explain the whole story. It may help to explain the initial policy design and why the programme was transformed, but it does not explain the further trajectory of the programme.

In terms of research method, the chapter builds on a combination of policy document analysis, archival research and elite interviews. With regard to policy documents, I analysed all primary legislation related to the programme. For the archival research, I surveyed parliamentary proceedings of key legislative changes, as well as government reports on the programme. Additionally, I scanned Turkish newspapers (primarily Cumhuriyet, Milliyet and Dünya) for relevant reports and looked at reports prepared and statements made by representatives of business and labour. I complemented this policy and archival research by interviewing two former Ministers of Labour and Social Security, who were in charge of the programme for nearly a decade.Footnote 1

The structure of the chapter is as follows. First, the literature on causal mechanisms in the development of unemployment insurance is surveyed. Then, I describe the creation of the programme in Turkey. Here, I trace how a draft bill prepared in the early 1990s became the basis for the unemployment insurance legislation. Next, I explore the transformation of the programme from 1999 to 2019 and show how the accumulation of a massive surplus and pressures from business facilitated comprehensive policy changes towards more active labour market policies. Finally, I conclude by discussing the implications of the findings for the comparative literature.

2 Causal Mechanisms in the Development of Unemployment Insurance

Unemployment insurance is a central institution of modern welfare states. It protects workers against job loss by paying regular cash benefits to the unemployed. This decommodifies workers to a certain degree (Esping-Andersen 1990), but also acts as an automatic stabiliser in case of economic recessions. It constitutes a significant intervention in the labour market, and this makes it one of the most contentious forms of social insurance (Sjöberg et al. 2010). This is reflected in its incomplete and lagged diffusion throughout the world. Compared to other forms of social insurance, insurance against unemployment has been adopted by far fewer countries. In terms of the adoption sequence, unemployment insurance typically comes last (ILO 2014, 4; Schmitt et al. 2015).

Nevertheless, around ninety countries in the world have adopted unemployment insurance. First implemented on a nationwide level in 1905 in France, the policy initially spread in Europe and was adopted by some non-European countries (e.g. South Africa, Uruguay) in the interwar years. In Southern Europe, the first adoptions came after the First World War in Italy and Spain, while Greece, Cyprus and Portugal created programmes only after the Second World War. In the Middle East, the first adoptions came after the Second World War (e.g. Israel and Iran). Compared to these countries, Turkey is a latecomer with legislation only being passed in 1999 (Sjöberg et al. 2010; SSA 2016).

The comparative literature has identified various drivers behind unemployment insurance. The policy offers tangible benefits to employees whose representatives—labour unions—are generally assumed to support the policy. In fact, in various cases, labour unions implemented unemployment insurance before the state did (Flora and Alber 1981, 152). That is why it is generally assumed that labour unions pushed for the introduction and expansion of unemployment insurance. However, the reality is more complicated. For various reasons, such as fear of co-optation, unions actually opposed unemployment insurance in some countries (Flora and Alber 1981, 153–54).

Business has many reasons to be lukewarm towards unemployment insurance. It not only increases non-wage labour costs, but also increases the bargaining power of labour through providing workers with a reservation wage (Paster 2013). However, the policy also has benefits for business. For instance, because workers can receive benefits in case of job loss, layoffs can be easier to manage (Carter et al. 2013, 4). Hence, it may even serve de facto flexibility of the labour market. Furthermore, unemployment insurance increases productivity, which is beneficial for employers (ILO 2017, 41–42). These different advantages and disadvantages for business may translate into inter-sectoral conflict among business, with high-risk sectors supporting and low-risk sectors opposing unemployment insurance (Mares 2003). Furthermore, business may interpret the policy depending on the prevailing ideational climate (Münnich 2010). Thus, although it is perceived as an arena of conflict between labour and business, the policy also offers an opportunity for “cross-class alliance” (Hellwig 2005).

In addition, even in cases where business opposes policy adoption, it might have an interest in sustaining or shaping the policy once it is implemented (Paster 2017). Therefore, it might propose reforms. This could involve programme downsizing through decreasing contribution rates and benefits, but also expansion of policy instruments from unemployment benefits to a broader set of labour market policies. In those cases where business is successful in convincing governments to adopt its proposals, one could speak of “business-led reforms”.

From the perspective of the state, the issue is complex. Within competitive regimes, policy adoption may constitute a response by policymakers and/or bureaucrats to demands from the broader electorate or from workers (Hicks 1999). Alternatively, policymakers may support unemployment insurance as a result of party competition. However, the popularity of the policy may be limited if the unemployed are generally seen as undeserving and few people expect to benefit from the insurance. Particularly in authoritarian regimes, the policy may also be created in anticipation of future demands by workers. Furthermore, unemployment insurance may also be attractive from a statist perspective as it promises to defuse labour conflict (Matsunaga 2017) and significantly increase state capacity in the realm of labour market policies.

Finally, unemployment insurance may not just increase the bureaucratic capacities of the state apparatus, but also strengthen state finances. As with any type of social insurance, unemployment insurance has the potential to be of “double benefit” to the state, as an instrument of social policy and as a financing device for the state. Although this “double benefit” is most relevant for pensions due to the peculiarities of old-age pension systems, it may also be relevant for unemployment. First, if the unemployment insurance system builds up a surplus, the funds allocated to the programme may be used to finance budget deficits. Often, social insurance (SI) entails regulations that funds may only be invested in certain areas, which in practice then translates into mostly buying public debt. Second, funds may be used to finance other projects. The comparative literature cites cases where an SI surplus is used to “finance projects of state building, offloading the costs of industrial restructuring and covering national debt” (Koreh 2017a, 117). Hence, although SI contributions are in principle “earmarked for specific SI schemes”, it would be wrong to understand “SI as a closed financial system” (Koreh 2017a, 117).

Third, the generation of a sizeable surplus may also shape the development of social insurance, particularly within competitive regimes. It may lead, for instance, to demands for a decrease in contribution rates. These demands would likely be raised by employers and right-of-centre parties, which are both sensitive to the level of non-wage labour costs and cautious about perceived detrimental effects of “decommodifying” social security. Conversely, a surplus may also facilitate programmatic expansion. The high level of contributions may create a need for “legitimation” which could underpin welfare state expansion. Alternatively, the surplus may also facilitate expansion as it gives policymakers the means to make social security more generous (the “surplus effect”, Koreh 2017b).

In recent decades, unemployment insurance has been increasingly complemented by active labour market policies (ALMP), such as retraining, that focus on increasing the employability and productivity of the labour force (Kenworthy 2010; Weishaupt 2019). In institutional and administrative terms, ALMP may be closely linked to unemployment insurance. In some cases, ALMP is at least partially financed by contributions made to unemployment insurance and managed by public employment services (PESs), which also administer unemployment insurance programmes, and eligibility to ALMP is tied to requirements for unemployment insurance. In terms of the drivers behind ALMP, business appears to be more supportive than labour, although labour unions have been shown to support some instruments such as training and employment assistance (Tepe and Vanhuysse 2013).

With regard to the Turkish case, research on the causal drivers behind the development of unemployment insurance has been limited. The arduous process of policy adoption has been analysed with a focus on how transnational cooperation with the International Labour Organization (ILO), the Organisation for Economic Co-operation and Development (OECD) and the World Bank shaped diffusion (Özkan 2011) and with a focus on how domestic political obstacles prevented an earlier adoption (Öktem 2020a). Much less is known, however, about how the programme has been transformed. While it has been argued that the introduction of unemployment insurance failed to alter the broader dynamics of unemployment compensation (Özkan 2016), effective access to unemployment benefits did increase (Öktem 2020b). Moreover, some researchers observed an increase in ALMP (Gün 2013; Lordoğlu and Koçak 2015; Kapar 2015), which some read as signifying a neoliberal transformation (Gün 2016). Still, the causal drivers behind this process remain to be unearthed. In this chapter, I aim to fill this gap in the literature and explore which mechanisms drove the introduction and transformation of Turkey’s unemployment insurance programme. I will particularly focus on how the policy was designed and how the policy design facilitated the programme’s transformation.

3 The Creation of Unemployment Insurance

3.1 A Failed Attempt to Introduce Unemployment Insurance in 1992

Policymakers first brought up the idea of creating an unemployment insurance programme in Turkey when they discussed the introduction of social insurance in 1935. From the late 1950s onwards, various governments prepared draft bills, often with the support of the ILO. Unemployment insurance became a bipartisan development goal with parties trying to “outcompete” each other with promises on this front but failing to live up to their promises when in power. In 1983, unemployment insurance legislation was even passed by parliament, but not implemented by the ruling military junta. In the context of a structural transformation of the economy from statism to neoliberalism in the 1980s, unemployment insurance was dropped from the government’s agenda. The centre-right Motherland Party (Anavatan Partisi, ANAP) governments of the 1980s opposed the policy (Öktem 2020a).

The situation changed in the early 1990s when ANAP lost power to a centrist coalition government of the centre-right True Path Party (Doğru Yol Partisi, DYP) and the centre-left Social Democrat Populist Party (Sosyal Demokrat Halkçı Parti, SHP). Both parties had made unemployment insurance part of their election campaign. The Minister of Labour, Mehmet Moğultay from the SHP, eagerly worked on drafting legislation. However, he immediately encountered resistance from business, which in the words of the former head of the Turkish Federation of Employer Associations (Türkiye İşveren Sendikaları Konfederasyonu, TISK), “went to war with the new government and the Minister of Labour Mehmet Moğultay over unemployment insurance” (Baydur 2006, 62).Footnote 2 The strategy of business was twofold. First, it insisted on a simultaneous cutback of severance payments, which had acquired the status of a functional equivalent of unemployment insurance (Başterzi 1995), as a precondition for introducing unemployment insurance—knowing full well that labour would not accept this. Second, it lobbied the centre-right DYP, which was the stronger party in the coalition. Eventually, this strategy succeeded.

However, Moğultay managed to prepare draft legislation, which became the standard on which subsequent drafts for legislation were based (Andaç 1999). The draft bill was guided by ideas from the OECD and the World Bank that entailed concerns about making unemployment insurance not too generous (Özkan 2011). It linked benefit duration to the contribution period. Employees who had worked for twenty/thirty/thirty-six months of the last three years would receive benefits for six/seven/eight months. This was slightly reduced to four/six/eight months in late 1992. Benefits were set at 45 per cent of the employee’s previous net wage, with a benefit floor set at half the minimum wage. Financing was to be based on employer, employee and state contributions, with each contributing 2 per cent of the employee’s gross wage (Öktem 2020a). This was equal to the contributions set in the ill-fated 1983 legislation that had not been implemented—even though the new proposal was markedly less generous than the 1983 legislation (Danışma Meclisi 1983).

Overall, the draft bill was geared towards labour market insiders, as people in intermittent employment could hardly fulfil contribution requirements, and benefits were not generous. Nevertheless, the plan encountered strong criticism from business, which maintained that severance pay reform should be a precondition for introducing unemployment insurance. Furthermore, business, represented by TISK, was afraid that unemployment insurance would raise non-wage labour costs and therefore demanded that the programme should not increase overall employer contributions. Similarly, labour unions, which generally supported the introduction of unemployment insurance but opposed severance pay reform, rejected any increase in employee contributions (Milliyet 1992b).

The question of financing therefore posed the government with a conundrum. It needed to raise funds for the programme, but social partners were unwilling to contribute. Initial funding was to come from the state, but this alone could not maintain the programme. As a solution, the idea to divert money from other funds was proposed, and by 1993, Moğultay supported this demand. To ensure that unemployment insurance did not pose any additional burden for social partners, other contributions that employers and employees had to make were to be reduced (Öktem 2020a). For this purpose, two funds were identified: the Homeownership Support (Konut Edindirme Yardımı, KEY) (Cumhuriyet 1993a) and the Mandatory Savings Fund (Zorunlu Tasarruf Fonu) (Milliyet 1993).

Both funds had been created by the previous ANAP government against the opposition of the DYP and SHP. Both policies had been devised in the heyday of neoliberalism in Turkey and aimed to achieve social security through financialisation—in an economic context of strong growth and high inflation. Arguably, both policies had been designed with the “double benefit” in mind. The idea behind KEY, legislated in 1986, was that workers would be supported when buying a house or flat (Resmi Gazete 1986). For this purpose, employers paid a fixed contribution for their employees into a fund. These contributions were invested mainly in public debt. After a contribution period, employees buying a home would receive a support payment based on their contributions plus interest. If employees did not buy such a home, their dependants would receive a payment on the worker’s death. KEY soon ran into problems. Critics argued early on that a large amount of money was collected without workers really receiving benefits. Eventually, contributions were stopped in 1996, and KEY was abolished in 1999.

The idea behind the Mandatory Savings Fund, legislated in 1988, was to create a provident fund-like savings account for employees (Resmi Gazete 1988). This fund was supposed to boost Turkey’s notoriously low savings rate, although observers questioned the likelihood of achieving this goal from the outset (OECD Economic Survey 1988, 55–56). The fund applied to private and public sector employees and complemented the public pension system. Akin to social security contributions, employers and employees paid contributions based on the employee’s gross wage. These contributions were invested in various ways. Employees would receive their savings on retirement, or their dependants would receive the savings in case of the employee’s death. If employees wanted to draw on their savings earlier, they could do so to a limited extent.

Initially, labour unions proposed to shift savings from KEY to unemployment insurance, and policymakers considered both KEY and the Savings Fund as potential resources for unemployment insurance. However, given that contributions to the Savings Fund were like social security contributions, and given that the government aimed to cancel the Savings Fund anyway (Milliyet 1992a), it made sense to simply repurpose the Savings Fund contributions for unemployment insurance (Cumhuriyet 1993b). The Savings Fund contribution rates were close to what the government had envisioned for unemployment insurance, with only employer contributions being 1 per cent higher. Curiously, the increased contributions apparently did not lead to adjustments in the generosity of unemployment insurance in the draft legislation.

The DYP-SHP government eventually failed to legislate on unemployment insurance. Instead of Moğultay’s proposal, a programme limited to state-owned enterprises was implemented. The privatisation of state-owned enterprises had been a central issue on the political agenda since the 1980s. The World Bank, which pushed for privatisation, proposed providing unemployment benefits for workers affected by privatisation. The SHP wanted this programme to be part of a comprehensive unemployment insurance policy, yet it failed to convince the DYP, and thus it was implemented as a separate programme.

While the DYP-SHP government did not implement unemployment insurance, it successfully brought the issue onto the political agenda. By the mid-1990s, nearly all parties promised to adopt the policy. Ministers of Labour in successive governments announced plans to introduce unemployment insurance, with all drafts circulated in the media remaining close to Moğultay’s draft bill (Öktem 2020a). However, none of the proposals came close to being implemented.

3.2 The 1999 Unemployment Insurance Legislation

In 1999, a centrist coalition government led by veteran leftist politician Bülent Ecevit came to power in the midst of a deep economic crisis. Ecevit had championed the cause of unemployment insurance already during his terms as Minister of Labour in the 1960s and Prime Minister in the 1970s. The new Minister of Labour and Social Security, Yaşar Okuyan, from the centre-right ANAP also strongly supported unemployment insurance. ANAP had a history of opposing the policy and was generally seen as a pro-business party, but Okuyan, who grew up in a gecekondu (shanty town), aimed to become a pro-labour minister.Footnote 3 Within three months the government managed to push unemployment insurance legislation through parliament.

Okuyan managed to pass the legislation by making it part of a comprehensive social security reform. In the 1990s, the social security system had started to run increasing deficits, mainly due to the maturing of a pension system with ample opportunities for early retirement and also due to the mismanagement of funds. Plans to ensure the sustainability of the system through raising the retirement age had been on the agenda of virtually all governments in that period. However, this met fierce resistance, especially from labour unions. A plan to link unemployment insurance to a comprehensive social security reform had been first brought forward in 1995. Yet, the reforms did not materialise, and therefore the deficits continued to increase (Yentürk 2018). This was seen as a major risk for the economy as a whole, and thus the perceived need for reform became ever more pressing (Özkan 2011).

The government immediately began to work on a comprehensive social security reform proposal. Initially, however, it did not include unemployment insurance in the bill. The reform featured a sharp increase in the retirement age, and therefore strong opposition from labour unions was widely anticipated. Okuyan presented his plan to include unemployment insurance in the bill only after labour unions voiced their criticism and in order to soften their stance. However, he failed to change the minds of labour with this manoeuvre. Moreover, by attaching the policy to the reform, he drew the ire of business, which opposed the new programme, unless severance pay was retrenched at the same time (Özbek 2006, 350–351; Öktem 2020a).

The unemployment insurance legislation that the new government prepared deviated slightly from earlier draft bills. It marginally reduced the contribution period by changing the requirement for continuous contributions over the last six months to four months. However, the focus on workers in “stable and continuous” employment remained (Öktem 2020a, 23). The benefit level was set at 50 per cent of the previous net wage. Importantly, while earlier proposals contained a benefit floor that guaranteed higher replacement rates for low-income earners, Okuyan’s proposal introduced a benefit cap. This cap was set at the net minimum wage. Hence, medium- and high-income earners would have a lower replacement rate. The instrument of a benefit cap had been demanded by business groups as early as 1992 but had apparently not played a role in previous draft laws. Finally, as in earlier proposals (Andaç 1999) beneficiaries would also be eligible for active labour market policies—a hitherto neglected policy area in Turkey.

The legislation also retained the plan to cancel the Mandatory Savings Fund contributions and repurpose them for unemployment insurance. The Savings Fund had continued despite government plans to abolish it. The 2 per cent employee contribution and the 3 per cent employer contribution made to the Mandatory Savings Fund would now be channelled to unemployment insurance and be complemented with a 2 per cent contribution by the state. Due to mismanagement, the Savings Fund was no longer seen as credible and legitimate. Raising the same contributions for unemployment insurance, which was a key demand of labour and a long-standing bipartisan development goal, would serve to legitimise the contributions. Furthermore, in contrast to the Savings Fund, unemployment insurance was not limited to medium and large enterprises but would also cover small companies. Therefore, in effect it constituted a new contribution for companies with fewer than ten employees. Civil servants, however, would not be covered by unemployment insurance, and thus the contributions made by the state as their employer would end.

Both business and labour voiced their concern about the reform. Yet, instead of seeking consensus with employers and employees, the Ecevit government simply chose to push through the reform against all odds. The draft legislation was debated and passed by parliament in August 1999. The parliamentary debates mainly focused on pension reform, with opposition MPs accusing the government of bowing to IMF demands for retrenchment. With regard to unemployment insurance, the opposition parties, DYP and the Islamist Virtue Party (Fazilet Partisi), voiced various criticisms. They argued that the policy would only protect labour market insiders at the expense of labour market outsiders, that unemployment insurance had to be coupled with severance pay reform and with job security provisions (a long-standing demand of business), that the public employment service lacked the capacity to administer the programme and that benefit levels were too low (Öktem 2020a).

Interestingly, the opposition also criticised the way that unemployment insurance would be financed. MPs from both the DYP and the Virtue Party argued that the workings of the Unemployment Insurance Fund (UIF), which would collect and manage the contributions, were insufficiently specified and not transparent. While the legislation foresaw the inclusion of employers’ representatives and employee unions in the management of the UIF, it left the question of how to invest the assets to a regulation to be issued later. Considering the negative experiences with the Mandatory Savings Fund and other policies, the MPs repeatedly warned that it was likely that the fund would go bankrupt and would have to be bailed out by the government (TBMM 1999a, 531, 1999b, 279, 1999c, 355, 372 and 390, 1999d, 493–494 and 504).

The opposition’s warnings were countered by government MPs who argued that contributions would be more than sufficient to pay for the benefits. Minister of State Fikret Ünlü argued that in the first twenty months after coming into force, a significant amount of money would be accumulated as no one would be eligible for benefits in this period. After this accumulation period, new contributions would continue to exceed benefits, Ünlü argued. Under a normal scenario, contributions would be nearly one and a half times the level of benefits (TBMM 1999c, 339). It is noteworthy that these calculations were made during a severe economic crisis, in which social security revenues tend to fall.

Yet, these arguments did not alleviate the opposition’s fears. The experiences with KEY and the Savings Fund were still very vivid. Furthermore, one key reason for the social security reform, of which unemployment insurance was but a small part, was the previous mismanagement of social insurance funds. Hence, opposition lawmakers feared mismanagement and a fund failure much more than they expected the massive accumulation of a surplus that would actually characterise the UIF.

To conclude, policymakers were overly cautious in designing the fiscal aspects of the programme (Alper 2019). Partly to alleviate opposition against the policy and partly as a reaction to the negative experiences with the management of social insurance and other funds, policymakers successively decreased the generosity of the unemployment insurance draft bills throughout the 1990s. At the same time, planned contribution rates were even increased as the Mandatory Savings Fund contributions were repurposed. A surplus was thus all but inevitable.

4 The Transformation of Unemployment Insurance (2000–2019)

4.1 Surplus Accumulation (2000–2007)

Unemployment insurance started in June 2000. Given that eligibility for benefits required a minimum twenty months of contribution period, the first benefits could be paid out only in early 2002. During this period, the UIF thus started to accumulate money, reaching 1 per cent of GDP. This was entirely in line with what policymakers and bureaucrats had envisioned for the accumulation period. Although no long-term actuarial planning documents are available for the early years, one can compare the actual surplus generated by the UIF to the expected surplus that was posted in official documents, such as the government’s annual programmes. Furthermore, letters of intent written to the IMF (which had a standby agreement with Turkey) in December 2000 and May 2001 provide information on how the government expected the UIF to develop before unemployment benefits started to be paid out (Önal and Erçel 2000; Bahçeli et al. 2001).

Figure 4.1 confirms that from the outset policymakers expected the fund to create a constant surplus. In fact, before the programme was rolled out, they underestimated the surplus that the fund would generate. Interestingly, there appear to be inconsistencies between the annual plans and the letters of intent to the IMF, with the latter being far more conservative in their outlook. Yet, both the letters to the IMF and the annual plans agreed that the UIF would create a sizeable surplus. Over time, this would lead to the accumulation of an enormous amount of money as Fig. 4.2 illustrates.

Fig. 4.1
A line graph of the unemployment fund surplus as a percentage of G D P annually from 2000 to 2014, for actual, plan, I M Fs in 2 different dates. The actual value rises for the first few years and gradually declines. The plan also has a similar trend to the actual value from 2005 to 2013. The other 2 lines increase in the beginning.

Annual surplus of the unemployment insurance fund as a share of GDP. (Source: “Plan” refers to annual programmes and illustrates estimates that were made in October of the previous year. So, the 2001 plan for instance, which was written in October 2000, foresaw an annual surplus of 0.9 per cent of GDP. The actual surplus in 2001 was 1 per cent of GDP. “Actual” refers to the revenue and expenditure posted in monthly reports on the UIF published by the PES. “IMF 12/2000” refers to the letter of intent sent by Turkey to the IMF on 18 December 2000, while “IMF 5/2001” refers to the letter of intent sent by Turkey to the IMF on 3 May 2001)

Fig. 4.2
A line graph plots the percentage of G D P as the annual unemployment fund reserves from 2000 to 2014 for actual, plan, and I M F 5 slash 2001. All 3 have an increasing trend. I M F ends in 2003.

Unemployment insurance fund reserves as a share of GDP. (Source: See Fig. 4.1)

Thus, it appears that policymakers were very much aware that the fund would generate a significant surplus, even in times of high unemployment. Benefit levels were low and qualification conditions were quite demanding, so benefit payments could simply not surpass contributions. Before unemployment benefits were paid out in 2002, the government realised that the surplus would be higher than necessary. In the context of a renewed economic crisis, it decided to reduce contribution rates by one percentage point for employers, employees and the state. Thus, overall contributions were decreased from 7 to 4 per cent of the gross wage with the aim of decreasing non-wage employment costs. Yet, as Figs. 4.1 and 4.2 illustrate and as policymakers quickly realised, even these lowered contributions were far above what was necessary to run the programme. Despite lower contributions, the fund continued to generate a massive surplus.

So, what happened to the money accumulated in the fund? Official reports indicate that the surplus was mostly used to buy bonds issued by the state. Hence, the fund bought public debt (Alper 2019). The accumulation of an ever-larger amount of money thus clearly served a fiscal function for policymakers. Unemployment insurance had a double benefit. Comparing the fund’s surplus to the size of the budget deficit would give an approximation of the fund’s importance for the financing of public debt. Figure 4.3 visualises this comparison by showing the size of the surplus as a share of the budget deficit. Note that this is just an illustrative comparison and does not reflect the amount of public debt actually financed by the UIF.

Fig. 4.3
A line graph of unemployment insurance fund surplus as a percentage of the annual budget deficit from 2000 to 2019. The line has 3 peaks, 90% in 2006, about 40% in 2001, and about 50% in 2014. Values are estimated.

Surplus of the unemployment insurance fund as a share of the budget deficit. (Source: Öktem (2020b))

Furthermore, it is also important to note that in the early 2000s, Turkey was under IMF supervision. The primary surplus of the public sector was monitored—and the UIF was included in the calculation of this surplus. Hence, the surplus in the UIF helped Turkey to fulfil the requirements of a stringent IMF programme. This would suggest that “fiscal concerns” (Koreh 2017b) played an important role in the development of unemployment insurance.

In any case, in the early 2000s, both the coalition government led by Bülent Ecevit and the centre-right Justice and Development Party (Adalet ve Kalkinma Partisi, AKP) government that succeeded it in late 2002 were far more concerned with ensuring that the new programme was properly implemented than they were with making substantive policy changes.Footnote 4 The Public Employment Service (PES) that had been put in charge of administering the programme was widely seen as lacking the capacity to do so. To ensure that the PES had the capacity to manage unemployment insurance, it was comprehensively reformed, receiving EU support in the process (Bölükbasi and Ertugal 2013).

In terms of policy innovation, the Labour Law passed in 2003 by the new AKP government foresaw the creation of a Short-Time Work Compensation (Kısa Çalışma Ödeneği) programme that would pay worker’s wages when companies reduced working hours during an economic crisis and a Wage Guarantee Fund (Ücret Garanti Fonu) that would pay wages for several months when companies went bankrupt. Both instruments were implemented in 2004 and started to pay out benefits from 2005 onwards. However, until the global economic crisis, these instruments remained extremely marginal. Thus, these instruments did not decrease the annual surplus, and the UIF was expected to surpass 5 per cent of GDP in the late 2000s.Footnote 5

However, the fund’s growing reserves as well as its continued surplus increasingly caught the attention of policymakers towards the mid-2000s. Opposition parties criticised the government for not supporting the unemployed through the fund. Emphasising that 97 per cent of the fund was invested in public debt, one MP argued in parliament in 2004 that “instead of supporting the unemployed, the fund has become a support for the state” (TBMM 2004, 651). Similarly, business was complaining that despite the large amount of money in the UIF, contribution rates were not decreased (Milliyet 2006, 2007a). If the fund continued to grow, business representatives feared, the likelihood of governments misusing the fund would increase (Birgün 2005).

The Minister of Labour and Social Security, Murat Başesgioğlu, agreed in principle that the programme needed reform, arguing that eligibility conditions were too strict and benefits were too low. Similar concerns were also voiced in the Second General Assembly of the PES (İŞKUR 2005b). Başesgioğlu repeatedly promised a reform that would have increased benefits and loosened eligibility criteria (Tan and Karslıoğlu 2010, 77–82). However, it was not made part of the government’s social security reform agenda. Unemployment insurance reform was neither mentioned in the white paper on social security reform (T.C. Başbakanlık 2005), nor to be found in the final reform legislation. As a result, the policy was left largely untouched in the AKP’s first term that ended in 2007. Instead of reducing contribution rates further or increasing programme generosity, the government tackled the surplus by starting to tax income on interest earned by the fund from 2006 onwards. However, on its own this did not have a large effect, and this policy was cancelled (more or less by accident) in 2008.

Overall, the first years of the unemployment insurance programme were thus shaped by a steady surplus accumulation that resulted in increasing fund reserves. In this way, the policy was shaped by the double benefit mechanism: it was not just a social policy tool but also served to finance public debt (Öktem 2020b).

4.2 Economic Crisis Triggers Policy Change (2008)

In 2008, the AKP government launched a major reform of unemployment insurance that set the path for the future direction of the policy. While the core statutes of the programme were virtually left untouched, a number of new instruments were added to the programme that gradually changed the logic of the policy.

A number of factors came together that facilitated policy change in 2008. First, economic growth had slowed. Turkey had grown quite fast after the severe economic crisis of 2001, but unemployment had remained stubbornly high. As the economy appeared to reach the end of a growth cycle, fears of a surge in unemployment mounted. In response, demands by business to make use of the UIF for employment creation increased. In early 2007, the Union of Chambers and Commodity Exchanges of Turkey (Türkiye Odalar ve Borsalar Birliği, TOBB) proposed the introduction of massive employer subsidies. The UIF should pay the social security contributions for all new employment generated for two years, TOBB demanded (Milliyet 2007c). TOBB had made the same demand in 2005, in response to a government call for more employment creation by the private sector. At that time, it had pointed to the treasury as the financing source (Milliyet 2005). Two years later, it asked for the UIF to fund the policy. TOBB’s proposal was supported by other business groups (Milliyet 2007e) but opposed by labour unions who argued that this was against the purpose of the UIF (Milliyet 2007d). The AKP government rejected TOBB’s proposal (Milliyet 2007b). Nevertheless, the episode showed that demands to make use of the fund were increasing.

Second, the accumulated surplus had surpassed what was warranted in terms of fiscal prudence. The fund had reached close to 4 per cent of GDP by 2005. At the same time the budget deficit had decreased, so that by 2006 the fund’s annual surplus was close to the annual budget deficit, as Fig. 4.3 illustrates. In theory, the fund could thus have been used to finance nearly all new debt. The decreasing budget deficit also relieved the IMF pressure on the government. Furthermore, the IMF programme was to expire in mid-2008. According to some sources, the IMF had previously opposed any substantial changes in the programme due to concerns about the budget deficit (Hürriyet 2008).Footnote 6 In this sense, from a fiscal perspective, using the money accumulated in the UIF became much more feasible by 2008.

Unemployment insurance reform became part of the official government agenda in the AKP’s second term starting in 2007. The AKP planned to make the policy more generous and accessible, as well as expanding ALMP measures, such as vocational training (Neziroğlu and Yılmaz 2013). Strengthening ALMP had long been demanded by business (Hürriyet 2003) and in the PES’ General Assembly Decisions, where social partners voiced their views (İŞKUR 2005a, 2005b, 2007). The draft bill presented by the government in 2008 proposed numerous changes to the programme. It envisioned making the policy more generous, expanding ALMP (including the creation of employment subsidies) and using the UIF to fund non-labour market policies. Of these three proposals only the latter two were eventually legislated. Still, the failed plan to make unemployment insurance benefits more generous deserves attention.

The draft legislation set unemployment benefit at 50 per cent of the gross wage instead of 50 per cent of the net wage. This would have resulted in substantially increased replacement rates of up to two-thirds of the net wage (Öktem 2020b). The draft bill had been submitted to parliament by the prime minister, and thus it was no surprise that the respective parliamentary committees, which were dominated by government lawmakers, accepted the draft. Yet, unexpectedly, government lawmakers decided to veto their own proposal for higher benefits in parliament. In its place they passed a watered-down version of the relevant article in the legislation that provided only a minor benefit increase. Government lawmakers justified this decision by stating that the benefit increase in the draft bill was too high. The increase should be more “reasonable”, they argued, to ensure that “unemployment benefits were not made attractive” for workers (TBMM 2008, 945). In other words, benefits should be kept low to ensure that they do not have a decommodifying effect on workers. The government thus shied away from making the programme more generous.

Instead of increasing unemployment benefits, the reform strongly expanded ALMP. This was done by expanding access to ALMP from just unemployment insurance beneficiaries to all unemployed registered with the PES. As a result, participation in ALMPs, such as vocational training or on-the-job-training, strongly increased. Crucially, the reform also included the creation of employer subsidies for the employment of young workers and women. This was received very positively by business (Dünya 2008), which had earlier demanded using the UIF for employer subsidies, as described above. Overall, ALMP expansion was thus in line with the demands of business. Moreover, it also conformed to policy ideas promoted as part of the Europeanisation of labour market policy. ALMP was one component of “flexicurity” that may be seen as the key concept behind the EU’s labour market policy (Bolukbasi and Ertugal 2013). Hence, the expansion of ALMP was driven by two causal mechanisms: the mechanism of business-led reform and the mechanism of transnational cooperation.

In addition to strengthening ALMP, the reform diverted money from the fund to a development project in southeast Turkey. This conforms to the double benefit mechanism, as the UIF was used for non-social policy purposes. In fact, these non-social policy-related expenditures surpassed the fund’s regular expenditures for passive and active labour market policies from 2008 to 2010, as can be seen in Fig. 4.4. Yet, this move was opposed by some section of business, which feared that this would mean a return to the era of “populist” policies when dedicated public funds were spent for unrelated policies (Cumhuriyet 2008). The move only received the open support of a government-aligned labour union (Dünya 2008).Footnote 7 This might explain why government was more cautious in using the UIF for totally unrelated policies afterwards.

Fig. 4.4
A 100% stacked bar chart of the share of unemployment insurance fund expenditure annually. Dominant categories are as follows, unemployment benefits from 2002 to 2007, others from 2007 to 2012, A L M P from 2012 to 2017, and employment subsidies from 2017 to 2019.

Share of unemployment insurance fund expenditure by category. (Source: Öktem (2020b))

To conclude, the 2008 reform triggered by the economic crisis constituted a significant change in the structure of Turkey’s unemployment insurance programme. Before the reform, the programme’s focus was nearly exclusively on providing unemployment benefits. Through the reform, new instruments were layered on top of the classic unemployment insurance programme. Moreover, the reform laid out the path for the following years in which ever-new instruments were added to the programme. These new instruments were often as much to the benefit of employers as they helped employees. This led to opposition criticism that workers’ funds were being used for employers. However, the government was unimpressed by this view. In parliament, the Minister of Labour and Social Security Faruk Çelik outlined the underlying argument for the new policy direction, showing that the government basically accepted the idea that what benefits employers would benefit society:

The argument has been made that “with this reform you [i.e. the government] give resource to the employers”. Respected friends, there is no longer any difference between employers and employees. Employees and employers are seen as inseparable. This is the point reached in industrial relations. This is the point reached today after the antagonistic perspective of the nineteenth and early twentieth century. (TBMM 2008, 877)

4.3 Frequent Policy Change (2009–2019)

Following the 2008 reform, the unemployment insurance programme saw a decade of frequent policy changes. These changes mostly focused on expanding ALMP and employer subsidies. Furthermore, new policies were increasingly devised as temporary instruments that were made permanent through subsequent changes. Table 4.1 summarises key policy changes made between 2009 and 2019. Let us look at the maze of reforms in the realm of employer subsidies and active labour market policies in more detail.Footnote 8

Table 4.1 Significant policy changes to unemployment insurance (2009–2019)

Active labour market policies, such as vocational training and on-the-job-training, were massively expanded following the 2008 reform. This was done through frequent and often temporary policy changes. In mid-2009, the amount that the UIF could allocate to ALMP (expenditure limit) was temporarily increased for 2009 and 2010. In 2011, a comprehensive reform broadened the aim of ALMPs and increased the expenditure limit permanently. In 2015, the expenditure limit was increased twice. In early 2017 and late 2018, this expenditure limit increase was temporarily extended. Thus, although temporary in nature, the expenditure increases in practice became permanent. As a result, by 2015 expenditure on ALMP had eclipsed spending on unemployment benefits (see Fig. 4.4).

ALMP expansion broadly followed the demands of business (TISK 2010, 114, 2013). The government justified this expansion by pointing to expected growth in employment levels, as well as employability and productivity of employees (TBMM 2011, 40). However, policy changes were also increasingly linked to the election cycle. For instance, the increases in the ALMP expenditure limit in 2015 coincided with general election campaigns. This suggests that the changes are also a case of the outcompeting mechanism. Particularly instruments such as public works could be flexibly deployed to offer tangible benefits to select parts of the electorate. Looking at beneficiary statistics confirms this assessment. For instance, monthly beneficiary numbers in the public works programme, which provides temporary, low-skilled and low-paid labour (Gün 2013), peaked before the 2017 referendum, the 2018 general elections and the 2019 local elections. Thus, by all accounts ALMP expansion went beyond following business demands for increasing employment to serve as part of the government’s election agenda.

In addition to classic ALMP, employer subsidies were massively expanded. This expansion also occurred through frequent and mostly temporary policy changes. Employer subsidies had been introduced in 2008 and consisted of reduced contribution rates for employers hiring women and young employees within the following one year. This instrument was extended for one more year in early 2009. In late 2009, new employment subsidies were launched that applied to broader parts of the unemployed. In early 2011, new (temporary) employment subsidies for unemployed were created, which were temporarily extended in 2015. In early 2015, special employment subsidies for participants in on-the-job-trainings conducted by the PES until the end of 2016 were launched. These subsidies were temporarily extended in late 2016 and 2017. In early 2017, new and more generous employment subsidies for unemployed were devised that covered hirings made in 2017. Finally, in early 2018, new employment subsidies were launched that were even more generous and applied to hirings made until the end of 2020. Thus, what started as a small, temporary crisis response in 2008 to encourage employers to hire women and young workers became an increasingly permanent policy that covered an ever-greater share of the labour costs for more and more new hirings. Tellingly, by 2018 employment subsidies had become spending item number one, as can be seen in Fig. 4.4.

Again, the increase in employer subsidies responded to demands made by business. Business organisations actively demanded that “employers who create employment should be subsidised” (TISK 2010, 111; TISK 2013, 104). Labour unions, on the other hand, opposed using the UIF for such employment subsidies, even if some unions were sympathetic to the idea of employer subsidies in principle (Dünya 2015). In a way, with the recent massive increases in subsidies, the government finally returned to the plan that had been first outlined by TOBB in the mid-2000s. TOBB had unsuccessfully demanded that all additional employment should be subsidised by the UIF. However, in the late 2010s employer subsidies basically resembled what TOBB had wanted all along. In fact, the latest employer subsidies were even formally announced as a cooperation between the government and TOBB (TOBB 2019). Furthermore, similar to ALMP, employer subsidies were timed to coincide with election campaigns with the goal of reducing unemployment just before the elections. Therefore, both the business-led reform mechanism and the outcompeting mechanism apply to developments in this area.

Beyond the expansion of ALMP and employer subsidies, the programme was also reformed to include a number of new instruments, which were often beyond the scope of classic labour market policies. In 2014 and 2015, payments for mining workers affected by some high-profile mining accidents were made. In 2015, some supports for employees and employers active in dangerous professions were created. In 2016, support payments for internships were launched. In addition to these mostly temporary measures, a kind of parental leave benefit was created, also in 2016, as a part of unemployment insurance. The Part-Time Compensation (Yarım Çalışma Ödeneği) provides benefits to parents of new-born children who decide to work part-time for up to six months. Yet, the compensation has had a very limited reach so far. Overall, the level of expenditure devoted to these non-labour market policies never reached the scale of the 2008–2011 period. Furthermore, these new policies were not as far away from the original function of the UIF as the development project financed in 2008–2011. It appears that the criticism of using the UIF to finance a development project (launched not just by labour unions, but also by business) made the government more cautious in this regard. The idea that the UIF should not be used to finance non-labour market policy and that social partners should always be consulted has become a mainstay of demands made by TISK (2013, 106) and has been voiced in the General Assembly of the PES (İŞKUR 2009, 2011, 2013, 2015).

To conclude, the second decade of the unemployment insurance programme was shaped by frequent reforms that established new policies. These reforms can generally be seen as driven mainly by three causal mechanisms: business-led reform, transnational cooperation and outcompeting.

5 Conclusion

In this chapter, I explored the introduction and transformation of unemployment insurance in Turkey with a focus on the causal mechanisms behind the developments. I focused in particular on the double benefit of social insurance as a social policy programme and as a financing device, as this double benefit mechanism shaped the early phase of the programme and facilitated subsequent reforms.

The analysis has shown that policymakers designed unemployment insurance overly cautiously with restrictive eligibility criteria and modest benefits. This was due to negative experiences with previous social insurance policies and related funds, and due to the dire economic and fiscal situation the country was in when unemployment insurance legislation was passed in 1999. As a result, the Unemployment Insurance Fund was bound to generate an ever-increasing surplus, which would in all likelihood be at least partly invested in public debt.

Soon after the first contributions were collected in 2000, policymakers became aware that the surplus would exceed their expectations. They reacted by nearly halving contribution rates, but this was not enough to create an actuarial balance and so the fund kept growing. The money accumulated in the fund was used to buy public debt, and this helped the country meet the IMF’s stringent fiscal criteria. In the early phase, unemployment insurance was thus as much a state-financing tool as it was a labour market policy.

However, the fund’s increasing reserves also increased demands for policy change. These demands were first brushed aside by the government, but as the state’s fiscal situation improved, the need for the UIF as a financing tool decreased. Thus, when the global economic crisis hit Turkey in 2008, the government comprehensively reformed the legislation. Responding to demands by business that promised increasing employment and productivity, the government strengthened ALMP and introduced employer subsidies. Furthermore, it used part of the money accumulated in the fund for a development project. Thus, in line with the double benefit mechanism, the fund was used for non-social policy purposes.

In the decade after the 2008 reform, the government frequently amended the unemployment insurance legislation. The focus of these reforms lay on ALMP and employer subsidies. As a result of repeated changes, the weight of the programme shifted first from UI to ALMP and then to employer subsidies. In terms of policy content, these changes mainly responded to demands made by business. In this regard, the changes are an example of the business-led reform mechanism. In addition, the expansion of ALMP is also part of the Europeanisation of labour market policy. Therefore, the transnational cooperation mechanism also explains the reforms. However, the different reforms were also increasingly tied to the election cycle. Especially insofar as they offered the opportunity to provide tangible benefits to voters, reforms were announced and implemented during election campaigns so as to “outcompete” opposition parties.

Overall, the findings on the significance of the double benefit mechanism resonate with the literature on the fiscal politics of the welfare state. Both “legitimation” and “surplus effects” (Koreh 2017b) can be observed in the development of unemployment insurance in Turkey. When the policy was introduced, contributions for the troubled and highly contested Mandatory Savings Fund were channelled to the new programme. Thus, the state was able to continue levying the same contributions from employers and employees by introducing unemployment insurance. The surplus effect is also clear: it was the very size of the UIF that allowed the government to attach new instruments, such as employer subsidies, to the programme.

Yet, it is important to emphasise that the double benefit mechanism, by itself, does not explain the whole story. In particular, the direction of policy change can be better explained by other mechanisms. The influence of business on government policy, the desire to outcompete the opposition in the context of competitive (even if increasingly unfair) elections and the cooperation with transnational actors all played a crucial role in shaping the development of unemployment insurance in Turkey.