Great research, Alex. I used to work in China’s hydropower industry. Another factor is snow melt off Tibetan plateau during summer months, biggest impact is in Yunnan, also Sichuan. Too much rain too fast can be bad as it gets released (wasted), instead of used by generators--so there’s a cap to usage ability based on water basin dams.
Yeah getting tibetan snow melt right is..... hard plus data availability is not great. Thats why I opted for tree based methods to pick up all these funky dependencies but I have no doubt it would be a better model if there was weekly or daily data granularity. Sadly there is not, for me at least.
Hi Alex, your piece comes hot on the heels of the authorities in Yunnan province issuing orders for aluminum smelters in the province to cut electricity usage through the coming months. An interesting aside to this is that the cuts were based on the % liquid metal transfers each plant did. In 2022, the cuts came a month later but were across the board, uniform. Can I deduce from your data that perhaps these announced cuts aren't quite the catastrophe they look to be from inside the aluminum market? (After all, it's only 3 months since the smelters restarted from last years' cuts.) The metal industry tends not to look at the detailed data as you have, thinking only in terms of what it means to metal price. Your analysis suggests this is putting some eggs in the basket more than some sort of emergency measure.
I should do a follow up piece but you can see Guangdong power output drop as Yunnan picked up in August-September. I am not sure what is motivating that but given generally much weaker demand and not great pricing would make cuts seem sensible. So the whole Southern Grid switch leading to lower GD burns and higher YN imports is well underway in Sept and continuing in Oct.
A fantastic research / analysis effort
Great research, Alex. I used to work in China’s hydropower industry. Another factor is snow melt off Tibetan plateau during summer months, biggest impact is in Yunnan, also Sichuan. Too much rain too fast can be bad as it gets released (wasted), instead of used by generators--so there’s a cap to usage ability based on water basin dams.
Yeah getting tibetan snow melt right is..... hard plus data availability is not great. Thats why I opted for tree based methods to pick up all these funky dependencies but I have no doubt it would be a better model if there was weekly or daily data granularity. Sadly there is not, for me at least.
Heard about this piece from Doomberg. Excellent work!
Great piece, amazing the scale of hydro there! I thought we had a lot of it here in Quebec...
I linked your post here:
https://www.libertyrpf.com/i/139721516/science-and-technology
Hi Alex, your piece comes hot on the heels of the authorities in Yunnan province issuing orders for aluminum smelters in the province to cut electricity usage through the coming months. An interesting aside to this is that the cuts were based on the % liquid metal transfers each plant did. In 2022, the cuts came a month later but were across the board, uniform. Can I deduce from your data that perhaps these announced cuts aren't quite the catastrophe they look to be from inside the aluminum market? (After all, it's only 3 months since the smelters restarted from last years' cuts.) The metal industry tends not to look at the detailed data as you have, thinking only in terms of what it means to metal price. Your analysis suggests this is putting some eggs in the basket more than some sort of emergency measure.
Paul
I should do a follow up piece but you can see Guangdong power output drop as Yunnan picked up in August-September. I am not sure what is motivating that but given generally much weaker demand and not great pricing would make cuts seem sensible. So the whole Southern Grid switch leading to lower GD burns and higher YN imports is well underway in Sept and continuing in Oct.
Hi Alex, great post. Would love to see a followup analysis / commentary to see any developments if you have the time!
Cheers.