In the debate over proposed market reforms in the NHS, opinions are plentiful but few commentators back up their opinions with evidence. So, what does the evidence say?
It is noteworthy that theoretical work in the field predicts that in markets with regulated prices, hospitals will compete along the lines of quality if the price set is above marginal cost. Yet, there are reasons why this assumption might not hold. For example, if information regarding quality is poor, there might not be any gains simply because patients are unaware of the best hospitals. Another problem arises if providers are induced to cream-skim patients, in which case efficiency could decrease overall unless the general equilibrium impact offsets it.
The NHS reforms of the 1990s were not subjected to much good research on the effects on quality/efficiency, but two papers (here and here) stand out. They find that competition actually led to slightly increased death rates following heart attack admissions, while it decreased waiting times. Yet, this is not surprising since the former quality measure was not publicly available while the latter was; producers focused on the measure with which consumers used to compare them. Furthermore, the reforms were short-lived and due to fears of segregation freedom of choice and supply was significantly curtailed. While the fears of inequality in health care due to the reforms appear to be unsubstantiated, the evidence from the NHS market reforms in the 1990s is mixed at best.
Turning to the more recent market reforms since 2006, the differences are striking. In these reforms, information on quality was publicly available, prices were set centrally and patients were given some actual choice. Regarding quality, three papers (here, here and here) find that competition between public hospitals decreased mortality rates among patients suffering from heart attacks. Additionally, one of these papers analyses all-cause mortality rates and finds that competition was positive also for this measure. We can thus safely conclude that the market reforms in the mid-2000s that increased competition between public hospitals improved quality.
So far so good. In a fresh paper from last month, however, some of the above authors find more heterogeneous impacts. Focusing on whether competition improved efficiency, measured by the average length of stay for patients undergoing elective surgery, the paper finds that competition between public hospitals was beneficial. Yet, when analysing the effect of private hospital competition on the NHS, it finds diametrically different effects: competition from private hospitals reduced efficiency. The authors explain this by showing that more private hospital competition increased the number of old and poor patients in the public hospitals. In other words, private hospitals appear to have been engaging in cream-skimming.
It is important to note, firstly, that the authors do not have data on the costs of patient care. In private correspondence, Cooper, the lead author, points out that if payments were sufficiently risk-adjusted their findings would indicate risk segmentation rather than a drop in efficiency. It is likely, however, that payments were not sufficiently risk adjusted, and, if so, this would indicate cream-skimming.
Of course, we also do not know whether or not the general equilibrium impact was positive; competition might have improved overall efficiency that swamped the cream-skimming effect. Such a hypothesis, however, must be subjected to research before any inferences can be made.
Two conclusions are in order. Firstly, and contrary to opponents’ arguments, the evidence from the NHS indicates that competition in publicly funded health care – when designed properly – can work well and improve both efficiency and decrease mortality. Secondly, it also points to the importance of ensuring that funding is appropriately differentiated, especially when allowing private market players to operate. This is something I have argued previously with regard to free schools. Markets in publicly funded services differ significantly from regular markets – due to third-party payments, information asymmetries etc. – making it crucial that the incentive structures are designed properly. It might be difficult to implement, but the evidence from competition among public hospitals in the NHS indicates that it is not impossible.