In this paper, we derive a structural model for commuting speed. We presume that commuting speed is chosen to minimise commuting costs, which encompass both monetary and time costs. At faster speed levels, the monetary costs increase, but the time costs fall. Using data from Great Britain, we demonstrate that the income elasticity of commuting speed is approximately 0.13. The ratio of variable monetary costs to travel time costs is estimated to be about 0.14. An implication of this is that as incomes rise commuters choose faster modes, despite their higher monetary costs. This has been an important factor in the growth of commuting by car in the past decades (for example, during the 90s the percentage of work trips made by car in Britain increased from 65 per cent to 70 per cent) and is anticipated to be relevant in the next decades for developing countries such as China and India. With increasing congestion, the time-advantage of car travel will decline, but unless faster public transport modes are available, there will be little incentive to switch to public transport (unless the monetary costs decline substantially in relation to car travel).