Statistical Methods & Applications ( IF 1.1 ) Pub Date : 2021-04-13 , DOI: 10.1007/s10260-021-00559-5 Claudia García-García , Catalina B. García-García , Román Salmerón
Despite the evidence, the correlation between environmental impact factors has mostly been neglected in econometric environmental models or treated with traditional methodologies such as ridge regression, which are recommended when the goal is prediction and the estimated parameters are not interpreted as causal effects. This paper addresses the existing collinearity with alternative methodologies, not only to mitigate the problem mechanically, but also to isolate the effects of the environmental impact factors with the main objective of designing better policies for countries. The methodologies are applied to analyze the CO\(_2\) emissions of 114 countries covering the thirteen most recent years with available data, and the results from the empirical and methodological perspectives are compared. The treatment of collinearity with the residualization or raise regression procedures allows the researcher to obtain a global vision of the relationship between the different factors affecting CO\(_2\) emissions, thus reaching alternative conclusions to those from traditional methodologies.
中文翻译:
在环境回归模型中面对共线性:来自世界数据的证据
尽管有证据,但环境影响因素之间的相关性在计量经济环境模型中大多被忽略,或采用传统方法(如岭回归)进行了处理,当目标是预测且估计参数未解释为因果关系时,建议使用此方法。本文以替代方法论解决了现有的共线性问题,这不仅可以从机械上缓解问题,而且可以隔离环境影响因素的影响,其主要目的是为国家设计更好的政策。将该方法应用于分析CO \(_ 2 \)利用现有数据对涵盖最近13年的114个国家的碳排放进行了比较,并比较了经验和方法学角度的结果。使用残差法或共回归法对共线性进行处理可以使研究人员获得影响CO \(_ 2 \)排放的不同因素之间关系的全局视野,从而得出传统方法的替代结论。