There has been a regular cycle of Government committees in Britain that have forecast 'shortages' and 'surpluses' in medical manpower. It is remarkable how little change there has been in the methods used in these forecasts despite criticism of their methods and the move towards competitive markets in the delivery of health care. The practice of medical manpower planning was criticised in the 1960s because of the failure to explore the implications of alternative skill mixes and incentives. Whilst these criticisms were ignored, the forecasters now pay more regard to budget constraints and produce more regular estimates. The 1992 forecast could be improved by sensitivity analysis (e.g. for GP/population ratios). However, unless the issue of scope (forecasting for all health care staff, not just for doctors alone) and the implications of skill mix and incentives identified in the 1960s are addressed effectively, the current recommendations may produce doctor 'surpluses' in the early twenty-first century. The Government's analysis should be challenged at the level of principle and practice, and this implies that the research activity proposed in the 1960s is carried out in the 1990s to facilitate efficiency in the twenty-first century. Furthermore, the issue of policy priorities and their ranking needs to be spelt out explicitly: does the Government want a competitive NHS internal market, and if it does, what is the role of central manpower planning? One implication of current policy choice and the maintenance of manpower planning is that cost containment is of more importance to policy makers than the internal market.