The COVID-19 crisis and policy responses by continental European welfare states.

Social protection in Germany, Belgium and the Netherlands share Bismarckian roots. Over time, these welfare states were however in constant flux and incorporated to a greater or lesser extend elements of both the Anglo-Saxon and Nordic models. While the Netherlands has from the beginning deviated from the Bismarckian model, in recent years this welfare state has undergone important reforms that have made it increasingly evolve into a "Bismarck cum Beveridge" model. Germany and Belgium also witnessed a dual transformation, with retrenched earnings-related benefits for long-term unemployed and an increasing number of atypically employed people on the one hand and expanded social security to the so-called "new social risks" on the other. It is against this changing institutional background that we can understand the similarities and differences in the extent to which these three continental welfare states used traditional social insurance systems to buffer the social and economic consequences of confinement. First, all three countries strengthened to varying degrees social protection systems for the active age population. So conceived, the policy responses were a response to the dual transformation of social protection that took place in recent decades without, however, changing its course. Second, the extent to which continental welfare states made use of existing social insurance schemes seems to be related to the extent to which these welfare states have moved in the Anglo-Saxon direction.

INTRODUCTION

In continental welfare states, social security has been a powerful instrument providing protection of individuals and families whose livelihoods have been threatened by unemployment, sickness and loss of economic activity during the prevailing COVID‐19 crisis. In doing so, social security also ensured systemic security by stabilizing consumption and helping businesses to bridge the crisis. There were however significant differences in the way social security was used as an automatic stabilizer. Belgium has to a large extent reverted to existing social insurance schemes; in Germany, current social insurance and also social assistance were used, while the Netherlands opted to use targeted benefits and to set up new protection mechanisms outside social security. The extent to which social security was deployed depended inter alia on the institutional context that existed before the crisis.

In this contribution, we look at social policy responses to the consequences of the lockdown in Germany, the Netherlands and Belgium with a focus on the use of social security for employees and self‐employed who had to temporaraly stop their work because of the health measures. Section 1 describes the main features of social protection systems in these countries. Section 2 deals with the social and economic situation before the pandemic. Section 3 zooms in on the asymmetrical way in which the crisis has affected different social–economic groups. Section 4 deals with policy responses and discusses similarities and differences. Section 5 concludes.

BISMARCK CUM BEVERIDGE: SOCIAL SECURITY IN CONTINENTAL WELFARE STATES

Germany, the Netherlands and Belgium belong to the group of affluent European democracies with a relatively comprehensive welfare state, including strong social protection and a statutory health insurance, providing quasi‐universal access to world‐class health services. Social benefit expenditure is high, at around 28% of GDP in all three countries.

Social protection systems encompass different, distinct and mutually complementary layers in which means‐tested social assistance, contribution‐based proportional benefits, flat rate contribution or tax‐based replacement incomes and occupational pensions co‐exist to varying degrees. Traditionally, three types of social security systems are distinguished. The Anglo‐Saxon or Beveridge like type of social insurance offers flat rate, contribution‐based benefits guaranteeing minimum income protection. In these welfare states, in the absence of adequate benefits, social assistance and private insurances have a relative high weight. The continental or Bismarckian social security systems are also insurance based. Contributions and benefits are linked to the employment status. There are different systems for employees, the self‐employed and civil servants. Social protection is part of the social dialogue and the social partners have an important role in the design and the operation of the schemes. That is why these systems are often labelled as “socio‐professional” or “corporatist.” They typically offer more than just minimum protection: the aim is to guarantee living standards through proportional benefits that replace income from work. In these settings, social assistance and private insurances remain rather marginal. The Scandinavian “demogrant systems” offer citizen's and tax based, flat rate benefits principally guaranteeing minimum income protection (at a comparatively high level). Traditionally, on top of social protection, the Nordic model also places great emphasis on the so‐called “enabling policies” and social promotion through universal and lifelong education, social services and active labour market policies (e.g., Cronert & Palme, 2019).

Although social security systems became more heterogeneous, to this day, the three welfare systems already described by Titmuss in Essays on the Welfare State (Titmuss, 1958) and later empirically reinvigorated in Esping‐Andersen's Three Worlds of Welfare Capitalism (1990) remain relevant. However, the Anglo‐Saxon, Continental and Nordic social security types never existed in pure form – the relative weight of social insurance, social assistance, social promotion and private insurances varies across countries and changes over time. In recent decades, for example, there has been a clear shift within Bismarckian systems towards activation, more social assistance and a greater emphasis on minimum income protection (see Palier, 2010; Hemerijck & Marx, 2012; Hinrichs, 2010). To a greater or lesser extent, over time, these Bismarckian systems incorporated elements of the Beveridge and the Nordic models.

Many typologies of welfare states categorize Germany, the Netherlands and Belgium in different groups (Arts & Gelissen, 2002): Esping‐Andersen (1990) placed Belgium and the Netherlands in the group of “social democratic welfare states” and Germany in the group of “conservatives”; Castles and Mitchell (1993) grouped Germany and the Netherlands in the “conservatives” and Belgium in the group of “non‐right hegemony” countries while Korpi and Palme (1998) assigned Belgium and the Netherlands to the group of corporatist welfare states and characterized the Netherlands as a “basic security” model. This indicates the intrinsic hybrid nature of these systems. They share the same Bismarckian origin but they are in constant flux and became more mixed from the side. Especially in recent decades, these welfare states have undergone major reforms that – to a greater or lesser extent – have given their social security characteristics of the Anglo‐Saxon system.

Germany and Belgium share the same Bismarckian roots: these welfare states are typical “social insurance states,” characterized by the predominance of institutionally segmented social insurance systems for different socio‐economic groups. Employees, the self‐employed and civil servants each have their own social protection system; protection is in principle earnings related and contribution based. In the Netherlands, on the other hand, from the onset, the social security system had a mixed Beveridgean and Bismarckian design structure. The Bismarckian component is made up of occupational social insurance provisions, providing earnings‐related benefits to workers and employees, financed through employers' and employees' contributions, insuring employees against sickness, disability and unemployment. The Beveridgean component pertains to universal basic pension supplemented by compulsory occupational pensions. Thus, in addition to Bismarckian features, the Netherlands has clear characteristics of the social democratic model (basic pension) and of the Anglo‐Saxon model (large share of assistance and private insurance) (Hemerijck & Marx, 2010). Moreover, the Netherlands entered the phase of “responsive corporatism” (see Visser, 1997, p. 106) earlier and more vigorously than its neighbouring countries: it is said to have begun with the Wassenaar Agreement of 1982, when unions, employers and government decided to revitalize the economy through shorter working times, wage moderation, privatization and budgetary restraints. In Germany and Belgium, since the second half of the 1990s, social protection systems also were subjected to sequential reform processes, giving them Anglo‐Saxon and Nordic characteristics (Hinrichs, 2010; Hemerijck & Marx, 2010): wage replacement schemes, traditionally aimed at status maintenance, were partly reoriented towards minimum income protection, protection shifted towards activation and social insurance contributions were replaced by an increasing higher share of tax‐funding. They remain however characterized by the architectural predominance of the socio‐economic segmented social insurance system stemming from the Bismarckian legacy (Hinrichs, 2010).

In the context of this contribution, the differences in social protection of the self‐employed, flexible workers and low‐wage earners are particularly important as the greatest impact of the COVID‐19 crisis on employment has been concentrated in these categories (see below). There are, moreover, major differences in the functioning of the labour markets in the countries under review. In the Netherlands and Germany, labour markets are more flexible than in Belgium. For instance, in the Netherlands and Germany, more than 52% of the workforce has a short‐time contract compared with 44% in Belgium. In the Netherlands, part‐time work is dominant, especially among women: according to OECD figures, 37% of employment is part time compared to 22% in Germany and 17% in Belgium (see Table 1). Greater flexibility in the Netherlands and Germany results in high employment rates. The less flexible labour market in Belgium translates into lower employment and higher unemployment rates, especially long‐term unemployment among the older and lower‐skilled part of the labour force. The following EUROSTAT‐figures are indicative of the importance of these differences: before the crisis, the rate of employment among the low‐skilled was 63.2% in the Netherlands, 61.8% in Germany and only 46.3% in Belgium.

TABLE 1

Public and private social spending, expenditures on means‐tested social benefits, share of temporary employment and part‐time work in the Netherlands, Germany and Belgium, 2019 or latest available year

	Netherlands	Germany	Belgium
Public social spending, 2019*	16.1	25.9	28.9
Gross private social expenditure, 2017 *	13.5	3.6	1.9
Expenditures on means‐tested social benefits as % of total social protection,2018**	14.9	12.4	5.4
Share of temporary employment, 2019 ***	20,2	12,0	10,8
Part Time Employment 2019 ****	37.0	22.0	16.9

Note: The high‐spending for means‐tested benefits in Germany is partly due to the fact that this includes the in‐work benefit for low‐wage workers.

Sources: * OECD Social Expenditure database, ** Eurostat (online data code: spr_exp_fto), *** https://stats.oecd.org/Index.aspx?DataSetCode=TEMP_Iource: OECD estimates based on national labour force surveys, **** https://stats.oecd.org/Index.aspx?DataSetCode=TEMP_I.

As in most countries, the initial income support for qualified unemployed in the three countries analysed is provided through contributory unemployment insurance systems in which benefits for eligible unemployed are a function of the previous wage. In Germany and the Netherlands, benefits are time‐limited. After the rights have expired, those who require continued income support (as well as those who do not qualify due to very low wages or an insufficient employment history) can fall back on unemployment assistance (in Germany) or on the generic social assistance programme (in the Netherlands). In Belgium, the duration of unemployment benefits is unlimited but benefits decline strongly over the unemployment spell (see for a discussion OECD, 2020). In recent decades, as a consequence of the persistence of structural unemployment, high benefit dependency and the ensuing perceived need for work‐related social security reforms, the access to and the generosity of unemployment benefits have been reduced. As a consequence, the number of beneficiaries of income tested social assistance has been on the rise (for Germany see Seeleib‐Kaiser, 2016 and for Belgium De Wilde, Cantillon, De Bie, & Vandenbroucke, 2017).

There are major cross‐national differences in social security for the self‐employed. In Belgium, social protection for the self‐employed is more developed than in Germany and the Netherlands (OECD, 2019). For self‐employed, who had to close down business, a specific social insurance scheme is in place, guaranteeing a flat rate income replacement and continued insurance for health care. Moreover, self‐employed can fall back upon the unemployment insurance for employees if they had a prior insurance record built up as wage earner (before they started up the self‐employed activity). The flat rate sickness income replacement benefits are at the same level as the guaranteed minimum benefits in the scheme for the employees.

In the Netherlands, where employees are insured for unemployment for a maximum of 2 years for 70% of their wage, the self‐employed are not insured for “unemployment.” Employees are insured for sickness and disability for 2 years through their employer, who is obliged to continue to pay their wages during the first 2 years of sickness. After 2 years, the WIA (an occupational disability insurance) covers this risk. The obligation to continue paying wages for employers during 2 years does not include the self‐employed and they are also not covered through the WIA. A specific social assistance for self‐employed is available, in particular for self‐employed who had to stop the business and/or where the business revenues due to old age or invalidity are below a defined minimum subsistence (Schoukens & Weber, 2020).

In Germany, unemployed self‐employed and employees who do not qualify for the contribution‐based unemployment insurance can access the means‐tested unemployment assistance scheme. Sickness earnings replacement benefits are “partial,” as self‐employed have to opt‐in to such coverage by the statutory sickness funds to be covered. In addition, the benefits are capped at EUR 109.38 per day. In all countries, self‐employed can always fall back upon the general social assistance schemes if their income is below the minimum subsistence threshold (e.g., due to closing down of the business).

Table 1 summarizes some key differences between the German, Dutch and Belgian social fabrics. The Netherlands stands out with the largest share of private social expenditures, a high proportion of means‐tested social benefits and a high incidence of part‐time work and temporary employment. Belgium, on the other hand, still seems to adhere most to the traditional Bismarckian model. As a result of the limited share of non‐standard work, a relative generous social protection scheme for the self‐employed and unemployment benefits that are not time‐limited, private insurance and social assistance remain marginal compared to its neighbouring countries. Arguably, these differences marked the divergent policy responses to the COVID‐19 crisis.

THE SOCIAL AND ECONOMIC SITUATION AT THE EVE OF THE PANDEMIC: A DUAL TRANSFORMATION OF SOCIAL SECURITY

The economic and fiscal development of Germany and the Netherlands since the 2008 crisis has been overall quite positive. The German government achieved fiscal surpluses and significantly reduced its public debt, a policy for which it was often criticized by the domestic opposition and international actors. In the Netherlands too, at the outbreak of the corona crisis, public finance and employment rates were good: public debt was low and unemployment rate was at an all‐time low level (3.2%). Belgium, on the other hand, recorded a persistently high public debt and relative low employment rates. Since the 2008 financial crisis, employment has risen significantly, but labour market participation remained relatively low, especially among the low‐skilled (Cantillon, 2018).

Poverty is relatively low though persistent, averaging between 16% in Belgium and Germany and 13% in the Netherlands (European Union Statistics on Income and Living Standards (SILC), EUROSTAT, 2018). Since the 2000s, at‐risk‐of‐poverty has followed different patterns in the countries under review: Germany has seen a reduction in at‐risk‐of‐poverty among the working‐age population and an increase among pensioners. In Belgium and the Netherlands, the opposite happened: at‐risk‐of‐poverty rose among the active population and decreased among the elderly. In the Netherlands, this was related to the increase of workers with a temporary contract and the rise of the so‐called “ZZP'ers” (self‐employed without personnel – staff). In Belgium, both in work‐poverty and poverty among jobless households increased (Cantillon, Goedemé, & Hills, 2019).

The impact of social security on poverty is generally considered to be very substantial. After the financial crisis of 2008, social benefit expenditure rose sharply, then fell back, but remained higher than before the great depression (ESPROSS, EUROSTAT). However, the protective capacity of social transfers decreased, especially among the working‐age population. If one assesses the distributional capacity of social transfers in the conventional way – by comparing poverty rates before and after transfers – in 2018 social transfers (excluding pensions) accounted for a reduction in poverty rates of around 33% in Belgium, 35% in Germany and 39% in the Netherlands (see Table 2). The corresponding figures in 2008 were 46, 37 and 47%, respectively. There are many factors that explain this significant weakening of social protection systems, including changes in the profile of social beneficiaries. However, an important explanation lies in the reduction of the generosity of social protection for the active population, especially in the case of long‐term unemployment.

TABLE 2

Impact of social transfers (excluding pensions) on poverty reduction: % reduction of at‐risk‐of‐poverty before social transfers, 2008–2018, EUROSTAT, last update: 17 December 2020

▼	Belgium	Germany	Netherlands
2008	45.56	37.19	47.24
2010	45.32	35.54	51.18
2012	44.77	33.74	50.97
2014	43.64	33.20	45.54
2016	41.51	34.78	42.53
2018	35.18	33.33	38.99

The flexibilization of the labour markets and growing pressures on unemployment benefits resulted in a growing division in social protection between workers with a permanent contract on the one hand and long‐term unemployed, flex workers, zero‐hour contracts and self‐employed for whom in most cases social protection is less generous. Since the 2000s, continental welfare states witnessed a dual transformation, which retrenched earnings‐related benefits for long‐term unemployed and atypically employed people and expanded social security to the so‐called “new social risks” (Bonoli, 1997) and “enabling” policies (Hemerijck, 2012). Work and family related spending – such as in‐work‐benefits and parental leave – has increased, while the generosity of traditional “passive” income support for the unemployed declined. This “dual transformation” (Bleses & Seeleib‐Kaiser, 2004) of the welfare state has included processes of “dualization” (Emmenegger, Häusermann, Palier, & Seeleib‐Kaiser, 2012), which continued to protect core workers at the expense of labour market outsiders, and may be at the root of the paradoxical increase in poverty among jobless households and the simultaneous further increase in social expenditure in recent decades (Cantillon, 2011). New social spending tends indeed to be less pro‐poor while passive income support for the unemployed became less protective. In Germany, only a minority – approximately a third – of the unemployed still receives earnings‐related benefits (see Bleses & Seeleib‐Kaiser, 2004; Seeleib‐Kaiser, 2016). Similarly, in Belgium, the Labour Force Survey reports a significant increase of the share of unemployed people without unemployment benefits, from 10% in the early 90s to 30% in 2018.

So conceived, on the eve of the COVID‐19 pandemic, social security systems in European continental welfare states appear to be weaker than at the time of the previous crisis: the dualization and flexibilization of continental welfare states in recent decades led to a social protection that at the eve of the pandemic had become less protective, especially for the bottom end of the labour market. Families, especially the less educated among them, were less resilient to bear the consequences of the COVID‐19 crisis than the labour market repercussions of the great depression in 2008. These families were also those most affected by unemployment, sickness and loss of economic activity during confinement.

THE ASYMMETRIC IMPACT OF THE LOCKDOWN ON EMPLOYMENT

Quicker than in the Netherlands and Germany, on 13 March 2020, the Belgian government announced far‐reaching measures to combat the spread of the coronavirus. All cafes, discotheques and restaurants were closed while all recreational, sporting and cultural activities, regardless of their size, were cancelled. From 16 March, schools were closed. From Monday 4 May, restrictions were gradually relaxed, first with the reopening of health care; on 11 May, the possibility of social contacts was extended and shops opened; from Monday 18 May, contact professions were restarted and there was a partial opening of schools and sports clubs; from Friday 5 June, primary schools and restaurants were reopened.

In the Netherlands, starting 9 March with instructions for hygiene, the limitations on public activities increased gradually. On 15 March, cafes, restaurants and fitness clubs were closed; on 16 March, schools and day care centres were closed. On 23 March, a so‐called “intelligent lockdown” – meaning that a high appeal is made on citizens' own responsibility to act in the collective interest – was declared. Groups of more than two persons were forbidden (except when people form a household) and people were asked to respect the norm of social distancing of 1.5 m. Enforcement of these rules was relatively lenient: citizens were primarily reminded to take their own responsibilities. People were asked to work at home as much as possible and work that involved interpersonal contact was forbidden – with the exception of health care work. In May, the lockdown was gradually relieved, starting with (o.a.) hairdressers and visits in nursing homes on 11 May, resulting in the opening of (o.a.) cafes, restaurant, primary schools and day care centres on 1 June. The social distancing norm of 1.5 m was however maintained. Only in public transport, a facemask was compulsory.

Germany entered the lockdown in a very coordinated fashion. On 22 March, the German Chancellor Angela Merkel and the heads of government of the 16 German states agreed to comprehensive lockdown measures. All hospitality venues and shops, with the exception of grocery stores, were closed and recreational, sporting and cultural activities cancelled. Schools and universities were also closed. Companies could continue to operate, but were obliged to adhere to strict hygiene regulations and social distancing. Many companies asked their employees to work from home. Already about 1 month later, however, Germany started the first measures to ease the lockdown, as individual shops and also libraries at universities could once again reopen in most Länder, provided they meet conditions for hygiene.

The literature studying the economic impact of the COVID‐19 crisis on employment, earnings and consumption (e.g., Adams‐Prassl, Boneva, Golin, & Rauh, 2020; Bachas et al., 2020; Chetty et al., 2020) documents substantial differences depending on socio‐economic status. The hospitality and arts sectors as well as manufacturing were most severely affected by the lockdown. In these sectors, relatively more people are at work with a vulnerable socio‐economic profile: students, young people, low‐skilled, single persons (with and without children), tenants, workers with flexible labour contracts and the self‐employed. In these hard‐hit sectors, wages are lower and workers typically live in families with a lower income, a higher poverty and less financial reserves to cover periods of declining income (Decoster, Van Lancker, Vanderkelen, & Van Heukelom, 2020; Horemans, Kuypers, Marchal, & Marx, 2020).

These are also sectors, where people often have very little financial buffer. In Belgium, approximately 60% of the typical affected workers do not have enough liquid assets to bridge 1 month wage loss without ending up in poverty. In a sector such as the hotel and catering industry, some 40% of employees have virtually no net assets (taking into account both liquid assets and other equity components). In addition, a lot of people, in, for example, the heavily affected cultural sector, are not adequately covered by the existing and newly introduced protection systems. Flexible and atypical work statutes are common here, they often do not benefit from compensatory measures (Horemans et al., 2020).

The coronavirus has also seriously challenged the work–life balance. During the lockdown, caring for the children and for the older people became more family based. This put a great deal of pressure on family life and on opportunities to remain economically active. Systems of parental leave were therefore extended in many countries.

POLICY RESPONSES

Continental welfare states responded to the consequences of the COVID crisis by strengthening existing or developing new schemes of income support to sick workers and their families, income support to quarantined workers who cannot work from home, helping dealing with unforeseen care needs, income support to persons losing their jobs or self‐employment income, helping firms to adjust working time and preserve jobs, financial support to firms affected by a drop in demand and helping economically insecure workers stay in their homes (EUROFOUND, 2020; OECD, 2020). In doing so, they repaired to a certain extent the restrictions that had been imposed on unemployment insurance in the recent past and provided additional protection mechanisms for non‐typical workers who had often fallen by the wayside of social protection before the crisis. But they did so in different ways.

In the Netherlands

Starting 12 March, the government took steps to ease the socio‐economic consequences of the corona crisis: employers and companies could get an extension of payment of taxes. On 17 March, the (first) Temporary Emergency Measures for the preservation of employment (NOW) and a Temporary Support Measure for Self‐employed (TOZO) were introduced. Both measures were taken for 3 months (march–may) and both have been extended for 3 months (July–September).

As in a number of – mostly but not exclusively English‐speaking – countries the Netherlands introduced a wage subsidy system. The NOW gives companies a compensation of 90% of wage costs, if it can be expected that they will at least experience a reduction of 20% of their business. Companies that applied for NOW were not allowed to fire workers. The wage costs include the costs of permanent and fixed contracts as well as temporary workers. For the protection of workers (and the companies where they work), the instruments of social insurances were not used, whereby workers in case of unemployment fall back on social benefits financed in principle by social contributions and which cannot be cumulated with earned incomes. In June, 284,000 applications by companies have been approved, on average each company received 70,000 euro and in total 19.8 billion euro has been paid out.

Contrary to Belgium and Germany, in the Netherlands NOW is a new system outside social security. The NOW wage subsidy is not contribution based and can be used by firms not only for hours not worked (like short‐term unemployment schemes) but also for hours worked (like standard wage subsidies). The reasons why the Netherlands opted for a new system and did not revert to the temporary unemployment scheme that was used as a response to the 2008 crisis are twofold. First, policymakers might have felt that due to the flexibilization of the labour market in recent years, the number of workers with flexible contracts who would not have been covered by insurance‐based short‐term employment had become too high. In the sectors most affected, such as hospitality, events, culture and tourism, employment relationships/contracts are indeed often of a flexible nature.

Wage subsidies are, secondly, a more flexible form of support for firms that can manage their hours freely without any reporting requirements. They also provide stronger incentives for firms to keep hours worked up and to increase them quickly when conditions improve. The system of wage subsidies has been used in many Anglo‐Saxon countries such as Canada, New Zealand and Ireland.

The TOZO, also a new system outside social security, gives social assistance to the self‐employed at a level comparable to the social assistance act. Approximately 300,000 self‐employed have received support. The TOZO differs from regular social assistance, in that it does not take property and capital into account (it does however take other income in the household into account).

Before the COVID‐19 crisis, in the Netherlands, a temporary unemployment scheme combined with a scheme for reduction of working hours was in place. With the introduction of the NOW and TOZO on 17 March 2020, these schemes were abolished. They were seen as too complex, the administration as demanding and therefore slow and the coverage as too limited because it only covered the risk of employees for whom the employer was obliged to continue paying wages in case of sickness. The risk of most of the workers with flexible labour relations or self‐employed (ZZP) was not covered by these schemes while they were also the categories most affected by the COVID‐19 measures. These developments coincided with an already ongoing discussion about dualization on the labour market, risk‐selection by employers and the limited social security for non‐standard workers and the self‐employed. In 2019, an advisory commission installed by the Minister of Social Affairs as well as the Scientific Council for Government Policy advised the Minister to take action and to reform social security and labour market policy in order to counter processes of dualization and to increase the social security of workers. This had already cleared the path for the abolishment of the existing schemes and the introduction of schemes with a broader coverage. What probably also facilitated these policy reforms was the fact that government and social partners were already intensively debating the implementation of policy reforms in the pension system.

In Germany

Germany has responded to the COVID‐19 crisis within the social policy domain in ways that were very similar to the responses during the financial crisis in 2009, but the response also included some additional measures, such as the extension of unemployment insurance payments for up to 3 months, the suspension of the wealth test for the unemployment assistance programme (Hartz IV) and the immediate support for the self‐employed and small companies. These social policy measures can be understood as part of a larger Keynesian response (largely due to the various fiscal measures and support payments for companies, for example, the huge package for Lufthansa). It was argued that Germany could pursue this policy, due to the very prudent fiscal policies in the previous years.

The negative impact of the restrictions on workers and their incomes were minimized by the existing short‐time work scheme (Kurzarbeit), access to which was made easier. Germany has a long tradition in utilizing the short‐time work scheme to mitigate economic shocks – it proved to be a very effective instrument also during the economic crisis in 2008. Proposals to instrumentalize the short‐time work scheme in order to mitigate the negative labour market effects of the pandemic were discussed already in early March. The scheme has the support of the social partners, but also of all major political parties. Having a Grand Coalition government certainly helped in arriving at the proposal to expand the programme; however, it needs to be emphasized that the Social Democrats took the lead in proposing the various expansions and liberalizations (including the introduction of some redistributive elements) of the scheme in Spring and Summer. In May 2020, more than 7 million workers or approximately 19% of the labour force were in receipt of short‐time work benefit. In core industry sectors, the public short‐time work benefit is complemented through additional benefits, agreed by the social partners through collective bargaining (Schulten & Müller, 2020).

The measures for easier access to unemployment assistance (Hartz IV) by the self‐employed became effective retroactively as of 1 March. As the scheme is not contribution based, solo self‐employed workers could always access the scheme. The most important new and time‐limited element was the suspension of the wealth test. Hence, the scheme became more middle‐class oriented as it was intended to support those who lost their income without having to access their savings or investments. According to figures from the Federal Employment Agency, approximately 81,000 unemployed self‐employed received unemployment assistance payments for at least part of the time in the months from April to September (“Corona Rettung kostet weitere 22 Milliarden,” in: FAZ, November 16, 2020, p. 15).

Working parents, whose children were unable to attend childcare or school due to closures, were entitled to receive an earnings‐related parental leave benefit with a replacement rate of 67% (capped at 2016 Euros) for the duration of 10 weeks for each parent (or up to 20 weeks for single parents). Families on very low incomes could apply for “additional emergency child benefits” of up to EUR 185 per month. In addition, all families were entitled to an additional EUR 300 for every child. The top 20% of income earners, that is, those with an annual income of more than 67,816 Euros, only benefitted partially or not at all, due to the relevant offsetting with child tax allowances.

Unemployment benefits for unemployed workers whose regular earnings‐related benefits would have ended between May and December 2020 were extended by 3 months. Benefits for the long‐term unemployed or those without entitlement to earnings‐related benefits, such as self‐employed or certain atypical workers, were not increased, although procedures to apply for the basic unemployment benefit were streamlined and the wealth test was suspended for applications until 30 September 2020.

In Belgium

The negative impact of the corona crisis on family incomes was largely cushioned by the operation and the reinforcement of existing social insurance schemes, most notably by the short‐term unemployment benefits for employees, the replacement income for the self‐employed who were forced to stop their activity and parental leave.

The existing temporary unemployment scheme for reasons of “force majeure” and for “economic reasons” provided Belgium with a strong instrument to protect workers while sustaining links between employers and employees. Temporary unemployment proved to be a very effective instrument during the economic crisis in 2008. Therefore, there was a large consensus to use the system, alternatives were never on the table at this stage (later on the Dutch wage subsidy system was put forward by employer organizations as an interesting alternative, especially with a view of activating the unemployed). The social partners played a crucial role in implementing very quickly the measures needed to adapt the existing systems to the unprecedented conditions. On 18 March 2020, the National Labour Council (NAR) concluded a collective agreement to extend the system of unemployment for economic reasons. A few days later, the government and social partners decided to scale up the also existing but little used system of temporary unemployment due to force majeure (so far limited to, e.g., fire or terrorist attacks). Subsequently, several agreements were concluded to extend the protection and to fill the remaining gaps. With more than 29% of the workforce, Belgium had the highest proportion of workers who have benefited from temporary unemployment benefits. The corresponding shares in Germany (Kurzarbeit) and the Netherlands (NOW) were respectively 16.8 and 22.3%.

Social insurance has also been used to support the self‐employed. The pre‐existing bridging right for self‐employed was extended to support self‐employed people who had to cease their activities because of corona. The bridging right is part of the social security system for the self‐employed. In the event of bankruptcy, it ensures the retention of the rights to child benefit and medical care for a maximum of four quarters and a monthly flat rate payment, equal to the minimum pension for self‐employed persons, for a maximum of 12 months. The system was extended to persons who are self‐employed as their main activity and forced to stop their activity, as soon as this interruption lasts longer than 1 week. More than 400,000 self‐employed persons received a bridging right. In addition, for the pre‐payment calculation of taxes, downward revisions were possible for self‐employed whose income is projected to be lower than the one currently used as reference. Self‐employed also had the possibility to postpone social security contributions.

The Belgian mandatory unemployment insurance is quite unique in Europe, in that benefits last in principle for the full duration of unemployment; yet the allowances decrease with the duration of unemployment, the so‐called “degressivity” of unemployment insurance benefits. As a response to the lockdown related labour market crisis, the social partners reached an agreement on the temporary freeze of the degressivity of regular unemployment benefits. As a result, the implementation was halted of the controversial measure dating from 2018, which combined higher benefits in the first period of unemployment and a stronger reduction of the benefits for long‐term unemployed with the aim to encourage return to work. Although further punctual measures were taken to support low‐income families, the relatively much more generous support to those who lost work as a result of COVID‐19 opened up new inequalities between those out of work because of the pandemic and pre‐pandemic social security claimants.

A corona parental leave was introduced by the federal government, the aim of which is to enable workers with young children to provide care of their children not able to attend school. It is not charged on the maximum duration of parental leave. The benefits are paid by the federal social insurance programme through a flat rate benefit. Workers who have to stay at home because a household member is infected can also get temporary unemployment benefits.

DISCUSSION

This article looked at policy responses to the social consequences of the first confinement in the COVID‐19 crisis in three continental European welfare states. Belgium, the Netherlands and Germany are all three robust and relatively well performing welfare states in which strong social insurance systems play a dominant role. Social protection in these countries share Bismarckian roots. Over time, these welfare states were however in constant flux and incorporated to a greater or lesser extend elements of both the Anglo‐Saxon and Nordic models. While the Netherlands has from the beginning deviated from the Bismarckian model, in recent years this welfare state has undergone important reforms that have made it increasingly evolve into a “Bismarck cum Beveridge” model. The labour market has become significantly more flexible while the share of social assistance and private social spending has increased. Germany and Belgium, albeit later and less pronounced, also witnessed a dual transformation, with retrenched earnings‐related benefits for long‐term unemployed and an increasing number of atypically employed people on the one hand and expanded social security to the so‐called “new social risks” on the other. Arguably, this evolution driving continental welfare states towards “Bismarck cum Beveridge” was greater in Germany than in Belgium.

It is against this changing institutional background that we can understand the similarities and differences in the extent to which these three continental welfare states used traditional social insurance systems to buffer the social and economic consequences of confinement.

First, all three countries strengthened to varying degrees of social protection systems for the active age population. This should not only be seen as an economic imperative for major supportive intervention but also as a response to the weakening of the protective capacity of social security systems in recent decades. In the three countries, COVID‐19 related risks were better protected than regular unemployment, although in Belgium the planned gradual reduction of the benefits for the long‐term unemployed was (temporarily) suspended while in Germany unemployment benefits were provisionally extended. As a consequence, the policies that increased the generosity and the accessibility of COVID‐related social risks were a response to the dual transformation of social protection that took place in recent decades without, however, changing its course.

Second, the extent to which continental welfare states made use of existing social insurance schemes in order to provide protection of individuals and families whose livelihoods have been threatened by the confinement seems to be related to the extent to which these welfare states have moved in the Anglo‐Saxon direction. Belgium, arguably the most Bismarckian welfare state, has relied entirely on existing systems of short‐term unemployment insurance for employees and the self‐employed. Germany also made use of existing systems, but self‐employed persons were covered by unemployment assistance and not, as was the case in Belgium, by social insurance. In the Netherlands, two new systems were introduced outside regular social security.

Third, a similar pattern is evident with regard to policy responses to the impact of confinement on the work–life balance. In Belgium and Germany, the existing insurance‐based parental leave systems were extended. In contrast, the Netherlands, where part‐time work is very widespread, did not take additional measures in this respect.

The various policy responses in continental European welfare states pointed to the importance of path dependency and existing social fabrics in paving the way for solutions to the unprecedented social and economic impact of the pandemic. They, however, also revealed a remarkable flexibility of welfare states to adapt to new, unprecedented circumstances. In all three cases, various social policy measures contributed to the mitigation of the negative labour effects and the stabilization of the economies. Future research will show which responses were most effective.