This paper develops a commuter location model able to explain and simulate residential location changes of commuters that result from transport improvements. The core model is based on the assumption of constant commuting time, while two extensions incorporate substitution possibilities having an upward effect on total commuting time. Estimation errors of the residential location of the working population with the existing transport system are limited to 7%. With the extended model, the impacts on commuting and residential choice are investigated for six higher speed rail connections between Amsterdam, located in the urban core of the Netherlands, and Groningen, located in its rural periphery. The model outcomes strongly influenced the public policy debate in the Netherlands.