Environmental regulation and safety outcomes: Evidence from energy pipelines in Canada

https://doi.org/10.1016/j.reseneeco.2020.101215Get rights and content

Highlights

  • We investigate the effects of more stringent environmental regulation on the safety performance of Canadian fedreally-regulated oil and gas pipelines.

  • We focused on two newly established policies: paying amortized abandonment fees as accrued and assuming absolute liability for adverse events by pipeline operators.

  • Our results suggest that imposing abandonment fees is positively and significantly related to increases in the number of pipeline adverse events.

  • In contrast, we find that the imposition of absolute liability is associated with the significant decreases in the number of pipeline adverse events.

  • Establishing absolute liability is more effective in reducing adverse events for Canadian major oil pipelines than non-major, smaller pipelines.

Abstract

This paper investigates the effects of more stringent environmental regulation—specifically, paying amortized abandonment fees as accrued and absolute liability for adverse events—on the safety performance of Canadian oil and gas pipelines. We construct a comprehensive dataset that includes pipeline adverse events, throughput, and regulatory intensiveness for fourteen federally-regulated energy pipelines in Canada. We find that abandonment fees are significantly and positively associated with pipeline events. On average, a 1% increase in abandonment fees is associated with 1.3 more events per month. In contrast, establishing absolute liability is associated with significant decreases in the number of pipeline events: on average, absolute liability is responsible for eliminating 80% of quarterly events for Canadian major oil pipelines and 20% for smaller pipelines.

Introduction

In 2017, more than 95% of Canadian crude oil and natural gas was transported to world markets through pipelines. Industry participants, regulatory authorities and the public have argued that inadequate pipeline capacity depresses Canadian oil prices and reduces public revenue (NEB, 2016, Heyes et al., 2017, Walls and Zheng, 2020). Inadequate transportation infrastructure places the Canadian energy sector at a considerable disadvantage (Dachis, 2018). However, building new pipelines or expanding existing pipeline capacity in Canada is constrained. Although proponents argue that pipeline construction would bring local economic benefits, mounting concerns from the public emphasized the safety issues associated with possible oil spills.

To reduce environmental risks and improve pipeline safety, the Canada Energy Regulator (CER) has imposed a set of new policies on all oil and gas pipelines under its jurisdiction.1 Since 2009, the CER has been working to establish a pipeline abandonment fund. Each pipeline company is required to evaluate and report future costs of abandonment, including all costs associated with environmental risks during dismantling. A pipeline company then amortizes its abandonment fee and makes payments annually to a third party fund. The Pipeline Safety Act, which came into force in June 2016, further clarified CER's jurisdiction over Canadian federally-regulated pipelines at post-abandonment stages. The Pipeline Safety Act also allows the CER to handle pipeline incidents or accidents by requiring pipeline companies to assume absolute liability. Absolute liability makes companies operating oil and gas pipelines responsible for all costs and damages, regardless of the origin of the fault. Federally-regulated pipelines are responsible for up to $1 billion absolute liability and are required to demonstrate sufficient financial resources to meet the newly established liability rules.

These new policies are intended to internalize the social costs of environmental impacts and to improve pipeline safety. Nevertheless, how effective the new regulations have been in improving safety outcomes remains open to debate. Intuitively, these new regulations may exert adequate deterrence, causing pipeline operators to increase safety efforts. However, the new regulations may impose significant financial costs on pipeline companies. As a result, they could reduce current expenditures by deferring maintenance and repair, increasing the risk of adverse pipeline events.2 This paper provides the first quantitative evidence on the effectiveness of the new regulations on pipeline safety in Canada. We estimate the effect of abandonment fees by exploiting the variation in abandonment fees across firms and time. To examine the impact of absolute liability, we estimate two sets of difference-in-difference models using various control groups. We first exploit the cross-firm variations in the timing of implementing absolute liability between major and non-major federally-regulated pipelines. We then estimate the effect of absolute liability on major oil pipelines by considering non-major, smaller pipelines as the control group when the non-majors were temporarily exempt from the regulation. Next, we explore the impact of absolute liability on the non-major pipelines by using provincial pipelines in British Columbia (BC) as the corresponding control group. To mitigate potential identification threats, we restrict the sample of treatment group to the non-major federally-regulated pipelines crossing BC. The limitation of this specification is that we may only obtain a lower-bound estimate of the impact of absolute liability.

We find that abandonment fees are significantly and positively associated with the occurrence of pipeline events. On average, a 1% increase in abandonment fees is associated with 1.3 more events per month. We also document that the imposition of absolute liability is associated with significant decreases in the number of pipeline events. The number of pipeline events for the major oil pipelines was reduced by 2 cases per quarter since the establishment of absolute liability, equivalent to a reduction in the average number of pipeline events by 80%. The absolute liability regulation also resulted in the non-major pipelines reducing events by 20% per quarter.

The paper is structured as follows. The next section discusses related literature. Section 3 gives a brief summary of the dataset. Section 4 presents the empirical analysis on the effect of abandonment fees. Section 5 presents the empirical analysis on the impact of absolute liability. Section 6 discusses the implications and limitations of our work and then concludes.

Section snippets

Related literature

The contribution of this article is three-fold. First, our work contributes to a broad literature on the effectiveness of deterrence on firms’ environmental compliance, specifically in the hazardous material industries. Exhaustive reviews of the empirical evidence on the effectiveness of deterrence are provided by Gray and Shimshack (2011) and Shimshack (2014). Numerous authors have investigated various monitoring and enforcement instruments for deterrence. For example, deterrence targeting

Pipeline events

We obtain information from the Transportation Safety Board of Canada (TSB) on adverse events for all federally-regulated pipelines, in Canada, from 2006 to 2018. The TSB reports the date and location of each adverse event and whether it is associated with rupture, injury, death, evacuation, or environmental impact. The TSB also describes the causes and the circumstances of an event. An incident represents an uncontrolled release of a commodity, with a minor effect on human lives and

Empirical specification and results

The empirical specification to analyze the effect of abandonment fees is as follows:yit=α+βAbandonmentit+λi+θXit+εit,where yit is the number of pipeline events (including incidents and accidents) for pipeline i at time t; Abandonmentit denotes the logarithm of abandonment fee paid by the pipeline firm i at time t; λi denotes the pipeline fixed effect; Xit represents a vector of time-varying covariates and εit is the error term. Eq. (1) is estimated as a negative binomial model because the

The effect of absolute liability regulation on pipeline safety performance

To examine the impact of absolute liability, we apply difference-in-difference (DID) models. We estimate the effect of absolute liability stratified by different types of federally-regulated pipelines, because the absolute liability took effect at different times for the major and non-major federally-regulated pipelines. Moreover, the effect of liability rule changes may differ by firm sizes (Alberini and Austin, 1999, Alberini and Austin, 2002). We first exploit cross-firm variations in the

Concluding remarks

In response to increasing public concerns regarding pipeline operating safety and environmental impact, the CER launched the Pipeline Safety Act to reduce environmental risks from pipeline operations and abandonment. We investigate the two most important features of the new regulation: collecting abandonment fees in advance and imposing absolute liability. We show that all else equal, pipeline firms with greater abandonment fees operated less safely, and absolute liability generally improved

References (31)

  • J. Boomhower

    Drilling like there’s no tomorrow: Bankruptcy, insurance, and environmental Risk

    Am. Econ. Rev.

    (2019)
  • J. Boyd

    Financial responsibility for environmental obligations: Are bonding and assurance rules fulfilling their promise

    In An Introduction to the Law and Economics of Environmental Policy: Issues in Institutional Design

    (2002)
  • B. Dachis

    Death by a thousand cuts? Western Canada's Oil and natural gas policy competitiveness scorecard

    The C.D. Howe Institute's Commentary No. 501, February

    (2018)
  • D. Earnhart

    Panel data analysis of regulatory factors shaping environmental performance

    Rev. Econ. Stat.

    (2004)
  • W. Gray et al.

    The environmental performance of polluting plants: A spatial analysis

    J. Regul. Sci.

    (2007)
  • Cited by (0)

    We thank two anonymous referees and the editor, Prof. Johnston, for comments that have helped to improve the content and clarity of this paper. All errors remain our own.

    View full text