One strand of the economic critique of government provision is that public providers face a soft bankruptcy constraint. If they operate inefficiently or extravagantly and run out of money then they can turn to government for a handout to cover any shortfall. If the government is short of money to bail them out they just put up taxes. Private providers, in contrast, operate in the face of a hard bankruptcy constraint. They must operate in the face of ever-present risk. If they don’t produce what consumers demand, and do so as efficiently as possible, then their continued existence is in doubt. Hence, the argument runs, a powerful incentive is missing from the sphere of public provision. This argument has been invoked as one component of the justification for shifting activities from public to private sector.

This argument came to mind this week with the co-incidence of two news items. The first is the report in Wednesday’s Guardian of Parliamentary scrutiny of the Government’s health reforms. The Public Accounts Committee has queried the fact that the NHS reform proposals do not appear to contain details of what happens if a provider fails in the new marketised health system. The Committee dismissed the Government’s response that they were not planning for failure. Quite right too. If it’s a market based system then there are going to be failures. It’s in the nature of the beast. The second news item was the arrival on Thursday of the deadline for written submissions to the Treasury Select Committee’s inquiry into the future of PFI (highlighted by a letter to the Guardian by Jesse Norman MP).

Both the NHS reforms and PFI raise the issue of risk and who ultimately bears it. And that raises the question of what type of bankruptcy constraint providers of (erstwhile) public services actually face in practice.

The economic argument, as set out above, is clearly simplistic. In an era of ingrained performance assessment, benchmarking and scrutiny, the idea that public organisations can be run without any concern for efficiency, and that government will simply make good any and every financial shortfall with tax revenues, is barely credible. So the bankruptcy constraint is not as soft in the public sector as it has been claimed. Whereas, as we have seen in the aftermath of the global financial crisis, some private providers are ‘too big to fail’. When failure appears a real possibility the state will step in to try to assure survival. The organisations are systemically too important. So effectively some private organisations do not face quite such a hard bankruptcy constraint after all.

However, the financial crisis represented a rather specific set of circumstances. There are perhaps few organisations apart from banks that cannot be allowed to fail because they are ‘too big’ in this systemic sense.

Yet, we can think about ‘too big to fail’ in other ways. In particular, most discussions of markets are aspatial. If, on the other hand, we think in terms of spatial location and markets having a spatial dimension the picture changes. And we can broaden the issue from ‘too big’ to ‘too important’ to fail. Then it does provide some pointers regarding the more general issue.

A hospital, for example, needs to operate at a sufficient scale to provide a wide range of secondary health care. Under a non-market health system the spatial distribution of hospitals would be guided by population density and each hospital is likely to effectively have a spatial monopoly. Even under an increasingly marketised system competition between health providers only starts to function in large urban areas because the distances involved become sufficiently short. And competition doesn’t really work for services such as accident and emergency where speed can be of the essence. Relatively few people want, or are able, to travel hundreds of miles for elective treatments, and spend time in hospital a long way from family and support networks. Those commentators who are frustrated by the lack of ‘proper’ choice behaviour by consumers in health markets fail to understand the factors taken in to account when making such decisions.

If a hospital is transformed into an independent or private entity in order to reap putative efficiency gains then the Government has to be prepared to let it go to the wall if it is inefficient or unpopular. That is the market way. It may be that a new provider will eventually enter the local market, having spotted anopportunity for a financial return, but in the meantime much of the local population will be without secondary health care.

If, on the other hand, a hospital is deemed too important – in order to ensure the resident population have continuous access to health care – to permit failure then effectively you have removed the hard bankruptcy constraint. So you don’t get the full benefits of the market discipline. But you do get a less integrated system, new information asymmetries, increased transaction costs, and an increased risk of incompatible incentives – hence making the system harder to manage.

You would like to hope that we occupied a world in which providing access to adequate health care took precedence over market dogma. So the second scenario seems more likely.

The principal rationale for PFI is to pass risk from the public to the private sector. Private infrastructure providers are, by assumption, more efficient. Because they are raising money commercially to fund their investment they are paying commercial interest rates and will be driven to be more efficient in order deliver returns to their shareholders. The history of PFI policy has been a recurrent questioning of whether such risk transfer actually occurs and whether claims to value for money are justified. The most recent instalment in the debate – which is inconclusive on the point – is the NAO report published yesterday.

The problem at the heart of the PFI debate is that ‘risk’ is a largely subjective phenomenon. The Treasury, in justifying and advocating for the policy, claim that there are considerable risks associated with development and operation of infrastructure facilities that the private sector has to bear. The wide range of critics lined up against them argue that PFI schemes are, in contrast, extremely bad value for money for the taxpayer. They are criticised as being both highly profitable and in practice very low risk for the private sector. Indeed, comments from the private sector sometimes suggest that infrastructure provision using this mechanism is seen as a a bit of a one-way bet with a high payout.

From the private sector perspective PFI represents, to all intents and purposes, a guaranteed and index-linked income stream over several decades. If the state wants changes to the infrastructure facilities provided it requires costly renegotiation. If the state wants to break the agreement then typically the breakage costs are very substantial. Even if the infrastructure is no longer required – for example, population movement means that there are not enough children locally to keep a school open – the state can often still be obliged to honour the contract.

At the moment the Government is trying to bring PFI providers back to the table to renegotiate the fees downward, given the broader climate of austerity. But the private providers are not obliged to respond to this initiative. The Government has publicly committed to honouring existing contracts. So they can only appeal to private providers’ better nature. Not a lot of progress appears to have been made yet. Meanwhile, a larger and larger proportion of budgets in sectors such as health is being taken up by the fixed cost of PFI fee payments. This has consequences for jobs and services – because something has to be cut.

So here again we have a narrative in which the private sector is preferred because it is subject to the discipline of the market, yet in practice the risks are very low. Unless, that is, the Government were to do something radical such as declare PFI debts to be odious debts and hence non-binding, which seems unlikely.

It would, of course, be possible to write future PFI contracts in a way that meant that risks were more genuinely moved to the private sector – so, for example, the fee the state is required to pay is correlated with usage but there are no perverse incentives to skew provision so that usage is directed towards PFI facilities. This would look a lot less attractive to private providers. Why? Precisely because there could be significant risk attached to signing up for a multi-million pound deal running for 30 years.

The point is not that the private sector cannot be used to provide services, but that when looking at alternative mechanisms for providing services we should do so in a way that recognises the situation for what it is.

If we are talking about a vital service – particularly with a spatial monopoly – that it would be politically unacceptable to cease providing then effectively the bankruptcy constraint is going to be pretty soft, whoever the provider is. Similarly, we need to recognise that the state can, if it negotiates poorly, create situations in which it basically insures the private sector against risk – as has effectively happened with many PFI deals. That largely renders the notion of a bankruptcy constraint a theoretically curiosity from the provider perspective.

If we are talking about a situation in which there is little risk and the harsh winds of competition cannot sensibly be expected to be blowing then there is little point in pretending otherwise. The hard bankruptcy constraint of the simple economic story is nowhere to be found. The consequence of pretending otherwise is simply to hand provision to organisations whose raison d’etre is to maximise the financial return for their owners rather than value for money for the taxpayer. The market narrative is that the two – financial return to owners and value of money for the taxpayer – are compatible. That can be made to be the case, but only under certain circumstances. And those circumstances do not obtain in relation to many areas of public provision.