A long-awaited private study by Cambridge University into the pricing of public sector information (PSI) by trading funds (Ordnance Survey, Met Office, Companies House, Land Registry et al) was published on the side with the 2008 Budget Report.
The study was commissioned by BERR following the OFT’s market study into the commercial use of PSI which found that the full benefits of public sector information were not being realised.
Entitled Models of Public Sector Information Provision via Trading Funds, the Cambridge study (154-page PDF) examines the cost and benefits for society, and the effects on government revenue, of different charging policies for PSI.
It’s worth first mentioning the standard economic argument set out in Appendix A.1 which rather dismisses the need for the study!
if the private sector is competitive and undistorted, then the public sector should sell any goods and services that it produces at the efficient price [which] is the marginal cost [and in the case of digital data, the marginal cost is near zero]. … If this argument is accepted then the case for selling PSI at marginal cost [ie near zero] is direct, and it is hardly necessary to estimate the extra social value that would be generated by moving from average to marginal cost pricing.
The paper is naturally filled with impressive arguments, charts and equations, but to sum up, it finds that:
socially optimal policy would involve leaving the charging regime for many [primarily “refined”] products unchanged, moving to marginal (zero) cost charging for [“unrefined” data]. [The trading funds] would now be in commercial competition with other suppliers as such suppliers would now have access to unrefined data at marginal cost [which] would immediately address the competition concerns raised by the OFT as, a fortiori, outside organizations would now have access to “unrefined” (“upstream”) data on the same terms as the trading fund itself.
So, in short, Government should Free Our (unrefined) Data so that we can all (business and community) get on with creating really useful applications, mixing and mashing it to our hearts content. Arguments are now likely to centre on what is and is not “refined” (or “value added”) data.
