“…Even when it is possible to include some of these effects, it is often not done because it increases the duration of modelling and thus adds time and costs to an already cumbersome and expensive practice (Johnston and Ceerla, 1996;MOTOS, 2007;Naess, 2011). The problem appears quite universal, and has been reported in Denmark (Nielsen and Fosgerau, 2005), Norway (Welde and Odeck, 2011), Sweden (Jonsson et al, 2011), Germany (Marte, 2003), the United Kingdom (Mackie, 2010) and the United States (Litman, 2012). In the case of do-nothing alternatives, a background trend of traffic development for different modes is often used to assess future levels of demand, regardless of network supply (Nicolaisen, 2012).…”