There have been various press reports recently about research by LSE academics on ID cards. Contrary to press reports, the authors have NOT issued or reached agreement on any figures other than those contained in the June report.
This short briefing note aims to clarify some key points.
In June 2005, following a six-month study, LSE published a report on the challenges, risks, opportunities, costs and implications of the government's identity card proposals called The Identity Project.
In the five months since then, the academics involved have continued research, consulted widely on numerous aspects of the original report, and clarified earlier questions with Home Office representatives - most recently at a meeting on Wednesday 9 November at the Houses of Parliament with the Rt Hon Andy Burnham MP, peers and industry guests attending.
Based on the statements of the minister responsible for the project, LSE academics now understand that:
The current government cost estimates have been ring-fenced around the Home Office and include no connections, implementation or interaction with other departments;
Other public sector organisations (local authorities, Department of Work and Pensions etc) will be involved in the scheme only on an 'opt-in' basis for which a business case and budget must be internally raised. According to government estimates, there are over 265 such agencies.
Therefore, the projected costs as stated in the Bill's regulatory impact assessment (RIA) relate only to the costs incurred by the Home Office to develop the National Identity Register and registering and issuing cards from the new national identity agency. The June LSE Identity Project approach was focused on the rollout cost of the system in accordance with the stated policy goals of combating crime, terrorism, identity theft, and benefits fraud. This difference of approach explains in part the difference in respective cost figures.
Expressed differently, the government had costed the scheme as a 'stand-alone' project, only for use by the Home Office. In order to fulfil the Bill's policy goals, the scheme requires a 'triangulation' infrastructure involving at the one edge the individual, on another the ID card infrastructure and register, and on the final edge the IT system of the relevant user organisation (eg a local council or GP). In this way the ID number and audit history can be recorded on the system of the user organisation while at the same time user organisations can be checked and verified by the register. Each organisation's IT system would have to be modified to incorporate the identity number. These integration costs across the 265 government agencies have not been factored into the RIA.
These integration costs may be reduced or even minimized by the upgrade of government IT systems over the next ten years that would be happening anyway. The government seems to take the view that additional functionality covering connections to the National Data Register can be built into new or replacement IT systems at little additional cost progressively over the next decade. But the cost of system implementation and integration across the public sector may be substantial, and even within the narrow terms of the RIA is likely to exceed the £5.8 billion originally suggested by the government.
While LSE's June costs model remains valid, LSE report authors are not prepared to estimate a single figure or a range of figures until the full specification is better understood.
Next step - Professor Ian Angell of LSE's Information Systems department has written to Mr Burnham seeking a detailed response on integration and cost points. The letter also asks about the likely take-up in the private sector, including the extent of legal compulsion and liability.
Given the government's assurances on a number of points (for example, that registration will involve a significant degree of automation), the authors believe it is likely that their original median cost projection will at first be reduced. However in the light of other new information there remains concern about the viability of biometric technology on the scale proposed, while the security risks outlined by industry leaders have increased.
New information from the Home Office provides assurance that the LSE's June report concerns regarding the scheme and its likely design implications and costs are valid;
In the June report LSE academics were conservative on the costs to the government as a whole, as we did not include the costs to all government departments, agencies and local authorities;
Continuing concerns about the performance of biometric technology will significantly affect the final cost assessment
LSE academics aim to publish a smaller, follow up report on the government's ID card proposals once Home Office clarifications are made, ideally before the end of December.
Thank you for your time and interest.
Professor Ian Angell, Information Systems, LSE
Professor Patrick Dunleavy, Government, LSE
Dr Edgar Whitley, Information Systems, LSE
Dr Gus Hosein, visiting fellow, Information Systems, LSE
Simon Davies, visiting fellow, Information Systems, LSE
on behalf of the LSE Identity Project group
17 November 2005