The economic Calculation Problem says that centrally planned economies are bad at efficiently allocating resources to production. The given reason for this is that prices reflect the need for any given product and that choosing the cheapest way to produce something is the most efficient possible solution, since the cheapest goods are those that are least urgently needed in other parts of the economy. (See the video for better explanation) Basically prices transmit valuable information about any given product that might otherwise be hard to obtain. So i've looked for some videos on youtube debunking this, but even FinBols video on this is kinda bad. He seems to misunderstand what the Problem is actually about.
Funny how porky got neural nets and algos that can find your normiebook page from seeing your on a street and predict your social status and mood just from how you move your mouse but calculating prices is somehow too hard for the worlds supercomputers. Not that I know anything about compsci, I'm sure it is too hard and we have to rely on efficiency of the markets instead forever and ever.
Luke Baker
Economic calculation problem doesn't seem to come up when i search it. But im gonna read it, ive been meaning to.
Well i think its more accurate to say it doesn't necessarily reflect actual need. But yes.
Aight so i know that a lot of socialists nowadays like to just say "lol supercomputers" in response to this and honestly its probably correct, however do these same socialists think that economic calculation was a hard problem during the Soviet era and that central planning has only been made possible nowadays?
Jackson Wright
I'm a brainlet so take this with a grain of salt, but I think the argument is not so much that planning was difficult in the Soviet days, it's that it took a massive bureaucratic structure to set goals and prices for all the industries which was prone to human error and political intrigue and what not, basically that part of Mises' critique of socialism was right at the time, so now with supercomputers that baggage is gone and what you need is people who can create a good planning algorithm, one with democratic inputs.
Kevin Powell
Love how ancap propaganda is always an abstract thought experiment completely divorced from reality. Fucking idealists.
Not during Stalin era for sure so they had to eyeball it, in the 60s and later it was totally possible but old brainlets in the party blew it and started to reintroduce market mechanisms.
Ethan Sanchez
The thing is that planning in the SU arguably worked better in the Stalin era than later.
Also a thing that bugs me about this argument is that while it does make sense in and of itself, Libertarians still have to demonstrate how this difficulty in obtaining information is so bad its necessarily fatal for any socialist economy, and how it is not offset by other benefits than central planning might bring. You can make up a 'efficiency calculation problem' for Capitalism: Lets say the road (if you've watched the video you know what i mean) that leads through the mountains will be better in the long term because its shorter so more stuff can be transported more cheaply from city a to city b. In a planned economy, this could be factored into the decision. However in a market economy only the cheapest option for the route will be chosen, with no regard for the future outcome for the rest of the economy. Climate change is actually such a problem in a way.
What i'm trying to say here is that while the economic calculation problem may be a real problem inherent in central planning, it might be more of a 'downside' than necessarily the 'reason why socialism fails'.
Kevin Richardson
TANS doesnt go much into it, read this cockshott paper instead
Except for the fact that price signals take a (long) time to propagate through the economy, having prices needed to be adjusted by all middlemen and inbetween products, as well as all affected downstream industries.
Sure Keeping track of how much of a product exists or is able to be produced is not hard in any way, modern corporations do it all the time. Keeping track of numbers in real time is easy to solve, its called all the shitty desktop pcs that even 3rd world bussiness have, and the internet.
Its fairly easy to keep track of all real production and it is comepletely possible to plan economies in kind. Keep in mind that if people try to throw the "planned economies are fixed in time so not as quick to react" that production for all industry is planned right now as well, internally, based on the imprecise prediction of what prices of input and output is going to be. Doing all the planning in the same step actually makes it so you know to a much larger extend how much supply and need there is going to be for a product, and thus how to best adjust production, than if you based it on purposefully imperfect information. Remember, imperfect information is incentivised by competition. Making sure your competition has imperfect competition improves your competitiveness, but on a societal scale it hurts all of us.
Dylan Smith
But also how much of it is being used in the rest of the economy at any given time. Nonetheless i will read that pdf, thx
Carson Gutierrez
Or to put it in different wording and context:
Yes, price signals convey informations about scarcity and the need for supply, and inform what option you ought to choose to be societally optimal. But these signals take time to propagate. The effects of a change in production in industry A will take a bit to affect their suppliers to the extend that they change their orders, and so forth. Doing all planning at once, and thus being able to run all the numbers on the changes and how they affect others, allows you to "propagate" all changes before ever enacting any change, in a much shorter (think, days) timespan than with market signals.
If industry and distribution is centrally planned then you can also super trivially keep track of how much stuff is being used as well. Its not any different from keeping track of production quantities. You know for each industry, the production rate of their products. You know for that production what the input is and what the output is. Multiply these two numbers and you know have the demand and supply. Add up, subtract the input from the output and you instantly have an overview of what products are being over or underproduced. (All O(n) operations, if you were wondering)
Alexander Rodriguez
The entire lolbert argument rests on planning organs not having any stats for planning, which is retarded and has nothing to do with the calculation problem.
Wyatt Cox
What Libertarians say: it's impossible to know what everyone needs and wants
What Libertarians mean: I want five 70-inch plasma screen TVs but the Neo-Cheka won't let me
Caleb Stewart
I doubt neo-cheka will have any use for their TVs. They will take away their 10 cars and put them on electricity ration, though.
Austin Hall
The ECP became a non-issue in the digital age. It was an alright argument back when modern technology would be considered scifi nonsense.
Kayden Myers
Reminder Mises was so buttblasted he couldn't refute Lange that the only thing him and Hayek could say was that socialism is mean.
You got some dank Lange pdfs? I've never heard of him now I want to read those polemics
Wyatt Wilson
This was literally won by the Marxists in its time (the 1930s) but ofcourse each side claims victory but only one wrote our history. The Austrians got BTFO by FACTS and LOGIC.
Lucas Long
This. You don't even need supercomputers, you just need feedback mechanisms from entities down the line to central planning institutions instead of it being a one way street - which is exactly how it worked in the Soviet Union! Central planners had trustworthy aggregated statistics from local units to work with and it worked better than with market reforms. Even a primitive economy can be planned.
Tyler Hernandez
I don't have many here, but this paper summarizes it, check the references at the end too
Sebastian Powell
Here, I'll write a labor price calculation program in shitty pseudocode right here: // Linear time, multiple iterations estimation of labor valuesmap laborValCalcIteration(map commodities) { for (c in commodities) { double deptotal = 0 for (depID in c.dependencies) { commodity dep = commodities[depID] total += dep.deptotal + dep.directlabor } c.deptotal = deptotal } return commodities}map laborValCalcForNIters(map commodities, int count) { for (int i in 0...count) { commodities = laborValCalcIteration(commodities) } return commodities} This solves half of the calculation problem. The other half is calculation in kind, which Cockshott's program solves.
Carter Wilson
Whoops, correction, it is log linear because the dependencies are proportionate to the log of the total commodities.
Kayden Reyes
also forgot to include multiplying the dep stuff by their ratio.
Isaac Myers
No they arent. The dependencies for production are not dependent on the amount of commodities in production. A simple loaf of bread takes the same input even if you create 20 new kinds of bread. Its relatively constant, so the algorithm is linear time.
Also you dont need to return your map since you change it in place, and you wrote "total" instead of "deptotal"also fuck one letter variable names tbh
Ian Taylor
C Y B E R N E T I C S
Kevin Williams
fuck those names, tbh. here's a simple take to make that shit code a little bit easier to read. (haven't read cockshott yet, interpretation might be wrong)
// calculates total dependecy labor cost for each commodity in a single iterationmap calculateLaborValueOfCommoditiesSingleIteration(map commodities) { for (commodity in commodities) { for (dependecy in commodity.dependencies) { commodity.totalDependecyLabor += dependecy.totalDependecyLabor + dependecy.directlabor } } return commodities}// calculates total dependecy labor cost for each commodity after a given number of iterationsmap calculateLaborValueOfCommodities(map commodities, int iterations) { for (int i in 0...iterations) { commodities = calculateLaborValueOfCommoditiesSingleIteration(commodities) } return commodities}
here's a brainlet attepmt at a haskell implementation
You can't really compare these two periods, during Stalin's industrialisation the economy was focused on raw production output and raw infrastructure, after Stalin services and consumer items became important and both these things must be addressed differently, the former is good with central planning the latter requires a bit of decentralised planning.
The thing is that more advanced economies have more commodities which are more complex. So while some commodities have below average dependencies (such as your bread), most vary around the log of the total number of economies. Some of Cockshott's students proved this.
Log linear is about as good as linear anyway, but there is still a distinction.
Well, I did say it was shitty pseudocode. However, you missed one important thing, which is that each commodity only holds a list of other commodities' IDs associated with the amount of that commodity (I missed the ratios at first too). The reason they need to hold the ID instead of the actual data is that duplicating the data will cause a huge increase in memory usage and will make it impossible for changes to propagate.
Please continue this discussion in the Cockshott thread, I reposted some slightly improved code there already:
Imagine how fucking stupid you'd have to be to think this makes sense.
Oliver Butler
i think its pseudocode that looks like c++
Andrew Peterson
yeah and it looks vaguely disgusting
Leo Ramirez
feels good mayn
Luke Gomez
Fair enough, but i think its important to distinguish that this is merely a trend at this moment and not the mathematical law of the calculation. I imagine that if you scale up a realistic economy that the relation between amount of products and ingredients per product starts to diverge away from nlogn towards linearity.
I was the guy you replied in that thread and I took the screenshots, it was a year ago and it really help me understand marxist economics, before that I was a very confused marksoc. So thanks again man.
