Global Health Research Center of Central Asia
Columbia Law School Scholarship ArchiveNot a member yet
5770 research outputs found
Sort by
Orthogonalizing Inputs
Get PDFThis paper examines an approach to algorithmic discrimination that seeks to blind predictions to protected characteristics by orthogonalizing inputs. The approach uses protected characteristics (such as race or sex) during the training phase of a model but masks these during deployment. The approach posits that including these characteristics in training prevents correlated features from acting as proxies, while assigning uniform values to them at deployment ensures decisions do not vary by group status.
Using a prediction exercise of loan defaults basedon mortgage HMDA data and German credit data, the paper highlights the limitations of this orthogonalization strategy. Applying a lasso model, it demonstrates that the selection and weights on protected characteristics are inconsistent. At the deployment stage, where uniform values for race or sex are given to the model, the variations between models lead to meaningful differences in outcomes and resultant disparities.
The core challenge is that orthogonalization assumes an accurate model estimation of the relationship between protected characteristics and outcomes, which can be isolated and neutralized during deployment. In reality, when correlations are pervasive and predictions are constrained by regularization, feature selection can be unstable and driven by the efficiency of the prediction. This analysis casts doubt on the continued reliance on input scrutiny as a strategy in discrimination law and cautions against the myth of algorithmic colorblindness
Reforming International Investment Law to Advance Tax Justice
Get PDFReforming International Investment Law to Advance Tax Justice highlights the detrimental impact of current international investment treaties on tax justice and sustainable development objectives. It argues that Investor-State Dispute Settlement (ISDS) mechanisms often impede states\u27 ability to implement effective tax policies by allowing foreign investors to challenge tax measures. The brief recommends a comprehensive reform of international investment law to ensure that investment treaties support rather than undermine tax justice. This can include eliminating ISDS provisions, drafting new treaties that safeguard the sovereign right of states to regulate taxation, and facilitating cooperation among states to reform tax systems at national, regional, and international levels. Restructured investment agreements, combined with progressive tax policies, could significantly enhance tax justice, particularly for developing countries vulnerable to tax avoidance and evasion by multinational enterprises and wealthy individuals
D-Hacking
Get PDFRecent regulatory efforts, including Executive Order 14110 and the AI Bill of Rights, have focused on mitigating discrimination in AI systems through novel and traditional application of anti-discrimination laws. While these initiatives rightly emphasize fairness testing and mitigation, we argue that they pay insufficient attention to robust bias measurement and mitigation — and that without doing so, the frameworks cannot effectively achieve the goal of reducing discrimination in deployed AI models. This oversight is particularly concerning given the instability and brittleness of current algorithmic bias mitigation and fairness optimization methods, as highlighted by growing evidence in the algorithmic fairness literature. This instability heightens the risk of what we term discrimination-hacking or d-hacking, a scenario where, inadvertently or deliberately, the selection of models based on favorable fairness metrics within specific samples could lead to misleading or non-generalizable fairness performance. We term this effect d-hacking because systematically selecting among numerous models to find the least discriminatory one parallels the concept of p-hacking in social science research of selectively reporting outcomes that appear statistically significant resulting in misleading conclusions. In light of these challenges, we argue that AI fairness regulation should not only call for fairness measurement and bias mitigation, but also specify methods to ensure robust solutions to discrimination in AI systems. Towards the goal of arguing for robust fairness assessment and bias mitigation in AI regulation, this paper (1) synthesizes evidence of d-hacking in the computer science literature and provides experimental demonstrations of d-hacking, (2) analyzes current legal frameworks to understand the treatment of robust fairness and non-discriminatory behavior, both in recent AI regulation proposals and traditional U.S. discrimination law, and (3) outlines policy recommendations for preventing d-hacking in high-stakes domains
Distinguishing Family Poverty from Child Neglect
Get PDFFamily courts and child protective services (“CPS”) agencies surveil, regulate, and separate hundreds of thousands of families for neglect annually. These families are overwhelmingly poor, and the history of this legal system reveals an expectation, if not an intention, to intervene in poor families. This raises the question whether family courts and CPS agencies are “confusing poverty and neglect” or if they intervene for more than “just poverty,” as a raging debate in the field is framed.
The law fails to help resolve this debate. Instead of distinguishing poverty from neglect, or providing nuanced examinations of what social science has long shown to be a complex relationship between poverty and neglect, the law assumes away the problem. The law asserts that neglect and poverty are distinct, so the legal system’s decision to label parental behavior as neglectful frames the case as about some parental fault or pathology and not about poverty. Consistent with that frame, the law has separated anti-poverty financial supports from interventions available in neglect cases, so neglect interventions largely avoid providing such supports, even though much empirical evidence shows they can reduce family court and CPS system involvement. Moreover, family court and CPS agency intervention can trigger a variety of steps which make poor parents poorer, undermining their ability to reunify with their children.
This Article identifies a range of changes which would improve the legal system’s ability to distinguish poverty from neglect, by both eradicating long-standing legal rules which confuse poverty and neglect, and establishing more radical rules that would reverse the historical division between neglect cases and anti-poverty financial supports. These proposals recognize how deeply intertwined poverty and neglect are currently, and the absence of any easy test to determine which families could stay safely together if they were not poor and which could not. Absent such a test, the best solution is to provide families the income or the supports that would replicate the experience of families who are not poor
BU S4E2
No full textPhoto of Xuan “Sharon” Di, Associate Professor in the Department of Civil Engineering and Engineering Mechanics at Columbia University.https://scholarship.law.columbia.edu/beyond_unprecedented_podcast/1043/thumbnail.jp
Performing Legality: When and Why Chinese Government Leaders Show Up in Court
Get PDFSince 2015, Chinese government leaders have been required by law to appear in court when citizens sue their unit or to designate an employee to take their place. We frame this policy as a demand on leaders to “perform legality,” sacrificing their time to demonstrate how seriously the government takes legal proceedings. Drawing on an original dataset of 127,529 administrative lawsuits decided between 2015 and 2018, we investigate how often government leaders appear in Chinese courtrooms, and for which kinds of cases. Overall, leaders attended 24.72% of hearings. Contrary to the State Council’s instructions to prioritize attendance in lawsuits that may spiral into protest or draw public attention, leaders are no more likely to attend cases involving multiple plaintiffs or repeat litigants. Instead, they appear more often in cases where plaintiffs are represented by lawyers. These findings demonstrate that leaders tend to sidestep the most contentious cases in the legal system, complicating the conventional wisdom that stability maintenance is the overriding concern driving officials’ behavior. Yet despite leaders’ conflict avoidance, China has continued to ask leaders to personally appear in court in contentious cases. This policy is unusual in comparative perspective, and highlights the key role played by courts in General Secretary Xi Jinping’s drive to re-train the bureaucracy to take law more seriously
It Takes a Movement: Lawyers, Litigation, and Legal Theory in the Attack on Campaign Finance Reform
No full textRecent Supreme Court decisions have hobbled campaign finance regulation and placed significant constitutional constraints on possible future reforms. But the Supreme Court has not acted alone. Major constitutional changes often require a substantial legal and political infrastructure, an ecosystem of actors, ideas, strategies, and tactics to instantiate new constitutional doctrine into law. In Big Money Unleashed, Ann Southworth examines the composition, development, organization, and intellectual underpinning of the legal movement that has worked to cripple campaign finance regulation. She chronicles the raising and testing of the key legal theories; the sources of political and financial backing for the movement; the combination of disparate groups into an antireform coalition; and the organization of what has proven to be a very effective litigation strategy. The result, she explains, has been the ascendancy of a powerful constitutional discourse that treats money as speech, campaign spending as freedom, and government regulation as censorship. The antireform litigation campaign on its own would not have been enough to transform the law; key appointments to the Supreme Court were essential. But the movement provided the Court with the cases, the arguments, and broader political support that are part of constitutional chang
Legitimate Objectives in Antitrust Analysis: The FIFA Regulation of Agents and the Right to Regulate Football in Europe
No full textThe paper considers FIFA’s regulation of the market for agents, which has been challenged in a number of proceedings. Following the judgments of the Court on the ESL, ISU and Royal Antwerp cases, if the proposed regulation involves a restriction by object, it cannot fall outside the scope of Article 101 TFEU because of its pursuit of a legitimate objective. Assuming that the regulation would involve restrictions by effects, the question arises of which legitimate objective can be appealed to in order for its restrictions to remain outside the scope of Article 101 TFEU. A review of the case law on the scope of legitimate objectives suggests that, even if the Court has not pronounced it clearly on the matter, those accepted by the Court are related to the public policy objectives that the undertakings under consideration are entrusted to pursue. Such a restriction appears to be a welcome safeguard against a potential dilution of the discipline of Article 101 TFEU. From this perspective, legitimate objectives that could justify restrictions of competition by effects associated with FIFA’s regulation of the market for agents should be found within the regulatory function entrusted to FIFA. Whether the objectives put forward by FIFA in terms of addressing moral hazard, adverse selection and hold up in the (labor) market for agents can be cast in those terms is far from clear
Opposition to Renewable Energy Facilities in the United States: June 2024 Edition
Get PDFAchieving lower carbon emissions in the United States will require developing a massive number renewable energy facilities at an unprecedented scale and pace. Although many renewable energy facilities are sited without any problem, local opposition often arises. This report updates and considerably expands three previous Sabin Center reports, published in September 2021, March 2022, and May 2023, which document local and state restrictions against, and opposition to, siting renewable energy projects, as well as energy storage and transmission projects that are closely tied to renewable energy generation. The time period covered by this report ranges from as early as 1995 to December 31, 2023. For purposes of this report, the authors do not make normative judgments as to the legal merits of individual cases or the policy preferences reflected in local opponents’ advocacy, nor as to where any one facility should or should not be sited. Nonetheless, the volume and nature of the restrictions and controversies catalogued in this report demonstrate that local opposition to renewable energy facilities is widespread and growing and that it represents a potentially significant impediment to achievement of climate goals.
The report finds that local and state restrictions, as well as controversies over individual projects, are numerous and widespread. In nearly every state, local governments have enacted laws and regulations to block or restrict renewable energy facilities, or project opponents have succeeded in forcing the delay or cancellation of particular projects, or both have occurred. This edition identifies at least 395 local restrictions across 41 states, in addition to 19 state-level restrictions, that are so severe that they could have the effect of blocking a renewable energy project. This edition also identifies 378 renewable energy projects that have encountered significant opposition in 47 states. The 395 local restrictions, 19 state-level restrictions, and 378 contested projects catalogued in this report represent a major increase over the totals in the May 2023 edition
Do Private Actors Have Rights under the WTO? The Motivation for and (Inadequate) Implementation of GATT Article X
Get PDFGeneral Agreement on Tariffs and Trade (GATT) Article X has been consistently overlooked in literature even though it is a unique provision in the GATT-regime: it acknowledges standing for private actors who can challenge trade-related administrative action by World Trade Organization (WTO) members. This provision is an unusual instantiation of the need to provide transparency for private actors about state policy. It is equally eccentric in seeming to provide a right for private actors to challenge governments, albeit only before domestic fora. How did the GATT end up with this provision? And are these apparent private rights meaningful in practice? Since implementation of Article X rarely arises in WTO disputes, this paper proposes an answer to the second question based on six case studies of the United States of America (US), Canada, the European Union (EU), Brazil, China, and India. This is a heterogenous group that comprises big trading nations. We find wide variance in state practice. We conclude with suggestions that would strengthen the WTO demos by providing better information through WTO monitoring on the rights accorded to private actors and the introduction of an explicit code of good practice