Data Set for the article Climent, Momparler, Carmona (2025). Predicting mutual fund performance with machine learning: Are ESG pillar scores relevant predictors of fund return?, Borsa Istanbul Review, 25, 1403-1419
Climent, Momparler, Carmona (2025)の論文用データセット:機械学習による投資信託パフォーマンス予測-ESGピラースコアはファンドリターンの予測因子か? (AI 翻訳)
Francisco Climent, Pedro Carmona, Alexandre Momparler
🤖 gxceed AI 要約
日本語
本データセットは、欧州のグローバル株式型リテール投資信託を対象に、2020年から2024年までの5年間のデータをRefinitiv Eikonから収集したものである。機械学習を用いてESGスコア(環境・社会・ガバナンス)がファンドのリターン予測に有用かどうかを検証するための変数(リターン、ESGスコア、TER、アルファ、ベータ等)を含む。
English
This dataset comprises European global equity retail mutual funds with five-year performance records (2020-2024) from Refinitiv Eikon. It includes variables such as fund return, ESG scores (environmental, social, governance), expense ratio, alpha, beta, and style matrix, designed to test whether ESG pillar scores predict fund returns using machine learning.
Unofficial AI-generated summary based on the public title and abstract. Not an official translation.
📝 gxceed 編集解説 — Why this matters
日本のGX文脈において
日本ではSSBJ開示が進む中、ESG情報の投資有用性への関心が高い。本データセットは欧州ファンドを対象とするが、機械学習によるESGスコアの予測力検証手法は、日本の運用機関や企業の開示実務に示唆を与える。
In the global GX context
Globally, this dataset supports research on the predictive power of ESG scores in fund performance, relevant to sustainable finance and disclosure frameworks like SFDR and CSRD. It provides a transparent, reproducible data foundation for ML-based ESG analysis.
👥 読者別の含意
🔬研究者:ESGスコアとファンドリターンの関係を機械学習で検証するためのデータセットとして利用可能。
🏢実務担当者:ESGデータを運用戦略に組み込む際の予測可能性の参考になる。
📄 Abstract(原文)
The data was collected in 2024 from the Refinitiv Eikon Database (https://www.refinitiv.com). We selected European mutual equity funds with a global geographic scope, focusing on actively managed share-based assets in euros, with continuous five-year performance records (january 1, 2020, to December 31, 2024), and a minimum investment requirement of up to €10,000, in order to concentrate on retail funds and exclude institutional funds. The variables used in the analysis are defined as follows: - Fund Return (ANNUAL RETURN): The annual total return of the fund, including both capital appreciation and dividends, calculated on a calendar-year basis. - ESG Score (ESGSCORE): A composite measure reecting the fund’s overall performance on Environmental, Social, and Governance criteria, based on peer-relative rankings. - Environmental Score (ENVSCORE): The fund’s score on environmental factors, such as emissions, resource use, and environmental management practices. - Social Score (SOCSCORE): The fund’s score on social factors, including community impact, human rights, product responsibility, and workforce practices. - Governance Score (GOVSCORE): The fund’s score on governance factors, such as board structure, transparency, shareholder rights, and executive compensation. - Total Expense Ratio (TER): The annual cost of managing the fund, expressed as a percentage of assets under management, covering management fees and other operating expenses. - Fund Total Net Assets (FUNDTNA): The total net value of the fund’s assets, calculated as total assets minus total liabilities, measured in thousands of euros. - Five-ear Annualized Standard Deviation of Returns (ANNUALSD5): A measure of the fund’s return volatility over the past five years, indicating the degree of variability in performance. - Alpha (ALPA): A risk-adusted performance metric showing the difference between the fund’s actual returns and its expected returns, given its level of market risk (beta). - Beta (BETA): A measure of the fund’s sensitivity to market movements, indicating how much the fund’s returns move in relation to the overall market. - Style Matrix (STLEMATRIX): A classification of the fund’s investment style and size, based on the Lipper Style Box, which categorizes funds by market capitalization and investment approach (e.g., value, growth, or core).
🔗 Provenance — このレコードを発見したソース
- openalex https://doi.org/10.5281/zenodo.21674825first seen 2026-07-31 06:16:52 · last seen 2026-07-31 06:16:54
🔔 こうした論文の新着を逃したくない方は キーワードアラート に登録(無料・3キーワードまで)。
gxceed は公開メタデータに基づく研究支援データセットです。要約・翻訳・解説は AI 支援で生成されています。 最終的な解釈・検証は利用者が原典資料に基づいて行うことを前提とします。