A Random Forest Approach for Bounded Outcome Variables

Quantifying the nonlinear relationships between environmental policy components and share energy from renewable sources

The mechanisms and nonlinear relationships between different components of environmental policies and the share of energy derived from renewable sources remain poorly understood. In this study, a comprehensive analysis of 13 policy components across three categories of environmental policies, namely, market-based policies, non-market-based policies, and technology support policies, in 19 representative OECD countries from 2004 to 2020, was conducted. Random Forest models were employed to explore the relationships between these policy components and the share of energy from renewable sources. The study successfully identifies the nonlinear relationships, and highlights the varying impact of environmental policy components at different levels. Results indicate that the carbon dioxide tax, expenditure on low-carbon research and development, and the particulate matter emission limit are the most influential components in the model. Analysis of interaction effects across policy categories showed that the combination of the carbon dioxide tax and solar energy support yields the most remarkable synergistic effects on increasing the share of energy from renewable sources. Conversely, the coupling of the carbon dioxide tax with the particulate matter emission limit exhibits strong antagonistic effects. These findings advance knowledge by quantifying relative policy importance, threshold effects, and interactions, bridging gaps in understanding how distinct policy components drive renewable energy transitions. This study identifies key policy stringency levels for achieving specific renewable energy targets and recommends that policymakers adopt a well-coordinated policy mix, considering the potential synergistic and antagonistic effects, to achieve more effective outcomes and mitigate policy conflicts.

Power logit regression for modeling bounded data

The main purpose of this article is to introduce a new class of regression models for bounded continuous data, commonly encountered in applied research. The models, named the power logit regression models, assume that the response variable follows a distribution in a wide, flexible class of distributions with three parameters, namely, the median, a dispersion parameter and a skewness parameter. The article offers a comprehensive set of tools for likelihood inference and diagnostic analysis, and introduces the new R package PLreg. Applications with real and simulated data show the merits of the proposed models, the statistical tools, and the computational package.