Valuing Public Investments to Support Bicycling

ing from leisure trips that serve no purpose of transportation, getting from A to B conveys disutility to people in the form of money, time, effort, fear of accidents and other costs, the sum of which are usually referred to as generalized costs. In the context of bicycling, the generalized costs are almost exclusively nonmonetary and thus internal to cyclists. The existence of internal costs follows readily from the fact that without an offsetting element, the presence of positive health benefits would make the bicycle the dominant transportation choice for all trips up to a certain distance. However, the bicycle mode share in Swiss cities is around 5%, with few cities reaching more than 10% (Federal Statistical Office, 2007, 2012), and similar or lower numbers apply to other countries (Pucher et al., 2010). Even though the presence of internal costs of bicycling may seem obvious, they are usually not considered when valuing the benefits of bicycle policies. Naturally, bicycling also conveys utility gains for some trips (i.e., negative internal costs), but the relatively low bicycle mode share implies that for many trips that could theoretically be carried out by bicycle, but are not, the internal costs are positive. The tradeoff between costs and benefits is implicit in the literature devoted to identifying the determinants of bicycle mode choice and/or mode share, most of which are non-monetary in nature. Rietveld and Daniel (2004) use generalized costs as a predictor variable, in which they include measures such as “costs of effort” or fear of accidents. Similarly, Hunt and Abraham (2007) report that bicyclists choose routes that are least “onerous”, and Broach et al. (2012) find that cyclists avoid high volumes of motorized traffic. In the following, we separate the determinants that have been empirically identified to affect the propensity to bicycle into the following groups: Valuing Public Investments to Support Bicycling 301 Swiss Journal of Economics and Statistics, 2014, Vol. 150 (4) 4 This study is based on GPS data of actual routes taken in Zurich, which are compared to constructed non-chosen alternative routes. The main result is that the observed bicyclists are willing to make only slight detours in order to improve along another dimension (such as the presence of a bike trail or fewer stops), implying large time and effort costs of bicycling. a.) The general environment General factors that have been shown to influence the propensity to bicycle include weather (temperature and precipitation), topography, city size, cultural and neighborhood characteristics including land use (Pucher and Buehler, 2012; Saelens, Sallis and Frank, 2003; Wardman, Tight and Page, 2007). Determinants that may be more easily influenced by city planners are the prices of alternative modes of transportation including parking costs. b.) Route and destination characteristics The bicycle mode choice/share is influenced by the presence of bicycle lanes/ paths; the lane width, the volume and speed of motorized traffic; competition for space between drivers and cyclists; the number of stops, traffic lights or other obstacles; the number of intersections and their characteristics, accident risk; and qualitative aspects about bicycle lanes such as continuity and connectivity or the presence of on-street parking (Berrigan, Pickle and Dill, 2010; Broach, Dill and Gliebe, 2012). Menghini et al. (2010) conclude that trip length is the dominant factor for route choice, which is consistent with high time costs and/ or costs of physical exertion. This category may also include trip end facilities such as locking stations or the presence of showers at work (Wardman, Tight and Page, 2007). c.) Personal characteristics In terms of personal characteristics, the choice of bicycling is influenced by a person’s age, race, gender, education, car ownership, aversion to driving, and the perception of bicycle-friendliness of the traffic environment or environmental preferences (Bauman et al., 2012; Heinen and Handy, 2012; Li et al., 2012). Furthermore, the translation between external factors and route characteristics into (dis-) utility of bicycling may vary across people: For example, the physical effort associated with a significant elevation gain may deter older people more than younger ones, all else equal. This means that even mode choice determinants that are the same for everybody (i.e., hilliness) can have a person-specific influence on the propensity to bicycle. 302 Götschi / Hintermann Swiss Journal of Economics and Statistics, 2014, Vol. 150 (4) 5 Such a linear relationship between km of cycling and health benefits requires the assumption that cycling – more specifically cycling trips of different utility-is fairly evenly distributed with regards to overall activity levels of cyclists, which arguably may be the case for utilitarian urban cyclists. This assumption, however, is unlikely to hold for occasional leisurely rides or long distance rides, and subjects which are either entirely inactive or extremely fit. To include these extremes of the dose-response curve, a non-linear relationship reflecting lower benefits per additional km (within subject) would better reflect the nature of the effect (Woodcock et al., 2011). However, subject-specific health benefit assessment would be beyond the scope of this analysis. 2.3 Externalities A share of the benefits from improved personal health accrue to the population as a whole in the form of lower health care costs, which leads to lower tax rates in nationalized health care systems or reduced insurance payments in systems relying on private health insurance. Better health may also increase the productivity of workers, which may be captured only partially by a wage increase. To the extent that bicycling substitutes for motorized traffic, it is also associated with a number of positive externalities, or more precisely, with the avoidance of negative externalities. These external effects include a reduction in air pollution, noise and congestion, lower demand for parking spaces, less wear and tear on roads, and intangible effects such as “livability” (Dumbaugh, 2005; Ellison and Greaves, 2011; Litman, 2004; Thakuriah et al., 2012). 3. Consumer Surplus from Public Investments in Bicycling The social value of bicycle investment is shown in Figure 1. On the horizontal axis we measure cumulative bicycle-km Q that take place in a population and over a particular time period. MC refers to the per-km net internal cost of bicycling, which we define to include all costs and benefits with the exception of health benefits. We ordered bicycle trips by their internal costs, such that the MC curve is increasing by construction. Bicycle-km on the left are associated with negative costs (i.e. benefits), representing the bicycle trips with the highest utility. As the overall level of bicycling increases, the internal costs increase because more trips take place during “bicycle-unfriendly” conditions (such as during inclement weather, over hilly terrain or in dense traffic), or are carried out by people with a stronger aversion to physical exercise or accident risk. Health benefits per bicycle-km are given by HB, which we assume to be constant across bicycle-km. The equilibrium level of bicycling is given by Q0; beyond Valuing Public Investments to Support Bicycling 303 Swiss Journal of Economics and Statistics, 2014, Vol. 150 (4) this point, the internal costs of an additional bicycle-km are larger than health benefits, and vice versa. The same solution could be obtained in a more traditional demand and supply framework by defining the demand curve for bicycling as D HB MC (dotted line), and intersecting this curve with the infinitely elastic supply of bicycling at a price of zero. We chose to separate health benefits from the remaining costs and benefits because we want to focus on their degree of internalization. Figure 1: Costs and Benefits from Bicycling