The public service ethos is in danger, and I am confused. I used to think that despite the increasing involvement of private providers, the ethos seemed set to continue. Now it appears less clear cut.

For the government, a service is 'public' if the public purse pays for it; the nature of its delivery is not part of the definition. Public funding should mean the public good is protected and equity is served. It should mean that everyone receives according to their need, while public health policy may be developed. The provider's needs come second to society's. As long as there is public scrutiny of how tax money is spent, why worry about who delivers the service?

This has definite appeal.

Publicly delivered services have sometimes been inefficient, complacent, self-serving, wasteful and inflexible.

Elsewhere in the economy, capitalism has developed entrepreneurial, financially efficient, user-friendly services in which the consumer drives the system, success seems directly rewarded and commercial failure appears both deserved and purifying, apparently purging inefficiency and sloth.

So why is the idea of publicprivate partnerships now being called into question? What makes health, and perhaps the other public services, so different?

First is profit. For the public sector, all profit is bad. It has become shorthand for exploitation, fat cats, efficiency at the expense of quality, and emphasis on the gloss of style at the expense of substance.

This reaction has created its own problems.Wringing the last bit of activity out of the health pound has certainly made the NHS appear efficient, but the lack of emphasis on the care environment or professional development and morale is now affecting customer satisfaction and staff recruitment and retention.

Good intent may be losing its power; people are no longer willing to drive themselves into the ground for the common good without tangible recognition.

This is the main argument for involving the private sector in public provision: give the workers appropriate incentives, and their work will improve.

The second issue is that the public sector doesn't carry out or understand performance management. Despite the reforms of the past dozen years, the NHS does not have the measures, ability, capacity or motivation to let contracts properly, or to monitor their performance meaningfully.

We have simplistic, reductionist measurements (waiting lists, emergency admissions and their ilk) which are at best irrelevant and at worst perverse.We reward with more resources providers who do not meet their targets, squeezing those who do.

We do little significant development work with providers.

In contrast, performance management in the best parts of the commercial sector includes remedy as well as punishment, a concept still foreign to the new NHS, despite the Commission for Health Improvement's attempts to be a developmental organisation.

The third and maybe most important issue is risk-sharing. If a service is paid for by one agency and provided by another, who carries responsibility for its effective delivery? How much are capital, equipment, financial and clinical risks to be borne by each of the parties?

Usually, the paying agency bears the brunt. If British Airways subcontracts its luggage-handling to another company which loses my bags, my complaint is still against BA. I do not care how the airline deals with the luggagehandler as long as my interests are maintained. The risks of providing the service will somehow be shared between them.

Similarly, the tax payer expects the government to manage public services fairly and efficiently.

The difference lies in what happens when things go wrong.

BA can replace its baggagehandlers, but can the government replace its private sector providers? If a commercial provider goes bankrupt, what happens to the service for which the government is responsible? If that provider simply chooses to go elsewhere, who is responsible for continuing the service? How fast can the public sector replace a commercial failure?

If the private sector is to be properly involved in providing public services, we will either have to improve our performance management vastly, or limit the commitment so that a failure is not disastrous. If catering services go wrong, the system will probably cope; if an entire private tertiary hospital is under threat of failure, the leverage held by the private provider over the government is probably unacceptable. Until we successfully resolve what happens in those circumstances, we should limit the extent to which we work with the commercial sector. For proof of what could happen if we do not, just look at the railways.