Health policy will be a major issue in Britain's next general election. The Labour government is committed to a substantial increase in funds for the National Health Service (NHS) and has eliminated tax relief for private health insurance. The Conservative Opposition party favors subsidizing private health insurance, though it has pledged to match the government's funding increases for the NHS. This paper develops and implements a methodology for estimating the welfare effects of increasing public and private health care in the United Kingdom, when these policies are financed either by distortionary taxes or by user fees for the NHS. User fees are currently minimal, and the national health market "clears" by creating waiting costs. In the private sector we assume that prices approximately reflect marginal supply costs, and there are no waiting lists. We find that the welfare change from increasing NHS output could easily be negative, particularly when extra spending is financed by distortionary taxes. In contrast, expanding private health care is always efficiency-improving in our simulations. In our central estimates, increasing private health care by a pound's worth of output produces an efficiency gain of 55 pence, but increasing national health output produces a net efficiency loss of 32 pence per pound! One reason for these results is that increasing the output of rationed health care has ambiguous effects on the total deadweight losses from waiting costs, but these costs unambiguously fall when the private health sector expands. Financing policies by user fees avoids the efficiency costs of raising distortionary taxes, and it also produces efficiency gains by reducing waiting lists. In fact, increasing national health care output produces an overall efficiency gain in most of our simulations, rather than an efficiency loss, when the policy is financed by higher user fees rather than by distortionary taxes. Still, the policy is generally less efficient than a user-fee-financed increase in private health care.