Skip to content

Is the UK Government Concealing ‘Very High’ Renewables System Cost Estimates?

Dr John Constable: GWPF Energy Editor

After an unexplained delay of a year since completion the UK’s Department of Business, Energy and Industrial Strategy (BEIS) has published (24.03.17) a report by Frontier Economics on the total system costs of uncontrollably variable renewable generators, a topic of crucial importance in understanding the cost-effectiveness of current climate policies. The study is not only very late, but entirely qualitative, and contains no quantitative estimates of additional system costs per megawatt hour (i.e. £/MWh), figures which would normally be considered the principal output of such work. However, examination of the peer reviews, which are published with the study, reveals that an entire table of numerical cost estimates, some of which were described by the external reviewer as “very high”, were in fact present in the version sent out for comment in mid 2015, but have been subsequently removed. This does not smell right and BEIS should release the original draft.

On Friday last week the United Kingdom’s Department of Business, Energy and Industrial Strategy (BEIS) released, without fanfare or detailed comment, a long study commissioned from Frontier Economics on the Whole Power System Impacts of Electricity Generation Technologies, which is principally concerned with the system costs of uncontrollable generators such as wind and solar. Accompanying the text are the comments of three peer reviewers, and an Addendum constituting the Department’s response to the reviewers.

Rumours of this work have been in circulation for some time, and it was trailed in a government statement before Christmas (Electricity Generation Costs (November 2016), 21), so it was hardly a surprise, indeed it was eagerly awaited.

Remarkably, the document is dated February 2016 (note: two thousand and sixteen), and the peer reviews are dated late July to very early September 2015. The Department has been in possession of the final work for over a year before releasing it to the public, and aware of its conclusions for something like a year and a half. Why the delay?

However, closer examination reveals that this suspicious cellar-ageing is almost the least surprising thing about the paper. The principal output of any complete study of the system costs of uncontrollable generation must be quantitative estimates of those costs per unit generated, say pounds per megawatt hour (£/MWh), but the study as published contains no such quantitative estimates at all, and consists almost entirely of qualitative reasoning concerning methodology, much of which is interesting and potentially valuable, but is entirely preliminary to quantitative calculation.

It appears that the consultants, Frontier Economics, are not to blame for this omission, since their draft as sent to the Department of Energy and Climate Change in mid-2015, did in fact include such cost estimates. This can be inferred from the peer review comment of Professor Lion Hirth, of the German consultancy Neon, and a well-known authority on system cost analysis. In his review for DECC, Dr Hirth wrote in July 2015:

Size of reported costs. I disagree with the statement that whole system impacts are “small relative to the direct costs”. For wind and solar power, they can easily [be] in the same order of magnitude as own generation costs at high penetration rates, if differences in displaced energy/capacity are included. Displaced generation cost, likely the most important category, is missing in table 3. I also have doubts about the numbers itself, which seem awkwardly high in some cases (see comment (1) regarding Section 5 below). (p. 3)

While the comment is interesting in itself (the system costs of renewables are high in relation to capital equipment costs) the point of most importance here is that Professor Hirth was basing his remark on a table of numerical data. This table, Table 3, is evidently not present in the text that BEIS has just released, since there are no numerical tables at all, but the point can be confirmed by reference to the BEIS Addendum, which helpfully provides a key to changes in pagination and changes in figure and table numbering to assist readers in following the comments of the reviewers. BEIS lists no equivalent for the original Table 3.

Table 3, then, is one of the tables to which BEIS refers when it writes (Addendum, p. 4) that the published text “excludes a couple of tables that were deemed unhelpful or unclear”. This is sloppy language betraying a bad conscience.

Given the sensitive nature of the subject, it is entirely reasonable to suspect deliberate suppression of embarrassing estimates. Professor Hirth gives us the clue, in a section that I will quote in full:

Major concerns with the size of estimates. You write that system costs are “relatively small” (p. 95), but then present numbers that are quite high, at least the upper bound. I think this is inconsistent.  More importantly, I have major concerns with Table 3 / 22. (i) It is not clear to which penetration levels these estimates refer. (ii) It is not clear if they represent marginal or average values. (iii) “Displaced generation costs”, likely the most important category, is missing. (iv) You should guide the reader how to interpret the sum of the cost components. (v) The upper end of the balancing and grid costs are very high; likely they represent extreme assumptions or “bad practice” examples. (p. 15)

BEIS should without delay publish the unredacted draft study as seen by Professor Hirth so that we can all make up our minds on these important points.