Decentralisation and data: Why they should go hand in hand

Policy makers in every country and at every level of government have access to more and better data than ever before. This should have greatly improved policy and public services and yet the results are underwhelming.

We can think of this as a narrower expression of the Solow paradox that “you can see the computer age everywhere but in the productivity statistics”. But in restricting ourselves to just thinking about public policy we can explore things more deeply and suggest some fixes.

Is data delivering?

While the promise of “smart cities” was never likely to live up to the unrealistic claims of its most excitable promoters, it also seems so far to have failed to live up to more reasonable predictions. Indeed, it is arguable that greater improvements to life in cities in recent years have come from the creation of bicycle lanes and the removal of car parking spaces by local government than from any smart technology.

In my city of Leeds, UK we can pay for bus and train tickets using a wide variety of cards and apps, which also provide timetables and journey planners. But journeys have got worse, ridership has fallen, real-time information is unreliable, and the data collected on journeys remains largely unused.

In a much more positive example, England’s national health service and its excellent integrated data platforms were crucial in proving and disproving the efficacy of Covid-19 interventions. This remains of huge global benefit. But it didn’t stop Covid outcomes in England itself being worse than its neighbours with less good data.

We should ask why data isn’t making things better by as much as we hoped and what we can do to achieve more.

Tom Forth

A centralising force?

More and better data has often enabled better policy and public services in about the same measure as it has enabled a centralisation of those services that leads to them getting worse. In the past, physical distance, variations in geography, and incompatible data made central planning hard to do well. Technological advancements today make many of these barriers smaller than ever.

More and better data increases the temptation to assemble national experts, design national dashboards and strategies, and plan from the centre. Centralisation sidelines the power of local democracy and the advantages of trial and improvement on small scales. It remains unclear if local experience can be replaced by newer forms of public input such as citizens assemblies and formalised user research.

The upsides of centralisation are obvious, we can see efficiencies of scale and imagine the benefits of the wider and faster application of proven methods.

The downsides of centralisation are less clear, yet threaten to undermine any technological gains. In particular, larger standardised datasets are more likely to either omit important local details or risk becoming unusably complex when all local details are included. And in disconnecting the responsibility for data collection from the power to act on that data we risk those with power in distant capital cities seeing problems well after the chance to act locally has passed.

Can industrial strategies be powered by local data?

With the return of active industrial strategy to fashion across much of the world we have a chance to avoid this pattern. Our work at The Data City classifies companies based on what they say they do on their websites rather than by the centrally-defined and increasingly poor statistical categories they are assigned. Our product enables the private sector, national governments, and local governments to identify promising companies within fast-growing sectors and clusters at both national and local scale. And we make sure that a part time researcher at a small local government can answer the questions they know matter locally just as easily as a large team of specialised researchers in a national capital city.

There is no reason that the availability of more and better data has to be followed by the centralisation of government. As more powerful tools become available more widely it can be followed by the opposite.

We already see early signs from the USA and France that industrial strategy works best where central government empowers rather than commands subnational units of government. Central plans work best when they are informed by local plans and are loose enough to allow local governments to deliver broader goals in their own ways.

More and better data can and should be a core part of industrial strategy and local growth policy, but to succeed we must remain vigilant that it does not lead to a policy of centralisation that may cost us many of the gains on offer.

Head of Data at Open Innovations | + posts

Tom Forth is Head of Data at Open Innovations, the CTO and cofounder of The Data City, and the boss at imactivate. He lives and work in Leeds, UK.

In 2013 Tom developed and maintained the Phonegap plugin for Moodstocks. He worked with Neil Clark on the 2014 release of Rusty the Squeaky Robot, an augmented-reality physical book for children with real-time translations and human voices. Later he worked with Jenny Lee on the release of Material Alchemy. He has specialisms in industrial structure and transport analysis and he has worked with Open Innovations. Tom writes for various publications about industrial strategy.

Tom is also doing The Data City, a spin-out from Open Innovations. He is using huge amounts of data to better understand economic development and the potential for innovation.