Path dependency and partisan interests: explaining COVID-19 social support programmes in East-Central Europe

ABSTRACT What factors influence governments' social policy responses to the COVID-19 crisis in East-Central Europe? We attempt to answer this question by analysing the social policy responses to the pandemic across three distinct institutional varieties and welfare states: Estonia, Slovakia, and Slovenia. Drawing on extensive analysis of qualitative and quantitative data, we argue that the constraints on government agency posed by previous, posttransition patterns of social policymaking and their underlying core institutional legacies have a distinct influence on governments' distributive choices. Governments' partisan interests are reflected in some of the enacted measures, albeit in less consolidated parts of welfare state structures.


Introduction
The economic crisis caused by the COVID-19 pandemic constitutes one of the most profound challenges for policymakers in recent memory. Most governments in Europe reacted to the crisis by providing considerable financial relief to workers and businesses to mitigate the social and economic effects of the containment measures. This is no less true in East-Central Europe (ECE), where governments drew upon existing social protection institutions, but also introduced a number of new policy measures. At the same time, the COVID-19 pandemic exposed vulnerabilities in existing welfare state infrastructures and created an opening for policy innovation. As Gourevitch (1986) noted, crises rattle systems of relationships and make politics and policy more fluid. They also reshape political coalitions and realign patterns of cooperation and conflict among key political and social actors. This leaves us with the question as to what factors influenced governments' social policy responses to the COVID-19 crisis in ECE, and whether there has been divergence from or convergence on existing institutional legacies.
In this paper, we argue that the COVID-19 crisis largely confirms and reinforces preexisting patterns of social policymaking and redistribution. Specifically, existing regimes of social protection institutions and the structural power of social partners, i.e. dominant business groups and trade unions, influence governments' distributive choices in social policymaking, resulting in considerable variation between countries. Nevertheless, governments use the remaining policy space to the extent that partisan ideology is reflected in some of the enacted measures. While the scale and budgetary impact of these measures is minor, their adoption illustrates the importance of looking beyond institutional legacies to analyse and explain crisis responses.
We probe this argument by analysing the social policy responses in three countries, representing three distinct institutional varieties: Estonia, Slovakia, and Slovenia. We show that, in line with their welfare state and social policymaking regimes, Estonia provided limited support in both welfare benefits and labour market measures, Slovakia focussed heavily on job retention, and Slovenia, in addition, made use of and extended pre-existing, generous social benefits (for policy details, see Tables A1-A3 in supplemental data). While the details of the variation between the three cases will be outlined in the individual case studies, the general trend, in particular the striking generosity of the Slovenian crisis response is illustrated by the aggregate change in social protection spending between 2019 and 2020 in Tables A4-A6. We thereby show that, even in such unprecedented circumstances as constituted by the pandemic, previously identified path dependencies in terms of past policy choices and social dialogue structures hold considerable explanatory leverage. In addition to these path dependencies, however, governments cater to electorally relevant groups by designing corresponding support programmes within the aforementioned constraints. We provide evidence for this argument on the basis of detailed accounts of the respective policy measures on the one hand and of publicly available documents and news reports on the other hand.
Below, we outline a theoretical framework that combines institutional legacies and political agency, and discuss our case selection, research design and data. We then proceed with our empirical analysis of the social policy responses to the COVID-19 crisis in our selected case studies. We conclude by discussing how the findings compare across the selected cases and the broader ECE region.

Explaining continuity and change in social policy
Path dependent legacies of institutions and past policy choices are one of the most powerful explanations in comparative political economy to make sense of national differences in social policy, even in times of crisis (Blum and Kuhlmann 2016;Hall and Thelen 2009). For the purposes of this paper, we are interested in traditional social protection policies, such as welfare benefits, old age pensions and labour market measures, which became extraordinarily significant after the economic contraction resulting from COVID-19 restrictions. Such social protection policies are institutionalised over longer periods of time as semi-permanent social compromises, in the form of rules, regulations, laws and organisational practices that answer to the demands of the dominant social groups and in turn constrain the behaviour of political actors (Amable et al. 2019). Due to the capacity of bureaucratic institutions, solidified legal codes and political coalitions, past policy choices tend to "lock-in" and influence contemporary preference distributions. When social policy reforms are proposed, they are not formulated and interpreted "as if the new institutions will operate on a blank slate", but in the context of existing institutions and past policy choices (Hall and Thelen 2009, 21;Inglot 2016). Hence, due to such constraints of past policies, new policies are most often only incrementally layered onto the core welfare state institutions. In a nutshell, previous policy choices restrict what types of policy outcomes are available within the space of theoretical alternatives (Beramendi et al. 2015, 12-13).
In Western European welfare states, different legacies of social policy can be said to be embedded in and constitute a key element of distinct national varieties of capitalism (Pierson 2001, 5;Amable 2003). In ECE, post-WWII social policy legacies have gone through major political and economic upheavals, first under communist rule and then during the post-communist transition from 1989 onwards. Despite these momentous changes, scholars have underlined the path-dependent nature of Bismarckian workrelated social insurance programmes lying at the core of most post-communist welfare states in ECE, with most variation in the residual, less consolidated (post-)socialist programmes (national healthcare, housing, family policy) and means-tested benefit schemes (Inglot 2008;Saxonberg and Szelewa 2007;Aidukaite 2009;Adascalitei 2012;Kuitto 2016). Furthermore, unlike in Western Europe, most ECE post-communist economies have built their comparative advantage on the basis of keeping labour cheap and docile, while becoming increasingly dependent on the decisions of foreign-owned TNCs (Nölke and Vliegenthart 2009). Different policy choices and patterns of integration into global supply chains have thus produced three different socio-economic models in ECE: (1) the neoliberal financialised model in the Baltic states with a weak welfare state, (2) the embedded neoliberal model in the Visegrád states that relies on exports of manufactured goods with a moderate level of social protection, and (3) the export-oriented neo-corporatist model in Slovenia with relatively more generous social protection (Bohle and Greskovits 2012).
These different institutional path-dependencies and past policy choices are supported by varied political coalitions in our three sub-regional groups. Although ECE has been characterised by falls in unionisation levels in the last three decades, and thus also labour's weaker associational power (Schmalz, Ludwig, and Webster 2018), and the concomitant increase in structural and instrumental power of business and managerial interests (Culpepper 2015), economic crises nonetheless offer opportunities for readjustments to existing social bargains (Adascalitei 2012). Such readjustments normally take place through social partnership institutions, but these vary in terms of their importance across the different socio-economic models found in ECEaffecting unions' institutional power. The Baltic countries have weak corporatist systems, which have only further declined after the GFC (Kalanta 2020). Although some instances of revitalisation of social partnership recently occurred, such as Estonia's tripartite negotiations since 2018 or Lithuania's public sector agreements of 2017-2018, the overall picture suggests stagnation of national level dialogue and the ongoing prominence of company-level bargaining. In Czechia and Slovakia, the success of national-level collective bargaining depends on the political support of governments, while sectoral and company-level bargaining are the preferred venue for resolving issues related to minimum wages, employment security and working conditions (Kahancová 2013). In Slovenia, despite the fact that both centreleft and centre-right governments bypassed the formal tripartite structures on numerous occasions after the GFC, the neo-corporatist institutions still remain a key feature of the Slovenian regime (Stanojević 2018).
However, even where social partnership institutions are weak, we still have to consider the structural positions of labour and business (firms or industry-wise sectors), which due to their power resources are important to the macroeconomy (Culpepper 2015;Schmalz, Ludwig, and Webster 2018). Structurally powerful actors that are key to sustaining economic growth are also likely to receive direct state support. Furthermore, we assume that welfare policy is not a key business interest to the extent that employers do not have to shoulder any additional labour costs. Likewise, the structural power of labour is most felt where workers are strategically located in key economic sectors as, for example, strikes in such sectors are particularly disruptive (Adascalitei 2012). Broadly in line with the logic of quiet and noisy politics (Culpepper 2010), we expect COVID-19 related social policy to be most salient to the electorate and unions, whereas measures such as tax deferrals and loan guarantees are mostly of interest to employers.
In addition to social partnership institutions and past policy choices, we introduce partisan interests to explain the overall shape of social policy responses. Governing parties are expected to develop policy in reference to the societal groups supporting them so as to retain, and potentially expand, their basis of support. Policy trajectories are thus also subject to strategic political calculations. However, we only expect policy shifts to take place in less consolidated welfare programmes and means-tested benefit schemes targeted at particular vulnerable groups. We therefore add partisan politics as an auxiliary, agency-based explanation to our structuralist institutional frameworkroughly in line with "electoral turn" perspectives such as the constrained partisanship model of Beramendi et al. (2015). Governments are assumed to attempt to maintain support both by directly serving the material interests of core constituencies and by preserving their partisan or ideological profile as regards the type of policy approach they represent. Sometimes these two dimensions conflict, for example, in cases where an otherwise neoliberal governing party has pensioners as a key electoral group, or a centre-left party restricts eligibility criteria for claiming unemployment benefits. To discriminate between the explanations of institutional path dependence and partisan choices, we investigate whether the social policy crisis measures align withor instead, deviate fromthe broader welfare state legacy within a country.
Furthermore, partisan interests have been especially influential during the first wave of the pandemic for two reasons. First, governments were seemingly less constrained by institutions of social partnership during the first turmoil as initial responses in Europe were generally less based on institutionalised dialogue and bargaining (Eurofound 2020a(Eurofound , 13-14, 2021. This opens space for partisan considerations in policymaking. Second, as the pandemic progressed and the economic crisis became deeper, social policy issues became highly salient, with voters having diverse policy stakes, largely depending on their employment status. While securing employment is important to all governments, we expect parties to provide more generous targeted support to groups central to their partisan strategies.

Case selection and research design
We show below how past policy choices and their institutional legacies affect government agency and thereby have an observable bearing on social policy responses during the COVID-19 crisis. We systematically investigate social policy responses in three cases with diverse institutional legacies: Estonia, Slovakia, and Slovenia. We choose these cases because they each represent one institutional variant of semi-peripheral capitalism in ECE, which we expect to have a distinct effect on social policy responses. This variation serves the strength of our argument to the extent that we show that institutional legacies affect social policy responses to the crisis regardless of the precise nature of said institutional legacies. In other words, our choice of cases increases our confidence that the link between institutional legacies and social policy responses to the COVID-19 crisis has some analytical value beyond the precise cases we study.
While Slovenia represents its own institutional variety, the choice for Estonia and Slovakia over the other Baltic and Visegrád states merits further clarification. First, both cases exemplify their respective institutional variety in a relatively coherent manner: Estonia has been categorised as emblematic for the Baltic mode of transition and patterns of policymaking (Bohle and Greskovits 2012). Slovakia, together with the Czech Republic, is a case that has not experienced a pronounced departure from the status quo in economic policymaking after the financial crisis. Both features allow us to analyse and single out the effect of the institutional legacies outlined above. Second, similar to Slovenia, both countries have experienced a change of government directly prior to or during the COVID-19 pandemic. This renders our cases least likely: policy continuity despite a change of government illustrates the strength and durability of the institutions we advance as our main independent variable.
To illustrate the observed variation in the dependent variable, i.e. the social policy response to then pandemic, Table 1 depicts the two described categories of spending in the three selected cases. While the following case studies will further detail the precise nature of each response, the present Table already reveals considerable variation as the response of Slovakia is considerably more generous than that of Estonia on all counts, with Slovenia spending most on labour market other than Kurzarbeit measures and being less generous in terms of welfare benefits this lack of generosity is, however, offset by the expansion of automatic stabilisers, as discussed in section 6.1.
To further delineate these responses and, crucially, explain the observed variation, the analysis proceeds in three steps: first, we show the social policy response to the pandemic and its fallout with specific regard to (a) measures targeting the labour market and In the case of Estonia, numbers do not include the waving of 2nd pillar pension contributions, the one-off increase of pensions in 2022 or the reduction in excise duties. Whereas for the former two there is no precise data available (note too that the second is rather small), the latter is not considered welfare even though it supports incomes. Source: see Tables A1-A3. employment, and (b) welfare benefits in more detail. Second, we analyse the extent to which the relative emphases of (a) and (b) go together with the institutional legacies of organised interests and thereby demonstrate the continuity in social policymaking throughout the crisis. Third, we demonstrate whether and to what extent party politics and government preferences are manifested in parts of the social policy response. Our aim is to comprehend which strata of society are served by these measures, why such groups have emerged as a relevant target of government intervention, and the general context of distributive decision-making in times of crisis.
The evidence we invoke chiefly comprises the details of the enacted policy measures (Tables A1-A3), corresponding news reports, publicly available documents and statements by government officials and social partners, as well as existing analyses of the respective welfare states. We do not provide authoritative within-case evidence linking, for example, a detailed assessment of a government's electorate with the distributive consequences of a given policy. Instead, we show the validity of our argument by comparatively showing the covariation of social policies in response to the crisis, institutional legacies, and partisan interests. The effect of the institutional legacies and partisan politics on the social policy response to the crisis is shown by revealing how the chosen cases differ on all three dimensions. Evidence demonstrating the link between these dimensions serves as a within-case addition to the cross-case nature of our study. We thus do not formulate prior expectations of voter preferences along various occupational classes, nor do we disaggregate producer groups and their interests based on industrial profile (Beramendi et al. 2015;Gourevitch 1986;Simons 2021). Rather, we focus on which interest governing parties serve, both regarding pivotal voter groups and those capital actors structurally important to the economy.

Estonia
Estonia's social policy response to the COVID-19 pandemic is weak on both the labour market and welfare state dimensions (Table 1). Although the former is more extensive, particularly the Kurzarbeit scheme to compensate worker wages, Estonia's labour market dimension is minor in cross-national terms. Moreover, social policy is also considerably smaller than Estonia's business support policies (Table A1).
Our primary institutionalist perspective of partnership institutions and policy legacies largely explains this response. Existing social arrangements eased the crisis' socio-economic repercussions, but did not provide an institutional precedent for additional generous welfare policies. Although social partners were partially involved in social policymaking, their general organisational weakness translated into a frail social policy agenda overall. Past legacies affected policy expectations, making small social gains seem like victories. The early 2021 change in government barely affected policy continuation, as the crisis seems to have been only partly used by governing parties to serve their partisan interests with additional policy.

A limited social policy package
In early 2020, Estonia responded quickly with health restrictions and measures to maintain employment and income levels as well as economic activity. First, regarding the labour market, the Kurzarbeit scheme, which supports 70 per cent of employee wages (60 per cent in 2021) for two months for those employed in severely distressed companies, is the most extensive measure (Eurofound 2020b). Further, the government decided to compensate the first three days of sick leave, which employees previously covered themselves (Eurofound 2020c). Other smaller wage support measures were implemented too, such as compensation for those active in creative industries and agriculture, and a two per cent wage increase for public sector employees in 2022 (Eurofound 2020e; Ministry of Finance of Estonia 2021, 23). Moreover, Estonia allowed participants in the second pension pillar to withdraw savings and temporarily suspend their contributionsin addition to making participation voluntary as well as temporarily halting government contributions (Piiritis and Laurimäe 2020). For subsequent years the government opted for active labour market policies with a new Employment Programme 2021-2023 for circa 20,000 unemployed (Eurofound 2020f).
Second, regarding welfare, the Estonian government adjusted the coverage level of unemployment benefitsthe insurance benefits increased from 50 to 60 per cent of the previous wage and the unemployment allowance, drawn upon after the former expires, from EUR 189 to EUR 292 in 2021 (Laurimäe 2020)and implemented a minor allowance for parents caring for children with special needs (Eurofound 2020d). Pensioners received a minor extraordinary pension increase on top of indexation (Ministry of Finance of Estonia 2020, 37).
In sum, the labour market side is dominant in Estonia's generally weak social policy approach. In fact, Estonia's social measures are small on their own terms but also compared to business support. Following the IMF's calculations of the 2020 supplementary budget (IMF 2021, 54), which included the bulk of COVID-19 related measures and amounted to 8.6 per cent of GDP, 5.6 per cent went to business support.

Limited corporatism and welfare legacies
Estonia's crisis response was largely technocratically designed by a government facing few institutional constraints on policymaking. Weak tripartism (despite an upsurge in 2018; Lulle and Reire 2017) and the low organisational capacity and representativeness of unions and, to a lesser extent, of employers as well as the general tradition of centralised governance isolated from societal contestation (Kattel and Raudla 2013) provided this autonomy. While social partners were consulted on several social policies, and unions (rather than employers) indicated their satisfaction with both the process and results (Eurofound 2021), it was the government that set the overall social agenda of limited welfare, and several of the discussed policies are the result of government agency rather than reflections of social partner preferences (Eurofound 2021).
On the one hand, social partners were consulted on the Kurzarbeit, sickness benefits, and unemployment benefit changes in context of the supervisory boards of the Unemployment Insurance Fund and the Health Insurance Fund. The sick leave adjustment is seen as a union victory since unions had been pressing for such changes in preceding years (Eurofound 2021;Kadarik, Laurimäe, and Masso 2021). On the other hand, most of the one-off income support measures are temporary and modest, thus the social partners' successes are limited. Policy legacy explains the perception of victory because Estonia's limited welfare state tradition lowers expectations on generosity. The marginal role in tripartite dialogue contributes to this too because limited involvement already is an improvement. Moreover, without an institutionalised permanent tripartite forum the social partners could not perform a significant agenda-setting role. The upward but minor adjustments to the unemployment insurance and allowance schemes, for example, were proposed by the government itself and the eventual policy changes were in accordance with its original proposal. Unions' and employers' explicit discontent with the government's untransparent policymaking and its late and inadequate response to the second wave of the pandemic (Kadarik, Laurimäe, and Masso 2021) reinforces the view of generally weak partnership and its corresponding policy effects.
In other words, such limited neo-corporatist constraints and low welfare expectations strengthen the autonomy of governments unprepared to stray from Estonia's weakly decommodifying path with welfare expansion and to rather opt for efficient and lowcost market-based measures. The Kurzarbeit scheme, as the single-largest social policy, supports this reading. Kurzarbeit is an effective means to retain employment and resume economic activity once restrictions are lifted. While serving labour interests and maintaining overall consumption, Kurzarbeit also eases employer concerns about employment. Estonia's neoliberal policy consensus legacy (Thorhallsson and Kattel 2013) structurally embedded preferences for such an approach (see below). Moreover, as Estonia initially was modestly affected by the pandemic and other countries successfully relied on pre-existing or improvised Kurzarbeit, the government grew aware of its efficiency. The comparatively small size of Estonia's Kurzarbeit itself can be explained from efficiency preferences too because with complex manufacturing less prominent than in Slovakia particularly, Estonia needs less to accommodate enterprises requiring scarce skilled labour.

Mixed partisan interests
What role do partisan interests play? To assess this, the analysis ought to consider Estonia's economic response in broad terms. Since Estonia's austerity response to the GFC has been partly explained by the neoliberal ideological leaning of the Ansip governments' leading member, the Reform Party (Kattel and Raudla 2013;Raudla and Kattel 2011), partisan profile seems an appealing explanation for the crisis response at hand.
What exactly are these profiles? The Ratas II government (April 2019-January 2021) was composed of a mixed coalition of the Centre Party, Conservative People's Party (EKRE) and Isamaa. The leading Centre Party has a centre-left, more redistribution and interventionist agenda that is supported by the lower incomes (and Estonia's large ethnic Russian minority; Saarts 2015). Previously, the Centre Party had, together with the social-democratic SDE, exited the Rõivas II government (2015)(2016) with an agreement to end neoliberalism and expand the welfare statethough this barely materialised. The increasingly popular EKRE is a nationalist-conservative party which is not redistributionist but does advocate for the protection of Estonian enterprise and disadvantaged rural areas (Petsinis 2019). Isamaa is nationalist-conservative too and besides liberal economic views it propagates population expansion and state survival (Parliament of Estonia 2021). In the subsequent Kallas government (January 2021-present), the Centre Party is the smaller partner in a grand coalition led by the Reform Party, which draws its support largely from urban and well-educated (upper) middle-class voters.
The Ratas II government's marginal social policy is somewhat inconsistent with the Centre Party's centre-left profile, thus supporting the institutionalist explanation of weak neo-corporatist constraints and strong neoliberal policy legacies. On the other hand, the Centre Party's economic policy profile does suggest a stronger inclination to follow a debt-based fiscal approach, as Ratas II initiated (see also Raudla 2021a). However, the Kallas cabinet has not substantially departed from this course, as borrowing continues in a similar fashion in the medium term (IMF 2021, 39; Ministry of Finance of Estonia 2021, 35). Although the Reform Party may have compromised with the Centre Party, more likely is that it realised that it could not suddenly de-couple from the past policy choice. Low debt levels and favourable borrowing terms plausibly eased hesitation to continue assuming debt.
Yet, some elements do suggest the relevance of partisan interests (see also Raudla 2021b). In early 2020, Ratas II was criticised for pushing through its preferences with the reduction in excise duties and the 2nd pillar pension reform (ERR 2020a). Income support through lower indirect taxation on energy is electorally attractive to the Centre Party (and presumably Isamaa) drawing on lower incomes and pensioners support. Excise duties fuelled discord in the Kallas coalition, with the Reform Party blocking further reductionsfor fiscal reasons, yet its higher income constituency also gains relatively less from such reforms than lower incomes (ERR 2022). Second, the pension reformparticularly the suspension of contributions and the possibility for account holders to access their savingswas conducted under the pretext of a crisis measure yet it has been long advocated for by Isamaa and the legislation was adopted prior to the pandemic outbreak (Velmet 2020). A constitutional conflict did not prevent Ratas II from proceeding with the reform (Piiritis and Laurimäe 2020).
Furthermore, the much larger business support seems partly related to the more interventionist policy paradigm of at least the two largest members of the Ratas II coalition. Among the more eye-catching subsidies, the government provided a EUR 125 million investment in an oil shale plant, which was partly criticised for being an indirect subsidy to the financially troubled state-owned Eesti Energia. Commentators suggested that the Centre Party pushed this investment project because many of its ethnically Russian voters work in the mining sector (Velmet 2020). One could argue that the Centre Party for similar reasons had an incentive to provide business support for the eastern county of Ida-Viru, although Ida-Viru together with Harja faced selective local restrictions to curb the pandemic (ERR 2020b).

Slovakia
The Slovakian social policy response to the COVID-19 crisis was heavily focussed on employment retention via its extensive Kurzarbeit scheme while the expansion of welfare benefits has been more meagre. The heavy tilt towards employment results from a model that, depending on different governments' policy objectives, gives little authoritative say to labour unions and persistently accommodates the needs of foreign (automotive) manufacturers. Welfare benefits diverging from the status quo reflect the current government's partisan profile.

Social policy geared towards employment retention
Immediate measures to combat the fallout of the pandemic in Slovakia were heavily focussed on the labour market with the aim to enable firms to retain their employees ( Table 1). The Slovakian government introduced a Kurzarbeit scheme on March 31 2020, covering up to 80 per cent of furloughed workers' wages and a flat rate contribution to the self-employed. The scheme was expanded in mid-April, successively amended, and finally made permanent from 2022 (The Slovak Spectator 2021; Simons 2022). More than EUR 700 million were allocated to the programme in 2020 and close to EUR 1.5 billion in 2021 (Table A2). Additional measures targeted family caretakers, quarantined or sick employees, and mortgage owners and renters, but these were more limited in scale compared to the Kurzarbeit scheme. These emergency measures compound the pronounced focus on the labour marketthe budgetary impact of labour market measures is four times bigger than that of other welfare benefits. To provide further relief, the government adopted a "social package" in July 2020 that included targeted benefits to pensioners and families worth EUR 500 million (ca. 0.5 per cent of GDP) in 2021. The package represents somewhat of a deviation from the previous all but exclusive prominence of employment in the government's policy response.

Export-dependence and weak social partnership
The emphasis of employment retention as opposed to the meagre one-off social benefits, and the distributive effects of the Kurzarbeit scheme indicate its embeddedness in the existing gearing of the social policy regime to the needs of the automotive industry as well as existing social partnership institutions.
Kurzarbeit schemes gained prominence due to the German response to the financial crisis and the measure is particularly suited as relief for exporting manufacturing companies that are dependent on the retention of a skilled workforce and, thus, "fits" the German export-led growth model (Brenke, Rinne, and Zimmermann 2013). A similar pattern holds true in Slovakia: while all Visegrád states converged on a model that is dependent on foreign-owned, exporting automotive firms, this dependence is most pronounced in Slovakia, where the state has played an essential role in accommodating such firms (Pavlínek 2016). This serves as the backdrop of its social policy regime (e.g. Drahokoupil 2007) as the pursuit of growth via foreign-owned car manufacturers relies heavily on the existence of a skilled and cheap workforce (Nölke and Vliegenthart 2009). Corresponding to this "match" between Kurzarbeit and Slovakia's growth model, an analysis of the Slovakian Kurzarbeit scheme by the European Investment Bank (Lalinsky and Pál 2021) reveals manufacturers of transportation equipment to be among the largest beneficiaries of the scheme. More than 50 per cent of firms in that category were supported (compared to 22.3 per cent across all sectors), and they received the biggest support relative to their size (two per cent of their total assets on average). Thus, while the Kurzarbeit scheme is well suited to support Slovakian automotive manufacturers by design, the actual uptake of the programme supports this intuition.
Moreover, the process of adoption of the scheme points to the general weakness of social dialogue: while social partners support the scheme, it does not appear to be a result of tripartite consultations. The manufacturing sector emerged as the main beneficiary of the government's measures not because of its influence through tripartite consultations but its ever-present relevance for economic growth, conditioned by past policy choices. This corresponds to developments prior to the crisis: as in other Visegrád states, tripartite structures in economic and social policymaking have been gradually weakened (Pula 2020); analogous to the overall dwindling organisational power of labour unions (see Tables A8 and A9). The influence of unions specifically hinges on partisan strategy: the biggest union association KOZ SR has close links with Smer-SD but this appears to be a function of coalition-building on the part of Smer-SD rather than institutionalised policy capacity (Simons 2021). Nevertheless, the relatively strong sector-and company-level bargaining structures in Slovakia (see Kahancová 2013) may have played its part in making employees' calls for the scheme heard.
During the crisis, the first calls for a Kurzarbeit scheme came from employers: the Association of Industrial Unions (APZD) advanced the necessity of state support for furloughed workers as early as March 16 2020 (APZD 2020). After the first package was introduced, all large employer associations took issue with the exclusion of larger companies in particular (Klub 500 2020); the Automotive Industry Association being an additional vocal player (ZAP SR 2020). The expansion to include large companies was adopted in mid-April (Simons 2022). KOZ SR also supported the idea of the scheme but, from the outset, complained about the lack of consultation with the country's tripartite body (Uhlerová 2020). As a similar disregard for consultations emerged in the efforts to make the scheme permanent in the summer of 2020, KOZ SR walked out of the tripartite body (Teraz 2020c;RTVS 2021). While it participated in the final consultations of the scheme, it concluded with various comments, e.g. the demand to include the self-employed, which, following news reports, was "promised" by the government (RTVS 2021). Therefore, while the scheme represents the preferences of all social partners, this congruence does not seem to be a function of the institutionalised influence of social partners on the government.

Political entrepreneurship and coalitional compromise
In contrast, the social package, announced and decided on in summer 2020, does not observably reinforce the Slovakian strive towards employment retention at all costs, but reflects partisan politics and the governing coalition's strategic concerns in multiple respects. Moreover, the lack of public statements by social partners on the social package indicates their little involvement in its adoption and design.
Slovakia experienced a change of government at the outset of the pandemic. The social-democratic Smer-SD, which has been characterised as the country's hegemonic party (e.g. Spáč 2014, 246), was replaced in March 2020 by a broad coalition consisting of the parties Ordinary People (OL'aNO -EPP), We Are Family (ID), Freedom and Solidarity (SaS -ECR) and For The People (EPP). Ideologically, the largest governing party OL'aNO has been deemed anti-elitist, moderately right-wing on the cultural dimension and decidedly centrist as regards economic policy (Havlík et al. 2020). Due to its heterogeneous electorate, the party attempted to attain voters from across the political spectrum (Rossi 2020). The most common feature of the typical OL'aNO voter appears to be their concern for corruption (Lysek, Zvada, and Škop 2020). In contrast, We Are Family has been characterised as more staunchly right-wing, with a conservative electoral platform (Marušiak 2017). The two smaller parties SaS and For The People display a centre-right, pro-business stance with libertarian traits (Marušiak 2017;Rossi 2020). Ideologically, this stands in contrast to previous Smer-SD governments which relied on an electorate characterised by left-wing views and old age (Simons 2021), corresponding to policy programmes marked by generous social policy, amongst others, benefiting pensioners (e.g. Rybář and Spáč 2017, 154).
The general direction of the measures included in the social package and the political rhetoric surrounding their adoption indicate the government's effort to cater to those constituencies previously served by Smer-SD while simultaneously distinguishing itself by emphasising the efficient and frugal allocation of social benefits. Policies include a 13th month pension, increased benefits for pregnant persons and students, and an increased tax bonus for children (while abolishing free school lunches), representing a rather conservative, family-focussed mode of social policy (The Slovak Spectator 2020). The package bears resemblance, not only in name, to previous packages adopted by Smer-SD, and prompted criticism from the Council for Budget Responsibility (Pravda 2020).
The government, specifically then Minister of Finance Eduard Heger (OL'aNO), proceeded to defend themselves by emphasising the targeted nature of enacted benefits and pointing to compensatory savings in administrative costs (Pravda 2020). This pattern indicates the effort to continuously emphasise efficiency and little state involvement, while, simultaneously, gaining a reputation as a generous benefactor in times of crisis. It also targets those strata of the electorate previously firmly in the hands of Smer-SD, partly on account of the more conservative parts of the governing coalition, For the People and We Are Family. This balancing act became apparent in the press conference to announce the package: while the speaker of parliament Boris Kollár (We Are Family) and Minister of Investment, Regional Development and Informatisation Veronika Remišová (For the People) stressed the package's effect on poverty reduction and the benefit for pensioners, Sulík and Heger both emphasised fiscal prudence and efficiency (Teraz 2020b). Similarly, maintaining the 13th month pension had been a political priority for Kollár specifically, who promised as much to the pensioners' union, but ultimately had to concede some adjustments, possibly due to pressure from the more austerity-prone parts of the government (Teraz 2020a;. This compromise echoes characterisations of the current government as "opportunistic", whereby "conservative elements [are adopted] into their programmes for electoral advantage instead of from actual conviction" (Rossi 2020, 235).

Slovenia
The Slovenian government enacted a well-rounded and generous social policy programme in response to the COVID-19 crisis, with the labour market-related crisis measures being more substantial (Table 1). With an eye on the upcoming parliamentary elections, the Janša government dispersed cash transfers to different social groups to broaden its appeal beyond its key electorate.
The path-dependency of a comparatively stronger welfare state tradition in Slovenia best explains the adopted social policy packages by an otherwise centre-right government. Even though the level of COVID-19 related welfare benefits as a percentage of GDP was similar to that in Estonia and Slovakia, the social policy crisis response came on top of increases in automatic stabilisers by 1.4 per cent in 2020 and 1.6 per cent in 2021, which include transfers for the unemployed, family and parental benefits, transfers for providing social security and rent subsidisation (see Government of the Republic of Slovenia 2021, 17). Nonetheless, to fully understand the shape that the social policy crisis packages took and the changes in the Kurzarbeit scheme, we also need to take into account the partisan considerations of the Janša government, which came into office in mid-March 2020, and the role of social partners.

A comprehensive social policy response
The Slovenian government responded quickly to the COVID-19 crisis, with labour market measures first in early April 2020, followed by additional welfare focused measures at the end of April 2020. On the labour market policy side, the government's crisis response consisted mainly of two key measures: (1) a temporary layoff scheme and reimbursement of related wage compensation to employers (EUR 539 million spent until March 2021); and (2) a short-time work subsidies scheme (or Kurzarbeit), which built upon the short-time work scheme adopted in the aftermath of the 2008 financial crisis (EUR 55 million spent until March 2021) (Eurofound 2020k). These employment retention measures guaranteed workers 80 per cent compensation of their gross wages. The employers were clearly more in favour of using the temporary layoff scheme because it was less restrictive to use, and they received higher wage compensation from the state. Furthermore, the government also introduced a temporary monthly basic income for the self-employed, clergy and farmers who received EUR 700 per month for the duration of the pandemic (Eurofound 2020g). Additionally, the government increased sick pay from the statutory 80 per cent to 90 per cent for workers who needed to self-isolate and 100 per cent for workers who got infected at workplace (Eurofound 2020i) and adopted temporary unemployment benefits for workers who are on irregular contracts or lost their employment during lockdown (Eurofound 2020h).
In terms of welfare measures, the government disbursed one-time solidarity allowances for low-income pensioners and disabled workers who work part-time (up to EUR 300), low-income and large families (up to EUR 200), new-borns (EUR 500) and students (EUR 150), and increased childcare allowance (Eurofound 2020j). These crisis cash transfers, most of which were paid out three times during each wave of the pandemic, came on top of an already generous welfare system in Slovenia, which was upgraded before the pandemic by the previous two centre-left governments under the pressure of the Left party and trade unions (Toplišek 2020).
When looking only at the COVID-19 related crisis measures, the labour market side of the Slovenian government's social policy response is clearly more substantial in budget terms. However, if we also account for automatic stabilisers in our analysis, which are activated as part of permanent welfare state institutions, this puts welfare policy measures at a similar level as labour market related measures in terms of expenditure, 1.8 per cent of GDP in 2020 and 1.66 per cent in 2021 (see Table 1 for comparison). The relaxed fiscal space as a result of the suspension of the EU's Stability and Growth Pact allowed for the implied budget variations in revenue and expenditure in response to the business cycle changes.

Explanation
6.2.1. Strong welfare state tradition despite weaker neo-corporatism Slovenia's crisis response in social policy came on the back of generous welfare institutional arrangements that have been shaped through tripartite social dialogue institutions for most of the three decades of Slovenia's independence (Stanojević and Klarič 2013). The post-2008 period has seen a deterioration in the quality of Slovenia's industrial relations system and only in 2015 the social agreement between employers and trade unions was re-established (Stanojević, Kanjuo Mrčela, and Breznik 2016;Mrčela 2021). The partisan pressure from the left-wing Left Party and increasing demands from public sector trade unions from 2016 onwards have pushed the centre-left governments of Miro Cerar and Marjan Šarec to gradually remove post-GFC austerity measures and raise the level of welfare benefits in line with living costs (Toplišek 2020). Social dialogue was again weakened when the Janša government took office on 13 March 2020 as it took until 30 April for the new government to appoint its representatives to the main tripartite institution, the Economic and Social Council (Eurofound 2021).
The two main social partners in the Economic and Social Council are the trade union umbrella organisation the Association of Free Trade Unions of Slovenia (ZSSS), representing public sector workers, manufacturing sectors such as chemicals, food processing, construction, metal and electronics, and the hospitality sector, and the Chamber of Commerce and Industry of Slovenia (GZS). The social partners were not institutionally involved in the design of the first anti-corona package, which was adopted at the beginning of April 2020 and covered both employment retention and welfare measures. The social dialogue for the second anti-corona package, which amended the first one and introduced liquidity support measures for businesses and additional welfare measures, was slightly improved. By the time of the third anti-corona package in early May 2020, the Economic and Social Council was re-established and the social partners were more actively involved in the subsequent anti-corona packages.
Despite the lack of sufficient institutional involvement of social partners in the adoption of the first two key anti-corona packages, the government still heeded the views of social partners (Eurofound 2021, 20-23). The fact that many of the proposals from social partners were still included in the adopted packages at the start of the pandemic points to their associational and structural power in the Slovenian economy. ZSSS welcomed the introduction of a generous temporary wage compensation scheme, exemption from payment of social security, pension and disability insurance contributions, and temporary financial assistance for the self-employed, while GZS underlined the widening of the eligible economic sectors for the temporary wage compensation scheme, state coverage of pension insurance contributions and state guarantee for moratoria on bank loan repayments (ZSSS 2020a(ZSSS , 2020bGZS 2020aGZS , 2020b. Furthermore, the amendments, such as the removal of the flexibility for employers to adjust the working time of workers and the ban on firing workers for the duration of claiming the government wage compensation subsidy, followed the demands of trade unions after the formal tripartite structures were re-established in May 2020. This is not to say that for the remainder of the pandemic the trade unions continued to have a productive relationship with the Janša government. By the time of the second wave of the pandemic, in October 2020, trade unions started to have serious concerns with the government proposals in new anti-corona packages as the government tried to secretly insert anti-social measures, such as forced retirement and generous subsidies for private healthcare providers (ZSSS 2020c).

Electoral and partisan considerations of a minority government
While the overall approach of the government's social policy response can be explained by the path-dependent nature of the Slovenian welfare state legacy, some of the temporary crisis measures cannot be explained without taking account of the conservative nationalist ideology of the main governing party Slovenian Democratic Party (SDS) and prime minister Janez Janša. Social policy was used to expand welfare benefits for larger families and new-borns, while catering to its electoral base of farmers and devout Catholics (SDS 2021a(SDS , 2021b(SDS , 2021c. The government clearly wanted to broaden its electoral appeal by offering generous income support also for workers and the self-employed, and increasing welfare benefits for pensioners. In the context of strong civil society mobilisation against the authoritarian nature of the Janša government and a weak parliamentary majority, SDS was keen on using the available fiscal space during the pandemic to implement measures that would maintain its lead in public opinion polls. While all business sectors benefited from the generous income support schemes and the temporary exemption from payment of social security contributions, the tourism sector received additional support through the tourist voucher scheme (EUR 200 per adult). The economy minister since 2014 and leader of the junior coalition partner the Modern Centre Party (SMC) Zdravko Počivalšek, previously the manager of the spa resort Terme Olimia, was the main advocate behind this measure (Government of the Republic of Slovenia 2020). It is also telling that the employers were more satisfied with the involvement in the shaping of the anti-corona packages and the outcome of social dialogue than trade unions (Eurofound 2021) and had less criticism for the government's authoritarian leaning tendencies. Despite the high congruence of the government's business-friendly ideology with the employers' interests, Slovenia's COVID-19 social policy response was moulded by the institutional legacy of a strong welfare state and pragmatic calculations of the Janša government before upcoming parliamentary elections.

Comparison and conclusion
The three case studies presented here indicate that social policy responses to the COVID-19 crisis can largely be explained by the path-dependent effects of pre-crisis social partnership institutions and social policy legacies. Differentiating between employment and income retention schemes in the labour market and income support through nonemployment related welfare arrangements as the two key dimensions of traditional social policy, the analysis shows that the structural effect of these institutions accounts for social policy continuation in these three cases, each representing different subregional capitalist varieties. Despite recent government changes in all three countries, their policy trajectories reinforce the pre-pandemic status quo regarding both the mode of governance and output of social policymaking.
In Estonia, social policy was modest on both dimensions. The generally marginal position of social partners, unions in particular, explains the weak and temporary expansion of social benefits and non-labour market income policies. Minor institutional constraints increase government autonomy and Estonia's neoliberal policy legacy reinforces government choice for weak social policy. The Slovakian social policy response relied more on the labour market dimension, especially the Kurzarbeit scheme. This is largely explained by the institutional role of social partners and, particularly, the weaker position of unions relative to employers. As expected, the influence of Slovakian social partners through tripartite dialogue is conditional on the government willingness to negotiate. Though diverse interests within the coalition led to a compromise on this front, opportunistic partisan motives especially explain the relative generosity towards pensioners. Finally, Slovenia's generous labour market and welfare policies are explained by the strong structural position of both social partners and a generous welfare tradition. While the government did not initially include social partners, because of their structural power in the Slovenian economy, including the associational power of trade unions, the expressed views of social partners could not be ignored. Moreover, despite its centre-right profile, the Janša government adopted generous welfare on top of existing arrangements.
Based on the case studies, we conclude that the type of social policy response largely conforms to specific institutional regimes. Social partnership and welfare institutions, partly interrelated, ensure continuity in policy approaches. The partisanship perspective provided an additional explanatory leverage, especially where the baseline structural path-dependent nature of institutions could not account for certain modifications in less structured parts of welfare states. Furthermore, the structural power of economic sectors fundamental to national growth models proved important as governments internalised the necessity to accommodate the interests of labour and business actors in such sectors as matter of a policy priority (e.g. Slovakia's Kurzarbeit for the automotive sector)though also observing partisan strategies as driving support for business interests close to governing parties (e.g. Estonia's mining industry or Slovenia's tourism sector). Finally, through inductive reasoning, the case analyses incidentally invoke perspectives which fall outside our theoretical framework. Among these are policy learning, which occurred in countries where governments were inspired by successful practices elsewhere (e.g. Slovakia's Kurzarbeit), and contextual factors (e.g. Estonia's unprecedented borrowing). These factors add explanatory complexity but are relevant mostly on the margins of our general argument.
We expect that these mechanisms are similarly present in other cases, and that the types of social policy responses will be largely similar in cases belonging to the same regime varieties as studied here. The dynamics observed in Estonia are likely to be present in the other Baltics too, although the balance of institutional and partisan factors may be differentfor instance Lithuania's more solidaristic response to the initial phases of the pandemic was largely driven by the ruling social-democratic party trying to remain in power in the October 2020 elections (Eihmanis 2022). Likewise, the dynamics of the Slovakian case are expectedly alike in the Czech Republic, where social policy was also more geared towards employment retention (Oellerich 2022). The similar but smaller reliance of Czechia on the manufacturing sector reinforces the same logic for an extensive Kurzarbeit scheme. Although belonging to the same capitalist variety, the case dynamics are, however, likely to be different in the more deviant and volatile cases of Hungary and Poland. It is not our aim to argue that policy continuity necessarily prevails throughout the regionfurther research should test our arguments in these and other ECE cases.
The findings of this paper have important implications for the development of the ECE region, which not only is structurally dependent on foreign capital, but also is known for having sparse social protection systems, relative to Western Europe, to adequately cushion asymmetric economic shocks (Bohle and Greskovits 2012;Bohle, Eihmanis, and Toplišek 2022). While making clear that significant changes in the ECE existing welfare structures are not likely in the near future, the paper shows how modest changes in social policy can result from governments' partisan strategies. Nevertheless, only if such social policy layering has intentional support of strong political coalitionsor the urgency of another period of economic turmoil necessitates their re-enactmentcan we expect gradual, if moderate, upgrading of the ECE welfare systems.

Disclosure statement
No potential conflict of interest was reported by the authors.