Abstract
Whether nationally distinctive regulatory arrangements can survive global liberalization pressure remains unsettled. The comparative financialization literature establishes that national trajectories hold out but says little about the internal political process that produces such divergence. This paper explains South Korea’s contestation over the institutional arrangement of Geum-San Bunri (GSB)—the statutory separation of industrial and financial capitals—over four decades. Using explaining-outcome process-tracing, the analysis reconstructs socio-political coalitions, resources, and framing strategies at work in each of the five key phases. Two findings emerge. Substantively, GSB’s regulatory reach has narrowed through accumulated exceptions—a trajectory that layering describes accurately in broad outline. However, this path has not been monotonic; the industrial-capital ownership ceiling rose in 2009 and returned to its prior level in 2013. Formally, its status has remained unresolved for forty-four years, the principle continuing to organize every liberalization debate without ever being settled in either direction. The paper theorizes this combination as contested adaptation, a trajectory-level pattern specifying the political conditions under which layering fails to cumulate into displacement. It argues that Korea’s developmental state legacy persists not as inertia but as a heterogeneous configuration of moral, epistemic, and mobilizational resources actively reproduced in institutional changes.
-
Keywords: Finance/Industry Separation; Developmental State; Neoliberalism; South Korea; Institutional Change; Contested Adaptation; Post-Developmental State; Asian Capitalism
Introduction
Since the 1990s, financial systems worldwide have been reshaped by neoliberal reforms that prioritized deregulation, market competition, and financial innovation. In the United States, this shift culminated in the repeal of the Glass–Steagall Act in 1999, dismantling the separation between commercial and investment banking that had anchored the New Deal financial order. In East Asia, IMF-backed reforms after the 1997 financial crisis opened financial sectors to foreign capital and accelerated privatization (
Haggard, 2000;
Harvey, 2005). The prevailing expectation was convergence: sustained liberalization pressure would gradually erode nationally distinctive regulatory arrangements.
However, South Korea’s experience complicates this narrative. Since 1982, the principle of Geum-San Bunri (GSB)
1—codified primarily in the Banking Act and the Financial Holding Companies Act, and restricting non-financial capital from acquiring controlling stakes in commercial banks—has repeatedly survived liberalization pressure without remaining unchanged. Scholarship on GSB has established its function as an instrument of chaebol discipline and its post-crisis reframing as an impediment to competitiveness (
Chang, 2006;
Kalinowski & Cho, 2009;
S. J. Lee, 2008;
Pirie, 2007), but has tended to examine single conjunctures rather than the trajectory as a whole.
Across four decades, contestation has produced a regime that is formally intact and substantively reconfigured through accumulated exceptions, partial carve-outs, and renegotiated thresholds. The chaebol have not captured the banks, but they have expanded their financial reach in ways the original legislation did not anticipate. The same basic conflict recurs in different guises: in the mid- 2000s, the Financial Industry Structure Law debates pitted progressive civil society against chaebol-aligned conservatives, and the revision failed in committee in April 2006 (
S. J. Lee, 2008;
PSPD, 2006); in the Lee Myung-bak era, a “financial big bang” agenda raised the industrial-capital ownership ceiling from 4 to 9 percent in 2009 before a subsequent National Assembly returned it to 4 percent in 2013 (
Banking Act, 2013;
FSC, 2008); the Special Act on Internet-only Banks, passed in September 2018 and effective January 2019, created a 34 percent ownership exception for information and communications technology firms while nominally preserving the separation principle (
Act on Special Cases Concerning Establishment and Operation of Internet-Only Banks, 2018;
FSC, 2015); and in 2025, the artificial intelligence investment imperative prompted a new liberalization push centered on corporate venture capital that remained unresolved in legislative deliberation (
Ha, 2025;
Lee et al., 2025). GSB has neither been dismantled nor preserved—it has been continuously renegotiated.
This paper asks how and why this pattern of iterative, inconclusive contestation has sustained itself across four decades. I argue that the answer lies in a configuration specific to post-developmental state economies: the chaebol’s structural appetite for financial expansion generates persistent liberalization pressure, while the embedded autonomy of developmental state institutions—the Fair-Trade Commission (FTC), civil society organizations forged in the democratization struggle, and technocratic regulatory culture—provides the organizational resources to prevent wholesale dismantlement. The result is what I term “contested adaptation”: a recurring cycle in which each liberalizing advance is met by a defending counter-mobilization, producing partial reconfiguration rather than decisive resolution in either direction.
The analytical puzzle is this: why has GSB not followed one of the two trajectories standard theory predicts? It has not been dismantled—four decades of persistent chaebol lobbying, four sympathetic governments, and successive waves of neoliberal reform have failed to produce formal repeal. But it has also not been consolidated—the principle’s substantive reach is narrower in 2026 than in 1982, eroded through accumulated exceptions. Neither the path-dependence literature’s prediction of persistence nor the financialization literature’s prediction of convergent liberalization captures this outcome.
Contested adaptation I propose in this study is not offered as a wholly novel construct. It builds directly on
Mahoney and Thelen’s (2010) account of incremental institutional change and on the institutional complementarity literature (
Hall & Soskice, 2001;
Kang & Moon, 2012), and extends them at the level of trajectory rather than proposing an unrelated alternative. What is claimed as original is the specification of the political mechanism governing the form and pace of change—not the discovery of a previously unrecognized mode of change.
This paper employs explaining-outcome process-tracing across five historical phases to identify the causal mechanisms through which contested adaptation unfolds. The paper reviews the relevant scholarship, sets out the theoretical framework and methodology, traces the five phases, discusses their implications, and concludes with the study’s limitations.
Literature Review
Financial Regulation, Neoliberalism, and the Developmental State
A substantial body of scholarship examines the rise of financialization and neoliberal financial governance since the late twentieth century (
Epstein, 2005;
Harvey, 2005;
Krippner, 2005). Neoliberal deregulation was not merely a technical adjustment but a mechanism enabling finance to become a dominant mode of capital accumulation—justified in the name of efficiency and innovation while frequently generating systemic vulnerability (
Helleiner, 1994;
Stiglitz, 2010). Comparative political economy has shown how these global pressures were refracted through national institutional structures, producing distinct trajectories of liberalization (
Hall & Soskice, 2001;
Vogel, 2018).
For East Asia, this literature intersects with the developmental state tradition (
Amsden, 1989;
Evans, 1995;
Johnson, 1982;
Woo-Cumings, 1999). Korea’s postwar growth under authoritarian governments exemplified the developmental state model: state-controlled banks channeled credit toward strategic chaebol through policy loans, export incentives, and administrative guidance that effectively made the state the financial intermediary of choice for large-scale industrial investment (
Haggard, 1990). The 1997 crisis disrupted this model but did not simply displace it: under IMF conditions, Korea liberalized substantially, yet key developmental state institutions survived and adapted (
Haggard, 2000;
Kalinowski & Cho, 2009).
The question of whether the developmental state model has survived the combined pressures of democratization, crisis, and global market integration remains contested.
Pirie (2007) argues Korea has undergone decisive neoliberal transformation.
Thurbon (2016) contends a “developmental mindset” persists among Korean technocratic elites.
Weiss (2003) identifies selective adaptation of developmental instruments while retaining the fundamental orientation toward governed economic transformation. This paper contributes a mechanism-level account of how developmental state legacies are reproduced through concrete regulatory struggles—not simply inherited from a prior era and not simply replaced by neoliberal alternatives.
Research on Korea’s GSB has highlighted its role in restraining chaebol dominance and safeguarding financial integrity (
Chang, 2006;
Haggard, 1990;
Pirie, 2007). Following the 1997 crisis, the principle was increasingly recast as an obstacle to global competitiveness and financial modernization—a reframing that escalated through successive iterations: internet-only banks in the 2010s, and AI investment in 2025 (
S. J. Lee, 2008;
Park et al., 2021;
FSC, 2015).
S. J. Lee’s (2008) analysis of the Financial Industry Structure Law debates provides the most detailed account of one phase of this contestation, documenting how progressive civil society forged an alliance with the ruling party to block chaebol financial expansion in the mid-2000s, and noting the intra-governmental split between technocratic regulators and elected politicians that shaped the outcome.
Kalinowski and Cho (2009) analyze the post-crisis political economy of financial liberalization, showing how the reform coalition temporarily regained state autonomy from the chaebol before market-friendly interests reasserted themselves under Lee Myung-bak.
These remain the most detailed treatments of GSB politics, but both predate the period in which the principle changed most. Recent scholarship points in two opposing directions. Studies of Korean financialization show that the chaebol's financial reach expanded after 1997 through non-bank affiliates and internal capital markets—channels GSB did not govern (
Lechevalier et al., 2019;
Park & Han, 2022). Work on the post-2017 reform administration shows the converse: an explicit reform mandate and a reform-oriented FTC produced incremental adjustment rather than structural change (
Doucette, 2024). Neither coalition, in short, has prevailed. What this literature does not supply is an account of the encounters between them—the episodes in which the two met over a specific regulatory threshold, and why those encounters have repeatedly ended without resolution.
That gap is one of mechanism. Historical institutionalism's typology of incremental change—layering, conversion, drift, displacement—is indispensable for classifying what happens to a rule's substantive content (
Mahoney & Thelen, 2010), but it does not address why erosion proceeds quietly and cumulatively in some settings while in others every increment is publicly contested, separately negotiated, and occasionally reversed. Specifying the conditions that distinguish these two politics of erosion is this paper's analytical objective.
The contribution is empirical as well as theoretical. Tracing the coalitions, resources, and framing strategies of each conjuncture from the 1982 codification to the 2025 AI debate, the five-phase analysis offers the most systematic longitudinal account to date of how GSB has been defended, eroded, reversed, and reconstituted.
Theoretical Framework
This study draws on historical institutionalism (
Mahoney & Thelen, 2010) and the post-developmental state literature to analyze financial regulatory change in Korea. Historical institutionalism emphasizes that regulatory outcomes are mediated by path-dependent institutional legacies: institutions are not neutral rules but sedimented arrangements shaped by prior political struggles, generating self-reinforcing dynamics that constrain feasible reforms (
Pierson, 2004).
Mahoney and Thelen's (2010) typology of incremental institutional change—layering, conversion, drift, displacement—is the baseline against which I specify my claim. It is not a rival account to be replaced. Layering describes one dimension of the GSB trajectory accurately; my argument concerns what it leaves unexplained.
The Varieties of Capitalism framework (
Hall & Soskice, 2001) usefully identifies institutional complementarities but sits uneasily with East Asian developmental states, where the state plays a directive rather than facilitative role irreducible to either LME or CME logics.
Kang and Moon (2012) propose “state-led market economy” as a more adequate category;
Witt and Redding’s (2013) Asian Business Systems framework identifies multiple distinct Asian capitalisms that diverge significantly from Western models. I therefore supplement VoC with the post-developmental state literature’s emphasis on how developmental state embedded autonomy—the capacity of technocratic institutions to act independently of the industrial capital they regulate—is reproduced through concrete political struggles rather than simply inherited from a prior era (
Evans, 1995;
Thurbon, 2016).
I define contested adaptation as a trajectory-level pattern in which a codified regulatory principle undergoes continuous substantive erosion while its formal status remains politically unresolved, because each increment of erosion is separately contested by a defending coalition with the resources to impose costs but not to prevent movement altogether.
Two dimensions must be held apart. Substantive scope refers to the reach of the rule: what it prohibits, at what threshold, with what exceptions. On this dimension GSB has narrowed, broadly cumulatively, and layering in Mahoney and Thelen's sense describes it accurately. Formal status refers to the political standing of the principle. On this dimension GSB has not moved in forty-four years: it remains the reference point around which every liberalization proposal is organized and against which every proposal must be justified—including those framed as leaving it untouched, such as the 2025 corporate venture capital measures.
The analytical claim is that the second dimension is not derivable from the first. Layering identifies what happened to the rule’s reach; it is silent on why that erosion was public, episodic, and separately negotiated rather than quiet, and on why a principle whose reach has narrowed still organizes the field it no longer fully governs. Contested adaptation addresses the second question. It is complementary to the Mahoney–Thelen typology.
The complementarity is of a particular kind. The typology is ordinarily used diagnostically: one asks which mode a given episode instantiates, and the four are treated as alternatives among which the analyst chooses. The GSB record does not sort that way. The 2009 amendment is layering. The 2013 reversal is none of the four, since no mode anticipates the restoration of a prior threshold. The 2018 Special Act is layering that simultaneously reaffirms the principle it departs from, which is the opposite of conversion. More than one mode is in play at once, and which one is visible in a given episode is an outcome of the contest rather than a property of the institution. Contested adaptation names that condition.
Mechanism: Legitimacy Contestation
The primary mechanism is contestation over institutional legitimacy. Mahoney and Thelen locate the drivers of incremental change in veto possibilities and in discretion over the interpretation and enforcement of rules—properties of the rule and the political system that permit change without formal amendment. The GSB case turns on something different: recurring public contests over what the institution is for, in which both coalitions must justify their position before an audience that treats the question as morally rather than technically charged.
Two further mechanisms sustain the oscillation. The first is an asymmetric cost structure. Openly advocating chaebol bank ownership activates the moral legitimacy frame and mobilizes civil society opposition; consolidating the strictest reading of GSB invites the charge of hostility to innovation and competitiveness. Neither side can win outright at acceptable cost, which produces an equilibrium of mutual containment. The second is frame renewal. A liberalizing frame that has been defeated cannot be redeployed without political cost, compelling the liberalizing coalition to develop a new justification in each conjuncture—financial hub competitiveness in the 2000s, internet banking innovation in 2015–2019, AI investment in 2025. Each new frame opens a window for advance; each is then contested and partially discredited, restoring the equilibrium.
Statutory codification is what makes this politics possible. A separation principle written into statute at a specified numerical threshold function as a
Schelling (1960) focal point: it gives both coalitions an unambiguous object to coordinate around, converts any proposed change into a discrete and publicly visible event, and supplies the defending coalition with a bright line whose crossing can be named. Where a comparable principle rests on discretionary supervisory practice, no such focal point exists, adjustment need not take legislative form, and the politics of erosion is correspondingly quieter.
The defending coalition is structurally heterogeneous rather than unified, drawing on three analytically distinct resources. Civil society organizations supply moral legitimacy, invoking anti-monopoly sentiment, democratic accountability, and the memory of the 1997 crisis to impose reputational costs on liberalization advocates. Technocratic agencies, particularly the FTC, supply epistemic authority, challenging the factual premises of liberalization arguments. Opposition parties supply political mobilization, translating diffuse public preferences into legislative obstruction and procedural delay. Because these resources sit with actors of distinct organizational logics, a liberalizing coalition that neutralizes one finds the others still active. This distributed resilience, not organizational unity, is the source of the coalition's durability.
The claim is testable. If durability derives from heterogeneity, the coalition should be capable of authorizing a carve-out through some components while others resist, and the resulting exception should be bounded rather than open-ended. The 2017–2022 conjuncture, examined below, is: a reform administration and a reform-oriented FTC presided over the widening of the internet-only bank exception, which makes it the strongest available test.
Data and Methods
Variant of process-tracing: I employ explaining-outcome process-tracing in
Beach and Pedersen's (2013) sense: the objective is a minimally sufficient explanation of a specific outcome—the survival of GSB in reconfigured form—rather than the testing of an existing theory or the generation of a portable causal law.
Case selection: GSB is selected not because its outcome is typical but because it is anomalous: the financialization literature predicts convergent liberalization, the path-dependence literature predicts persistence, and GSB has done neither. Nor does the design select on an invariant outcome, since the dependent variable varies within the case: the ceiling rose in 2009, was returned to its previous level in 2013, was bypassed by a bounded exception in 2018, and remains contested in the unresolved 2025 proposals—four legislative episodes, four outcomes, including one outright reversal.
Materials: The primary evidentiary base is the legislative record: bills introduced, committee proceedings and voting records of the National Assembly's Finance and Economy Committee and its successors, enacted amendments and their statutory text, and audit transcripts, obtained from the National Assembly Bill Information System and the Korean Law Information Center. To these are added policy documents, press releases, and briefings of the Financial Services Commission and the Fair-Trade Commission; position statements and bulletins of the People's Solidarity for Participatory Democracy and the Citizens' Coalition for Economic Justice; and contemporaneous coverage from national dailies and the financial press.
Phase demarcation: Phases are bounded by the conjunction of two criteria: a change in governing coalition, and the emergence of a new justificatory frame deployed by the liberalizing coalition. Both must be present—a change of government that reproduces the previous frame does not open a new phase, and neither does a new frame advanced within a stable governing coalition.
Contested Adaptation: Five Phases of GSB
Developmental Origins (1980s)
GSB was codified in Korea’s Banking Act in 1982, limiting any single shareholder to 8 percent of voting shares in commercial banks. Its origins lay in the developmental state logic that had governed Korea’s political economy since the 1960s. Under Park Chung-hee’s government, the banking system was nationalized and credit was directed toward strategic chaebol through policy loans, export incentives, and administrative guidance that effectively made the state the financial intermediary of choice for large-scale industrial investment (
Amsden, 1989;
Woo, 1991). By the early 1980s, the chaebol’s appetite for bank ownership had become a systemic threat—controlling banks meant channeling depositors’ funds into affiliated firms, insulating inefficient subsidiaries from market discipline, and consolidating family ownership across industrial and financial domains.
The 1982 Banking Act revision was precipitated by the Jang Yeong-ja scandal, in which a large-scale private lending network exposed the systemic dangers of unregulated Finance-Industry entanglement and implicated figures close to the Chun government, making GSB codification politically urgent (
Schopf, 2001). Against this backdrop, the Economic Planning Board codified GSB not merely as a financial stability measure but as a mechanism of state power: a durable institutional constraint that would preserve developmental bureaucracy’s leverage over the chaebol regardless of which political coalition held power (
Milhaupt & Pistor, 2008).
GSB, state-directed credit allocation, and administrative oversight of the chaebol formed what
Hall and Soskice (2001) would call a similarity-based institutional complementarity: all three elements were organized around the shared logic of the developmental state’s capacity to steer industrial transformation. The state needed financial leverage over the chaebol to steer the economy; GSB provided that leverage by keeping banks out of conglomerate hands. This complementarity made GSB politically and institutionally expensive to remove—doing so would require reconfiguring the entire edifice of state-led financial governance, not merely amending a single statute.
The form of codification mattered as much as its content. By fixing an explicit numerical threshold in primary legislation, the 1982 revision created the focal point described above. Every subsequent contest over industrial ownership of banks would take the form of a proposal to move a specified number, and every such proposal would require legislative action visible to all parties. This is the structural precondition for the pattern traced in the four phases that follow.
The 1997 Crisis and the First Fracture (1997–2007)
The Asian financial crisis of 1997 was the first major shock to GSB's institutional configuration. Under IMF bailout conditions Korea liberalized sweepingly: interest rate ceilings were dismantled, foreign investors gained access to domestic capital markets, and non-performing banks were restructured or sold (IMF, 1997). Yet the crisis reinforced GSB rather than dismantling it. President Kim Dae-jung, long critical of the chaebol–state nexus, framed the reforms as an opportunity to build a "democratic market economy" that would constrain conglomerate power (
Kalinowski & Cho, 2009); the decision to sell nationalized banks to foreign investors rather than domestic conglomerates was in part a deliberate strategy against chaebol capture.
The defending coalition drew primarily on moral legitimacy in this phase. Two civil society organizations forged in the democratization struggle—the People's Solidarity for Participatory Democracy (PSPD) and the Citizens' Coalition for Economic Justice (CCEJ)—framed chaebol bank ownership as a question of democratic accountability, which resonated with a public that attributed the crisis to crony capitalism. Their organizational capacity had grown enough through the 1990s to make the frame stick: in 1998 alone the PSPD filed shareholder derivative suits against former managers of Korea First Bank, SK Telecom, and Hyundai Heavy Industries, establishing itself as an institutional actor in financial governance rather than a protest movement. One senior chaebol executive described it as "like another branch of government" (
Hamlin, 2001).
The Financial Industry Structure Law debate is the revealing episode. The bill proposed strengthening restrictions on chaebol financial ownership and failed in the National Assembly's Finance and Economy Committee in April 2006—a defeat the PSPD recorded as a victory for the coalition defending Finance/Industry separation (
S. J. Lee, 2008;
PSPD, 2006). The FSC, created in April 1998, proved more receptive to free-market arguments than the ruling party's politicians, an intra-governmental split that recurs in every subsequent phase (
FSC, 2008;
J.-T. Lee, 2008).
This phase produced neither substantive movement nor any resolution of formal status. It is the one phase in which the defending coalition prevailed on both, and it sets the baseline against which subsequent erosion is measured.
The Lee Myung-Bak “Financial Big Bang” and Its Reversal (2008–2013)
The Lee Myung-bak era produced the most sustained and explicit attempt to dismantle GSB. The Financial Services Commission's "Policy Directions for Financial Industry Development" (
FSC, 2008) laid out a three-stage liberalization blueprint, the most comprehensive challenge to the principle since its 1982 codification. Entering office with strong chaebol backing and a business background of his own, Lee pursued an openly “business-friendly” agenda (
FSC, 2008). The National Assembly passed a modified version in May 2009, raising the ceiling to 9 rather than the proposed 10 percent (
Banking Act, 2009). The Financial Investment Services and Capital Markets Act, in force from February 2009, simultaneously dismantled sectoral barriers between banking, securities, and insurance (
Financial Investment Services and Capital Markets Act, 2007).
The phase's distinguishing feature was ideological. Under Kim Dae-jung and Roh Moo-hyun the chaebol had been the objects of reform, their bank ownership restricted to prevent them from undermining the emerging post-developmental order. Under Lee they became the subjects of financial modernization—the actors whose capital and global ambition would build the "financial hub" (
FSC, 2008).
Yet legislative majority and ideological coherence did not produce dismantlement, and the reason lies in the heterogeneity of the opposition. The FSC, operating under a market efficiency logic, was partially aligned with the liberalizing agenda; the FTC, operating under an anti-concentration mandate, resisted any expansion of chaebol control over financial institutions regardless of efficiency arguments (
Fair Trade Commission, 2009;
S. J. Lee, 2008). Civil society framed relaxation as a reversion to pre-crisis crony capitalism, its moral legitimacy frame reinforced by the Samsung slush fund revelations of late 2007 and the special counsel investigation that followed (
Kalinowski and Cho 2009); opposition legislators deployed procedural delay. The Samsung case shows the combination at work:
2 the government drafted Insurance Business Act amendments that would have enabled the Lee Jae-yong succession structure through the Everland → Samsung Life → Samsung Electronics chain, and FTC opposition together with progressive legislators blocked them (
J.-T. Lee, 2008). The Financial Holding Companies Act was revised in 2009 (Act No. 9784) to facilitate holding company formation, but the core prohibition survived. What defeated the liberalizing coalition was its inability to answer three logics at once—anti-concentration, moral legitimacy, and political mobilization—no one of which could be neutralized by addressing another.
The phase did not end there. In 2013, under the economic democratization agenda that had dominated the preceding presidential campaign (
Doucette, 2015), the ceiling was returned from 9 to 4 percent, restoring the pre-2008 position (
Banking Act, 2013). The collapse of the Tongyang Group that year lent the reversal additional force: an industrial group had used its affiliated securities arm to sell the commercial paper and corporate bonds of failing subsidiaries to retail buyers, and the supervisory authority put the losses at roughly 1.6 trillion won across some 41,000 individual investors, after which civil society organizations pressed for extending the separation principle beyond banks to non-bank financial institutions (Financial Supervisory Service, 2013).
This episode is analytically decisive. Elsewhere in the sequence, liberalizing advances are contained or bounded—outcomes consistent with slow layering against friction. The 2013 amendment is not a containment but a reversal, returning the rule to the position it held before the most determined liberalization attempt in its history, and a unidirectional reading cannot accommodate it.
Fintech, Reform Government, and the Innovation Exception (2015–2022)
The fourth phase introduced a new liberalizing frame: technological innovation. ICT conglomerates—Kakao and KT—argued that GSB was structurally incompatible with digital banking (
FSC, 2015). The frame differed from its predecessors in asking not for general relaxation but for a bounded exception, on the grounds that a specific new business form could not operate under the general rule. The government's own 2015 plan proposed raising the ceiling for internet-only banks from 4 to 50 percent while leaving the general principle intact (
FSC, 2015); the figure that survived legislation three years later was 34. This is frame renewal in its most effective form: conceding the principle while requesting exemption from it denies the defending coalition the target a general relaxation would have offered.
The Special Act on the Establishment and Operation of Internet-only Banks passed in September 2018 and took effect in January 2019, permitting non-financial major shareholders to hold up to 34 percent of an internet-only bank subject to suitability review (
Act on Special Cases Concerning Establishment and Operation of Internet-Only Banks, 2018). Its carve-out was deliberately bounded: fintech entrants could participate, but groups formally designated as large business groups under the Fair-Trade Act could not. The boundary proved hard to hold. Kakao, operating at the scale of one of Korea's largest corporations by market capitalization, was not so designated at the time of licensing (as its total assets remained below the 10-trillion-won threshold for cross-shareholding restrictions in 2016–2017, placing it initially under a lower asset classification) and could therefore hold a 27 percent stake in Kakao Bank while its competitors remained subject to GSB (
Lee, 2021). By 2021 traditional banks and civil society groups were challenging the asymmetry as regulatory favoritism reproducing, under a new label, the preferential financial access GSB was designed to prevent (
Lee, 2021).
The 2020 revision of the Fair-Trade Act extended the pattern to a different instrument, permitting general holding companies to establish corporate venture capital subsidiaries
3 subject to limits on external fundraising and overseas investment (Monopoly Regulation and Fair Trade Act, 2020). The structure is identical—a bounded carve-out on an innovation rationale—and the 2025 debate concerns the boundaries of this provision rather than its creation.
Both exceptions were legislated under an administration elected on an explicit chaebol reform mandate, with a former civil society campaigner for chaebol reform heading the FTC.
4 This makes the phase the strongest available test of the heterogeneity claim advanced above. Had the defending coalition's durability rested on unity of purpose, this should have been the moment of collapse: those best positioned to defend the principle instead created exceptions to it. What happened instead is that the exceptions were bounded at every point—the 34 percent ceiling, the exclusion of designated business groups, the suitability review, the CVC fundraising and investment limits—and each boundary was negotiated separately, with civil society organizations opposing both measures as the administration advanced them and contesting their implementation afterward (
Citizens' Coalition for Economic Justice, 2020;
People's Solidarity for Participatory Democracy, 2020). The component that conceded did not carry the components that did not.
On the two dimensions, substantive scope narrowed more in this phase than in any other, while formal status was not merely left unresolved but explicitly reaffirmed: the Special Act was drafted as an exception to GSB rather than an amendment of it. The principle was invoked in the act of departing from it.
The AI Conjuncture (2025–)
On October 1, 2025, meeting OpenAI's Sam Altman alongside Samsung's Lee Jae-yong and SK's Chey Tae-won, President Lee Jae-myung signaled openness to relaxing GSB for the AI and semiconductor sectors, subject to safeguards against monopoly harm and spillover into other sectors (
Lee et al., 2025). The proposal centered on CVC regulation—a proposed increase in the external fundraising ceiling from 40 to 50 percent and in the overseas-investment ceiling from 20 to 30 percent, neither of which had been enacted at the time of writing—rather than on the bank ownership prohibition itself (
Jang, 2025). Both ceilings are boundary conditions written into the 2020 carve-out, and the defending coalition read the proposal to lift them as tailored to a single group: Solidarity for Economic Reform noted that the relaxation had been pressed most insistently by the chairman of SK, and that non-holding-company venture capital was already free to fund AI investment (
Jeong, 2025). At the National Assembly audit of October 28 the FTC chair acknowledged that targeted CVC relaxation was under consideration while maintaining that the separation principle remained valid (
Choi, 2025), and at a regulatory reform meeting on December 19 the Commission endorsed a conditional version with a five-year review mechanism, again reaffirming the principle (
Ha, 2025). The PSPD and CCEJ characterized the proposal as chaebol special treatment rather than AI investment policy (
Jeong, 2025). The measures remained in legislative deliberation at the time of writing.
The episode reproduces the structural features of the preceding phases. The liberalizing frame is new and the previous one is not redeployed; the proposal targets a bounded instrument rather than the core prohibition; the FTC combines conditional accommodation with normative defense, as in 2009 and 2018; civil society mobilized on the legitimacy frame within days; and the five-year review is a boundary condition attached as the price of concession, in the same form as the 34 percent ceiling before it.
Because the episode is unresolved, my reading is provisional. Contested adaptation predicts a bounded outcome: relaxation confined to a specified instrument, review or sunset conditions attached, the core prohibition reaffirmed in the enacting text. Early-stage dismantlement would look different—relaxation extending from CVC to the ownership ceiling, no boundary conditions, and, most diagnostically, no sustained counter-mobilization.
Discussion
Begin with the concession. On the dimension of substantive scope, layering is a largely accurate description of the GSB trajectory: exceptions were added around a formally unaltered core, the core's practical reach narrowed as a result, and no single amendment constituted repeal. The 2013 reversal complicates this reading—cumulative accounts do not anticipate the restoration of a prior threshold—but qualifies rather than defeats it.
What layering does not supply is an account of the politics. Mahoney and Thelen's mechanisms operate through veto possibilities and through discretion in interpretation and enforcement; they explain how change occurs without formal amendment. The GSB case is the opposite phenomenon. Change here required formal amendment at every point, was publicly contested at every point, and was separately bounded at every point, and there was little discretionary drift because a statutory threshold leaves little to interpret. The question is therefore not which mechanism produced the erosion but why the erosion took this political form.
The answer has three components. Codification at an explicit threshold creates a focal point that converts every proposed change into a discrete public event. Legitimacy contestation, rather than enforcement discretion, is the medium through which such events are fought, because the principle carries normative weight beyond its regulatory function. And a defending coalition whose resources sit with organizationally independent actors cannot be defeated by neutralizing any one of them, which yields bounded concessions rather than victory or defeat.
Table 1 separates the two dimensions accordingly.
This bears on how the typology is used. Applied diagnostically, it asks the analyst to select one mode; applied to the GSB trajectory, it yields layering, resistance to conversion, and an episode of re-consolidation, none of which characterizes the whole. What determines which is visible at any moment is not a property of the institution but the standing of the contest at that moment. Contested adaptation therefore does not compete with the typology for the same explanatory slot. It specifies the conditions under which the modes operate jointly rather than one prevailing: a principle codified at an explicit threshold, a balance in which neither coalition can secure complete victory, and a principle carrying public normative weight beyond its regulatory function. Whether those conditions travel to other post-developmental economies is a question for comparative work this paper does not undertake.
The case speaks to the debate over whether the Korean developmental state has survived neoliberal globalization.
Pirie (2007) reads the post-crisis period as a decisive rupture in which market-oriented governance largely replaced developmental infrastructure;
Thurbon (2016) finds a developmental mindset persisting among technocratic elites;
Weiss (2003) occupies a middle position, with instruments adapted and orientation retained.
Contested adaptation offers a mechanism-level resolution. The five phases show the developmental legacy persisting neither as continuous ideational inheritance, nor as stable adapted infrastructure, nor as disappearing residue, but as a configuration of political resources reproduced through concrete regulatory struggles. The FTC's anti-concentration mandate, the organizational memory of the PSPD and CCEJ, the expertise of regulators who recall why GSB was established—these are deployed, contested, partially depleted, and partially reproduced in each round.
Conclusion
This paper has traced South Korea's Geum-San Bunri principle across five phases—the developmental origins of the 1980s, the post-1997 conjuncture, the Lee Myung-bak "financial big bang" and its 2013 reversal, the fintech and reform-government innovation exceptions, and the ongoing AI debate—and argued that the trajectory exemplifies contested adaptation: continuous substantive erosion alongside an unresolved formal status, because each increment is separately contested by a coalition able to impose costs but not to prevent movement.
Three implications follow. First, contested adaptation specifies a political dynamic the standard typology leaves unaddressed. Layering describes what happened to GSB's substantive reach reasonably well; it is silent on why erosion is public and separately negotiated in some settings and quiet in others. That difference matters, because an institution eroded through public contest retains a capacity for reversal—demonstrated in 2013—that an institution eroded through drift does not. Second, developmental state legacies operate as active political resources rather than passive residues:
Evans's (1995) embedded autonomy persists not as administrative command but as organizational capacity, epistemic authority, and moral legitimacy deployed in concrete regulatory struggles. Third, the defending coalition's durability rests on organizational heterogeneity rather than unity of purpose, as the 2017–2022 conjuncture shows most clearly: components of the coalition authorized carve-outs that other components simultaneously contested, and the resulting exceptions were bounded at every point.
Whether this produces better outcomes than dismantlement or consolidation remains open. What is clear is the political form it takes: recurring mobilization and counter-mobilization, bounded concessions satisfying neither side, and long-run erosion that may eventually undermine the protective purpose without ever formally abandoning it—while remaining, as 2013 showed, susceptible to reversal in a way quiet erosion is not.
Four limits bound these claims. The design is single-case explaining-outcome process-tracing, so the framework’s portability to other post-developmental economies is untested. The 2025 conjuncture remains unresolved, and my reading of it is provisional pending the observable implications set out earlier. The evidence is the public record, adequate for reconstructing positions and outcomes but not the deliberation behind them—elite interviews with regulators, legislative staff, and civil society organizers, together with internal ministerial records, would be the single most valuable extension.
Two extensions follow. Within Korea, the framework can be assessed against circular shareholding regulation, which shares GSB's origin in economic-concentration policy and its FTC anchoring, and against labor market dualism, where the defending resources are configured differently and the test is correspondingly harder. Comparatively, the most informative cases are those where codification form and coalition composition vary independently—Taiwan, with a comparable developmental legacy and a differently organized civil society.
Notes
Table 1.Contested Adaptation in Relation to the Mahoney-Thelen Mechanisms
Table 1.
|
Mechanism |
Substantive scope |
Medium of change |
Defending coalition |
Formal status |
|
Layering |
Cumulative narrowing |
Addition of new rules |
Weak or bypassed |
Progressively irrelevant |
|
Conversion |
Stable, redirected |
Discretion in interpretation |
Acquiescent |
Reinterpreted |
|
Drift |
Narrowing by non-adaptation |
Non-enforcement |
Absent |
Quietly voided |
|
Displacement |
Superseded |
Substitution |
Circumvented |
Formally removed |
|
Contested adaptation |
Narrowing, non-monotonic |
Public legitimacy contest; formal amendment required |
Active, resourced, heterogeneous |
Unresolved; remains the organizing reference point |
REFERENCES
- Act on Special Cases Concerning Establishment and Operation of Internet-Only Banks [인터넷전문은행 설립 및 운영에 관한 특례법], Republic of Korea, Act No. 15856, October 16, 2018 (effective on January 17, 2019).
- Amsden, A. H. (1989). Asia's next giant: South Korea and late industrialization. Oxford University Press.
- Banking Act [은행법], Republic of Korea, Act No. 9784, June 9, 2009.
- Banking Act [은행법], Republic of Korea, Act No. 12101, August 13, 2013.
- Beach, D., & Pedersen, R. B. (2013). Process-tracing methods: Foundations and guidelines. University of Michigan Press.
- Chang, H.-J. (2006). The East Asian development experience: The miracle, the crisis and the future. Zed Books.
- Choi, S.-J. Choi, S.-J. (2025, October 28). Gukgam '25: “AI geumsanbunri wanhwa geomto … GP bangsigeun sinjung jeopgeun” [2025 parliamentary audit: Easing finance-industry separation for AI under review … cautious on the GP model]. EBN. https://www.ebn.co.kr/news/articleView.html?idxno=1684041
- Citizens' Coalition for Economic Justice. (2020, October 22). Statement of opposition to permitting corporate venture capital within general holding companies [일반지주회사의 CVC 허용에 대한 반대 의견서].
- Doucette, J. (2015). Debating economic democracy in South Korea: The costs of commensurability. Critical Asian Studies, 47(3), 388–413. https://doi.org/10.1080/14672715.2015.1057025
- Doucette, J. (2024). The postdevelopmental state: Dilemmas of economic democratization in contemporary South Korea. University of Michigan Press.
- Epstein, G. A. (Ed.). (2005). Financialization and the world economy. Edward Elgar Publishing.
- Evans, P. (1995). Embedded autonomy: States and industrial transformation. Princeton University Press.
- Fair Trade Commission (2009). Annual report on competition policy in Korea. Seoul: KFTC.
- Financial Holding Companies Act [금융지주회사법], Republic of Korea, Act No. 9784, June 9, 2009.
- Financial Investment Services and Capital Markets Act [자본시장과 금융투자업에 관한 법률], Republic of Korea, Act No. 8635, August 3, 2007 (effective on February 4, 2009).
- Financial Services Commission. (2008). Policy directions for financial industry development [금융산업 발전방향].
- Financial Services Commission & Financial Supervisory Service. (2015, June 18). Plan for introducing internet-only banks to promote IT-finance convergence and new growth engines [IT·금융 융합 및 신성장동력 창출을 위한 인터넷전문은행 도입방안]. Financial Services Commission.
- Financial Supervisory Service (2013, November 21). Comprehensive measures to prevent recurrence of similar incidents following the Tongyang Group crisis. https://www.fsc.go.kr/no010101/70914
- Ha, S.-R. (2025, December 19). Bandochae 'geumsanbunri' wanhwa teungnye, 5nyeonmada Gongjeongwi simsa badaya [Semiconductor exemption from finance-industry separation to require FTC review every five years]. Edaily. https://marketin.edaily.co.kr/News/ReadE?newsId=04237766642400160
- Haggard, S. (1990). Pathways from the periphery: The politics of growth in the newly industrializing countries. Cornell University Press.
- Haggard, S. (2000). The political economy of the Asian financial crisis. Institute for International Economics.
- Hall, P. A., & Soskice, D. (2001). An introduction to varieties of capitalism. In P. A. Hall & D. Soskice (Eds.), Varieties of capitalism: The institutional foundations of comparative advantage (pp. 1–68). Oxford University Press. https://doi.org/10.1093/0199247757.003.0001
- Hamlin, K. (2001, March 31). Scourge of the chaebol. Institutional Investor. https://www.institutionalinvestor.com/article/2btfm6ll7mr13f68vzh1c/home/scourge-of-thechaebol
- Harvey, D. (2005). A brief history of neoliberalism. Oxford University Press.
- Helleiner, E. (1994). States and the reemergence of global finance: From Bretton Woods to the 1990s. Cornell University Press.
- International Monetary Fund (1997, December 3). Korea—Letter of intent, memorandum on the economic program.
- Jang, Y.-M. (2025, October 3). OpenAI yeonghyangnyeok i jeongdo? … AI-e jinsimin I Jae-myeong, 43nyeon mugeun geumsanbunri 'wanhwa' chujin [OpenAI's influence this great? AI-committed Lee Jae-myung moves to ease the 43-year-old finance-industry separation]. ZDNet Korea. https://zdnet.co.kr/view/?no=20251003110201
- Jeong, S.-M. (2025, October 2). I Jae-myeong daetongnyeong geumsanbunri gyuje wanhwa sisae siminsahoe banbal 'hwaksan' [Civil society backlash spreads after President Lee Jae-myung signals easing of finance-industry separation]. Korea NGO News. https://www.ngonews.kr/news/articleView.html?idxno=215444
- Johnson, C. (1982). MITI and the Japanese miracle: The growth of industrial policy, 1925-1975. Stanford University Press.
- Kalinowski, T., & Cho, H. (2009). The political economy of financial liberalization in South Korea: State, big business, and foreign investors. Asian Survey, 49(2), 221–242. https://doi.org/10.1525/as.2009.49.2.221
- Kang, N., & Moon, J. (2012). Institutional complementarity between corporate governance and corporate social responsibility: A comparative institutional analysis of three capitalisms. Socio-Economic Review, 10(1), 85–108. https://doi.org/10.1093/ser/mwr025
- Krippner, G. R. (2005). The financialization of the American economy. Socio-Economic Review, 3(2), 173–208. https://doi.org/10.1093/SER/mwi008
- Lechevalier, S., Debanes, P., & Shin, W. (2019). Financialization and industrial policies in Japan and Korea: Evolving institutional complementarities and loss of state capabilities. Structural Change and Economic Dynamics, 48, 69–85. https://doi.org/10.1016/j.strueco.2017.08.003
- Lee, J.-T. (2008). I myeongbak jeongbu sidae, geumyung bikbaenggwa Hanguk gyeongjeui unmyeong [The Lee Myung-bak era: The financial big bang, and the fate of the Korean economy]. Simingwa Segye, (13), 82–100.
- Lee, K.-M. (2021, August 10). KakaoBank being granted 'special treatment.' The Korea Times. https://www.koreatimes.co.kr/business/banking-finance/20210810/analysis-kakaobank-beinggranted-special-treatment
- Lee, S. J. (2008). The politics of chaebol reform in Korea: Social cleavage and new financial rules. Journal of Contemporary Asia, 38(3), 439–452. https://doi.org/10.1080/00472330802078519
- Lee, Y.-J., Park, K.-S., Shin, Y.-A., & Park, S.-Y. (2025, October 3). I "geumsanbunri wanhwa, jehandoen yeongyeok" jeonmyeon chujin seon geutgo uiji jaehwagin [Lee: "Easing only in limited areas," rules out wholesale push but reaffirms intent]. Seoul Sinmun. https://www.seoul.co.kr/news/economy/2025/10/03/20251003012006
- Mahoney, J., & Thelen, K. (Eds.). (2010). Explaining institutional change: Ambiguity, agency, and power. Cambridge University Press.
- Milhaupt, C. J., & Pistor, K. (2008). Law and capitalism: What corporate crises reveal about legal systems and economic development around the world. University of Chicago Press.
- Monopoly Regulation and Fair Trade Act [독점규제 및 공정거래에 관한 법률], Republic of Korea, Act No. 17799, December 29, 2020 (effective on December 30, 2021).
- Park, C., & Han, G. (2022). The financialization of non-financial firms: The case of South Korea, 1997-2015. Current Sociology, 70(7), 1085–1105. https://doi.org/10.1177/00113921211012734
- Park, Y. C., Kim, J.-K., & Park, H. (2021). Financial liberalization and economic development in Korea, 1980-2020. Harvard University Asia Center.
- People's Solidarity for Participatory Democracy (2006, April 4). Revised Kumsanbup fails to pass in the Assembly's Finance and Economy Committee. Economic Reform Center.
- People's Solidarity for Participatory Democracy (2020, December 9). Nonpyeong: Jinjeongwa hangye dongsie namgin Jeongmuwiui gongjeonggeoraebeop gaejeongan [Commentary: The National Policy Committee's Fair Trade Act amendment: Progress and limits]. https://peoplepower21.org/economy/1750912
- Pierson, P. (2004). Politics in time: History, institutions, and social analysis. Princeton University Press.
- Pirie, I. (2007). The Korean developmental state: From dirigisme to neo-liberalism. Routledge.
- Schelling, T. C. (1960). The strategy of conflict. Harvard University Press.
- Schopf, J. C. (2001). An explanation for the end of political bank robbery in the Republic of Korea: The T + T Model. Asian Survey, 41(5), 693–715.
- Stiglitz, J. E. (2010). Freefall: America, free markets, and the sinking of the world economy. W. W. Norton & Company.
- Thurbon, E. (2016). Developmental mindset: The revival of financial activism in South Korea. Cornell University Press.
- Vogel, S. K. (2018). Freer markets, more rules: Regulatory reform in advanced industrial countries. Cornell University Press.
- Weiss, L. (2003). Guiding globalisation in East Asia: New roles for old developmental states. In L. Weiss (Ed.), States in the global economy: Bringing domestic institutions back in (pp. 245–270). Cambridge University Press. https://doi.org/10.1017/CBO9780511491757.013
- Witt, M. A., & Redding, G. (2013). Asian business systems: Institutional comparison, clusters and implications for varieties of capitalism and business systems theory. Socio-Economic Review, 11(2), 265–300. https://doi.org/10.1093/ser/mwt002
- Woo, J. (1991). Race to the swift: State and finance in Korean industrialization. Columbia University Press.
- Woo-Cumings, M. (Ed.). (1999). The developmental state. Cornell University Press.
