How overwhelming is the evidence in favor of Road Diets? A note on the cost-benefit methodology proposed by Noland et al. (2015)

- This Critique is part of a debate on how to evaluate the impacts of Road Diet conversions.
- Questions are raised regarding the methodology proposed in (Noland et al., 2015).
- This Critique focuses on the specification of Costs and Benefits, not whether or not Road Diets are a good idea.

Writing in these pages, Noland and colleagues recently proposed a methodology for cost-benefit analysis of 'Road Diets' (re-design of the cross-section of a four-lane arterial). On the basis of the proposed procedures, the authors conclude that a Road Diet of an empirical case study (in New Brunswick, New Jersey, USA) would provide 'overwhelming' benefits. The study team has employed similarly unambiguous language in statements in other public fora. In this paper, I do not take a view of whether Road Diets are in general desirable or not (which will depend heavily on local context). Rather, I demonstrate here that the specification of benefits and costs in the proposed methodology is systematically biased (upwards in the case of 'benefits' and downwards in the case of 'costs'). I also show that the authors mis-interpret public opinion regarding the proposed Road Diet; general public opinion cannot be known on the basis of the evidence that is presented to readers. Finally, it has been previously shown that transportation planners tend to, on average, systematically under-estimate costs and over-estimate benefits. While the authors' motivations during the study at issue cannot be known, the systematic bias in specifying costs and benefits is consistent with Flyvbjerg's 'political-economic hypothesis', in which it is theorized that planners strategically misrepresent costs and benefits in order to increase the likelihood of a politically-preferred project being advanced. Flyvbjerg suggests making independent peer reviews publicly available as part of a strategy to encourage accountability through transparency; this is the spirit in which the present paper is disseminated.