Tuesday 20 August 2013

Measuring the UK's digital economy with big data

a research report by Max Nathan and Anna Rosso with Tom Gatten, Prash Majmudar and Alex Mitchell for the National Institute of Economic and Social Research

Key Findings
  • The digital economy is poorly served by conventional definitions and datasets. Big data methods can provide richer, more informative and more up to date analysis.
  • Using Growth Intelligence data on a benchmarking sample, we find that the digital economy is substantially larger than conventional estimates suggest. On our preferred measure, it comprises almost 270,000 active companies in the UK (14.4% of all companies as of August 2012). This compares to 167,000 companies (10.0%) when the Government’s conventional SIC-based definitions are used.
  • SIC-based definitions of the digital economy miss out a large number of companies in business and domestic software, architectural activities, engineering, and engineering-related scientific and technical consulting, among other sectors.
  • Companies in the digital economy have a similar average age to those outside it. Shares of start-ups (companies up to three years old) are very similar. Given the popular image of the digital economy as start-up dominated, this may be surprising to some. As digital platforms and tools spread out into the wider economy, and become pervasive in a greater number of sectors, so the set of ‘digital’ companies widens.
  • Inflows of digital companies into the economy have always been relatively small, given its sectoral share. However, using our new definitions of the digital economy, inflow levels are substantially higher.
  • As far as we can tell, digital economy companies have lower average revenues than the rest of the economy, but the median digital company has higher revenues than the median company elsewhere in the economy. Revenue growth rates are also higher for digital companies. However, these results come from a sub-sample of older, likely stronger-performing companies, so there is some positive selection at work.
  • Switching from SIC-based to Growth Intelligence derived measures substantially increases the digital economy’s share of employment, from around 5% to 11% of jobs. Digital economy companies also show higher average employment than companies in the rest of the economy (this reverses when we use conventional SIC-based measures of the digital economy). Looking at median employees per firm, the digital/non-digital differences are always a lot smaller. Our employment results should also be treated with some care, as not all companies report their workforce information.
  • The digital economy is highly concentrated in a few locations around the UK: Growth Intelligence software provides a fresh look at these patterns. In terms of raw firm counts, London dominates the pictures, but Manchester, Birmingham, Brighton and locations in the Greater South East (such as Reading and Crawley) also feature in the top 10. Location quotients show the extent of local clustering, which for the UK’s digital economy is highest for areas in the Western arc around London, such as Basingstoke, Newbury and Milton Keynes. Areas like Aberdeen and Middlesbrough also show high concentrations of digital economy activity.
Full text (PDF 43pp)


No comments: