Is it wise for DfE to be investing in a “bad bank”?

One of the things I have been thinking about for a few years is how policy makers ameliorate some of the structural and policy issues around academisation. Indeed in 2016, having proposed full academisation while at Policy Exchange, I wrote about a way in which the “schools no one wants” problem could be addressed:

Government could require the RSCs to be a provider of last resort, via its own arms length trust, probably with a duty to make the school viable for re-brokering into another chain over time. Although I dislike the analogy, given that the schools in this instance haven’t done anything wrong, this is similar to the ‘bad bank’ approach taken by governments to take on or otherwise financially guarantee debts or other financial products which cannot be held by the private sector (and which can on occasion include whole banks in the case of RBS and Lloyds), improve them, and in time return them to the market.

The reason this issue comes about is because in any quasi market, like academy schools, there are elements of the market – certain geographies or certain services or certain customers – which are less attractive to a provider. (As a brief aside here – yes, this is marketised language. No, I don’t think schools are like tins of beans. Yes, school trusts should and indeed are motivated by a whole range of issues that are not financial. No, this isn’t a discussion about privatisation. But yes, this does use language of economic theory because this sort of partnership and quasi market is exactly what the academy model looks like, in common with many other public services in England, and this is the easiest language to use to explain the principles behind policymaking in it.)

There are, I think, three broad ways in which regulators can address differing attractiveness of elements within a market:

• Price mechanism – either paying a greater unit price to a service provider for these elements of the market, or accepting a lower bid for a service that covers these areas in contracts that are designed this way. This is effectively how a lot of the health service operates via different pricing for different patients and procedures.
• Regulation – compelling providers, in some way, to take on the less attractive elements of the market. For example, there’s something called a universal service obligation in sectors including telephony, postal services, and some utilities, which means in short that these companies have to service rural and isolated communities within the same price (more or less) that they do for serving Islington. If they don’t want to do that, they can’t have a contract.
• Isolation – effectively only operating a quasi market for some bits of the service and keeping others in house – either for political reasons, or because the market won’t take it up (for example when probation was outsourced, the higher risk elements were kept in house. Child protection is also often not outsourced because the risk to providers is too high.)

In 2016, I argued that the “bad bank” approach – model no 3 – was the right one for academisation.

So I was initially delighted to see the bad bank proposal being taken up by government, and the announcement today about the chair – even though there are still a lack of details as to how it will work. I’m still mostly keen on it – and I’ll expand on why below. But I’ve also had a couple of conversations in the last few weeks with some very smart people who have made me reconsider my point. And as the Twitter meme goes:

No-one:

Me: ok, ok, let me write you a couple of thousand words on my emerging and changing views on bad banks and academisation.

The case for

In a sense, it starts with what I dont’t like about the other two options:

• I like the principle of regulation – essentially telling MATs if they want a large, high performing secondary they have to take on a different school with a bunch of challenges alongside it – but I worried that we didn’t have enough providers in the system (ie enough MATs) to make it work, and that ultimately if too many of them declined to expand on this regulated basis, the government would be stuck.
• For the same reason – excess power of limited suppliers against large demand – I was cautious of a price mechanism, because I worried that would in some senses put the government over a barrel – if only a few large providers are interested in taking a challenging school, the “clearing point” of the market (ie what the government would have to pay) could be very high indeed.
• So, I reasoned, the easiest thing to do is to create a different trust of some sort, held entirely by government, and facing different incentives, with a duty to take on the more challenging schools.

Although I didn’t expand at the time on what those incentives are, I think in truth it has to be around some elements of finance, and Ofsted. These are the things which put many MATs off taking small schools, or those with big deficits, or those with PFI, or those which will take several years to get above Ofsted RI. There has to be an acceptance that these levers will bite less – that for a trust made up solely of the most challenging schools, recognition must be given of the special circumstances. In very concrete terms, this might mean not giving this trust an immediate financial notice to improve, pausing Ofsted inspections for a period of time, allowing some deficit budgets, constructing a mechanism to sort out PFI debt, giving it priority for capital investment, allowing small schools to close or merge within the trust without the same level of regulatory oversight etc. But there must still, of course, be some bite – otherwise all government will do is take underperforming schools from one custodian and give them to another with little prospect of improvement. (And if you want to skip ahead, one of the doubts which has crystallised in the last few weeks is exactly where you set this level of ambition for such a bad bank, and whether it’s feasible)

The case against

• Are we going to spread talent too thin? The whole premise of MATs, after all, is one of pooling talent to address problems at scale. If DfE starts to create a number of these bad bank trusts on a national or regional basis, is this going to suck in a number of people from other MATs because of moral purpose, or salary, or anything else – and hence weaken other, sustainable MATs?
• Can a bad bank actually be run? One of the strengths of a MAT is that you can more or less control your pipeline of schools. You only expand when you’re ready to do so (or maybe well after you think you’re ready) But a bad bank won’t have that luxury. New schools could come in potentially very frequently, and at short notice. This makes planning for where you allocate your finances and the time of your improvement team very challenging. A lack of stability could really hamper a reform programme.
• Where is the biting point? If you don’t loosen the restrictions on what this trust can do, then even with all the money in the world, it will be hard going. But if you loosen all restrictions too far – if you give in to the soft bigotry of low expectations and have a tacit acceptance that the schools in the band bank MAT won’t ever really improve – then that’s a terrible outcome. Judging where to loosen incentives and to what level is almost impossible, because trial and error doesn’t work – there won’t be a large number of competing bad banks out there to allow DfE watch how the sector responds and tweak the incentives over time. I have low confidence that government could place the regulatory hurdles at the right level.
• What does this mean for nudging growth of existing MATs? I think this is my biggest concern now about a bad bank. What is the incentive on an RSC now to have a number of difficult conversations with MATs in their patch about how they can collectively improve standards and take on challenging schools? Indeed, it’s now in everyone’s interests – the RSCs, the local MAT, the LA, the other schools – to “throw the school over the wall” into the bad bank, and let someone else deal with it under a looser regulatory regime. Similarly, any prospect of using methods 1 and 2 to grow MATs – financial incentives and regulatory nudging – get much worse. An element of deal making is a crucial part of government “making the market” here – to eliminate debt, to allow a good school to be rebrokered if a challenging one is too. If the MAT can stay strong and resist taking this school, knowing that ultimately the bad bank will step in, it’s harder for them to be compelled to take a deal, and to be pushed to grow in a way which the system needs.

So while I’m now not opposed to a bad bank entirely, I have more questions than I did previously. Increasingly, I’m more attracted to model 1 – the financial model. As noted above, this happens anyway now, just in bilateral discussions between the department and the MAT. A purer and more effective model, perhaps, is to make this much more open. Essentially, and without wishing to get all IEA-neoliberal about this, this is an auction issue. The market is not clearing at the correct price. Some schools are to challenging to take on for the financial support available to MATs. That’s why we have an excess of supply of such schools over demand to take them. The DfE has spent the past few years implicitly demonstrating that it thinks it is being price gouged by MATs, and that they can drive the price down through tougher negotiating by Agnew. But seemingly, they can’t – or rather, they can, but only by making some schools effectively un-sponsorable. Raise the price of taking on such schools accordingly – and offer such schools to multiple providers at the same time – and there will always be a sum of money at which a school can be rebrokered. Making this transparent also allows for greater efficiency as everyone knows after a while what the “price” for a certain type of school is.

Again: the language here is deliberately one of a market and of a product, and no, I don’t think schools are akin to products. But this is how many other public services operate. Commissioners – normally local government, and very often Labour ones – set a specification for services they want (“empty these bins once a week”, “run this care home” “provide catering for this hospital”) and providers offer a price. Commissioners choose the provider on a combination of price and quality. But the price element does vary depending on the service being offered, and after a while a reasonably clear shared understanding of what that price is emerges. There is no in principle reason why schools couldn’t be treated the same way.