Calculating benefits of infrastructural investment : paper presented at ERSA-NECTAR Special Session on Accessibility and Spatial Patterns ERSA 50th Conference Jønkøping (Sweden), August 19-23, 2010, Sustainable Regional Growth and Development in the Creative Knowledge Economy

The paper continues the ongoing discussion about the concept of accessibility and its applicability to infrastructural appraisal. David Metz states that travel time savings disappear in a long term perspective and to empirically untested "constant travel time budget" literature. Metz therefore concludes that it is inadequate to measure the benefits of an infrastructural investment by means of travel time savings only. Further it is noted that the inadequacy results from longer term decisions - such as location choice - which arise in the long run. These longer term decisions should therefore be included in the modeling of transport - or better of the comprehensive economic - systems for public infrastructure appraisal respectively. Some authors hint at the possibility to measure consumer surplus applying the concept of the expected maximum utility (EMU). If we accept this argument the question arises to what extent we produce errors with "short term" models. To examine this issue we implement a proof of concept model with the purpose to simulate long term equilibrium with different decision possibilities in a minimal urban system. We use random utility maximisation and discrete choice theory with an agent-based modelling approach to respect the discrete nature, consistency with discrete choice theory and to what seems to many to be the way forward in the field. We simulate equilibrium conditions letting the agents chose their commuting situation allowing for different combinations of decision dimensions. The results show that improving accessibility results in an increasing and expected demand for more peripheral locations. Thus we conclude that profiteers of infrastructural measures are property owners whose properties get higher accessibility. Concerning the utility indicators we find that EMU is more consistent. Further we note substantial differences when applying different decision spaces. The appropriateness of using travel time savings as indicator for utility gains very much depends on considered time horizons because trade-offs between utility out of short- and long-term decisions is likely to occur. In case of assessing long time periods we should operate with models incorporating long-term decisions such as location choice. This is a strong argument in favor of LUTI-models.