Idea and power

Historical institutionalism has underscored the resilience of institutions, highlighting how path dependency and self-reinforcing mechanisms make policy frameworks resistant to change (Pierson, Reference Pierson2000; Kay, Reference Kay2005). However, an institution is transformed, often through critical junctures triggered by external shocks, such as economic crises or natural disasters. The punctuated equilibrium model posits that these shocks disrupt institutional inertia, thereby enabling transformative change (Krasner, Reference Krasner1988). Yet, attributing institutional shifts solely to exogenous shocks has limitations, as change is rarely automatic. Scholars increasingly recognise that endogenous factors, including actors and agency, can shape the trajectory of policy transformation (Streeck and Thelen, Reference Streeck, Thelen, Streeck and Thelen2005; Mahoney and Thelen, Reference Mahoney and Thelen2010).

Within this broader debate, the role of ideas has gained prominence in understanding institutional change (Steinmo, Reference Steinmo, Porta and Keating2008; Béland, Reference Béland2009). Ideas influence the policy cycle, from defining social problems at the agenda-setting stage to framing solutions in the policy design phase and justifying decision-making reforms (Béland, Reference Béland2016). In times of crises, the role of ideas becomes more pronounced as actors face greater uncertainty. According to Blyth (Reference Blyth2002), ideas mitigate uncertainty by providing cognitive frameworks for interpreting crises, delegitimising existing institutions, and offering institutional blueprints. In doing so, ideas play a central role in enabling institutional change.

Early research on ideas often assumed that persuasive arguments and logical coherence were sufficient for policy adoption (see Béland, Reference Béland2016). However, as Parsons (Reference Parsons2016) note, “Ideas exercise an influence of their own but not completely by themselves” (448). Therefore, increasing attention has been directed to the role of power. Hay (Reference Hay2002) conceptualizes power as “the ability of actors (whether individual or collective) to ‘have an effect’ upon the context which defines the range of possibilities of others” (185), a definition that allows us to view power as encompassing both structural and agential dimensions. In other words, power may operate in “an indirect form … in which power is mediated by structures,” and in “a direct sense when A gets B to do something that s/he would not otherwise do” (Hay, Reference Hay2002, 186).

In this vein, the distinction between instrumental power and structural power is useful. These two concepts were introduced and developed in the analysis of business power (Winters, Reference Winters1996; Gough and Farnsworth, Reference Gough, Farnsworth and Gough2000; Hacker and Pierson, Reference Hacker and Pierson2002). Instrumental power refers to a firm’s capacity to influence policymakers through direct means, such as lobbying and sponsorship. In contrast, structural power refers to the influence that firms exert, even without direct action, arising from their positions within political and economic institutions. For example, governments often adopt business-friendly policies not because firms explicitly demand them, but because of their significant role in investment, employment, and state revenue. Further, ideological hegemony, “the ability to colonise sectors of social life through the shaping of underlying preferences” (Gough and Farnsworth, Reference Gough, Farnsworth and Gough2000, 87), is another critical source of structural power.

This distinction can also be applied to the exercise of citizens’ power. Citizens can exercise instrumental power by participating in elections, campaigns, or lobbying activities that directly influence policymakers (see Teorell et al., Reference Teorell, Torcal, Montero, van Deth, Montero and Westholm2007). At the same time, citizens may exercise structural power through their contributions to the economy and business, or simply by the threat of a strike that can influence decisions by both the state and firms. Therefore, while ideas may possess inherent generative capacity, their transformative potential in the policy arena is fundamentally contingent on the structural positions and resource endowments of actors.

Power over ideas is particularly relevant to understanding crises, as it highlights who has the authority to define problems and dictate the terms of debate (Carstensen and Schmidt, Reference Carstensen and Schmidt2016). The ability to control the framing of a crisis determines which policy solutions are considered legitimate. Existing studies on power and ideas have primarily focused on knowledge elites (Mandelkern and Shalev, Reference Mandelkern and Shalev2010) or expert networks (Helgadóttir, Reference Helgadóttir2016), examining how technocrats and policymakers shape policy debates. While these actors are important in examining instrumental power, focusing exclusively on elites overlooks broader power dynamics. In democratic systems, citizens also influence policy debates, and socio-economic and political hierarchies often stratify their impact on these debates. Some groups possess greater political and economic visibility, more substantial organisational structures, and closer ties to decision-makers, making them more effective at promoting their preferred ideas. Others, particularly marginalised groups, struggle to ensure that their concerns are even acknowledged in policy discussions.

This article shifts the focus to the power of citizens as political actors, arguing that their power determines the extent to which their interests are reflected in policy change. While citizens do not typically generate new policy ideas, they play a crucial role in determining which ideas gain traction by influencing elections, public debates, and policymakers’ incentives. However, not all citizens have equal access to policy-making processes. Policymakers tend to be more responsive to politically influential groups, while less organised citizens struggle to translate their preferences into policy outcomes. Therefore, we shed light on how unequal power resources across citizen groups influence policymakers’ responsiveness, demonstrating that policy change is not merely about identifying problems or generating solutions, but about who has the power to ensure that their ideas become reality.

Unequal responsiveness in the institutional context

The political responsiveness literature highlights how policymaking prioritises certain groups’ interests while overlooking others, a phenomenon known as ‘unequal responsiveness’ (e.g., Elkjær, Reference Elkjær2020). This concept refers to the fact that some groups’ concerns are more likely to be considered not only in policy outcomes but also in the agenda-setting process in government discussions (Traber et al., Reference Traber, Hänni, Giger and Breunig2022).

Much of the literature on unequal responsiveness has focused on economic class as a determining factor. Research from the US has demonstrated that income inequality directly influences policymaking, with the preferences of wealthier groups reflected in policy more than those of lower-income citizens (Gilens, Reference Gilens2012). This pattern is also found in Germany (Elsässer et al., Reference Elsässer, Hense and Schäfer2021), the Netherlands (Schakel, Reference Schakel2021), and even Nordic countries (Mathisen, Reference Mathisen2023; Persson, Reference Persson2024). Scholars suggest both instrumental and structural mechanisms that could drive these disparities. The instrumental mechanism is the ability of wealthy citizens to influence policy through donations and lobbying, as documented in the US (Gilens and Page, Reference Gilens and Page2014). Information asymmetry also plays a role, as wealthier citizens generally possess greater greater information, leading their policy preferences to more closely reflect prevailing political and economic conditions (Elkjær, Reference Elkjær2020). Indirect and structural mechanisms also reinforce unequal responsiveness, as higher-income individuals tend to participate more actively in elections and advocacy.

While unequal responsiveness is common across countries, its patterns and intensity vary significantly depending on institutional contexts. Scholars have analysed these variations through Esping-Andersen’s (Reference Esping-Andersen1990) welfare regime framework. In Norway, a social-democratic welfare state, high-income groups influence policy; however, institutionalised social citizenship ensures that other voices are also considered, thereby reducing disparities in policy outcomes (Mathisen, Reference Mathisen2023). In contrast, in Germany, a conservative welfare model, higher-grade employees, civil servants, and business owners have far greater influence over policy outcomes than lower-status groups (Elsässer et al., Reference Elsässer, Hense and Schäfer2021). This occupational stratification is reinforced by Germany’s contributory welfare system, which links benefits to employment history and income level, thereby further institutionalising unequal responsiveness.

Although research on unequal responsiveness in East Asia remains scarce, Korea’s case can be understood through the concept of developmental citizenship (Chang, Reference Chang, Chang and Turner2012a, Reference Chang2012b). Unlike Western welfare states, East Asian developmental states historically distributed rights and resources based on economic contribution rather than citizenship alone. This model emerged in the 1960s–1970s, when the state prioritised export-led growth, providing policy-driven financing and tax incentives to chaebols (Korean conglomerates) while linking social benefits to economic productivity (Chung, Reference Chung2007). Employment security and welfare benefits were concentrated among regular workers in large firms and the manufacturing sector, who were regarded as pivotal contributors to economic growth.

The stability of the manufacturing regime also depended heavily on subcontracted workers directly tied to these core enterprises. The share of subcontractors among small and medium-sized manufacturing firms rose dramatically – from 23.2 per cent in 1981 to 53.3 per cent in 1989, and further to 63.2 per cent by 1992 (Kim, Reference Kim1996), reflecting a deepening reliance on multilayered subcontracting networks. The 1988 decision to extend the coverage of the National Pension Scheme to workers in firms with 10 or more employees can thus be interpreted as an institutional response to this structural transformation, aimed at consolidating a stable and resilient production base. In contrast, the self-employed, non-standard workers, and marginalised populations outside the manufacturing sector were largely excluded.

Despite Korea’s democratisation in 1987, developmental citizenship remained entrenched, reinforced by neoliberal restructuring after the 1997 Financial Crisis (Chang, Reference Chang2012b). While social policy expanded, its benefits concentrated on regular workers in the manufacturing sector. In 1989, although manufacturing employment accounted for approximately 28 per cent of total employment (Yoo et al., Reference Yoo, Kim, Kim, Kim, Kim, Shin and Jeon2004), manufacturing workers comprised 2.4 million out of 4.5 million National Pension subscribers, nearly 54 per cent (National Pension Corporation, 2000).

During this period, Korea developed a two-tiered system of citizenship. Political elites occupied the core of policymaking and determined which issues received attention. Surrounding them were first-tier citizens, white-collar and manufacturing workers in large firms, as well as workers in subcontracting manufacturing firms. At the outermost periphery were second-tier citizens, including workers in non-manufacturing small and medium size enterprises (hereafter, SMEs), the self-employed, and non-standard workers. Over time, the composition of these two tiers has evolved in line with socioeconomic changes. For example, workers in subcontracted manufacturing firms have lost their position in the first tier, while platform workers, freelancers, and dependent contractors have come to be newly included in the second tier. Positioned far from the core of policymaking, these groups face significant barriers to conveying their voices to political elites and are consequently more likely to be excluded from policy benefits. This study argues that this stratified model of developmental citizenship offers insight into why the 1997 Financial Crisis and the COVID-19 pandemic produced vastly different policy responses, despite generating similar employment crises and policy proposals.

Many studies on unequal responsiveness, which predominantly employ quantitative methods, reveal the existence of unequal responsiveness; however, few provide insight into the underlying mechanisms. Therefore, this article adopts a case-oriented approach to examine how these power structures and dynamics influence policy outcomes. To this end, we demonstrate that policy ideas are not implemented in a vacuum but are mediated through existing hierarchies of power in the specific institutional context.

Methodology

This article employs a case study design using the Comparative Process Tracing (hereafter, CPT) method to compare Korea’s EI reforms during two major crises – the 1997 Financial Crisis and the COVID-19 pandemic. Process tracing is a qualitative methodology developed to address the limitations of quantitative research, which often relies on statistical correlations but fails to unpack the causal mechanisms – the so-called “black box” – between independent and dependent variables (Trampusch and Palier, Reference Trampusch and Palier2016). By closely examining event sequences, process tracing reveals “what is going on in the causal arrow in-between” (Beach, Reference Beach, Wagemann, Goerres and Siewert2020, 699), providing detailed evidence of how ideas and power interact throughout each stage of policy development.

A key strength of process tracing lies in its capacity to trace instrumental power, enabling researchers to identify how actors directly mobilise resources, exert influence, and shape decision-making through lobbying, coalition building, and participation in negotiations. However, tracing structural power with this method faces significant challenges: structural power often functions indirectly through institutional alignments and position-dependent constraints, rather than through observable sequences of actions. Therefore, CPT is increasingly recognised as a valuable tool for capturing structural power, particularly when carefully designed to analyse how social and economic structures condition actors’ capacity, set policy boundaries, and favour certain groups over others even in the absence of visible lobbying or contestation (Fairfield, Reference Fairfield2015).

CPT builds on traditional process tracing, which typically focuses on single-case analysis, by introducing a comparative dimension (Bengtsson and Ruonavaara, Reference Bengtsson and Ruonavaara2017). This approach compares processes across cases to examine how causal mechanisms operate in differing institutional and political contexts. Specifically, CPT employs a two-step analytic strategy, within-case analysis and structured comparison between cases, which enables broader theoretical generalisations about causal mechanisms in crisis-driven policy development (Bengtsson and Ruonavaara, Reference Bengtsson and Ruonavaara2017). In this context, CPT is particularly well-suited to the comparative analysis of the two crises examined in this study. By systematically comparing the sequence of events and the institutional and contextual constraints across the two periods, CPT enables the identification not only of direct forms of actor intervention but also the latent institutional advantages and disadvantages embedded in Korea’s political economy.

Data collection involved reviewing policy reports, government documents, news articles, and academic literature to gather relevant information. As this study focuses on structural power embedded across citizen tiers and instrumental power, we analyse the sequence of key events in each crisis rather than a bill-by-bill tracing of legislative bargaining processes. For the 1997 Financial Crisis, the analysis spans 1997–2000, concluding with President Kim Dae-jung’s announcement in December 2000 that Korea had fully repaid its IMF loans. For the COVID-19 crisis, the study examines the period from 2020 to 2023, when the pandemic ended.

Two crises and two responses

The 1997 Financial Crisis

The 1997 Financial Crisis was an unforeseen shock to Korea, a nation experiencing rapid economic growth. The collapse of major conglomerates, once deemed “too big to fail,” left a deep economic and social scar. Of the top 30 conglomerates, 17 collapsed or were sold, while financial institutions underwent significant downsizing (Cho, Reference Cho2007). The shock of restructuring among conglomerates serving as principal contractors precipitated a wave of bankruptcies among SMEs, with approximately 23,000 SMEs failing in 1998 alone (Ministry of Economy and Finance, 1999). As a result, Korea’s economic growth rate plunged from 6.2 per cent in 1997 to −5.1 per cent in 1998 (Lee, Reference Lee2020a). In response, the government injected substantial public funds into financial institutions destabilised by conglomerate insolvencies (Kim, Reference Kim2006) and implemented emergency measures for SMEs, including loan repayment deferrals and restructuring support (Ministry of Economy and Finance, 1998).

The unemployment rate, which had remained in the 2 per cent range for the decade up to 1997, soared to 6.8 per cent in 1998 (Lim, Reference Lim2000), peaking at 8.8 per cent in early 1999 (Lee, Reference Lee2020a). The employment shock affected white-collar workers in large corporations, that is, the core of first-tier citizens. In 1998, the number of workers losing EI coverage in large firms (1,000 or more employees) increased by over 70 per cent compared to 1997 (Central Employment Information Office, 1999). Also, white-collar workers experienced particularly steep increases, with losses rising by 81.6 per cent among clerical support workers, 61.9 per cent among managers, and 58.7 per cent among professionals, substantially exceeding the overall average increase (41.9 per cent) (Central Employment Information Office, 1999). Many white-collar employees faced downward mobility after the crisis: only 41.7 per cent retained their original position, while 38.4 per cent moved into lower-status sales jobs and 19.9 per cent into manual labour (Shin et al., Reference Shin, Lee and Cho2005). Moreover, in the manufacturing sector, both conglomerates and their subcontracted firms, which were vertically linked to them, were severely affected. By the end of 1998, a total of 1.12 million jobs had been lost, with 53 per cent of these losses concentrated in the manufacturing sector (Choi and Lee, Reference Choi and Lee1999).

At the time, Korea’s social policy was inadequate to absorb the economic shock. EI, introduced in 1995, was the only institutional mechanism available, yet coverage was highly limited. Initially, only workplaces with 30 or more regular employees were eligible for the unemployment benefit, while employment stabilisation and skills development programmes applied only to firms with 70 or more employees. As a result, in 1997, only around 20 per cent of the workforce was insured, leaving over four-fifths without protection. In sectoral terms, EI was concentrated primarily in manufacturing, which accounted for 48.1 per cent of all insured workers, the highest share across industries (Central Employment Information Office, 1999). Yet, while the manufacturing sector experienced severe disruption, with employment declining by approximately 600,000 between 1997 and 1998 (Ministry of Data and Statistics, 1999), unemployment benefit recipients remained at only around 200,000 (Central Employment Information Office, 1999).

This illustrates that EI failed to safeguard even the unemployed within Korea’s pivotal industrial sector. The crisis thus laid bare the structural vulnerabilities of the system in maintaining developmental citizenship, underscoring the urgent need for reform, and, indeed, the idea of expanding its coverage was implemented. This requires a careful tracing of how the reform unfolded during this period. However, Jeong (Reference Jeong, Yang, Kim, Jo, Kwon, Woo and Jung2008), who closely traced the history of the EI system, noted that ‘Little was achieved in collecting basic data… In most cases, requests were turned away after the relevant officials stated that the materials had been lost, leaving an undeniable sense of dissatisfaction’ (192–193). This suggests that EI reform can be understood only in the context of the broader dynamics that shaped its development during the crisis.

Based on the available data at the time, the post-crisis expansion of EI resulted from a combination of government efforts to address widespread employment and income insecurity, labour unions’ demands, and tacit consent from employers’ associations. Initially, the government proposed expanding EI as an emergency economic countermeasure at a meeting in early December (Song and Seo, Reference Song and Seo1997). Following his election later that month, President Kim Dae-jung announced detailed plans, including expanding coverage to businesses with five or more employees by July 1998 and to temporary and part-time workers by July 1999 (Kang, Reference Kang1998). Behind the scenes, as a recipient of International Monetary Fund (hereafter, IMF) Structural Adjustment Loans, Korea was required to implement reforms as a condition of the loan. The IMF agreement of 3 December 1997 explicitly called for the expansion of EI to enhance labour market flexibility.

Kim’s administration attempted to introduce a European-style welfare model, shifting away from growth-oriented developmentalism and neoliberal approaches. Central to this strategy was social dialogue, leading to the establishment of the Tripartite Commission in January 1998. This consultative body, comprising representatives from labour, business, and government, became the primary forum for negotiating crisis response policies. Within a month, the Tripartite Commission brokered the Social Pact of 6 February 1998, formally committing to expand EI coverage to businesses with five or more employees by mid-1998 (Ministry of Employment and Labour, 2016). As rising unemployment pressures accelerated the implementation timeline, by March 1998, the coverage had expanded ahead of schedule (Kim, Reference Kim1998).

The reform continued to evolve, and the idea of extending EI coverage immediately to all workplaces, including those with fewer than five employees, soon emerged. The origins of the idea remain contested between the government and organised labour (Jeong, Reference Jeong, Yang, Kim, Jo, Kwon, Woo and Jung2008). Despite differing claims regarding its origins, by late 1998, the worsening unemployment crisis led labour unions, civic groups, and academics to push for a faster expansion, arguing that workers in businesses with fewer than five employees were especially vulnerable (Ministry of Employment and Labour, 2001). Public opinion surveys indicated strong support, with 61.5 per cent of regular employees in small businesses and 58.7 per cent of temporary workers favouring the expansion (Hur and Hwang, Reference Hur and Hwang1998). The conservative opposition party did not oppose the expansion.

Consequently, the government moved forward with a Cabinet-approved expansion in August 1998, followed by parliamentary approval in September (Ministry of Employment and Labour, 2001). By October 1998, EI coverage was extended to all businesses (Table 1). The number of covered workplaces increased eightfold, from approximately 47,000 at the end of 1997 to 400,000 by 1998 (Ministry of Employment and Labour, 2001). Notably, the number of manufacturing firms covered by EI rose from 18,305 in 1997 to 97,044 in 1998 (Ministry of Employment and Labour, 2001). This expansion was groundbreaking, as key provisions of the Labour Standards Act, including severance pay, still did not apply to small workplaces at that time (Ministry of Employment and Labour, 2016). The swift and comprehensive expansion of EI in just over a year remains an unprecedented global case, marking a critical turning point in strengthening Korea’s social safety net.

Table 1. Changes in mandatory EI coverage by firm size

The COVID-19 crisis

The COVID-19 pandemic triggered a severe economic shock comparable to the 1997 Financial Crisis. In Korea’s recent history, negative economic growth has only been recorded in 1980, 1998, and 2020 (Han, Reference Han2021), underscoring the severity of the downturn. However, while the 1997 crisis primarily affected manufacturing and export-driven industries, the COVID-19 crisis disproportionately impacted service sectors. Strict social distancing measures disrupted industries that relied on face-to-face interactions, including wholesale and retail trade, hospitality, and education (Oh, Reference Oh2020).

As a result, the labour market impact differed markedly from that of 1997. While the 1997 crisis disproportionately affected permanent workers, job losses during the pandemic were concentrated among temporary and daily workers, as well as the self-employed (see Figure 1). An analysis of income changes between 2019 and 2021 shows that the self-employed without employees experienced an 18.2 per cent decline in income, and those with employees saw a 6.9 per cent reduction; by contrast, permanent employees experienced only a slight decline in earnings (Kwak, Reference Kwak2022). These patterns indicate that precarious workers bore the brunt of the crisis, in contrast to regular workers.

Figure 1. Employment trends during economic crises

Source: Sung (Reference Sung2021, 183).

Note: Translated from Korean by the authors.

The crisis also exposed the heightened vulnerability of newly emerged non-standard workers, including platform workers, freelancers, and dependent contractors, who were not widely recognised in 1997. The digitalisation of work has created a new class of precarious labour, but their lack of employment security has left them particularly exposed during the pandemic. According to a survey of freelancers and dependent contractors, average income fell by 69 per cent relative to the pre-COVID period (Seon, Reference Seon2020). Reductions in working hours were also prevalent among non-standard workers. Based on a survey conducted among workers in Seoul, 52.7 per cent of non-standard workers experienced reduced working hours, compared with 14.9 per cent of regular employees (Shin and Woo, Reference Shin and Woo2020).

Significant coverage gaps remained despite the expansion of EI since 1998. By 2018, only around 48 per cent of all employed individuals – both wage and non-wage – were enrolled, leaving large segments of precarious workers unprotected (Kim, Reference Kim2020). These gaps stemmed from a system designed for traditional employer-employee relationships, leaving self-employed individuals and non-standard workers ineligible for coverage. Additionally, many temporary and daily workers who were legally required to enrol were excluded due to high job mobility and weak enforcement (Joint Task Force of Relevant Ministries, 2020).

Recognising the urgency, President Moon Jae-in declared COVID-19 an “economic emergency” and introduced relief measures (Jeong, Reference Jeong2020). Among these, temporary cash transfer programmes were rapidly implemented to support workers excluded from EI. In April 2020, the government launched ‘emergency employment security payments’, providing 500,000 KRW (about 340 USD) per month for three months to dependent contractors, freelancers, and self-employed individuals who lost income during the pandemic (Presidential Office, 2020). Additionally, the government introduced disaster relief payments for all citizens, though the second round in 2021 was restricted to the bottom 88 per cent of income earners (Choi et al., Reference Choi, Noh, Han and Gentilini2024). However, as the pandemic persisted, short-term cash transfers were insufficient to address the structural vulnerabilities of precarious workers, underscoring the need for a more inclusive employment protection system.

President Moon formally proposed the “Universal Employment Insurance (hereafter, UEI)” in May 2020, aiming to cover all workers who have an income, regardless of their employment type (Do, Reference Do2020). In July, the Economic, Social, and Labour Council adopted a tripartite agreement to establish an implementation plan by the end of the year (Economic, Social, and Labour Council, 2020). This culminated in December, when the government unveiled the UEI Roadmap, outlining a transition from a workplace-based to an income-based EI system (Joint Task Force of Relevant Ministries, 2020). The plan proposed a gradual expansion, starting with artists in December 2020, then dependent contractors and platform workers from 2021, and self-employed workers in a later phase.

The idea of income-based EI did not suddenly emerge in response to the COVID-19 pandemic. In 2017, the Moon administration announced plans to reform the EI system by basing eligibility on income, thereby mandating enrolment for dependent contractors and artists (Kim, Reference Kim2017). Subsequently, in 2018, the Employment Insurance Committee, comprising representatives from labour, business, and government, approved a plan to expand its scope (Ministry of Employment and Labour, 2018) and submitted related legislation to the National Assembly. However, the proposed bill stalled in the National Assembly, leaving the agreed reforms largely unimplemented (Park, Reference Park2020). Later, amid the escalating COVID-19 outbreak and in the run-up to the April 2020 general election for the 21st National Assembly, radical left-wing parties put forward UEI as a campaign pledge (Lee, Reference Lee2020b). Yet major political parties did not adopt it as a core campaign commitment, keeping the initiative on the margins of the policy agenda.

Public opinion on the introduction of UEI has been broadly favourable, with strong support expressed not only by the general public but also by second-tier citizens. According to a survey conducted by the Presidential Committee on Jobs (2020), approximately 70 per cent of all respondents and 66.8 per cent of the self-employed expressed support for the adoption of such a scheme. In the same survey, 72.3 per cent of current non-enrolled indicated a willingness to enrol even at the cost of paying premiums. A survey conducted by the Ministry of Employment and Labour (2020a) targeting dependent contractors also revealed that about 85 per cent were willing to join the EI system.

Despite widespread support, concerns arose regarding the slow, phased approach and the exclusion of self-employed workers from early implementation (Nam, Reference Nam2020). The labour-oriented left-wing party welcomed the introduction of UEI but criticised the government’s gradual approach as “complacent” and “unduly passive,” arguing that it would leave those in the EI blind spot unprotected for an extended period (Yoon, Reference Yoon2020). By 2022, the scope of EI had been expanded to include artists, along with certain categories of dependent contractors and platform workers; however, coverage still remained partial, discussions on the inclusion of self-employed individuals stalled, and no further progress on income-based EI was made (Nam, Reference Nam2024). Following the election of conservative President Yoon Suk-yeol in May 2022, the momentum for the UEI weakened further. While Yoon pledged to expand an income-based EI system, his administration failed to advance institutional reforms, and efforts to develop a Real-Time Information (hereafter, RTI) system were either halted or disbanded.

Ironically, the stalling of the UEI plan was in part driven by the Democratic Party, one of its principal advocates. In December 2022, the Income Tax Act was amended to require that, from 2024 onwards, the Simplified Statement of Payment of Earned Income for all regular wage earners be submitted monthly rather than every 6 months, as part of establishing the RTI system. However, in November 2023, a Democratic Party lawmaker introduced a bill to defer this change until 2026 (Lee, Reference Lee2023).

The lawmaker defended the postponement by citing the increased compliance burden on small business owners. However, when pressed by the media on whether the perspectives of vulnerable groups had been taken into account, no clear response was provided (Ko, Reference Ko2023). In fact, the Korea Association of Certified Tax Accountants, together with the Korea Federation of SMEs, the Korea Federation of Micro Enterprises, and the Korean Taxpayers Association, took the lead in submitting a formal statement of opposition, arguing that the system would place an excessive burden on SMEs and microbusiness owners (Oh, Reference Oh2023). Ultimately, such political decisions left the UEI agenda unrealised. By late 2024, EI enrolment stood at 15.31 million, well short of the 2025 target of 21 million (Ministry of Employment and Labour, 2025).

Discussion: three-gap mechanism

Both periods were marked by unprecedented economic and employment crises, and in both cases, the proposed EI expansion faced common administrative challenges and burdens. Following the 1997 Financial Crisis, incorporating all workers into the system proved extremely challenging. In fact, bureaucrats at the time opposed such an expansion, as it would have required a substantial increase in administrative personnel and infrastructure (Jeong, Reference Jeong, Yang, Kim, Jo, Kwon, Woo and Jung2008). Also, during the COVID-19 era, extending EI to non-standard workers and the self-employed was regarded as a highly complex and institutionally difficult task.

Politically, in both periods, progressive governments were in power. However, the 1997 Financial Crisis presented a less favourable environment for reform, as the administration operated under a divided government, with opposition parties controlling the National Assembly. In contrast, during the COVID-19 crisis, the ruling party held a comfortable majority, creating far more conducive conditions for pursuing reform. Nevertheless, the 1997 Financial Crisis led to rapid institutional reform, whereas the UEI failed to be fully institutionalised during the COVID-19 crisis. This divergence raises fundamental questions about who controls ideas, how different actors shape the policymaking process, and how power translates into institutional change. To address this, we introduce the developmental citizenship framework, rooted in a manufacturing-centred growth model, classifying citizens into two tiers based on their level of power. Here, power includes both instrumental power and structural power (e.g., Hacker and Pierson, Reference Hacker and Pierson2002).

The findings reveal that during the 1997 Financial Crisis, the primary victims were first-tier citizens, white-collar and manufacturing workers in large firms, as well as workers in subcontracting manufacturing firms, who wielded strong power. There was little resistance to the swift expansion of EI. In contrast, the COVID-19 pandemic hit second-tier citizens the hardest: service-sector workers, the self-employed, and non-standard workers, who had significantly weaker power in the Korean economy. In this context, while UEI had been extensively discussed, policy progress came to a standstill. This article argues that the three-gap mechanism, comprising (1) structural presence gap, (2) representation gap, and (3) political participation gap, shapes and sustains power inequality among citizens through both instrumental and structural means. While the structural presence gap reflects structural power, the representation and participation gaps encompass both instrumental and structural dimensions of power.

First, a structural presence gap separates first-tier and second-tier citizens. As discussed earlier, Korea has consistently maintained an economic growth-oriented discourse, with the manufacturing sector serving as the foundational industry and the core of national competitiveness to date. Since the 1980s, the manufacturing sector has consistently accounted for between 20 and 30 per cent of Korea’s GDP, a level markedly higher than the OECD average (World Bank, 2025). In this context, Korea’s distinctive manufacturing production network – anchored in large conglomerates and their subcontracting chains – became firmly established, with these actors regarded as pivotal to the national economy and occupying a long-standing dominant position within the broader socioeconomic structure. This endowed their workers, known as first-tier citizens, with structural power, such that any loss or hardship they experienced was not perceived merely as an individual misfortune but as a threat to the viability of the national economy.

The severity of the 1997 crisis vividly revealed this structural dependence. When the manufacturing ecosystem faced a potential collapse, safeguarding this sector was essential for economic recovery. Instability within conglomerates inevitably spilt over to subcontractors, whose collapse in turn had severe repercussions for the conglomerates themselves, thereby underscoring the necessity of an inclusive social safety net that encompassed not only conglomerate employees but also workers in subcontracted SMEs. The Tripartite Commission moved swiftly to expand EI at the price of labour market flexibilisation.

This institutional compromise left behind a dual legacy for Korea’s society. While the expansion of EI marked an unprecedented strengthening of the social safety net, the flexibilisation of the labour market simultaneously accelerated the rise of non-regular workers in manufacturing, thereby reducing the relative size of first-tier citizens. Still, large-firm workers remained in a comparatively privileged position. They retained greater organisational power and bargaining capacity, enabling them to resist some of the harshest forms of flexibilisation that SME workers endured. In contrast, service workers, non-regular employees, and the self-employed were not prioritised as part of the nation’s core economic sectors and essential labour force, leaving their structural presence weak.

This socioeconomic structure did not remain confined to the past; it continued to exert a path-dependent influence in subsequent crises. During the COVID-19 pandemic in 2020, the hardships of second-tier citizens were not treated as urgent issues warranting immediate policy intervention. As a result, despite ongoing demands for institutional reform, the crisis experienced by second-tier citizens was comparatively easy to neglect. This illustrates how the structural presence gap between the two citizen groups has contributed to the long-term reproduction of inequality in crisis response.

Second, a representation gap highlights the unequal ability of citizens to convert interests into policy influence. Korea’s single-member district electoral system and limited proportional representation create structural barriers to including diverse voices in policymaking. Between 1987 and 2020, over 90 per cent of elected lawmakers were male university graduates (Kim, Reference Kim2021), underscoring the chronic underrepresentation of women, individuals with lower educational attainment, platform workers, dependent contractors, and self-employed citizens – all key constituencies of the second-tier population.

This difference is also clearly reflected in the contrast between the manufacturing and service sectors. In 1995, Korea had approximately 1.61 million union members, of whom manufacturing accounted for about 39 per cent. By contrast, among major service sectors, wholesale and retail trade, food and accommodation services accounted for only around 5 per cent, while social and personal services represented around 14 per cent (Lee, Reference Lee1997). Although there has been a recent rise in unionisation in the service sector (Kwon, Reference Kwon2023), precarious second-tier citizens continue to face significant barriers to organising unions due to their vulnerable legal status.

During the 1997 Financial Crisis, the two major trade unions, the Korean Confederation of Trade Unions and the Federation of Korean Trade Unions (hereafter, FKTU), joined tripartite negotiations, ensuring that the challenges faced by first-tier citizens were actively addressed. However, although second-tier citizens were the most severely affected by the COVID-19 pandemic, the FKTU, whose organisational base weakly represents this group (Jeon and Go, Reference Jeon and Go2021), served as the sole labour representative in the 2020 tripartite agreement. Moreover, the UEI Task Force organised by the government lacked explicit provisions ensuring the participation of relevant stakeholders, namely second-tier citizens, in its organisational structure, stipulating only that experts may be appointed when necessary (see Ministry of Employment and Labour, 2020b).

This representational gap inevitably exerts a profound influence on the discourse surrounding the UEI. Expanding EI to previously uncovered groups would require higher financial contributions from those already within the system. Given the relatively higher unemployment risk faced by second-tier citizens, expenditure on the unemployment benefit would likely increase substantially, leading to higher contribution rates (Kim and Kim, Reference Kim and Kim2021). In particular, regular employees in large firms, whose incomes are relatively high and whose tax bases are more transparent, would be likely to bear a disproportionate share of this increased financial burden. Such dynamics could weaken first-tier citizens’ support for UEI, even when they understand its underlying objectives. Without a strengthened political voice for second-tier citizens, policy changes will likely be decided by the interests of first-tier citizens, thereby delaying policy reforms and further entrenching the exclusion of second-tier citizens from the social protection system.

Third, a political participation gap further reinforces unequal responsiveness. A survey conducted in 2021 revealed that 86.5 per cent of regular workers reported participating in most or all national elections in Korea, followed by self-employed individuals (82.99 per cent), temporary workers (76.26 per cent), and daily workers (65.87 per cent) (Korea Institute for Health and Social Affairs, 2021). Despite introducing an early voting system in 2013 to enhance participation opportunities, voter turnout among second-tier citizens remains relatively low. Particularly, dependent contractors, freelancers, and platform workers face restrictions in exercising their voting rights, as they fall outside the classification of ‘workers’ under the Labour Standards Act (Lee, Reference Lee2024). Furthermore, workers in small businesses with fewer than five employees are not entitled to paid leave on election day, which significantly impedes their political participation (Kim, Reference Kim2022).

In the domain of non-institutional political participation, such as signing petitions, engaging in protests, or boycotting, first-tier citizens also exhibit higher levels of engagement. According to the 2021 Social Integration Survey, 33 per cent of regular workers reported participating in at least one non-institutional political activityFootnote ¹ in the past year, followed by temporary employees (31.8 per cent), the self-employed (26 per cent), and daily workers (24.2 per cent) (Korea Institute of Public Administration, 2021). Overall, first-tier citizens are more politically active, ensuring that their voices are more effectively represented in the political landscape. This imbalance contributes to the persistence of unequal responsiveness and helps explain why policy change was slowed or blocked in 2020, despite widespread recognition of the need for reform. Table 2 provides an overview of the three-gap mechanism.

Table 2. Comparative analysis of EI reforms during the 1997 Financial Crisis and the COVID-19 pandemic