Beyond the public sector

What’s the problem? Open Data Camp 6 in Aberdeen discussed how to get SMEs to publish and use open data. And session leader Prateek Buch from the DCMS said that was important.

But he also wanted to discuss the bigger issue of how public policy could support other kinds of organisation to release data as open data.

Public commons

One participant suggested that companies that benefited from public policy should release data as open data almost as a quid-pro-quo for the support they received. A participant from the oil industry said this was common in his sector, where firms had to release information after a set number of years in operation. Another suggested that companies that bought public land might be required to release information about what had happened to it.

Prateek said this pointed to an issue of public commons. If there is a public good to be derived from data, should be collected and published as open data? As an example, he discussed information about anti-microbial resistance.

This, he pointed out, is held by the NHS and the pharmaceutical industry. The Wellcome Trust and others want the pharmaceutical industry to start publishing some of its data. It is meeting resistance from a large, well funded industry; but there is clearly a public good in getting it out.

Levers

Jack Hardinges from the Open Data Institute said he had been thinking levers for encouraging non-public bodies to publish information, and identified at least four:

  • Do nothing: Uber has published its movement platform (possibly under the pressure of having it published)
  • Compel transparency through statutory accounts as part of a contractual obligation: New York has obliged Uber to share data as part of its condition of licencing
  • Compel publication to avoid the formation of natural monopolies – or to mitigate their ill effects
  • Encourage publication because there is a wider social benefit.

The problem, participants pointed out, is that while the argument is compelling, companies may respond by threatening to withdraw from a market: it is far from clear how US IT companies are going to react to the plans unveiled in the Budget to impose new taxes on them.

Also, small competitors forced to publish information might lose out to larger competitors; or be bought out by them. Or the requirement to publish information, and the cost of doing so, might act as a barrier to market entry.

To address the last point, existing schemes that require companies to collect and publish information put thresholds on the size of enterprise covered: only companies with more than 250 employees have to identify their gender pay gap, for example. However, this kind of banding can still create “cliff edge” barriers to entry – or expansion.

Unintended consequences

The session went on to discuss some of the unintended consequences of publishing and not publishing information. One participant argued that ‘allowing’ companies to hold and use sophisticated data-sets might increase inequality: by enabling them to target services on profitable sectors and areas. Others argued that publication might have the same effect: by exposing issues or enabling competitors to target the same sectors.

The session also discussed what the users of services might think of their data being published in this way [session notes are here]. Participants felt that individual users trying to remove their data from a large data set would make little difference to it; but publication might well affect the perception of the organisation in question.