Valuing Reductions in the Risk of Traffic Accidents Based on Empirical Studies in Sweden

At the beginning of the 1990s the Swedish National Road Administration (SNRA) adopted the willingness to pay (WTP) approach for estimating the benefits of safety improvements. However, since then there have been discussions about how to find the most valid and reliable empirical methods to conduct WTP studies. The purpose of this thesis is to describe and analyze approaches used to estimate the value of safety and to present empirical results that might be considered for future revisions of SNRA’s value of safety. Two contingent valuation (CV) studies from Sweden are presented, in which individuals estimate their own subjective risk and their WTP for an individual risk reduction. The two studies each include two sub-samples, with questions to test for different kinds of bias. Results from the first study conducted in 1993/1994 provided an estimate of the value of a statistical life (VOSL) of SEK13 million in 2001 prices. Results from the other study from 1998 indicated an estimate of VOSL of about SEK21 million in 2001 prices. The VOSL adopted by the Swedish authorities in 2002 was 16.3 million in 2001 prices. Results from an international survey show that the Swed-ish accepted VOSL was lower than the accepted VOSL in the US, but of the same magnitude as in the UK. Estimates also indicate that a safety regulation which costs more than SEK116 million on average per life saved will be counterproductive by causing more deaths than it intends to prevent. A method is presented for estimating the value of reducing the risk of non-fatal casualties. The value of preventing an average severe injury and average slight traffic injury was estimated to about 16% and 1% of the VOSL, respectively. This thesis also provides a new insight into how sensitive results of WTP are for subjects’ inability to consider small differences in terms of the scope and scale of the physical risks. Results from the CV studies show the impor-tance of providing information to respondents, on risks and costs for alternative goods and services on the market, in order to improve their ability to make rational decisions on hypothetical markets. In addition, estimates from long-term follow-up studies of the value of resources used to treat and rehabilitate non-fatal traffic injuries are presented. These estimates are based on the incidence approach. Results from incidence studies indicate that about 2/3 of the long-term costs appeared 2-8 years after the accident. These results show the importance and the difficulties of long-term follow-up studies for collecting resource use data. Conclu-sions yield both the interpretation of estimated values and the design of future studies.