Observant readers of the financial press will have noticed that the LINK board, which manages the country’s main ATM network, has announced phased reductions in interchange fees. These are paid by banks when their customers use ATMs owned by another bank or by an ‘Independent ATM Deployer’ – IAD. The proposed reductions total 5% a year for four years, starting in July 2018, with an exemption for ‘remote’ ATMs and a strengthened Financial Inclusion Program – see http://bit.ly/2s334mg for the full story.

I say ‘observant readers’ advisedly, as those of us enjoying ‘free in credit banking’ are not of course impacted by this change. Our own bank absorbs the costs of us using an ATM owned by a different bank or IAD.

Even more observant readers will have noticed some fallout from this announced change – one source quotes ‘experts’ as warning of ‘cash machine deserts’; and the FT reports ‘ATM operators in turmoil over Link move to cut fees’.

So did the LINK board get it right? Or are we facing the end of our beloved cash machines?

In theory, of course, the answer is easy – the marginal net benefit (MNB) of an extra ATM in a given area will decline with the number of ATMs, and presumably the marginal cost curve for additional ATMs rises – non-bank ATMs are more costly than bank ones, and some locations are more costly to serve than others. So the number of ATMs in an area should grow until the marginal cost of an additional ATM exceeds its MNB. And if we’ve done Econ 101, we might think that the pricing mechanism should have some role to play in bringing about this happy equilibrium.

But this is not of course how either the LINK scheme works. For most customers, free access to LINK ATMs is part of their current account bundle, so there is no price signal for customers if the costs of ATM provision is ‘too high’. And on the supply side, the interchange fee is set on the basis of a cost study, which effectively allows for the recovery of the total costs of operating the LINK network. Cost-based recovery mechanisms, widely used in regulated utilities, are known to cause problems, as they weaken incentives to cost control. But more seriously, for the ATM network, they offer no mechanism for limiting ATM numbers: if an operator can put in a new ATM with below average costs, and with little impact on transaction volumes, it is in their interests to do so. This is why, in regulated utilities using cost-based pricing (usually cost-plus to allow for a profit element), the regulator normally has some control over the volume of costs being remunerated – so, for example, how many nuclear power stations we build.

But even if equating MNB and marginal cost is not that simple, the methodology does tell us something – in particular, if the NMB curve is shifting to the right – ie an additional ATM is less valuable to consumers than it used to be – the ‘optimal’ number of ATMs will fall. And there can be little doubt that, with the rapid growth of contactless payments, this shift is occurring. Visa Europe reports, for example, that, by 11 May 2016, three billion contactless contactless transactions had been made in Europe in the last 12 months – nearly tripling the volumes in the previous year.

However, as the LINK documentation shows in a revealing graph, while the volume of cash payments in the UK has been declining, the number of free to use ATMs has been growing – a trend particularly pronounced over the 2013 to 2016 period, and one which is expected to continue. This could only be ‘optimal’ if the number of ATMs in the UK was, prior to this shift, significantly sub-optimal. But there is no evidence that this is the case. International comparisons, for example, show the UK as being ‘well stocked’ as far as ATM numbers are concerned.

So as I look at the evidence I conclude that the LINK board have done a good job. There is clear evidence that the continuing growth in ATM numbers is unlikely to be warranted. The only lever they can pull to change the number of ATMs is the interchange fee, and they have made modest adjustments (5%) to that for this year, with further changes to come. This is after published consultations and specialist enquiries from LINK itself and the PSR, and buttressed by comprehensive monitoring arrangements and a ‘hard review’ in 2019 before the rate cut exceeds 10%. And the Financial Inclusion Programme (the one area where serious detriment might occur) has been strengthened.

The reduction in interchange also helps mitigate the competitive threat from other networks who can offer lower costs. Arguably, the LINK board would be failing in its duties if it had not taken some action on this.

There are of course winners and losers, as the consultation shows – the major losers being the IADs. But it is not the job of an industry regulator (the PSR) or the Treasury Select Committee to ossify the payments landscape in the face of significant changes in payments media and competitive threats. Not all changes are changes for the worse.

And the golden ATM? The picture is of the first ATM opened in the UK in 1967, in Enfield, now painted gold. Perhaps regulators with an innovation remit might like to ask why the functionality is still much the same?

SaveSave