costs

This page considers the cost of national identifier and registration schemes.

It covers -

• introduction
• information
• cards
• devices
• databases
• cost/benefit analysis

    introduction

We have sometimes been asked "how much will the Australia Card cost?", with one activist in 2005 requesting a specific figure on a per capita basis. In practice there can be no definitive answer to such questions. That is for two reasons.

The first is that there is no agreement about what a future Australia Card (or overseas counterpart) might comprise and how it would be funded.

Would, for example, consumers be expected to make some payment towards a standard/enhanced card (in the same way that fees are charged for driver licences and passports)? Would costs be bundled - critics say disguised - in government spending on universal entitlement schemes such as an enhanced identity/payment card for pharmaceutical benefits and medical services? The UK government thus controversially claimed in 2005 that there would be some 'cost recovery' from bearers of its proposed card but that many costs would be absorbed by using existing passport system infrastructure.

The second reason is that there is disagreement about the identification of costs.

Some proponents for example claim that the only 'real' cost is the price of the bit of plastic or paper. Critics, often as disingenuously, assert that national identity schemes necessarily cost billions without commensurate benefits.

Such assertions may not fully recognise that some costs are sunk in data collection/processing programs that will be conducted by government irrespective of whether a national identifier regime is introduced.

In practice, assessing the likely cost of a national identity card regime is akin to going on a random walk with a set of very large numbers. The following paragraphs highlight some costs and concerns about forecasting.

The experience of Australian and overseas governments with recurrent problems through poor specification, major project creep, system integration difficulties, institutional empire building, vendor capture and tacit acceptance of open-ended costing or 'compromises' means that some observers accept proposals for a new card in principle but are wary of what that regime might cost on an ongoing basis.

    information

We have commented that one concern regarding contemporary identification schemes is that there appear to be substantial numbers of identity papers/cards that have either forged, corrupted or merely improperly issued. Some people, for example, have two drivers licences - both of which have a correct photo but are issued to two identities. A fundamental cost of an Australia Card scheme reflects the need to ensure that the information on a card correctly represents the individual.

Collecting biometric information such as fingerprints or retinal scans from all adult Australians and incorporating that data in cards and databases is a daunting task, one that is often understated by card enthusiasts. It would involve, for example, physical contact between data subjects and an official or government agent (such as a general practitioner). That contact would be more demanding than traditional arrangements - such as the 100 Points identity verification scheme - where a clerk quickly scrutinised an individual's papers and compared a face in front of the counter with a Polaroid snap. Accurate collection of fingerprint or other identifiers would take more than five minutes.

That data would then need to be validated if a card regime is not to embody 'garbage in, garbage out'. The cost of that validation to individuals is rarely quantified. The cost to government in data collection (paying officials and/or agents) and verification would necessarily be much greater than current expenditure on maintaining the passport system. Some critics in Australia and overseas have questioned whether

    cards

The cost of the card that embodies each individual's information is dependent on factors such as the sophistication of that device, its robustness and the number issued.

Characterisation of the card as a device reflects calls for the government to adopt a smart card: one that has some encryption capability, that features discrete digital 'wallets' or files (eg separate public/private health information and other information) and that holds several biometrics. Such a card would go some way to addressing concerns in parts of the community but would be more expensive than traditional bits of plastic with a magnetic stripe or merely the bearer's details and photograph.

Robustness is an issue, as faulty cards and readers will need to be replaced as they wear out. Some allowance needs to be made for cards that are lost or stolen. Overseas proponents and critics of such cards suggest that annual replacement rates are between 0.1% and over 5%.

Depending on the particular scheme costs would be reduced through economies of scale if there were was substantial takeup, whether on a compulsory or voluntary basis

    devices

Proponents of 21st century national identity verification regimes have typically argued that effectiveness is dependent on cards being electronically checked in real time against one of more databases. In essence, there is little point in extending the existing driver licence regime, which as noted earlier in this note relies on observers intuiting that the card is authentic because it "looks right" and capturing the name, birthdate, address, registration number for matching with a register.

Some technology will therefore be required to read the card. The sophistication of the reader (a magnetic stripe device, an RFID reader?) and its deployment would be dependent on expectations about how the regime would operate.

Enthusiasts often forecast that readers would be ubiquitous, in use for example in all hospitals, medical practices, government offices, universities, tax agents, real-estate agents, banks, airports and any other location that requires identification for public or private purposes. Others have been more cautious, highlighting concerns if devices proliferate until they are in use in cafes and video rental shops.

    databases

For most specialists the key concern about an Australia is how it relates to one or more databases, rather than the shape and cost of the bit of plastic that might be carried by all or most residents of the country.

Will the card embody a set of discrete files, each linked to databases that are quarantined in order to inhibit inappropriate data-matching and reduce security risks. Will such databases be new or an enhancement/extension of existing data collections. Will information instead pass through a central gateway? Will different databases communicate with each other, as proposed by some advocates for the purposes of data verification and the identification of offenses or predictive activity?

We do not know the answers to such questions. They are pertinent because they affect the cost of an Australia Card regime. They also reflect uncertainties about its effectiveness. Successive reports by the Australian National Audit Office, along with embarrassing system failures and criticism by prime contractors, demonstrates that federal government agencies have experienced difficulty with specification and development of large scale systems. Performance has often been underwhelming and costs have often blown-out.

That history is understandable, given difficult relationships between some agencies and contractors, large-scale project creep that reflects ongoing changes in legislation and evolving ministerial visions, the difficulty of designing complex systems without a clear sense of what they are meant to achieve (and how they are to be measured), conflicts between different agencies, and management inadequacies in particular agencies (including a tendency to embrace glitzy offerings without a deep understanding of what's being offered).

One conclusion is that specification, development and maintenance of the 'back end' of a ubiquitous national identifier regime is likely to be significant more expensive than projected by most agencies.

    cost/benefit analysis

A traditional - and often disingenuous - response to concerns about the cost of national identity schemes is to ask "what price national security".

In practice it is difficult to offer a coherent and comprehensive cost/benefit analysis because

• details of costs are unavailable or are disputed
• details of benefits are unavailable or are disputed
• forecasts often prove to be pessimistic or optimistic

   next page (overseas models)