17 Comments

I'm not sure the assumption of stagnant electric demand is a good one. Despite the RE bust, demand continues to grow at over 3% (through Aug) this year. Moreover, don't we have to assume the Chinese population's standard of living continues to approach Western levels? This would imply continued increases in urbanization, as well as associated increases in per capita energy usage, including electricity. We're talking more air conditioning, washing machines, dryers, misc. appliances, etc. etc.

The real conundrum with your premise is the planned growth in coal power generation in China. There is 243 GW under construction/permitted, potentially reaching 392 GW when including planned but unpermitted projects. This suggests a potential 23% to 33% increase in coal capacity from 2022 levels.

Total power gen demand CAGR's of 3% at China's scale requires an all in on every generation source type of approach. The massive solar buildout, in tandem with the increased coal gen construction, aligns more with the continued growth outlook from policy makers.

Expand full comment
author
Sep 18, 2023·edited Sep 18, 2023Author

You see the prevalence of hand wavy loose stuff like this is why detailed modelling is required. This isn’t analysis. Household power demand is dropping hard in a lot of places because of efficiency upgrades, and thats before you consider what China’s demand might converge towards. McMansions in the Midwest - a wildly more heating intensive climate and massive sqft per person or, say, Italy? Because its about at Italy levels already and not far off Singapore (1MWh/cap/yr in China vs 1.28MWh/cap/yr) which has aircon running all year around. Have fun with this heating / cooling demand model. https://renewables.ninja/

Expand full comment

3% growth during a RE bust is not hand wavy loose stuff. Nor is the decision to build 243 GW of extra coal gen if, as you say, the demand does not warrant it.

Based on 2019 data, China is over 28% of global manufacturing, which requires energy. If China is going to to do most of the heavy industrial lifting for the developed world, it makes complete sense for them to have higher energy requirements per capita. The industrial sector in the US still accounts for 33% of total end-use energy consumption, and manufacturing is 76% of that. I'm sure these numbers are are much higher in China. But you're modeling it, so you tell me?

There's not a lot of smelting, refining and steelmaking going on in Italy.

Expand full comment

My assumption is always that any Chinese published statistic is hard to accept without a grain of salt. I belhebe the 243GW or extra coal gen in part because they dont gain much from lying about it (either optically or otherwise) but the energy demand numbers could easily be data massaging.

That being said, yes China is a giant in manufacturing and that doesn't seem to be changing any time soon. The question is whether these new coal plants will have the same coal appetite as their aging fleet does, if the fleet is purposeful overbuild (which seems likely given the hydro issues this summer), whether EV adoption increases electricity demand faster than their non-fossil energy output increases, and whether urbanization continues for the remaining 400M rural Chinese.

There's an extremely large set of conflicting factors so I think it's hard to draw any conclusions from published Chinese statistics and the only answer seems to be doing your own modeling with reasonable error bars on every unclear input.

If I had to guess, despite the RE bust, I would imagine energy demand will grow over the next 5-7 years before leveling off for a variety of reasons. I think peak coal for China is coming sooner than that in part because it doesn't seem like their total industrial output can grow much more and they are hell bent on driving the rest of energy demand from more reliable sources (whether that's solar, LNG from russia/central Asia, nuclear, etc). China is going to do everything humanly possible to avoid importing oil and coal.

Expand full comment

The more conspiratorial viewpoint flying around (from very serious investors I might add) is that China is watching/allowing the "West" to increase their manufacturing input costs via higher energy prices while China continues to buildout it's cheaper electric gen capacity via the coal route. This could allow them to remain competitive and maintain their manufacturing dominance for many decades to come.

A sidebar to this is that Chinese dependence on foreign crude oil imports is a state risk during any potential future conflict and I'm sure we all agree geopolitical tensions have been escalating recently. Switching to EVs or NEVs is being directly promoted by the Chinese govt for this reason. If you then begin modeling potential EV/NEV grid demand, it begins to make perfect sense why they would construct these extra coal plants.

It really doesn't make sense to increase your coal-gen capacity in 2023 if you're simply going to begin retiring the older portion of the fleet in 2026. And choosing to assume they're going to abide by official commitments after such a buildout is just plain naive.

Expand full comment

I think China's "all of the above" energy generation policy mix is to give themselves maximum options to handle as many unknown future scenarios as possible. I absolutely believe (it doesn't even seem conspiratorial IMO) to say that China wants to keep coal as an option for very cheap energy production for manufacturing. That being said, coal is still an external resource that they generally have to source from geopolitical rivals (Indonesia, Australia) and this is a strategic risk they are unlikely to rely on long term especially given how things went with their kerfuffle with Canberra in 2021.

As you noted, the exact same risk (and an even more important one than coal) is their foreign crude oil reliance. That being said, its worth remembering that road fuel demand represent less than 1/3 of their crude oil demand so even if they zero out their entire road fleet (which will take well over a decade) there's still an incredible amount (almost 10M bpd) of non transport oil use. Its unclear when their total oil demand will peak but I would expect that to happen in the next 5ish years but it will be a structural risk for them for decades unless they somehow find more oil domestically (unlikely).

I do not doubt that the retirement plans are merely suggestions not hard targets that can not be missed. China will see what percentage of the new fleet they end up using and make decisions about the existing fleet based on that. You have to understand that their decision making here is purely driven by risk avoidance, they had to shutter billions of dollars worth of manufacturing assets during the summer heat wave and that's an untenable situation for them given their mounting economic issues. Their coal fleet is already almost entirely loss making and barely at 50% utilization. The amount of new coal plants they are adding is only going to decrease that utilization rate whether or not they retire any of the old fleet.

My strong belief is that coal is an insurance policy not a planned usage and that their coal consumption is likely going to be in structural decline within the next 2-3 years if not sooner. Curious to hear your thoughts though.

Expand full comment

That's well said. I tend to think the Sept/Oct 2021 energy crisis probably created a "soft" policy pivot that favored economic growth and social stability over their "Double Carbon" goals of 1) Carbon Peak by 2030 and 2) Carbon Neutral by 2060. Since then we have seen expansion of coal production/supply and now we're seeing local officials deciding to build out local coal-gen capacity & infrastructure before the carbon-peak "deadline" of 2030.

The 13th FYP (2016-2020) was when they decided to "dissolve excess coal production," and the 14th FYP (2021-2025) was when they decided to prioritize clean and low-carbon energy sources, optimize use of fossil fuels and enhance power-gen, transmission & distribution coordination.

We've already seen the reversal of the 13th FYP supply-side coal reforms. And you could interpret the new coal-gen buildout we're witnessing as the reversal of the 14th FYP non-carbon prioritization. I'll let the policy wonks take up that debate. But ever since Li Keqiang chaired a meeting of the National Energy Commission on Oct 9, 2021 - coal has become an increasingly more important part of China's long term energy requirements.

Expand full comment

Agree that per capita power consumption in China will almost certainly rise as the country continues to urbanise and grow a more consumerist middle class. I think the recent buildout of coal power plants is partly supply-side stimulus, partly hedging on volatile hydro after the droughts of recent years hurt both industrial production in provinces like Sichuan and caused rolling blackouts in some southern cities. Looks like they are just going for full energy autarky/abundance and want to keep their options as open as possible.

But China has official commitments to start cutting coal use by 2026 and peak emissions by 2030 — despite the new coal buildout they are on track for that, even though the targets are insufficient to keep global warming below 1.5C (but that's a world-wide failure of coordination). I think they will adhere to those commitments, partly because they are realistic targets, and partly because it makes them look more responsible than geopolitical competitors. That would be a good baseline assumption for any model looking to analyse Chinese coal consumption as a share of global use. But you would also need to take into account rising demand from rapidly urbanising and industrialising countries like India.

As an aside — I'm guessing the new coal plants are more efficient than older ones that will start being retired from 2026 onwards.

Would also agree that heavy industrial processes will continue to drive demand for coal and LNG regardless of how decarbonised the electricity grid becomes. Not sure if China's official LNG data includes all the new imports from Russia. There will also likely be even more Russian LNG flowing to China fairly soon: https://carnegieendowment.org/politika/89552

Ultimately China is just going to have an incredibly diversified energy portfolio that will be relatively resilient to supply shocks in the coming decades.

Expand full comment

The big unknown is how the grid actuary works - of course this is not 1 grid is mode like 9 - I have looked at this in the past and know some people who have modeled it - I'll connect you

Expand full comment

Super interesting - but whats the read-through here? Much lower commodity demand from China going forward? You say the NBS/China Electricity Council data is non-sensical - but does this suggest they are ahead of where we would expect them to be in terms of renewables rollout, or are they massaging the data (eg are only producing a small amount of the electricity they should be able to - 3.8% seems low in terms of capacity?)

Expand full comment
author

In theory it should be weaker already and a lot weaker going forward but it is undershooting what the data for installations would indicate so.... I am as curious as to why as you are. If this is some connection backlog and it all clears suddenly that will be quite a shock.

Expand full comment

Great read! David Fishman may be worth reaching out to regarding getting higher frequency provincial power output data: https://twitter.com/pretentiouswhat?lang=en

Other thoughts:

- Could low capacity factors from solar be due to using that energy for PSH? China counts PSH output as hydropower, which may mean that energy generated by solar directed towards PSH is not counted as solar generation and is instead lumped in with conventional hydropower output. I'm far from being an expert but that's one possible explanation for solar's strangely low capacity factor. Or it could just be yet another example of spotty official data.

- I wrote an overview of PSH that may be of interest: https://chinapologist.substack.com/p/the-solution-to-the-green-transition

Expand full comment

Gasoline peaked last month, thanks to EVs.

Expand full comment

Have you tried JAX? It absolutely rips, and comes with auto-diff to make your optimisations easier. That said, it's more like playing with bare linear algebra, and I don't know these packages you mentioned.

Expand full comment
author

Its great, and JuMP is great for OR / linear program type stuff. Have not hit any incremental challenges yet and if its good enough for NREL its probably good enough for some fairly chunky problems.

Expand full comment