국제투자분쟁해결절차상 심사기준에 관한 연구 = A Study on the Standard of Review in International Investment Dispute Settlement Proceedings

국문 초록 (Abstract)

국제투자중재에서 ‘심사기준(standard of review)’은 단순히 사법적 개입의 강도를 조절하는 기술적 기제를 넘어, 외국인 투자자 보호와 투자유치국의 규제 자율성이라는 상충하는 가치 간의 균형점을 모색하는 핵심 판단기제로서 기능한다. 이는 ‘투자자-국가 간 분쟁해결(Investor-State Dispute Settlement: ISDS)’ 제도가 직면하고 있는 정당성 위기(legitimacy crisis)를 타개하고 체제의 신뢰를 회복하기 위한 실효적 방안이 될 수 있다는 점에서 그 중요성이 크다. 그러나 기존 연구들은 심사기준을 체계적인 분석 틀이 아닌, 수사적 도구로만 활용하거나 심사방법(methods of review)의 선택에 치중하는 등의 한계를 보였다.
이에 본 연구는 심사기준을 그 자체로서 공법적 차원의 권한배분(allocation of power)을 구현하는 규범적 판단기제로 재정립하는 것을 목적으로 한다. 이를 위하여 심사기준을 심사 주체와 대상에 따라 국제투자중재판정부가 국가의 규제 조치를 심사할 때 적용하는 ‘실체적 심사기준(substantive standard of review)’과 중재지 법원이 국제투자중재판정의 효력을 심사할 때 적용하는 ‘절차적 심사기준(procedural standard of review)’으로 이원화하고, 이를 토대로 주요 판정례에 나타난 심사기준에 관하여 체계적으로 고찰한다.
ISDS 주요 판정례 분석 결과 심사기준의 적용에 있어 일정한 경향성이 있다는 것을 확인할 수 있었다. 우선 실체적 심사기준과 관련하여, 국제투자중재판정부는 환경보호나 공중보건과 같은 민감한 영역의 공공정책에 대해서는 비교적 존중적인(deferential) 태도를 취하면서도, 해당 공공정책이 형성·집행되는 과정에서의 절차적 정당성, 합리성, 자의성 여부에 대해서는 보다 엄격한 심사기준을 적용하는 경향을 보인다. 이는 국제투자중재판정부가 국가의 정책적 선택 그 자체를 대체하려 하거나 재단하려는 것이 아니며, 오히려 국가가 그러한 결정을 내리는 과정에서 자의적이거나 불공정한 행위를 하지 않았는지를 통제하는 데 집중하고 있음을 시사한다.
반면에 절차적 심사기준과 관련하여 중재지 법원은 국제투자중재판정의 실체적 판단에 대해서는 개입을 자제하고 존중(deference)하는 태도를 확인할 수 있다. 다만, 국제투자중재판정부의 권한이 국가의 동의에 기반한다는 점을 고려했을 때, 권한의 범위를 설정하는 ‘관할권’ 문제나 변론권과 같은 ‘절차적 정의’와 결부된 문제에 대해서는 엄격한 심사기준을 적용하는 경향을 보인다. 이는 본안 판단에 대한 개입은 최소화하여 국제투자중재판정의 최종성을 보장하되, 관할권 유월과 같은 중대한 하자는 엄격히 바로잡아 ISDS 제도에 대한 신뢰를 제고하려는 것이라 할 수 있다. 그러나 이러한 경향성에도 불구하고, 심사기준에 대한 통일된 기준이 부재함에 따라 여전히 다수의 국제투자중재판정에서 판정 결과의 불일치가 반복되고 있으며, 이는 궁극적으로 ISDS 제도의 예측 가능성과 정당성을 저해하는 요인으로 작용하고 있다.
이러한 구조적 모호성을 해소하기 위한 구체적인 대안으로서 본 연구는 ‘단계별 심사기준 적용’ 모델을 제안한다. 이 모델은 심사의 대상을 ‘중재판정 단계(arbitral phase)’와 ‘판정 후 단계(post-award phase)’로 이원화하여, 각 단계의 기능적 특성에 맞는 상이한 심사기준의 적용을 핵심으로 한다. 즉, 중재판정 단계에서 국제투자중재판정부는 국가의 규제 권한을 기본적으로 존중하여야 한다. 이를 위하여 기존의 엄격한 ‘비례성 심사’에서 절차적 ‘합리성 심사’로 전환하고, 고도의 과학적 전문성이 요구되는 사안에 대하여 규제 당국의 판단을 존중하는 ‘인식적 존중(epistemic deference)’을 부여할 것을 주장한다. 반면, ‘판정 후 단계’에서는 국제투자중재판정의 최종성을 보장하기 위하여 관할권 존부 및 절차적 하자에 대하여 엄격한 심사기준을 적용하되, 실체적 분쟁에 관한 법률적·사실적 오류에 대한 개입은 엄격히 배제되어야 함을 주장한다.
그 밖에도 현행 체제 내에서 즉시 적용 가능한 대안으로서 국제투자협정에 심사기준을 명문화할 것을 제안하며, 또한 과도기적 조치로서 체약국 간 공동해석선언의 활성화와 국제기구 차원의 심사기준에 관한 실무 가이드라인의 제정 등 구체적인 이행 전략을 함께 제시하였다. 이러한 조치들은 모호한 심사기준으로 인하여 발생하는 규제위축(regulatory chill) 효과를 완화하고, 투자자 보호와 국가의 규제 권한 사이의 균형을 위한 현실적이고 효과적인 방안이 될 것이다.
물론 ISDS 체제의 비상설적(ad hoc) 성격과 규범의 파편화(fragmentation), 그리고 실체적 심사기준과 관련하여 제기될 수 있는 국내법원의 사법주권 침해 가능성에 대한 우려 등을 고려할 때, 일관된 심사기준의 정립이 현실적인 난관에 봉착할 것이라는 비판은 여전히 유효하다. 그러나 국가의 공공정책을 대상으로 하는 현 ISDS 체제의 정당성 위기를 극복함에 있어, 심사기준의 정립은 상설투자법원이나 상소제도 도입과 같은 급진적 구조 개혁보다 현실적이고 즉각적인 대안이 될 수 있다. 이는 불확실한 현 체제 내에서 투자자 보호와 국가의 규제권한 사이의 균형을 모색하고 체제의 지속 가능성을 담보하는 실효적인 규범적 나침반으로 기능할 것이다.

다국어 초록 (Multilingual Abstract)

I. Introduction

The primary mechanism for resolving disputes between foreign investors and host states has shifted from domestic remedies and diplomatic protection to the Investor-State Dispute Settlement (ISDS) system, which allows investors to file arbitration directly against sovereign states. While the number of ISDS cases has surged to over 1,401 as of December 2024, confirming the system's effectiveness, the increasing number of claims challenging public policy measures in areas such as public health and the environment has raised significant concerns regarding ‘regulatory chill’ and inconsistent arbitral awards. Despite various state-led reforms, including treaty amendments and withdrawals, the lack of a unified ‘standard of review’ continues to undermine the system's legitimacy and predictability.
Therefore, this study aims to fill the recent academic gap by establishing a theoretical framework for the standard of review and empirically analyzing its application in case law. By examining the standards applied by both arbitral tribunals during the arbitral phase and oversight bodies during post-award proceedings, this research seeks to derive a ‘Differentiated Standard of Review Application’ Model. This proposed model intends to provide a practical solution to balance investor protection with state regulatory autonomy and to enhance the consistency of the ISDS regime without waiting for long-term structural reforms like an appellate mechanism.
This study focuses on the standard of review regarding the allocation of power and deference between the decision-maker and the reviewer, specifically within ISDS cases filed under major rules such as the ICSID Convention and UNCITRAL Arbitration Rules. The scope of analysis extends beyond the arbitral award stage to include the standard of review applied in post-award remedies, such as annulment or set-aside proceedings by domestic courts at the seat of arbitration. To achieve this, the study conducts a comprehensive analysis of primary sources, including arbitral awards and decisions, as well as secondary academic literature.
The dissertation is structured to first establish the theoretical basis and typologies of the standard of review in Chapter 2, followed by an examination of tribunal practices regarding state measures in Chapter 3. Chapter 4 analyzes the standards applied by courts and committees during post-award reviews. Based on these analyses, Chapter 5 synthesizes the factors considered by adjudicators to propose a structured standard of review applicable to different functional phases, and Chapter 6 concludes by presenting substantive directions for establishing a consistent standard of review to restore trust in the international investment law system.

2. Understanding the Standard of Review in International Investment Dispute Settlement Proceedings

The ‘standard of review’ functions as a fundamental mechanism for coordinating relationships between various authority holders within a public law order, primarily by determining the ‘allocation of power’. In the context of international investment dispute settlement (ISDS), this concept plays a pivotal role in establishing the boundaries of authority between arbitral tribunals, sovereign states, and domestic courts. As international investment law governs disputes arising from state regulatory measures that impact public interests, the choice of review standard is critical for balancing the state's regulatory autonomy with the protection of foreign investors.
To determine the appropriate standard of review, one must first comprehend the ‘hybrid’ nature of international investment law, which blurs the traditional distinction between public and private law. Structurally, ISDS mimics international commercial arbitration, utilizing ad hoc tribunals and rules like the UNCITRAL Arbitration Rules. Functionally, however, it adjudicates public disputes involving sovereign acts, legislation, and administrative measures based on international treaties. This duality creates a tension: if the ‘private’ aspect is emphasized, tribunals act as independent adjudicators with broad powers to interpret contracts; if the ‘public’ aspect is prioritized, the principle of subsidiarity suggests they should grant deference to the democratic decision-making of states. This hybridity necessitates a careful allocation of power. Tribunals possess the authority to determine their own jurisdiction (Kompetenz-Kompetenz) and adjudicate substantive treaty breaches. Conversely, domestic courts play a supervisory role, primarily in the enforcement and set-aside of awards. The standard of review thus serves as the gauge for how much the international tribunal should respect the primary determinations of the state, and subsequently, how much the domestic court should respect the findings of the tribunal.
Conceptually, the standard of review is defined as the intensity of scrutiny or the degree of deference a reviewing body applies to a decision. ‘Deference’ refers to the attitude of judicial restraint adopted by the reviewer, which is inversely related to the intensity of review. The standard exists on a continuous spectrum. At one extreme lies ‘de novo review’ (strict scrutiny), where the tribunal re-evaluates the facts and law entirely from scratch, substituting its judgment for that of the state. At the other extreme is ‘total deference’, where the tribunal accepts the state's decision without substantive review, barring bad faith. Between these poles lies ‘intermediate review’, often manifested as a reasonableness test. Crucially, the standard is bifurcated based on the phase of proceedings. The ‘substantive standard of review’ is applied by the tribunal during the arbitral phase to assess whether a state measure violates treaty obligations like Fair and Equitable Treatment (FET). The ‘procedural standard of review’ is employed by oversight bodies during the post-award phase to review the validity of the award itself, focusing on procedural legitimacy rather than substantive correctness. While most treaties are silent on the standard, some modern agreements like CETA explicitly mandate deference regarding domestic law interpretations, and international norms like the WTO's standard of review provide comparative models.
The standard of review is operationalized through specific ‘methods of review’ derived from comparative public law. ‘Proportionality analysis’ is a rigorous method that examines whether a measure is suitable, necessary (least restrictive means), and proportionate strictly sensu. It typically represents a strict standard of review, allowing tribunals to deeply scrutinize the state's policy choices. In contrast, the ‘reasonableness test’ serves as a more lenient method, inquiring whether the state’s action has a rational basis and is not arbitrary. It respects the state's policy discretion as long as the decision falls within a range of reason. The ‘margin of appreciation’ doctrine, originating from the European Court of Human Rights, is another tool for deference, granting states a breadth of discretion in fulfilling obligations, particularly in sensitive areas like national security or public morals. Finally, ‘procedural review’ focuses on the fairness and transparency of the decision-making process-such as due process and the right to be heard—rather than the substantive correctness of the outcome. By verifying procedural propriety, tribunals can respect the state's epistemic expertise while effectively controlling arbitrary exercises of power. Ultimately, the standard of review is not merely a technical rule but a constitutional device that determines the depth of international supervision over sovereign acts, balancing the need for control with the imperative of respect for regulatory autonomy.
The standard of review, deference, and method of review are interconnected. The level of deference determines the standard, which in turn dictates the appropriate method. For instance, a high level of deference (lenient standard) is typically implemented through methods like ‘Procedural Review’ or ‘Margin of Appreciation’, where the tribunal avoids second-guessing the state's policy choices. Conversely, a low level of deference (strict standard) calls for ‘Proportionality Analysis’, where the tribunal rigorously examines the necessity and balance of the measure. ‘Reasonableness’ sits in the middle but leans towards deference. However, in practice, the application is fluid; ‘reasonableness’ can be applied strictly (New Reasonableness), and ‘procedural review’ can be rigorous regarding due process. This complex relationship necessitates a careful empirical analysis of case law to understand how these theoretical constructs are applied in reality.

3. Standards of Review Applied by International Investment Tribunals to Domestic Measures

In the Investor-State Dispute Settlement (ISDS) regime, the arbitral tribunal plays a pivotal role in determining whether a host state’s legislative, administrative, or judicial measures comply with its obligations under International Investment Agreements (IIAs). While tribunals rigorously adhere to explicit treaty provisions, they have often been compelled to establish de facto standards of review in areas where treaties are silent. This interpretive process effectively concretizes the normative framework of international investment law. Although the authority to set such standards is not expressly granted by treaty texts, the standard of review functions in practice as a critical mechanism for balancing the host state's regulatory autonomy with the protection of foreign investors. This balance is particularly crucial in cases involving sensitive public policy issues, scientific expertise, emergency measures, and the interpretation of domestic law. However, the history of ISDS adjudication reveals a lack of consistency in the application of these standards, with tribunals often reaching divergent conclusions in cases involving similar facts and identical treaty obligations. Such inconsistency has raised concerns regarding the legal stability, predictability, and overall institutional legitimacy of the ISDS system. Against this backdrop, this chapter conducts a systematic analysis of major arbitral awards to classify the standards of review applied across different issues and to identify the key determinants influencing these judicial choices.

3.1. Factors Considered in Selecting the Standard of Review

The determination of the applicable standard of review in ISDS proceedings begins with the text of the relevant treaty. Consequently, tribunals must first derive the appropriate standard through the interpretation and application of relevant provisions. However, since the vast majority of IIAs do not explicitly prescribe a standard of review, the decision effectively falls within the tribunal's discretion. To mitigate the risk of arbitrary judgments, tribunals resort to extraneous considerations, weighing factors such as the allocation of power between the tribunal and the regulatory authority, information asymmetries, and relative professional competence. Like other international treaties, IIAs must be interpreted in accordance with customary international law as reflected in the Vienna Convention on the Law of Treaties (VCLT). Article 31 of the VCLT mandates interpretation in good faith according to the ordinary meaning of terms in their context and in light of the treaty's object and purpose. Yet, the inherent ambiguity of treaty texts—products of complex diplomatic compromises—often sparks interpretative disputes, such as the longstanding debate over the scope of the Minimum Standard of Treatment versus Fair and Equitable Treatment. In resolving these disputes, the determination of how much deference to accord to the host state's measures constitutes the core of applying the standard of review.
Given the rarity of explicit treaty provisions on standards of review, tribunals comprehensively evaluate extraneous factors reflecting the specificities of each case. This study identifies three decisive variables derived from fundamental questions regarding the tribunal's role: what the tribunal can know (expertise), what it has the authority to judge (regulatory power), and how it should control the measure (method of control). First, scientific complexity and expertise act as primary variables. In disputes involving environmental or public health issues that demand high-level scientific knowledge, tribunals face the dilemma of whether to respect the regulatory authority's expert judgment through epistemic deference or to conduct a de novo review of the scientific evidence. Second, economic and social crises and the public interest significantly influence the standard. In situations of national emergency, such as financial collapses or energy shortages, the degree of discretion allowed to the state becomes a central point of contention. The divergent outcomes in the Argentine financial crisis cases illustrate how the tribunal's assessment of ‘public interest’ and ‘necessity’ directly shapes the standard of review. Third, the procedural legitimacy of administrative and judicial acts serves as a crucial determinant. Tribunals increasingly shift the focus of their review from substantive outcomes to the decision-making process itself. If a state measure lacks transparency, consistency, or due process, tribunals are more inclined to apply a stricter standard of review to correct the procedural deficit.

3.2. Application Examples Based on Major Factors

The application of standards of review in ISDS is most clearly observed in cases involving public policy, expertise, and regulatory authority. State regulations concerning environmental protection and public health frequently conflict with investor rights, and because these regulations often rest on complex scientific knowledge or sensitive policy choices, tribunals must define the scope of deference owed to the regulatory authority. In modern society, states are duty-bound to take preemptive regulatory measures against scientific uncertainties such as environmental pollution and public health crises. Recognizing their comparative lack of scientific expertise, tribunals have developed the doctrine of ‘epistemic deference’. This doctrine posits that if a state measure is supported by rational scientific evidence or international standards, the tribunal should refrain from substantive re-evaluation and instead grant a broad margin of appreciation.
The case of Methanex v. USA is illustrative of this approach. The tribunal adopted a ‘reasonableness review’, explicitly rejecting a de novo review of scientific evidence. It held that because California's ban on the gasoline additive MTBE was based on a peer-reviewed report by the University of California and a transparent public process, it constituted a rational exercise of police powers, notwithstanding ongoing scientific debate. Similarly, in Chemtura v. Canada, the tribunal stated it would not second-guess the specialized determinations of Canada's Pest Management Regulatory Agency regarding the pesticide Lindane. Instead, it applied a ‘procedural review’, focusing on whether the agency had followed proper scientific review procedures and provided the investor with an opportunity to be heard. The Philip Morris v. Uruguay tribunal further solidified this trend by showing high deference to Uruguay's tobacco packaging regulations. Noting that the measures were implemented in accordance with the WHO Framework Convention on Tobacco Control (FCTC), the tribunal applied a ‘flexible proportionality’ test, requiring only a ‘rational connection’ between the measure and the public health objective rather than searching for the ‘least restrictive means’.
State measures for environmental protection are often justified under the ‘Police Powers’ doctrine, which holds that non-discriminatory, bona fide regulations for public purposes do not constitute compensable expropriation. However, recent jurisprudence reveals a trend of ‘bifurcation of liability’. While tribunals may exempt states from expropriation claims under the police powers doctrine, they may still find violations of the Fair and Equitable Treatment (FET) standard if the regulatory process was arbitrary or unfair. In both Infinito Gold v. Costa Rica and Eco Oro v. Colombia, tribunals recognized that bans on mining for environmental protection were legitimate exercises of police powers and did not constitute indirect expropriation. Yet, in Eco Oro, the tribunal found an FET violation due to administrative inconsistencies and prolonged delays in defining the protected area's boundaries, which caused ‘damage without purpose’ to the investor. This nuance demonstrates that while substantive regulatory goals are respected, administrative conduct is strictly scrutinized for procedural legitimacy. Conversely, tribunals strictly scrutinize measures that use environmental protection as a pretext for protectionism. In S.D. Myers v. Canada, while acknowledging the high measure of deference owed to public policy regulations, the tribunal applied a ‘reasonableness test’ to find that Canada's ban on PCB waste exports was actually motivated by a desire to protect the domestic industry, thereby violating National Treatment and FET standards due to its discriminatory nature and lack of scientific basis.
The balance between regulatory power and investor protection is also dynamic in the context of economic and social environmental changes. Tribunals apply differing standards depending on whether the situation involves an ‘extreme crisis’ threatening the state's existence or merely ‘normal’ policy changes. The Argentine financial crisis of 2001-2002 triggered a wave of cases centering on the ‘state of necessity’ defense. Early tribunals in cases like CMS, Enron, and Sempra applied strict scrutiny, equating the treaty's necessity clause with the customary international law standard under ILC Article 25. They ruled that because theoretical alternatives such as dollarization or spending cuts existed, the measures were not the ‘only way’ to resolve the crisis. This approach was widely criticized for hindsight bias and for substituting the tribunal's judgment for that of policymakers. In contrast, the Continental Casualty and LG&E tribunals adopted a more deferential approach. The Continental Casualty tribunal treated the necessity clause as a specific treaty exception, applying a ‘reasonableness’ test rather than the strict ‘only way’ requirement, while the LG&E tribunal recognized a period of necessity where the state's liability was exempted, showing deference to the state's crisis management.
In non-crisis situations involving fiscal or financial stability, tribunals generally grant states broad discretion, rejecting the notion that investors have a right to a frozen legal framework. In cases like Paushok v. Mongolia and Saluka v. Czech Republic, tribunals held that legislative and administrative judgments regarding taxes or banking regulations should be respected unless they are ‘confiscatory’ or ‘manifestly arbitrary’. However, the renewable energy cases in Spain illustrate a fragmented landscape. While the Charanne tribunal applied a deferential standard finding no violation absent a stabilization clause, the Eiser tribunal applied strict scrutiny to the ‘effect’ of the radical regulatory change, finding a violation. The RREEF tribunal adopted an intermediate standard based on the ‘reasonable rate of return’ concept, balancing regulatory power with economic reality.
When reviewing technical administrative regulations or final judicial decisions, tribunals tend to avoid de novo review of substantive outcomes and instead focus on ‘procedural review’, examining transparency, consistency, due process, and the absence of arbitrariness. Tribunals do not act as courts of appeal for domestic judicial decisions. In cases like Mondev v. USA and Eli Lilly v. Canada, tribunals emphasized that a domestic judgment constitutes a treaty violation only if it amounts to a ‘denial of justice’ or is egregiously unjust, reflecting a high level of deference to judicial sovereignty. For administrative acts, however, scrutiny of the decision-making process is more intense. In Tecmed v. Mexico, the tribunal applied a strict ‘proportionality analysis’, finding that the revocation of a license due to community pressure was disproportionate. Similarly, in Crystallex v. Venezuela, the tribunal found an FET violation by identifying procedural defects, noting that the permit denial was based on a flawed report lacking scientific data and was driven by political directives rather than technical criteria. Furthermore, when administrative discretion is abused for political reasons or corrupted, deference is withdrawn entirely. In Elliott v. Korea, the tribunal found the National Pension Service's vote on a merger was manipulated by political pressure, and in Lone Star v. Korea, the tribunal ruled that the financial regulator's delay was motivated by a desire to avoid political criticism. These cases demonstrate that while policy discretion is respected, political interference in administrative processes triggers strict scrutiny.

3.3. Determinants and Application of the Standard of Review

While the primary basis for the standard of review remains the treaty text, the silence of most IIAs leads tribunals to rely heavily on extraneous factors. This chapter has identified three key determinants: scientific complexity and expertise, emergency and public interest, and procedural legitimacy. The application of these factors reveals a consistent pattern. First, in areas of scientific uncertainty such as environment and health, tribunals generally demonstrate ‘epistemic deference’ to the specialized knowledge of national authorities, provided the process is rational and non-discriminatory. Second, in the context of economic crises, there has been a shift from strict scrutiny to a more deferential ‘margin of appreciation’, recognizing the state's primary role in crisis management. Third, regarding administrative and judicial acts, tribunals apply a ‘bifurcated’ standard: they defer to substantive outcomes but exercise ‘strict control’ over procedural legitimacy, transparency, and consistency. To implement these standards, tribunals employ various methods. ‘Proportionality analysis’ allows for balancing interests but can be intrusive; ‘reasonableness review’ checks for a rational connection and is more deferential; ‘procedural review’ focuses on the fairness of the process rather than the result; and the ‘margin of appreciation’ grants states breathing room in sensitive policy areas. In conclusion, while ISDS lacks a formal doctrine of precedent, a functional jurisprudence is emerging where deference is accorded to substantive public policy and scientific determinations, while strict scrutiny is applied to procedural fairness and jurisdiction. This trend underscores the necessity of establishing a consistent and predictable standard of review within the current system to address the legitimacy crisis.

4. Post-Award Phase and Standard of Review in International Investment Arbitration

While the previous chapter examined the ‘substantive standard of review’ applied by arbitral tribunals to the legislative, administrative, and judicial measures of host states, Chapter 4 shifts the analytical focus to the ‘procedural standard of review’ employed during the post-award phase. This stage involves the review of arbitral awards by oversight bodies, specifically domestic courts acting as the seat of arbitration or ad hoc Annulment Committees under the ICSID Convention. Unlike appellate review, which allows for a comprehensive re-examination of the merits, these post-award mechanisms are restricted to reviewing limited grounds such as procedural defects or jurisdictional excesses as stipulated in relevant statutes or treaties. However, even within this limited scope, significant controversy remains regarding the appropriate intensity of review. A distinct dichotomy emerges depending on whether the subject of review is ‘jurisdiction’ or the ‘merits’ of the dispute. Regarding jurisdictional issues, many courts tend to conduct a de novo review to ensure legal legitimacy, whereas for the merits, they prioritize the finality of the award by respecting the tribunal's judgment. This chapter aims to clarify this spectrum of review standards to enhance the predictability of the Investor-State Dispute Settlement (ISDS) system and proposes a balanced approach that harmonizes judicial restraint with necessary control.

4.1. Post-Award Remedies in International Investment Arbitration

The mechanisms for challenging an investment arbitral award are bifurcated into two primary regimes: the ICSID Convention system and the non-ICSID system (governed by the ICSID Additional Facility Rules, UNCITRAL Rules, etc.). This structural distinction is crucial because it determines whether the oversight is internal or external. Under the ICSID Convention, the review process is a self-contained system where domestic courts have no power to review or annul awards; instead, an internal ad hoc Annulment Committee reviews awards based exclusively on the narrow grounds listed in Article 52 of the Convention, such as manifest excess of powers or serious departure from a fundamental rule of procedure. This design limits the review to legitimacy and procedural regularity, explicitly excluding a review of substantive correctness to protect the finality of the award.
In contrast, non-ICSID awards are subject to the supervisory jurisdiction of the national courts at the seat of arbitration. These proceedings are governed by the arbitration laws of the seat, which are often modeled after the UNCITRAL Model Law or the New York Convention. Consequently, non-ICSID awards face ‘external control’ by domestic judges who may annul awards based on grounds such as invalid arbitration agreements, procedural irregularities, or violations of public policy. The ‘public policy’ ground, in particular, opens a door for courts to potentially review the substantive content of an award if it conflicts with fundamental legal principles, creating a variable standard of review depending on the jurisdiction.

4.2. Standard of Review for Jurisdictional Determinations

The review of jurisdictional determinations represents the area where the tension between arbitral autonomy and judicial control is most acute. Although arbitral tribunals have the authority to rule on their own jurisdiction under the Kompetenz-Kompetenz principle, this authority is generally viewed by courts as provisional rather than final. Major common law jurisdictions, including the United Kingdom and Singapore, have established a clear precedent of applying a ‘strict standard of review’, often characterized as de novo review or a standard of ‘correctness’ regarding jurisdictional challenges.
The United Kingdom Supreme Court's decision in Dallah v. Pakistan is a paradigmatic example. In this case, the court refused to enforce an award against the Pakistani government, conducting a full re-hearing of the evidence and foreign law to determine that there was no valid arbitration agreement. The court explicitly rejected the notion of deference to the tribunal's jurisdictional findings, asserting that the tribunal's decision carries no special weight and that the court must independently determine the ‘correct’ answer regarding the existence of jurisdiction. Similarly, the Singapore Court of Appeal in Sanum v. Laos affirmed that jurisdictional review is a de novo hearing, not an appeal, and that the court is the final arbiter of whether state consent exists, rejecting arguments that deference should be owed to the tribunal's expertise in international law. This approach treats jurisdiction as a matter of ‘legal fact’ that must be objectively established.
Conversely, the United States Supreme Court's decision in BG Group v. Argentina illustrates a more deferential approach rooted in a pro-arbitration policy. The Court distinguished between substantive arbitrability (which courts review independently) and procedural gateway issues (which are presumed to be for arbitrators to decide). By categorizing a treaty's local litigation requirement as a procedural precondition rather than a jurisdictional one, the Court applied a highly deferential standard, upholding the tribunal's decision to excuse non-compliance. However, this approach has faced criticism for potentially conflating admissibility with jurisdiction and expanding state consent beyond its intended limits.

4.3. Standard of Review for Substantive Findings and Procedural Defects

While jurisdictional review often invites strict scrutiny, the review of the ‘merits’ (substantive findings of fact and law) is governed by a principle of strict non-intervention or deference. ICSID Annulment Committees have consistently held that their role is not to correct errors of law or fact. As clarified in MCI v. Ecuador, a distinction must be drawn between the ‘failure to apply the proper law’ (a ground for annulment) and the ‘misapplication of the law’ (which is not). Even if a tribunal's legal interpretation is arguably incorrect, it does not constitute a manifest excess of powers as long as the tribunal attempted to apply the correct governing law. This high threshold is designed to prevent annulment proceedings from turning into disguised appeals.
However, recent jurisprudence indicates a shift towards a ‘zero tolerance’ approach regarding ‘procedural integrity’ and ‘arbitrator ethics’. Unlike substantive errors, violations of fundamental due process or arbitrator impartiality are increasingly viewed as fatal defects that warrant annulment regardless of their impact on the final outcome. The 2025 annulment decision in Lone Star v. Korea serves as a critical example. The Committee annulled the award because the tribunal had relied on evidence from a separate commercial arbitration to which the Korean government was not a party, thereby denying the state its fundamental right to be heard. The Committee ruled that such a breach of due process constituted a serious departure from a fundamental rule of procedure, emphasizing that a conclusion reached through a flawed process cannot stand, even if the substantive outcome might have been correct.
Similarly, in Eiser v. Spain, an ICSID Committee annulled an award due to an undisclosed professional relationship between an arbitrator and the claimant's expert witness. The Committee established that the ‘appearance of bias’ alone-without proof of actual bias or material impact on the result—was sufficient to undermine the legitimacy of the proceedings. This reflects a rigorous standard where procedural justice is prioritized over finality. Domestic courts also exercise strict scrutiny over procedural fairness. In Commonwealth Coatings, the US Supreme Court held that arbitrators must avoid even the appearance of bias, and in Iran Aircraft v. Avco, enforcement was refused because the tribunal had misled a party regarding evidence submission, violating due process.

4.4. Establishing Issue-Based Bifurcated Standards

Synthesizing these findings, the current landscape of post-award review is defined not by the type of oversight body, but by the nature of the issue under review, leading to an ‘Issue-based Dichotomy’. For jurisdictional issues, courts in jurisdictions like the UK and Singapore apply a de novo standard to ensure the tribunal acted within the scope of state consent. For procedural justice issues, such as the right to be heard and arbitrator independence, oversight bodies apply strict scrutiny to safeguard the integrity of the system. Conversely, for substantive determinations on the merits, a standard of judicial deference or restraint is firmly maintained to preserve the finality of awards.
Although the decentralized nature of the non-ICSID system and the influence of the lex arbitri lead to some fragmentation, a convergence is observable regarding these core principles. To resolve the legitimacy crisis of ISDS, a consistent application of this bifurcated model is necessary. Oversight bodies must act as ‘gatekeepers’, rigorously policing the boundaries of jurisdiction and procedure while respecting the tribunal's mandate to decide the merits. This functional allocation of review standards offers a realistic pathway to enhancing the stability and legitimacy of the international investment law regime without requiring immediate structural reforms like a permanent investment court.

5. Conclusion and Suggestions

This study identifies the root cause of the ‘legitimacy crisis’ facing the Investor-State Dispute Settlement (ISDS) regime as the failure to establish a consistent ‘standard of review’. As ISDS has expanded beyond investment protection into sensitive areas of national public policy—such as environmental regulation, public health, and crisis management—arbitral tribunals have struggled to reconcile state regulatory sovereignty with investor property rights. The resulting inconsistency in arbitral awards has undermined the legal stability and predictability of the system. To address this, this dissertation comprehensively analyzed the ‘arbitral phase’ (tribunal review of state measures) and the ‘post-award phase’ (oversight review of awards) to propose a normative direction for the future.
Chapter 3 empirically analyzed the arbitral phase, revealing that tribunals oscillate between ‘strict scrutiny’ and ‘deferential review’. In cases like CMS and Enron regarding the Argentine financial crisis, tribunals applied strict proportionality, effectively substituting their judgment for the state's policy choices through de novo review. While this strengthened investor protection, it faced criticism for infringing upon democratic policy-making. Conversely, cases like Methanex, Chemtura, and Philip Morris demonstrated ‘epistemic deference’, where tribunals respected state decisions based on scientific expertise and rational administrative procedures. This analysis confirms that the substantive standard of review functions as a constitutional mechanism defining the scope of arbitral authority.
Chapter 4 analyzed the post-award phase, finding that review standards are converging toward an ‘issue-based dichotomy’. Regarding ‘jurisdiction’, courts in the UK (Dallah) and Singapore (Sanum) have firmly established a standard of ‘strict scrutiny’ (or de novo review), treating jurisdiction as a legal fact to be independently determined by the court rather than a matter of arbitral discretion. Regarding the ‘merits’, however, oversight bodies maintain strict ‘judicial deference’, refusing to annul awards for simple errors of law or fact. Notably, recent jurisprudence (Lone Star, Eiser) demonstrates a ‘zero tolerance’ approach to ‘procedural justice’ and ethics; violations of the right to be heard or arbitrator independence warrant annulment regardless of the substantive outcome.
Based on this analysis, this study proposes the ‘Differentiated Standard of Review Application’ Model. This model bifurcates the object of review into ‘substance’ and ‘procedure/jurisdiction’, applying distinct standards suited to the functional nature of each phase.
First, regarding substantive issues in the arbitral phase, the governing principle must be ‘deference’. Arbitral tribunals lack the democratic legitimacy and technical expertise to second-guess complex public policy choices. Therefore, they should respect national regulatory judgments unless the measure constitutes manifest arbitrariness or unjustified discrimination. To achieve this, the study argues for replacing the ‘strict proportionality’ test (specifically the intrusive ‘least restrictive means’ requirement) with a ‘procedural reasonableness’ test. This approach limits the tribunal's role to verifying whether a rational connection exists between the policy objective and the measure, rather than seeking the ‘optimal’ policy solution.
Second, regarding jurisdiction and procedural fairness in the post-award phase, oversight bodies must apply ‘strict scrutiny’. Because an arbitral tribunal's authority is derived solely from state consent, the scope of that authority cannot be left to the arbitrators’ discretion. Furthermore, procedural integrity is the bedrock of the system’s legitimacy. Therefore, domestic courts and Annulment Committees must act as rigorous ‘gatekeepers’, applying de novo review to jurisdictional challenges and strictly sanctioning procedural defects to ensure the system operates within the bounds of the rule of law.
To implement this model within the current international legal order, a two-track strategy is required. The hard law approach involves legislative control through treaty modification. Long-term efforts should focus on codifying the standard of review in International Investment Agreements (IIAs), similar to CETA Article 8.9. Specifically, treaties should include a ‘Standard of Review Clause’ explicitly stating that reasonable regulatory measures for public welfare do not violate treaty obligations unless manifest arbitrariness is proven. The soft law approach focuses on interpretative guidance. Given the difficulty of amending thousands of existing treaties, immediate stability can be achieved through ‘Joint Interpretive Notes’ issued by contracting states to regain control over treaty interpretation. Additionally, international bodies like UNCITRAL or ICSID should develop ‘Practice Guidelines on Standards of Review’ to provide arbitrators with authoritative, de facto binding checklists for applying deference.
The standard of review is not merely a technical tool but a constitutional mechanism that balances state sovereignty and capital protection. The current crisis stems from a structural failure to calibrate this balance. The proposed ‘Differentiated Standard of Review Application’ Model restores equilibrium by securing ‘policy flexibility’ through deference on merits and ensuring ‘legal stability’ through strict control over procedure and jurisdiction. Rather than waiting for distant structural reforms like a permanent investment court, the international community must focus on establishing consistent review standards within the current framework. When tribunals recognize their institutional limits and oversight bodies strictly safeguard procedural justice, the ISDS regime can shed the stigma of ‘unchecked power’ and function as a legitimate pillar of the global economic order.

목차 (Table of Contents)

• 그림 차례 ⅳ
• 표 차례 ⅴ
• 국문 요약 ⅵ
• 제1장 서론 1
• 제1절 연구의 배경 및 목적 1
• 제2절 연구의 범위와 방법 8
• 제2장 국제투자분쟁해결절차상 심사기준에 대한 이해 14
• 제1절 국제투자중재의 성격과 심사기준 15
• 1. 국제법상 공법적 요소와 사법적 요소 16
• 2. 국제투자중재의 혼합적 성격 17
• 3. 국제투자중재에 있어 권한배분과 심사기준 21
• 제2절 심사기준의 개념 및 유형 29
• 1. 심사기준의 정의 29
• 2. 심사기준의 유형 및 규정례 31
• 제3절 실체적 심사기준과 심사방법 38
• 1. 심사방법의 유형 39
• 2. 심사기준과 존중, 그리고 심사방법의 관계 48
• 제3장 국제투자중재판정부의 국내조치에 대한 심사기준 52
• 제1절 심사기준 선택 시 고려요소 53
• 제2절 주요 고려요소에 따른 심사기준 적용례 59
• 1. 국가의 공공정책에 대한 전문성과 규제권한 59
• 2. 경제·사회적 환경 변화와 규제권한 71
• 3. 행정 및 사법 작용의 절차적 정당성 81
• 제3절 심사기준의 결정 요인과 적용 88
• 제4장 국제투자중재판정에 대한 사후적 통제와 심사기준 93
• 제1절 국제투자중재판정에 대한 사후구제수단 95
• 1. 국제투자중재규칙별 취소 및 무효화의 체계 96
• 2. 주요 쟁점에 따른 심사기준의 이원화 101
• 제2절 관할권 판단에 대한 심사기준 103
• 1. ICSID 체제의 ‘명백성 요건’에 따른 제한적 심사기준 104
• 2. 비-ICSID 체제의 엄격한 심사기준 적용 기조 106
• 제3절 실체적 판단 및 절차적 정의에 대한 심사기준 118
• 1. 실체적 판단에 대한 심사기준 119
• 2. 절차적 정의와 윤리 규정에 대한 심사기준 123
• 3. 판정 이유 불기재와 공공질서 위반 131
• 제4절 쟁점별 상이한 심사기준의 정립 133
• 제5장 국제투자분쟁해결절차상 심사기준의 정립 방안 138
• 제1절 심사기준 정립의 현실적 제약과 대응 방향 139
• 1. 심사기준 정립의 필요성 140
• 2. 구조적 개혁의 이상과 현실적 한계 141
• 3. 현실적 대안으로서의 ‘단계별 심사기준 적용’ 모델 145
• 제2절 단계별 심사기준 적용 모델에 따른 실체적 심사기준 147
• 1. 심사방법의 전환: ‘엄격한 비례성’에서 ‘절차적 합리성’으로 149
• 2. 과학적 증거에 대한 존중 153
• 3. 적법절차 중심의 실체 심사 전환 155
• 제3절 단계별 심사기준 적용 모델에 따른 절차적 심사기준 157
• 1. 관할권에 대한 심사기준 159
• 2. 적법절차에 대한 심사기준 161
• 3. 법적·사실적 오류에 대한 심사기준 163
• 제4절 현행 체제 내에서의 실질적 구현 방안 165
• 1. 조약 문언의 구체화 166
• 2. 공동해석선언의 활성화 168
• 3. 국제기구 차원의 ‘심사기준 실무 지침’ 제정 171
• 제6장 결론 173
• 참고 문헌 178
• 영문 요약 194