By Manos Schizas, senior economic analyst, ACCA
As I write this post, it’s been a year and a day since the Independent Commission on Banking (ICB) published its report on the future of UK banking regulation. The establishment of the ICB was a response not only to the financial crisis but also to the unique challenge of maintaining the UK’s status as a global financial centre without taking on literally a world of risk.
Since then the Government has published its response to the ICB last December, as well as a white paper on financial stability in June. Within the banks the great groaning wheels of compliance have been set in motion. We’ve got seven years to go now until all UK banks are fully compliant.
Most important among the ICB proposals was the proposed ‘ring-fencing’ of activities crucial to the real economy, namely the taking of deposits and provision of overdrafts. This did not come as a surprise to anyone; many around the world had already called for the separation of retail and investment banking. So did ACCA, back in September 2008. The idea is that, if losses incurred in the world of wholesale finance cannot spill over into retail, then governments can rest easy that they will never have to bail out a bank brought to the brink by excessive risk-taking. Even better, banks that can no longer expect a bailout should adjust their appetites for risk accordingly, making more money available to lend to businesses.
Ring-fencing applies to SME credit in a fairly straightforward way (explained in detail on p. 70 here). In principle, it splits each bank into a ring-fenced and a non-ring-fenced portion, each legally separate from the other and with their own structures for managing risk. Any service likely to impact an SME’s viability if disrupted will have to be provided within the ring-fence. That’s current accounts, deposits and overdrafts by the looks of it, as well as some simple hedging products. SME loans, on the other hand, will have to be provided outside the ring-fence. The ring-fenced bank will be subject to higher capital requirements and can only deal with the non-ring-fenced bank on an arm’s length basis – i.e. on commercial terms as a completely separate entity. Small banks are exempt, but in fairness their current share of SME credit in the UK is tiny.
As a result of this structure, nearly all of the information relevant to an SME loan application will have to be generated within the ring-fence: behavioural scoring, for instance, is only possible on the basis of transactions; bank managers can only exercise discretion if they have a relationship with the client based on face-to-face interaction over time. This is the only proprietary information the banks have, since credit scores and published accounts are easily obtainable by their competitors. So far, there’s nothing threatening about this.
However, upon closer inspection it is clear that to assume any credit risk inside the ring-fence would be extremely difficult. Without capital regulation, business overdrafts could in principle be easily financed from deposits, but the cost of capital imposed by the ICB proposals and CRDIV (the European regulatory package governing capital, liquidity and leverage requirements for banks) would make this very expensive – banks would incur the substantial regulatory costs associated with a small business overdraft (in the form of additional capital and liquidity) even if the customer never used their facility.
Banks could still, of course, operate both within and outside the ring-fence at once, and might even be able to cross-subsidise their business within the ring-fence from what they make elsewhere, though not in real time. For instance, a division of BankCorp called, say, BankCorp Commercial, could provide loans outside the ring-fence and a division called BankCorp Retail could provide overdrafts and current accounts within the ring-fence. The two could be able to share information on a commercial basis, subject to the customer’s permission, and could even be made to appear to offer a seamless service, but would in all other respects have to be separate businesses.
But the question remains: if BankCorp Commercial can’t tap into deposits in order to lend, why would BankCorp want it to permanently cross-subsidise BankCorp Retail? The answer is: it wouldn’t. In fact, people at the Treasury clearly know this:
Banks may face a loss of diversification in the long term as banks will no longer be able to cross-subsidise or cross-sell services between the ring-fenced and non-ring-fenced bank. There will also be upfront transitional costs (such as establishing new subsidiaries) and ongoing costs of operating two entities rather than one (such as operating separate IT platforms).
This model would essentially split all major banks into two parts: a ‘Smart Bank’ (BankCorp Retail) and a ‘Dumb Bank’ (BankCorp Commercial). What makes a ‘Smart Bank’ smart is that it owns proprietary customer information and is therefore able to make informed decisions about their creditworthiness with some hope of a competitive advantage; however, it cannot carry any loans on its balance-sheet. The Dumb Bank has access to the wholesale markets and their cheap(er) funding, lower capital requirements, expertise in managing portfolios of loans and pooled risks, and access to financial engineering on an epic scale; but it cannot decide on the creditworthiness of individual SMEs. Essentially, it’s an investment fund that specialises in business loans.
So far, so good, one might think. Except for that fact that, as we saw earlier, the Smart Bank’s small business operation is almost guaranteed to be unprofitable unless it is able to charge someone a fee or commission for its services – either the customer or the Dumb Bank. This is because the vast majority of small and medium sized enterprises have very few dealings with their bank beyond maintaining a current account and possibly an overdraft. The Smart Bank could, in theory, charge all of its customers a fee for these services. But given the reluctance of most SMEs to pay for this, it would be more likely to make its money out of the few candidates that turn out to be suitable for a loan – either by charging them for the privilege of bringing their case to the Dumb Bank or by charging (gasp!) the Dumb Bank in return for originating loans. Or a bit of both.
Science fiction? Absolutely not. It’s already happening – SME loan funds are paying banks to originate loans for them as we speak. Elsewhere, peer-to-peer lenders are taking advantage of the fact that, as originators, they don’t need to hold on to regulatory capital (and the individuals carrying the risk don’t have to either) in order to undercut the banks. The incentives provided by regulation are as discreetly irresistible as gravity.
But if that’s the case, then why stop there? The Smart Bank could, in theory, originate loans for any Dumb Bank. Similarly, the Dumb Bank could buy loans originated by any Smart Bank, or even non-bank lenders, if the risk and return profiles are good and allow it to build a diversified portfolio offering decent returns. It could even buy securities made up of multiple loans in this way. Some Dumb Banks would have the same risk appetite as today’s high street banks. But newcomers need not.
In the medium term, there could be substantial advantages to this model in concentrated SME banking sectors such as the UK’s. Unlike today’s banks, competitors to the Smart Banks could spring up overnight, and risk carriers (the Dumb Banks) would have to compete aggressively for access to the best SME debt with each other and any other major investor interested in the asset class. Customers could switch easily between originators by simply moving their transaction data (which, I am told, is already possible) and without changing the carrier of existing loans.
Government could intervene in order to resolve market failure in ways that have hitherto been impossible. Its major disadvantage (a lack of know-how in assessing credit risk) would become irrelevant as it would only need to set itself up as another Dumb Bank – the best-capitalised and most cheaply funded of them all. If recent announcements of a British Business Bank ever produce a useful policy tool, it will almost certainly be under this model.
The problem is that this model works in exactly the same way as the ‘originate-to-distribute’ model that failed so spectacularly in the US mortgage market and beyond. The Smart Bank has only a weak incentive, especially in good times, to provide the Dumb Bank with high quality loans; the Dumb Bank has little incentive to perform the kind of expensive due diligence that would address this (theory here). The model also invites all of the problems associated with advisers working on commission that the FSA’s Retail Distribution Review was set up to address.
A government Dumb Bank could distort the UK SME loan market with its cheaper financing, like Fannie Mae and Freddie Mac did for the mortgage market in the US – although thankfully EU State Aid regulations should help rein it in.
But worst of all, the Dumb Banks, lacking access to a retail deposit base, would have to finance all SME loans almost entirely from the wholesale markets. A sudden tightening there, as in 2007, would bring the entire SME loan market crashing down in an instant.
This last point brings us to the heart of the matter. When people ask for retail and investment banking to be kept separate, regulators need to be able to see beyond the simple phrasing and realise that this means keeping the retail and wholesale financial markets from influencing each other. To be fair, they almost certainly do, but are not prepared to deal with the reduction in credit supply this would lead to. That’s one thing. But in the case of SME loans, the ring-fence just might achieve the exact opposite of what was intended. Wouldn’t it be ironic if the regulations enacted in reaction to the financial crisis were to end up producing some of the exact same incentives that contributed to the crisis in the first place?