Nonlinear analysis of employment in waste management

ABSTRACT

This paper contributes to the ongoing debate on the employment effects of waste management policy. We examine the dynamic linkage between waste tonnage and employment in the region of Paris (France), at the point when the waste management policy by delegation of service is adopted. To account for the presence of possible policy shifts, we propose the implementation of nonlinear causality tests based on the smooth transition autoregressive regression (STAR) framework. Using weekly data for four waste streams over the period January 2015–June 2017, the linearity tests reveal the presence of nonlinearity in most of the data series. When applying nonlinear Granger tests, our results provide strong support for nonlinear dependencies between waste tonnage and employment across different waste streams. For instance, we find strong statistical evidence that the causal relationship is consistently bidirectional in the miscellaneous and outdoor garbage waste streams. From a policy point of view, our findings suggest that waste management practices should factor the presence of these strong linkages and how they would affect environmental jobs creation.

KEYWORDS:

• Waste management sector
• jobs creation
• nonlinear causality

JEL CLASSIFICATION:

• C22
• Q52
• Q53

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ More sophisticated technologies for waste processing may lead to labour substitution and a trend for lower employment.

² Most of documents in this area are non-academic papers, and to the best of our knowledge, Sell et al. (1998) is the only peer-reviewed source.

³ The private operator has been designed to manage the collection of waste burdens through competition in a tender offer with other operators. For that reason, we cannot give the name of this private operator. In addition, data are confidential because the private operator uses them to predict waste production and answer the public tender offer.

⁴ As argued by Koop and Potter (2000), a lot of evidence for nonlinearity in economic time series might, in fact, be due to structural changes.

⁵ Lowercase letters denote the logarithmic values of the variables.

⁶ As most of employment series are level stationary, the possible presence of a cointegration relationship between variables in the level is excluded.

⁷ The optimal lag length to be introduced in model (1) could be also determined with different information criteria, including Akaike Information Criterion (AIC), Schwarz Information Criterion (SIC), etc.

⁸ For details about linearity test procedure see Teräsvirta (1994).