ID cards and investment appraisal

Home Office minister Andy Burnham reckons ID fraud costs us £1.7bn a year. Let’s assume – contrary to Home Office practice – that this figure is remotely plausible. It allows us to treat ID cards as an investment project, and appraise it accordingly.

Let’s regard this £1.7bn a year as the amount we would save if we had ID cards. Government practice is to discount the gains from investment projects at a real rate of 3.5% a year. Over 10 years, then, the net present value of stopping ID fraud is £14.1bn.

The Home Office reckons ID cards will cost £5.8bn – that’s £93 per person for a card that lasts 10 years.

Orthodox NPV analysis, then, says the Home Office should regard ID cards as a good investment.

If, though, you take the LSE group’s median estimate of the cost - £14.5bn – they’re a bad investment. And remember, there’s a precedent for cost over-runs of much more than 140% (14.1/5.8) in public sector IT projects.

All this rests upon loads of assumptions, among them:

1. ID cards will wipe out all ID fraud.

2. We put zero value upon the civil liberties’ cost of ID cards.

3. We ignore real options effects. Even though the NPV of ID cards is positive in Home Office numbers, there may still be a case for postponing the project, if doing so allows us to explore ways of cutting the costs of the scheme.

4. ID fraud without ID cards will not grow in real terms.

5. We ignore cost/benefit case beyond 10 years. It’s quite possible these will be significantly positive – because there’ll still be ID fraud to stop but the costs of renewing ID cards will be lower, as the start-up costs will have been covered.

These assumptions (except 3) can be easily accommodated within the NPV framework; for example, if you think ID cards will stop only half of ID fraud, the NPV of the benefits just drop by half.

The ID card scheme, then, can be analyzed within an orthodox economic framework; we can see more clearly what assumptions we need to make to support or oppose the scheme.

That's the technocrat in me speaking. The rest of me thinks Burnham should shove his ID card up his jacksy sideways.

Comments

[The Pedant-General]

"The ID card scheme, then, can be analyzed within an orthodox economic framework; we can see more clearly what assumptions we need to make to support or oppose the scheme.

That's the technocrat in me speaking. The rest of me thinks Burnham should shove his ID card up his jacksy sideways. "

Guffaw!

[Jack]

Even if the costs are a wash there is a change in who is paying. That seems important both in terms of the initial redistribution of wealth from taxpayers to banks and in terms of incentives thereafter.

Fraud costs are currently born by the banks It is their job to minimise fraud. In other countries, Poland for instance, banks pay for customers to have unique ID dongles and proper password schemes without government intervention. Apparently it is not deemed worth the investment in the UK. Why subsidise the banks?

[rjw]

A bit of an aside, but when I started work as a government economist some 13 years ago, we used real rates of 6 and 8%. How times change.

[Chris Lightfoot]

"Why subsidise the banks?"

It's not clear how much of a subsidy it would be. The £5.8 billion includes only the running costs of the scheme which would be borne by the Home Office; if the banks wanted to use the thing, they'd (presumably) have to pay to integrate their systems with the Home Office's. Absent political pressure, I'd be surprised if they bothered; firstly, the Home Office scheme is unlikely to offer them any tangible security, and secondly the banks have never been that worried about (e.g.) ATM or credit-card fraud unless they threatened to be a significant embarrassment to them.