Friday, June 10, 2011

BDUK second wave funding allocations - doing the maths

It's interesting to analyse the funding allocated to the three successful Broadband Delivery UK (BDUK) second wave bidders. According to the Department for Culture, Media & Sport's press release, issued two weeks ago today, funding allocations "are in the region of" £30m for Devon and Somerset, £15m for Norfolk and £4m for Wiltshire, with exact figures to be finalised in coming weeks.

BDUK's bidding guidance for the second wave, published earlier this year, had this advice for bidders:
"The Local Broadband Plan will need to identify the potential phasing of any capital funding sought from BDUK. The bid for funding should sit within a BDUK advised notional grant of £60 per premise, which can be flexed to take account of factors such as topography, population density and network architecture. BDUK is currently agreeing a funding allocation approach and expects to be in a position to confirm an indicative budget for BDUK funding allocations to local bodies on submission of their bids. BDUK would also be able to provide an indicator of what type/scale of requirement its modeling suggests may be feasible, within an overall budget that includes an assumption of additional public sector funding (arranged by the local body) and private sector funding (from a successful supplier) available."
So what's the origin of this notional grant of £60 per premise? Well, according to official government figures, there are currently 26m households and 2.1m businesses ("enterprises registered for VAT and/or PAYE in March 2010") in the UK. As we know, BDUK funding is intended for the final third of premises which won't be reached by the market's roll-out of next generation broadband. From the BDUK bidding guidance again:
"It is estimated that the private sector, given their currently published plans, will cover approximately two thirds of premises with access to superfast broadband by 2015. This leaves approximately one third of premises (or approximately 9 million premises) potentially within scope for public sector intervention in superfast broadband. We estimate that up to 1 million of these would prove too costly to rollout superfast broadband by 2015 and so are potentially out of scope for intervention in superfast broadband."
The total amount of funding available to 2015 is £530m, which a potential extra £300m available to 2017 if required; £530m divided by 9 million premises (the final third) gives approximately £60 per premise.

So how does this fit with the funding allocations made to the three second wave winners? According to official government statistics, the number of households and business premises in the three successful regions break down as follows:
  • Devon & Somerset: 882,000 households + 52,691 business premises = 934,691 total premises
  • Norfolk: 370,000 households + 21,569 business premises = 391,569 total premises
  • Wiltshire: 188,000 households + 11,049 business premises = 199,049 total premises
Some caution required, as the above figures relate to 2008; the household figures are available here  (Table 406: Household projections, by district, 1991-2033) and the business premises figures are here (go to Topics > Physical Environment > Commercial and Industrial Floorspace and Rateable Value Statistics). The business premises figures are based on five different types of commercial and industrial premises: retail premises, offices (disaggregated into commercial, and 'other' office types), factories, warehouses and other bulk premises.

So, if we apply the final third principle to these regions, we end up with 311,564 final third premises in Devon and Somerset, 130,523 in Norfolk and 66,350 in Wiltshire. Which is very interesting, as if you divide the respective funding allocations (to recap: £30m for Devon & Somerset, £15m for Norfolk & £4m for Wiltshire) by these figures, only one region (Wiltshire) matches the £60 per premise figure suggested by BDUK. The others are much higher - £96.29 per premise for Devon & Somerset and £114.92 per premise for Norfolk.

Of course, another (and much simpler) way to analyse this is to divide the funding allocations by £60 to indicate (?) the number of premises the grant is intended to support. For Devon and Somerset, this results in 500,000 premises and 250,000 premises for Norfolk (for Wiltshire, the result is the same as the previous calculation, obviously).

Which leads me to two possible conclusions: either that connectivity per premise is significantly more expensive in Devon, Somerset and Norfolk (doubtful), or (more likely) that significantly more than a third of properties in Devon, Somerset and Norfolk are beyond the reach of commercial superfast broadband roll-outs. Wiltshire provides a good fit with the final third model, whereas more than half of premises in Devon, Somerset and Norfolk will require intervention if superfast broadband services are to reach them. Percentage-wise, this equates to 53% of premises in Devon and Somerset and 64% in Norfolk. Final Two Thirds, anyone?

So, do BDUK's funding allocations provide a metric to illustrate differing degrees of rurality across regions? Interesting food for thought, you can find a spreadsheet summarising my calculations here.


  1. I arrived at the £60 (ish) some months ago

    I think it's fairly obvious that the final third of households isn't going to be a uniform statistic across all local authorities. If you look at current cable, FTTC and LLU penetration in Norfolk and Lincolnshire compared to Greater Manchester it gives you a fair clue.

    I do think it will cost more per premise in sparse or remote areas, whatever the technology. Equally there are some counties with a lot more sparse and remote than others.

  2. Anonymous10:39 am

    Intersting analysis. I think you might need to dig down to a lower geographic level to get to the bottom of the story. The Analysys Mason risk assessment done for CLG uses LSOAs You might also want to refer to DEFRAs categorisations of urban and rural areas too. Each of the winning areas contains some sizeable towns and cities which need to be factored into. I suspect that these numbers are all subject to change right through to the procurement process anyway.