Urban environment and travel times to work

The increasingly long time spent on daily travels to work causes heavy economic and non-economic losses in many cities worldwide. Possible determinants of the length of urban trips to work can be found at different levels of spatial focus, based on demographic, socioeconomic, and transport infrastructure characteristics. Following a brief review of two broad hypotheses, the paper discusses theoretical fundamentals and results of alternative duration models applied to average work travel times across 92 cities in 66 countries. Estimates are fairly robust to sample size and specification of the hazard rate. Relatively smaller and /or wealthier cities tend to have higher hazard ratios, thus implying reduced average time spent in work trips. In two out of five parametric models, the shape of the hazard function is non-monotonic, with duration dependence changing from positive into mildly negative at a long spell stage. This outcome may be interpreted as a sign of possible 'stickiness' of travel time beyond a certain threshold of high urban congestion.