Tl;dr: He's seeing the effect of immature operators and mature technology. Onshore installations in Denmark, where operators are mature, show a performance decline with age which, though statistically significant, is perfectly acceptable.

Long version:
His statistical methods are perfectly standard for panel data analysis. The specification search is kind of sloppy, but his robustness checks seem thorough and come out fine, so I don't think that's a problem. Overall, I like his statistical methods and would probably use similar analysis myself.

The problems start to show up in the interpretation. The correct interpretation of his data is that:

1. Onshore wind is a mature technology, with perfectly satisfying performance when operated by mature operators (performance degrades by only four percentage points, or 18 % (2.0 % annualized) over a ten-year period - this is not alarming).

2. British onshore operators are not mature, as evidenced by their considerably steeper degradation of performance.

3. Either offshore wind technology or the Danish offshore operators are immature. Since this study demonstrates a strong effect of operator experience it is not possible to discriminate between the two, as Denmark ceded leadership of offshore installation to Germany early in the present century.

4. One worrying effect is that in the British data, capacity-weighted estimates display worse performance curves than unweighted estimates. This is a reverse of the expected effect (and indeed a reverse of the effect observed for both Danish on- and off-shore), and may indicate a major problem with British operators' management of large projects.

5. The policy analysis is utter garbage on at least two levels: It displays either nearly complete ignorance or mendacious mischaracterization of the existing policy regime, and further bases policy recommendations on the assumption that both the technology and British operators are mature, and that there are therefore no further infant industry economies to harvest.

6. The executive summary of the paper may be gainfully replaced by Figures 1, 10 and 14, which contain all the conclusions that are both interesting and accurate.

(Incidentally, nuke fans should take heed: Complicated technology depends on mature operators as well as mature technologies, and developing a mature operator and supply infrastructure takes decades.)

- Jake

Friends come and go. Enemies accumulate.