Fiscal policy responses during the Great Recession through a gender perspective

The impacts of fiscal policy responses on the recent crisis are an emerging field of exploration. A limited number of studies have appeared thus far, in part owing to the difficulty in researching an ongoing and continually evolving event. ^{Footnote 1}

The research on policy responses to date displays the great variety in reactions and policy mixes implemented during the current global crisis across the developed and developing countries, utilising different combinations of monetary and fiscal policies and support from international financial institutions (IFIs). In fact, policy responses to the recent economic crisis have been crucial in either limiting or aggravating the first wave of adverse effects. Although it is difficult to generalise about differences in policy formulations, the overview of the initial policy responses highlights two major approaches to fiscal policy: expansionary and contractionary. ^{Footnote 2} It appears that countries with relatively stronger fiscal positions and without debt payment problems were able to implement expansionary fiscal policies, as has been the case in many developed and developing countries, particularly Russia and some Asian and Latin American countries.

Looking more closely at the fiscal measures taken to boost domestic demand, it can be observed that responses from the developed countries, with their relatively strong social protection systems, primarily operated through ‘automatic stabilisers’ built into their fiscal policies (Reference Lahey and de VillotaLahey and de Villota, 2013; Reference Ortiz and CumminsOrtiz and Cummins, 2013). In contrast, the developing countries’ responses have tended to rely on discretionary spending, and measures have been more directed towards infrastructure and, to some extent, social transfers, owing to the lack of comprehensive social security systems (Reference EspinoEspino, 2013; OECD, 2010; Reference Ortiz and CumminsOrtiz and Cummins, 2013). In the developed countries and in the OECD countries in particular, tax measures appeared to be the dominant form of fund transfers to people affected by the crisis (OECD, 2009).

As a body of evidence accumulated, the first wave of policy responses attracted serious criticism, especially of the effectiveness of fiscal stimulus packages in protecting the most disadvantaged segments of society (Reference Belser and LeeBelser and Lee, 2011; Reference Berg and TobinBerg and Tobin, 2011; Reference SaithSaith, 2011). The impact of expansionary policy packages on increasing GDP growth was obvious (Reference Arnold, Brys and HeadyArnold et al., 2011; Reference Beetsma and GiuliodoriBeetsma and Giuliodori, 2011), but job recovery remained dismal and the jobs crisis persisted. In this respect, the limited funds allocated to job creation came in for some of the strongest criticism (Reference Leschke and JepsenLeschke and Jepsen, 2011; Reference McCarty, Corner and GuyMcCarty et al., 2009). In addition, some studies have argued that the labour market and social protection measures were doomed to ineffectiveness because of their low coverage and the strictness of eligibility conditions ^{Footnote 3} (OECD, 2010; World Bank, 2011). Increasing informalisation of labour markets further aggravates the coverage and eligibility problems. Moreover, infrastructural investments and public works programmes in the developing countries, which acted as a cushion during the crisis for those without access to social security or unemployment insurance, have been criticised for not sufficiently addressing community needs, often clashing with environmental concerns and failing to provide jobs for women (Reference Green, King and Miller-DawkinsGreen et al., 2010).

Despite evidence on the gendered impacts and outcomes of the crisis, the policy responses formulated by public authorities to prevent financial and economic collapses were notoriously gender-blind (Reference SeguinoSeguino, 2010). Public funds allocated to employment creation have been inadequate, and the limited funds typically addressed male-dominated sectors, as in the case of Germany, where the fiscal stimulus package had a strong male bias due to its focus on the automotive industry (Reference Lahey and de VillotaLahey and de Villota, 2013; Reference Leschke and JepsenLeschke and Jepsen, 2011).

Infrastructural investments such as roads, dams and water systems have become a ‘saviour’ for many governments in the developing countries as a way to boost domestic demand during the recession and generate employment, mainly for male workers. In the same way, fiscal stimuli mostly addressed capital-intensive large enterprises that employ few workers. This is also likely to negatively affect women, who disproportionately work in labour-intensive sectors such as textiles and clothing and services. Those sectors were largely neglected in the fiscal packages (Reference King and SweetmanKing and Sweetman, 2010; Reference McCarty, Corner and GuyMcCarty et al., 2009).

Direct job creation in labour-intensive sectors, which unfortunately was not a primary component of most fiscal packages, was found to be more equitable, growth-enhancing, poverty-reducing and affordable (Reference AntonopoulosAntonopoulos, 2009a; Reference Antonopoulos, Kim and MastersonAntonopoulos et al., 2010; Reference Lahey and de VillotaLahey and de Villota, 2013). In their study of the United States, Reference Antonopoulos, Kim and MastersonAntonopoulos et al. (2010) show that the social care sector creates more jobs than the physical infrastructure and green energy sectors, in addition to having positive income distribution effects. Job creation in the social care sector is found to benefit the most vulnerable segments of the society, those with low skill levels and low incomes, with women being among the major beneficiaries. In this respect, South Korea’s experience is worth mentioning: Increasing investments in and support for care work during the 2000s were continued during the 2008 crisis as a way of coping with the recession, initiating economic growth and incorporating women and the elderly into the labour market (Reference Peng and İlkkaracanPeng, 2010).

In the recent crisis, however, women had less access to job activation, unemployment insurance and social protection measures, as seen in the case of Germany and the United Kingdom (Reference Leschke and JepsenLeschke and Jepsen, 2011). These policy tools and social security systems are, in general, linked to formal-sector employment and the level of past earnings. As women tend to work in the informal sector in greater numbers and/or have non-standard conditions of employment, usually with lower wages, many of them are therefore likely to be excluded from contributory social security systems (Reference Cichon, Behrendt and WodsakCichon et al., 2011; International Labour Office (ILO), 2010).

Tax incentives were another major component of fiscal stimulus packages. While there is a prevailing assumption that these are gender-neutral, they could be shown in this case strongly to benefit high-income group. As men earned disproportionately higher incomes, the tax incentives favoured them (Reference Lahey and de VillotaLahey and de Villota, 2013; Reference Leschke and JepsenLeschke and Jepsen, 2011).

These few examples of policy implementation with diverse consequences for men and women, nevertheless, provide sufficient evidence from which to argue that the estimated impacts of measures to enhance societal welfare and eradicate poverty will be deficient without considering the disparate gendered outcomes of crisis management on welfare, income distribution and the labour market.

Turkey’s fiscal stimulus package

In October and November 2008, Turkey’s Central Bank (CBRT) and Banking Regulatory and Supervisory Agency (BDDK) led, through monetary policy, the first series of measures to combat the initial impacts of the crisis. The CBRT incrementally reduced the overnight policy rate from 16.75% in October 2008 to 6.50% in November 2009 and eased liquidity conditions both in terms of foreign exchange and the Turkish Lira. The BDDK placed limits on banks’ profit dividends to strengthen equity capital (Reference AkyüzAkyüz, 2010; Reference Öniş and GüvenÖniş and Güven, 2011).

The government announced Turkey’s fiscal stimulus package in March 2009. It rested on four pillars ^{Footnote 4} : tax incentives, employment supports, investment incentives and regulations on personal loans and credit cards (Table 1). Tax measures included a list of concessions and reductions provided to individuals and the business sector, with the aim of increasing private consumption and investment, as well as reducing production costs. Employment supports consisted of active and passive labour market policies intended to protect the already employed, to compensate those who became unemployed due to the crisis, to increase the employability of workers by enhancing their skills and knowledge and to generate new employment. Investment incentives mainly involved subsidies reducing the share of insurance premiums paid by employers and, to a lesser degree, reductions in corporate loan interest rates, land allocation for investment, financing availability for small- and medium-size enterprises (SMEs) and agricultural enterprises, incentives to exporters and credit guarantee support to SMEs. Finally, regulations on personal loans and credit cards were intended to keep credit channels open to the corporate sector and to households, by relaxing conditions on foreign-exchange-based and foreign-exchange-linked loans and by restructuring non-performing credit card debts, respectively.

Table 1. Fiscal stimulus measures taken in Turkey to combat the global crisis.

Source: Undersecretariat of Treasury (2009).

VAT: value-added tax; SME: small- and medium-size enterprise.

The size of Turkey’s fiscal package was significantly larger than those of the OECD countries as a group, but modest compared to those of some of the countries in the Asia-Pacific region (China, Malaysia, Singapore, South Korea and Vietnam) that responded to the crisis rapidly and comprehensively (Figure 1). In Figure 1, the magnitude of the various fiscal packages is shown as the net change in consolidated general government balances during the period 2008–2010 as a percentage of 2008 GDP (OECD, 2009). Only discretionary spending in response to the crisis, namely, those expenditures not mobilised automatically depending on the performance of the economy, was taken into account in the calculation of announced fiscal packages. ^{Footnote 5} Negative lines show expansionary fiscal policy, while positive lines show contractionary fiscal policy implemented during the 2008 crisis. Turkey’s sizeable fiscal package, corresponding to 5.5% of 2008 GDP, indicates the relative ease with which the government was able to announce expansionary fiscal policy in the midst of the crisis. In contrast to countries such as Hungary, Iceland, Ireland and Portugal that had to respond via contractionary policies, Turkey was not constrained by a high public-debt-to-GDP ratio, weak fiscal balances or a weak banking sector prior to the global crisis. These advantages provided sufficient room for increasing public expenditures without endangering financial stability and public finances. In this respect, the size of Turkey’s public expenditures being noticeably larger than most of the OECD was an important element in preventing the most adverse consequences of the global crisis.

Figure 1. Economic growth and size of the stimulus packages (2008–2010) in OECD and Asia-Pacific countries (as a percent of 2008 GDP).

Source: OECD (2009).

Magnitude of the fiscal packages is shown as the net change in consolidated general government balances during the period 2008–2010. The size of Turkey’s fiscal package has been calculated by the author, as the cumulative net change in the period 2008–2010, expressed as a percent of 2008 GDP, using the data in the Pre-Accession Economic Program (DPT, 2009). Asian fiscal packages are derived from Reference AkyüzAkyüz (2010) and estimated for 2009 as a percent of 2009 GDP (GDP from IMF, 2011).

Disaggregating Turkey’s fiscal package ^{Footnote 6} into spending and tax measures, it can be seen that the share of taxes in the planned stimulus package is 14%, while the remaining 86% constitutes the spending items. As a percentage of 2008 GDP, tax incentives correspond to a very limited figure, 0.8% (Table 2). This feature of Turkey’s fiscal package, relying on expenditure items rather than tax incentives, is not very different from many developing countries, where the policy responses mainly relied on infrastructural investments and social transfers as a quick way of reallocating funds, instead of utilising tax and social security system channels.

Table 2. Breakdown of Turkey’s fiscal package (I).

Source: DPT (2009); author’s calculations.

GDP: gross domestic product.

A sector-by-sector (households, labour and business segments) breakdown of Turkey’s fiscal package shows that public infrastructural investments in the Southeast Anatolian region and public highways received the lion’s share with 34%, followed by funds directed at boosting household consumption with 26% (Table 3). Therefore, Turkey’s crisis measures gave priority to boosting domestic demand by means of large public investments and consumption. Then followed the non-negligible 20% of the funds allocated to labour-related measures (1.1% of 2008 GDP), while business sector incentives to be used directly in business activities amounted to 9% of allocations.

Table 3. Breakdown of Turkey’s fiscal package (II).

Source: DPT (2009); author’s calculations.

GDP: gross domestic product.

Table 4 depicts the breakdown of the non-negligible labour-related measures announced in the fiscal package. The biggest item among these measures appears to be the reduction in employers’ social security premium contributions, constituting 18% of the total fiscal package for a measure that primarily favours employers. The remaining labour market measures intended either to generate or to protect employment (active employment measures, short-term working arrangements, unemployment insurance and public works programmes) amount to a meagre 2% of the total package (or 0.1% of GDP), a very dismal amount transferred directly to workers.

Table 4. Breakdown of Turkey’s labour market measures.

Source: DPT (2009); author’s calculations.

GDP: gross domestic product.

In summary, Turkey’s fiscal plan prioritised public infrastructural investments and domestic consumption, while employment creation/protection was expected to result from the private sector being given incentives in the form of sizeable reductions in social security payments. As a result of this structure, the expected major channels through which economic growth was to be triggered were not the automatic stabilisers (taxes, social protection and unemployment insurance expenditures, in particular) but direct expenditures in construction and one-off tax relief. Therefore, Turkey’s fiscal package cannot be labelled as pro-labour but rather pro-business (Bağımsız Sosyal Bilimciler (BSB), 2011: 35–36), with the public sector’s role in employment creation and protection assumed to be limited.

Employment policies through a gender lens

In comparison with many other fiscal stimulus packaged announced elsewhere, it would be unjust to claim that Turkey’s fiscal package was gender-blind. Women’s employment received special attention among the labour market measures announced in the fiscal plan, with TL 219 m in allocations in the 2008–2010 period designated for encouraging new employment of women and young people (Devlet Planlama Teşkilatı (DPT), 2009: 38). This goal is expected to be achieved through the reduction in the employer share of social security premiums, a measure corresponding to 0.02% of GDP or 1% of all labour market expenditures.

The encouragement of women’s employment is not a new phenomenon that only materialised during this recent crisis. Women’s participation in the labour markets in Turkey has been very low by all measures over the last two decades (e.g. 31% in 2013). In public speeches and documents, special emphasis has been placed on increasing women’s employment, and efforts to this end coincided with Turkey’s EU convergence process in the late 2000s. One such commitment to increase women’s employment was the new employment law (Employment Act #5763) enacted in May 2008. This involved a reduction in the employer share of social security premiums for new female employees (at a descending rate from 100% in the first year to 20% at the end of 5 years). Appraised in the Progress Report published by the European Commission (2008: 61) this measure was later incorporated into the fiscal stimulus package announced in 2009, extending the law until 31 December 2015. ^{Footnote 7}

The government can in fact be an influential actor in the labour market in terms of creating jobs, facilitating worker employability (by increasing skills and education) and reducing or eliminating gendered structures in labour markets (such as segregation in jobs and discrimination against women). On the other hand, public policies can also serve to reinforce gendered structures and further discriminate against women. In fact, the labour market in Turkey is already heavily segregated against women, which is evident in the sectoral and occupational distribution of women employees (Reference İlkkaracanİlkkaracan, 2010). The agricultural sector employs more than 40% of all women, mostly as unpaid family workers, followed by the services sector.

In this respect, Iskur emerged as an important actor, given its increasing role in Turkey’s labour market over the past decade. This role was further strengthened during the 2008 crisis, as evident in the surge in applications to the institution by both men (up 74%) and women (up 218%) between 2007 and 2011. With women’s employment level in Turkey hovering around 25%–30% for decades, Iskur received EU funds during the 2000s to reduce this gender gap and rapidly became a focal point for women, who face multidimensional obstacles to entering the labour market. As of the end of 2011, work facilitated by Iskur accounted for 1.8% of the total non-agricultural employment for men and, even more importantly, 2.5% of the total for women. ^{Footnote 8}

Figure 2, constructed with data from Iskur’s bulletins and reports, shows separately the percentage of women job seekers among total Iskur applicants, among job seekers meeting employers referred by Iskur (represented in the figure as ‘presentation to the employer’) and among job seekers placed in a job (represented in the figure as ‘placement in a job’). Since the outbreak of the crisis, there has been a significant increase in the ratio of female applicants to the institution and a rather gradual but obvious increase seen in women’s share of job seekers meeting employers and work placements. An important part of the rise in the percentage of female applicants to Iskur from around 30% to 41% can be attributed to the ‘added worker effect’, which is in line with other research findings on Turkey (Reference İlkkaracan Ajas and Değirmenciİlkkaracan Ajas and Değirmenci, 2012). Female applicants, according to 2011 figures, are mostly young (62% were between the ages of 20 and 34 years) with a primary or secondary school education (65%). The added worker effect is expected to be short-term in nature, as household members enter the labour market and/or increase their working hours to compensate immediately the income losses of other household members. In 2009, overall job losses among men in Turkey totalled 192,000, mainly in the non-agricultural sectors, whereas employment for women grew by 276,000. In the following years, employment growth for men was resumed.

Figure 2. Impact of Iskur policies on women’s employment (women as a percent of total applicants to Iskur, as a percent of job seekers meeting employers referred by Iskur and as a percent of total job seekers placed in a job, seasonally adjusted).

Source: Türkiye İş Kurumu (İŞKUR, 2007–2012). http://www.iskur.gov.tr/KurumsalBilgi/istatistikler.aspx#dltop.

Grey shaded area is the first quarter of 2009, the most intense period of the 2008 crisis.

The rise in the other two categories in Figure 2 falls short, however, of the pace of the increase in female applicants, apparently indicating a situation in which Iskur could not sufficiently meet women’s demand for employment. Over the period 2007–2012, the moderate increases in job placements and in the number of applicants presented to employers imply an evident effort on the part of Iskur to increase women’s employment. But despite those efforts, women constituted only 28% of people referred to employers, and only one-third of those who actually were placed in occupations, as of June 2012.

Addressing the question of which types of occupations are promoted and discouraged by Iskur, one needs to look more closely at occupations according to the levels of economic activity. Table 5, derived from Iskur’s data, shows the gender composition of employment by economic activity during the crisis period between 2007 and 2009, and the following recovery phase. The year 2007 has been chosen primarily because it was the last pre-crisis year. As in many other countries, the economic crisis in Turkey disproportionately affected men’s employment opportunities compared to women’s in the period 2007–2009, especially in the manufacturing, trade and utilities sectors. ^{Footnote 9} In this particular period, the increase in women’s employment during the crisis took place largely in female-dominated sectors, with a clear shift away from occupations treated as ‘male jobs’ to those seen as ‘female jobs’. Notable increases in women’s occupations facilitated by Iskur can be observed in education, social and community services; the hotel and restaurant sector; construction and public works; real estate and renting activities and transportation, health and agriculture, all of which (except for construction) are female-dominated sectors. On the other hand, the crisis conditions led to a sharp contraction in women’s employment, similar to that of men, in the manufacturing, wholesale and retail trade sectors, followed by declines in international organisations, mining, banking and utilities. In all these sectors, except the manufacturing industry, the rate of contraction in women’s employment surpassed that of men, reflecting the relatively more vulnerable positions of women in those segments of economic activity.

Table 5. Gender composition of Iskur occupations, Turkey, 2007–2011.

Source: Türkiye İş Kurumu (İŞKUR, 2007–2012).

A few sectors, namely, the financial, wholesale and retail trade sectors, thus seem to have acted as a buffer during the crisis, with the female share of employment declining during the bust and increasing during the recovery. The financial sector accounts for around 50% of official female employment in Turkey. As Iskur figures show, the rate of contraction in employment was sharper for women than for men during the crisis and notably higher during the recovery period, in line with the Buffer Hypothesis.

During the recovery between 2009 and 2011, the total rate of growth of women’s employment was notably higher than that of men (245% vs 195%, respectively), evidence of Iskur’s continued efforts to promote women’s employment (Table 5). In 11 of the 17 sectors of economic activity, the employment growth rate among women was higher than that of men, leading to significant increases in women’s share of employment in the agriculture and major services sectors. The 6% increase in women’s share of manufacturing sector employment appears promising, although the concentration of women in industries related to textiles and clothing is in line with the usual gendered segmentation of labour markets. Therefore, major advances were achieved in areas thought to be suited to ‘women’s nature’, showing the prevalence of a gendered structure of occupations and job segregation in Turkey’s labour market – a structure Iskur appears to be facilitating or accommodating.

Moreover, consecutive increases in women’s share of total job placements during both the bust and recovery periods indicate that women were proportionately better-protected in agriculture, real estate and renting activities, social and personal services, the manufacturing industry, transportation, communication and the wholesale and retail trade sectors. These findings serve to confirm gender segregation in Turkish manufacturing industry and the resilience of these sectors to the crisis. However, it is difficult at this stage to disentangle to what extent the increase in the female share of employment is due to the substitution effect of employing women instead of men. Research that goes beyond the meso-level analysis presented here, and based on Iskur’s job placements, is required to sufficiently verify the separately protective and substitution effects of gender segregation in the Turkish labour market.

Iskur’s tendency to direct women towards gender-segregated jobs is apparent not only in job placements but also in the agency’s training courses: those with predominantly female participants provide skills and knowledge for ‘women’s jobs’. More formally, women are trained in the skills that enable them to search for job opportunities in a limited range of sectors with lower pay and status. Table 6 shows the number of beneficiaries in Iskur’s courses in 2011 ^{Footnote 10} and women’s concentration in terms of sectors: women’s participation is highest (i.e. 75%–100% of participants are women) in courses such as cosmetics, care services, textiles/clothing and secretarial services. The second-most female-dominated group of courses (with 50%–74% concentration ratios) are again those requiring what are thought of as ‘womanly’ skills, such as office-related work, accounting, cooking and pastry-making, housekeeping and marketing. It can be observed that women are less likely to receive training in the manufacturing, machinery and maintenance sectors. Therefore, Iskur’s active labour market policies (ALMPs) do not seem to be challenging the pattern of gender segregation in the labour market but rather to be reinforcing them.

Table 6. Gender-disaggregated composition of Iskur training courses (2011).

Source: Türkiye İş Kurumu (İŞKUR, 2007–2012).

http://www.iskur.gov.tr/KurumsalBilgi/istatistikler.aspx#dltop.