Interesting how so much I/O stuff is recent.
Regardless, what's the potential efficiency gains in a planned economy if you only have to track enough production to properly reconstruct the whole economy?
https://publications.jrc.ec.europa.eu/repository/handle/JRC120513
Finally got around to reading this one. I was thrown off by its use of the word "projection" when they actually mean extrapolation. For planning, completely disaggregated data are crucial, as demonstrated by the Soviet experiment and as pointed out by @wpc and others. The bit about the energy sector is interesting because I had the same problem when I was toying with I/O planning in Workers & Resources. Square matrices are bad for planning.