Causal inference for quantifying displaced primary production from recycling

Abstract

Recycling only creates environmental benefits when it displaces other material production. Without displacement it only delays rather than prevents ultimate disposal. It is therefore critically important that we improve our understanding of the causality between recycling and other material production. This research focuses on estimation of the causal link between an increase in recycling and a reduction in primary material. We first review how structural models of supply and demand, for both the primary material and the recycled material, can be used to identify a causal link. The supply and demand approach suffers from issues of endogeneity, which require the use of advanced regression techniques. These techniques, in turn, require detailed and large datasets, which are often hard to obtain. Here, we introduce the Difference-in-Differences estimator to the industrial ecology literature, as an alternative approach to quantifying the causal effect between additional recycling and primary material production. The Difference-in-Differences estimator is based on a quasi-experimental approach, in that it classifies data into treatment and control groups. We introduce the method, analyze the data structures and assumptions needed for identification of causal effects, and discuss the advantages relative to the supply and demand framework. A hypothetical application of each method to aluminum recycling is provided, along with a simulated quantitative example of the Difference-in-Differences technique. Our proposed method will help to better understand, measure, and promote the conditions under which recycling creates environmental benefits.