preprints.org > business, economics and management > economics > 201905.0058.v2

Working Paper Article Version 2 This version is not peer-reviewed

Version 1 : Received: 3 May 2019 / Approved: 6 May 2019 / Online: 6 May 2019 (11:37:39 CEST)
Version 2 : Received: 5 September 2019 / Approved: 6 September 2019 / Online: 6 September 2019 (04:32:59 CEST)
Version 3 : Received: 20 November 2019 / Approved: 21 November 2019 / Online: 21 November 2019 (03:52:45 CET)

Analyses of the Environmental Kuznet's Curve (EKC) hypothesis have largely focused on economy level data with occasional analyses exploring sector level data. This paper exploits a new data set which contains sector level data on greenhouse gas emissions from the US energy sector as well as subsector data from six disjoint subsectors which together comprise the entire energy sector. The data contained in this data set is annual data at the state level from 1990 through 2011. By using differenced data we specify an econometrically sound EKC model and compare it against a model containing only a linear GDP per capita term. We find that by using a subsector level modelling approach, evidence for the EKC hypothesis is virtually nonexistent. Moreover, we find that aggregated subsector level estimates outperform sector level estimate on in-sample accuracy. These estimated models are then used to forecast emissions for the energy sector. We find evidence that US greenhouse gas emissions from energy production are at or near a peak.

emissions; energy; environmental Kuznet's curve; panel data; subsector analysis

Business, Economics and Management, Economics