We don't need *a* Public Data Corporation, we need a plurality of competing Public Data Corporations

Yesterday we submitted evidence (pdf) to the Government's consultation Data Policy for a Public Data Corporation.

Update 15-Dec-11: We welcome the government's announcement today at an RSA/2020 open data round table where Cabinet Office Minister Frances Maude said "we're moving away from a model where government tries to gain in the short term by extracting value from public data, and moving towards long term benefits [to the UK economy]."  Open Digital understands the plans for a Public Data Corporation have been replaced with a Public Data Office, and it's not just a change in name.

Although the Cabinet Office consultation asked some specific questions about its policy of creating a Public Data Corporation - a central agency to facilitate release of public data sets - including questions about what services the corporation should be allowed to charge for, I decided to argue for not one Public Data Corporation but a plurality of competing Public Data Corporations.

The idea from government seems to be (phase 1) create a Public Data Corporation; (phase 3) profit.  The government seems to think direct returns can be made either by floating the Public Data Corporation and/or charging for access to certain data sets and passing-on some of the charges to the originating departments.

Our submission outlines many of the secondary benefits to free and open public data.  It will keep the barrier to data innovation low; tinkervation - tinkering with data sets despite there being no apparent business model to do so, will only happen if data is free and open.  The UK economy will benefit if UK businesses profit from open data, and the public will still benefit through a range social, democratic and educational uses of data in cases where no taxable revenue is created.

I then argue that the idea behind an investment-led approach to public data is flawed, mainly because we don't know enough about (a) how much of the data held will be used - that's the job of data innovators; and, (b) how online content will be funded in future. The online ad sector is booming, with stacks of cash being spent, but this isn't - as yet - trickling down to online publishers.

There is a real danger that data sets will be under- or over-valued. We simply don't know what future uses will be found for much of the data, plus there will be a large amount of uncertainty over future ad revenues until the online ad market matures; meanwhile, we will have based important public policy decisions on a nascent market valuation.  Some products won't happen, whilst others will lead to profiteering.

I also highlight a secondary problem. the Public Data Corporation will hold a de facto monopoly over data processing, creating a whole series of problems with pricing and tendering for the necessary IT infrastructure.

To solve many of the problems I propose encouraging the establishment of many Public Data Corporations, with input data sets being available to any qualifying organisation, and a series of public subsidies paid on a per-megabyte basis for data served.  Competition should make the Public Data Corporations lean and cost effective, helping the Government serve free and open data cost effectively.  Plus, a per-megabyte subsidy should reward services with fast transfer rates, as data consumers will shun slower providers.