You see that, as Simon points out in his piece, in their needing to revise China forecasts upwards for the post-subsidy era. It's not "new policy surprised the IEA". It's that they still don't get how fast solar / storage / EVs are dropping in price, or that it will continue.
1
1
4
Good points, not going to disagree. Though, can I throw a question back (I do not follow solar markets that closely). Were you (& others in the industry) surprised by the growth in China in solar (in 2020)? Or was it, nah, whatever, nothing to see there? Old news...
1
2
China's solar growth in 2020 surprised me. Especially given Covid. In a way the 2020 China surge is an artifact, driven by a policy cliff. We've seen similar (if smaller) things in the US - a lot of new projects coming online just in time to make a calendar year policy deadline.
1
1
1
I think missing China's 2020 solar surge, by itself, isn't a sign of bad forecasting. But the long-term pattern of the IEA is pretty clear. And not only on solar, but also EVs, energy storage, and offshore wind. There's a deep conservativism in their modeling of the future.
1
4
I'm not actually sure that cost assumptions are the key to this any more. IEA uses capital costs from IRENA and In last year's WEO it made major updates to assumptions on finance costs, which had been an issue. iea.org/reports/world-energy…
1
4
They now use VALCOE, some variation of LCOE that includes system costs. I am not sure how that effects results. There was a commentary on it, but I can't find it. Some info here, towards the bottom. iea.org/reports/projected-co…
2
2
IEA says they use IRENA cost numbers. But when you dig in, IEA uses a much more conservative learning rate than IRENA, thus underestimating future cost declines. And that's just solar. Bigger issue in batteries & EVs.
2
2
well, I haven't looked at learning rates so that could be an issue; not sure I've seen IEA explicitly expressing assumptions for this?
1
1
In WEO 2019, IEA stated in the text that they used a 20% learning rate for solar. (Though their actual cost inputs look even below this.) IRENA shows a >30% learning rate for solar. I'll have to check WEO 2020.
2
2
I looked at IEA's future solar cost forecast (WEO 2019) vs a simple 30% learning rate in the second half of this: rameznaam.com/2020/05/14/sol…

May 12, 2021 · 8:22 AM UTC

1
1