This work evaluates the effectiveness of the federal government of Mexico's programme of industrial estates, a programme implemented to bring about a radical change in the geographical distribution of industry away from Mexico's three largest metropoles. Such decentralisation was expected to induce economic development in lagging regions. This research shows that the Central Region of Mexico City and its surroundings are the major beneficiaries of the limited decentralisation that has taken place. However, neither the net employment directly created by the programme, nor the formation of local linkages, provide the bases for large multiplier effects in the local economy. Nonetheless, other long-term benefits can be recognised. These findings are framed within a 'process evaluation' perspective emphasising the complexity of policy formulation and implementation in the field in order to understand how outcomes are generated.