In Soviet Union, Optimization Problem Solves You (2012)

http://crookedtimber.org/2012/05/30/in-soviet-union-optimization-problem-solves-you/

by mpweiher • 8 years ago

92 108 8 years ago

crookedtimber.org

dang• 8 years ago

Discussed in 2015: https://news.ycombinator.com/item?id=10679582

baursak7• 8 years ago

For those more interested in this, there's a follow-up to comments, which results in this (still on-going?) exchange:

- The original article (linked by OP): http://crookedtimber.org/2012/05/30/in-soviet-union-optimiza...

- A comment made by Paul Cockshott (a Marxist economist and computer scientist, most known for "Towards New Socialism" book): http://crookedtimber.org/2012/05/30/in-soviet-union-optimiza...

- Shalizi's response to Cockshott, among his other responses to other comments: http://bactra.org/weblog/919.html

- Cockshott's response to above: https://paulcockshott.wordpress.com/2018/06/25/plan-balancin...

petermcneeley• 8 years ago

Central planning is the response to Von mises argument about economic calculation in socialist countries.

https://wiki.mises.org/wiki/Economic_calculation_problem

kingkawn• 8 years ago

If you link to the wikipedia article instead it actually discusses criticisms of Von Mises instead of simply supporting it without counter-argument: https://en.wikipedia.org/wiki/Economic_calculation_problem

jayalpha6• 8 years ago

The funny thing about Mises is:

He understood socialism and its' problems much better than capitalism. He never understood capitalism but got stuck with his understanding in something reflecting a middle-ages barter economy.

UncleEntity5• 8 years ago

Citation please?

averros4• 8 years ago

What von Mises has proven logically in 1920 (and the practical experience has confirmed it later over the course of 20th century - repeatedly, dozens of times, in different countries, with different cultures) is that a centrally planned (i.e. socialist) economy ALWAYS SELF-DESTRUCTS.

It has nothing to do with computers and collection of information and planning algorithms. The issue (completely misunderstood by the author of the article discussed) is that the very fact of coercing people destroys information about their subjective utility functions which could only be gleaned by observing FREE exchange. No amount of fancy software or social engineering will fix it. Garbage in - garbage out.

Socialism is a very bad idea. No ifs and no buts about it. And, yes, Marx was a kook. His entire economic and social theory is a humongous monument to the principle of logical explosion: if you start with nonsense (labor theory of value), you end up being able to derive anything from it. (The fact that labor theory of value is nonsense is easily established by observing "value" created by someone digging a hole and filling it in - any number of times, so as to put in arbitrarily large amount of labor).

fouc• 8 years ago

It may not matter if you start from good first principles or bad ones. Deriving a system from logical reasoning seems to always fall short in reality.

(At least for natural languages as opposed to something like mathematics).

averros4• 8 years ago

Ghm. Mathematics and sciences are rather successful in reality. Natural language can be used in a rather rigid and formal manner - and mathematical language can be used carelessly and in error. (In fact, using natural language carefully may be preferable to playing with formal symbols - as it makes gaffes and logical gaps more visible to the readers.)

claudiawerner8• 8 years ago

Not to speak for or against Marx's law of value,

The fact that labor theory of value is nonsense is easily established by observing "value" created by someone digging a hole and filling it in - any number of times, so as to put in arbitrarily large amount of labor

Is rubbish. In fact, if you read less than the first chapter of Capital Vol. 1, Marx specifically rebuts this argument on the third page, in which he restricts himself to 'socially useful' labour performed at the rate necessary to complete it. Not even anti-Marxists use this argument any more because it simply doesn't hold for Marx's theory, nor even Smith and Ricardo's. Mises himself doesn't use this argument.

averros4• 8 years ago

Of course, anyone who is friends with logic will immediately spot that "socially useful" cannot be defined in any way which does not depend on (drumroll) value to the people.

In other words, all Marx said is that value is created when labor creates value. Duh. Circular reasoning of this sort is a hallmark of kookery, and it takes a rather brainwashed mind (or absence thereof) to miss this obvious logical fallacy.

There are other ways to demonstrate the idiocy of the labor theory of value: for example by constructing a realistic scenario when clearly "socially useful" labor (for example, work at a factory) destroys value of a "free for all" resource which wasn't created by labor at all - i.e. by polluting the air). Incidentally, Soviet Union was full of grossly polluting factories.

I think a personal anecdote is in order: unlike Western Marx fanboys I happen to be formally educated in Marxism, at the top university in the USSR, no less, (this was mandatory, I wouldn't poison my mind with that junk willingly). The professor reading the course of Political Economy was the well-known Marxist/Leninist textbook author and certainly was quite smart. So after his first or second lecture (I don't remember exactly which, it was 30 years ago), I asked him to clarify exactly the point above: how to define which labor is useful and which it is not. He called me into his office and told me in no uncertain terms to never ask questions like that again if I wanted to stay out of trouble (didn't help), and that the best answer was "whatever Party directs us to work on". This pretty much clarified everything regarding Marxism to me, and I took his advice to heart, so I still have my transcript from MSU with "with distinction" ("отлично") marks for Marxist/Leninist Philosophy, Political Economy, and Scientific Communism. I guess I should thank him for a sound advice on surviving cult indoctrination.

Animats• 8 years ago

He sets up his own straw man, and demolishes it. Big deal.

If you really had to solve this problem, you'd need to divide it into sub-areas, like "clothing" and "heavy machine tools", and solve it hierarchically. The solution will be somewhat suboptimal but should come close.

The trouble with a capitalist model is that it's a huge collection of feedbacks and delays, so it oscillates around some optimum. That's a basic result from control theory. Delay causes instability.

The USSR's Gosplan had a monthly reporting cycle and a yearly planning cycle. That was pre-Internet and too slow to work. WalMart has a daily reporting cycle (probably faster now) and a weekly planning cycle, and runs a consumer economy bigger than the USSRs.

Not that industrial optimization is the controlling issue today. That's a problem for when you don't have enough manufacturing capacity. We're past that.

avivo• 8 years ago

I think this frame of market capitalism as a set of feedbacks and delays is crucial. Though I expect that there will just be different ones with alternative systems in practice.

petre4• 8 years ago

The reporting was mostly BS or too late or they tried to adjust reality to meet their quotas. They doctored the numebrs to look good. When only the state has a stake in all of companies and everybody is an employee of the state, there is zero interest for things to work out the way they were intended to. And this is why capitalism works better practice: because it encourages individual enterprise while it also encourages cooperation amongst individuals when things get rough.

https://en.wikipedia.org/wiki/Pandora%27s_Box_(TV_series)#Pa...

7sidedmarble5• 8 years ago

When only the state has a stake in all of companies and everybody is an employee of the state, there is zero interest for things to work out the way they were intended to.

This can easily be rephrased to:

"when all the major politicians have stakes in all the major companies, there is zero interest for things to work out the way they were intended to."

And now there's no actual argument to what you're saying. Both systems have flaws. That does not prove one's better, you need more then that.

Seeing the flaws is the first step to asking yourself the question, "is there a better way to do this?"

Do you think we're better off because our corporate funded politicians have all the power, and we're just living in their fiefs? Is that the sign of a good system?

Capitalism uses minor successes in it's lower classes like bread and circuses. They're distractions. Don't say to yourself, "at least I'm not starving in some South American country", ask yourself "why can't we all be doing better?" Is this system really the best mankind can do?

tvanantwerp2• 8 years ago

I don't think the planning in the USSR and Walmart are remotely comparable, and not just in terms of timeliness. A central planner is deciding who needs what, how much, when, and where. But Walmart is discovering who needs what, how much, when, and where. They aren't decided well in advance what people need; they are working to understand what people want and how to get it to them. How much you can sell and at what prices drive their forward-looking decisions around what people value. Without those signals, a central planner can't really know who values what. It's just guesswork.

claudiawerner8• 8 years ago

Must it be guesswork? Most of us have devices (computers and mobile phones) that could very easily be used to register demand for various products and to show what's available. In what way precisely does Walmart work to understand what people want which a planned economy could not possibly?

rileyphone• 8 years ago

What the author fails to address is the most recent research into the planned economy that Towards a New Socialism advocates. It advocates for an economy that is still socialist in nature (in seeking to eradicate the commodity form) while still aggregating opinion into production as market forces do. This is accomplished by using a labor token while using market forces to influence prices and planning. Sure, it can't be optimised perfectly, but neither can our current one, and that is never the goal of a socialist economy anyways.

davidgould• 8 years ago

I'll add that the book Red Plenty is entertaining and thought provoking and quite fun to read. The historical and cultural scenes depicted are interesting as well however you feel about the case for central planning.

js87• 8 years ago

It seems to me, contrary to what is often claimed, the problem of central planned economies (not just state-wide ones, but also company-wide ones) is too much optimization at the expense of robustness.

In top-down planned systems, there is always somebody who says something like "we only need this one thing for everybody". For example, in communist Czechoslovakia, there was only one factory designated to produce certain electronic chips. But then everybody becomes dependent on that one factory, and if that one factory slips in schedule, everything that depends on it will slip.

There is a fundamental trade-off to be made. You can have an efficient system but it's robustness (ability to withstand a failure or adapt to change) will be very low.

But the planners are often willing to take risk of being less robust for the sake of efficiency, because it usually gives you a free promotion in human status hierarchies.

(If you ever worked in a large company, I am sure you can recognize the above effect, somebody hyper-optimizing through a mandate, and it has 2nd order effects in decrease of efficiency.)

Compare that with market economy, which actually has huge overabundance of production capacity, and multiple producers of almost everything. This is less efficient but works really well in practice. (However, it has the other kind of problem; there is no way the free market can recognize that there is too much companies, without the agony of "malthusian" overcrowding and associated nasty effects such as more stress for everybody. Things like bullshit jobs are probably an end result.)

eesmith2• 8 years ago

"too much optimization at the expense of robustness"

That is not unique to central planned economies. Consider how the 2011 Thailand floods lead to a global shortage of hard disk drives, or the saline solution shortage in the US when the plants in Puerto Rico were damaged by Hurricane Maria.

I think the author expressed it clearly that the planning algorithms depend on linear programming, and that non-linear effects - and I believe 'robustness' counts as such - are beyond what can even hope to achieve. The author also points out:

If it’s any consolation, allowing non-convexity messes up the markets-are-always-optimal theorems of neo-classical/bourgeois economics, too. (This illustrates Stiglitz’s contention that if the neo-classicals were right about how capitalism works, Kantorovich-style socialism would have been perfectly viable.)

Hence why capitalist systems may also have robustness problems.

"You can have an efficient system but it's robustness (ability to withstand a failure or adapt to change) will be very low."

I think that's an observation in how centralized systems are implemented. I don't see why robustness can't, at least in theory, be included as part of the optimization process.

Beyond the factors highlighted in the essay, of course, like the inability to decide how to weight those factors, and the non-convexity.

js87• 8 years ago

"That is not unique to central planned economies."

Yes, however, when it happens in market economies, it is not caused by planning for greater efficiency (saving money). Nobody really decided there to be only one factory that produces hard drives; it is a consequence of hard drives being particularly difficult to produce.

"I think the author expressed it clearly that the planning algorithms depend on linear programming, and that non-linear effects"

I disagree with that. I think planning itself (algorithms) is not the issue here, I think the objective is.

"I don't see why robustness can't, at least in theory, be included as part of the optimization process."

In theory, perhaps. But in practice, it's difficult to prove to the people in charge (whoever that is) that you made the system more robust, compared to proving that you made the system more efficient. So there will always be an inherent organizational bias towards efficiency.

Airlines are an interesting example of industry which strives for efficiency, yet has strong robustness guarantees through regulation enforced by central agency. (Although in this case, it is not economic robustness.)

Maybe the robustness of the economy as a whole is given by how much choice you have in being conservative or take an experiment. So there really cannot be "robustness" as an objective, since it depends on continual experimentation with different models of operation, where most of them fail.

tonyedgecombe• 8 years ago

Yes, imagine the chaos if Walmart failed tomorrow.

3rdAccount4• 8 years ago

Interesting to hear you say this.

Power energy markets are run using LP & MILP algorithms and a huge topic right now is accounting for non-convexivities.

petre4• 8 years ago

If you lived in communist Czechoslovakia you probably saw the amount of bullshit jobs in a (small) centrally planned economy. We had communist party activists in all institutions, secret police listening all day to people's conversations and the list could probably go on and on.

js87• 8 years ago

These might be considered useless but are not necessarily bullshit according to Graeber's definition. The definition of bullshit job is the person doing it knows that it is bullshit, yet they have to do it because there is no better job available for them.

And to be honest, I am not really sure there is less bureaucrats (which are only sometimes bullshit jobs) in the market system. I actually suspect the opposite, because there are much more hierarchies, and there is less people needed in agriculture and industry.

vbezhenar• 8 years ago

The inefficiency of market economy is terrible. Also it leads to production of bad products where repairability, quality is sacrificed to reduce costs and increase sells (who would make a good light bulb, which will last years if you can make a bad light bulb which will last months and sell dozens of them). I'm living in post-USSR country and a lot of old consumer goods which were "made in USSR" are an etalon of quality compared to modern goods.

I still think that central planned economy is superior to market economy. USSR failed because of human factor and probably because computers were not good enough to produce proper plan on a USSR-scale, but even with those factors it was able to compete with the rest of the world in science, space, military. It's quite fascinating.

vkou• 8 years ago

There's some survivorship bias at play here. The crappy, poor quality goods made in the USSR went to the scrap heap decades ago. The stuff you're seeing today lasted the test of time, but a lot of other stuff didn't.

Some Soviet categories of goods, and brands were pretty decent. My grandparents' fridge has lasted a good 40 years, without any issues. Their television was broken more then it was working.

ekianjo2• 8 years ago

even with those factors it was able to compete with the rest of the world in science, space, military.

Competing on a few key sectors is easy when you sacrifice the livelihood of your whole nation, spoil their property, and send millions to the Gulag to deter any political opposition.

And the Soviets did not compete in "Science". They competed on military and space exploration mainly, but most of the other areas of Science were left dry by the Soviets, or left to crazy experimentation based on pseudo science in agriculture. Most of the genetic discoveries were done in the US, most of drug development happened in the US, computer development was way ahead in the US as well - there is no world where the Soviets were competitive across the spectrum.

And let's not forget that the US continuously shipped grains to the URSS because they could not produce enough food for their own people consistently. You can't do Science if you can't even grasp agriculture.

https://www.washingtonpost.com/archive/politics/1979/10/04/h...

badpun• 8 years ago

I'm living in post-USSR country and a lot of old consumer goods which were "made in USSR" are an etalon of quality compared to modern goods.

Maybe the simpler ones, like blow dryers. Cars made in communist Poland usually broke down within first couple of hundred kilometers of purchase. People used to take a new car to a (private, as in not-state-owned) mechanic, who would dissasemble it and then reassemble, to make sure everything was put together correctly. Not to mention that communist Poland, a nation of 40 million, was unable to design its own car and had to licence it for production from capitalist corporations.

nine_k2• 8 years ago

If you ever lived in the actual USSR, you must also remember how poor average people were, how hard was it to find and buy anything decent, or simply anything. Including food. Want a piece of meat? Spend two hours in a line. Want a piece of Bologna sausage, or butter? Here's your food stamp for a minimum monthly ration.

USSR just did not have enough resources to properly feed everyone using planned economy. It had to admit the very smallest pieces of market economy (sell apples from your orchard, don't go to jail) to keep people somehow fed.

claudiawerner8• 8 years ago

USSR failed because of human factor and probably because computers were not good enough to produce proper plan on a USSR-scale

I think this too. There has been recent work by Paul Cockshott (a computer scientist) on new models of central planning with modern computers. I'd urge anyone interested to check it out.

v_lisivka9• 8 years ago

a lot of old consumer goods which were "made in USSR" are an etalon of quality compared to modern goods

You just cannot afford high-quality product.

icebraining2• 8 years ago

I'm living in post-USSR country and a lot of old consumer goods which were "made in USSR" are an etalon of quality compared to modern goods.

If you replace USSR with "period up to the late 80s", that statement is repeated in many countries, so maybe the central planning part is not the distinguishing factor.

Fins2• 8 years ago

Strictly speaking, Soviet goods (very few of them, though) were of somewhat decent quality when compared to the cheapest crap made in a non-name sweatshop for dollar stores. Or export to the Eastern Europe and ex-Soviet countries.

Even in the space, USSR achieved much more launches than the US for a simple reason -- Soviet satellites crapped out in months instead of years.

marmaduke2• 8 years ago

only need this one thing for everybody

Not too diminish the aspect which is particular to communist Czechoslovakia, but it is a hard earned tenet of highly available systems best practices to have redundancies throughout a given stack. For example, a storage system that has disks in RAID, dual RAID controllers, dual NAS servers, dual connectivity to switches, etc is more expensive but worth it in the long run if you depend on it.

Maybe that lesson hasn’t been absorbed more widely. Market economy perhaps gets redundancy for “free” (or cost subsidized by losers)

7sidedmarble5• 8 years ago

Market economy perhaps gets redundancy for “free” (or cost subsidized by losers)

That's a platitude, not a fact. We just are moved to believe that the market is naturally redundant by propaganda, our governments in the west urge us not to look at the instances where it's not, like economic collapses caused by banks or other government investing groups that are "too big to fail."

peisistratos• 8 years ago

In the three decades Stalin was general secretary, the Soviet Union had enormous economic growth. Russia went from being a backward peasant country ruled by a czar, to being the first country to send a satellite to space. Its success and its example were what was feared. While unemployment was 25% in the US in 1933, the USSR was charging forward, and US and European firms and workers were going there to work.

On Stalin's death, Khrushchev announced economic growth was no longer a priority, and other such policies, and stagnation set in.

China also had enormous economic development under the communist party from 1949 to 1978, and from 1978 to now (although the party took a right turn under Deng Xiaoping in 1978, to where some question if China was socialist after then).

The average inflation-adjusted hourly wage in the US is below what it was 45 years ago, leading to anti free trade Trump and Bernie Sanders and Alexandria Ocasio Cortez. Also Corbyn, Syriza, Brexit, ping-pong from Lula to Bolsonaro, TARP bank bailouts plus the carbon externalities destroying the planet.

Of course, I don't think a traditional Marxist worldwide communist revolution of factory workers led by Leninist parties will happen.

The present economic system might be eradicated though. It has not been all that long since European kings were having their heads chopped off or were being bayoneted to death in basements.

averros4• 8 years ago

Of course, the Stalinist "economic growth" was built on the bones of millions forced into slavery. Do you really consider this kind of growth desirable or sustainable?

claudiawerner8• 8 years ago

Not to excuse the brutalities of the Stalinist regime, but isn't our own economic growth built today on almost intolerable conditions in sweatshops, child labour and qualitatively bad ways of living life in various parts of the world and even closer to home?

x2203• 8 years ago

The USSR wasn't charging forward. It was charging into starvation, gulags, and genocide.

StudentStuff3• 8 years ago

The USSR accomplished quite a bit coming from a nation of malnourished serfs that were formerly used as cannon fodder.

Breaking up the existing communes and murdering the Soviets (then using their name) were terrible things, but its impressive that a landlocked country with minimal infrastructure and an uneducated populace could fight off Germany and liberate most of Northern Europe, push the bounds of science, and be a world power for a number of decades despite having half the populace of the other world power of that era.

StudentStuff3• 8 years ago

Note that South Korea, Japan and others also went through rapid industrialization and economic improvement during this time.

I don't think China is a great model, their development was supercharged by demand from the USA for products made with low cost labor. What Cuba accomplished wrt education & lifespan (they live longer than us Americans!) is quite impressive.

The USSR managed quite a feat too (being a world power for quite a few decades) for essentially being a landlocked country with minimal infrastructure and an uneducated population of serfs that was half as numerous as us Americans. Breaking up the existing communes and murdering the Soviets (then using their name) were terrible things though.

gandhium• 8 years ago

In the three decades Stalin was general secretary, the Soviet Union had enormous economic growth. China also had enormous economic development under the communist party from 1949 to 1978

Suprisingly (or rather unsurprisingly) they've also managed to starve tens of millions along with their 'growth'.

atemerev• 8 years ago

For me, the principal problem of central planning is that you cannot plan for innovation. Innovation works best in competitive capitalistic environments, where there is a significant prize awaiting innovators.

claudiawerner8• 8 years ago

Haven't there been many innovators and scientists who come from the most unlikely walks of life, and die penniless? Did Einstein, for instance, not have to divide his time between work at the patent office and groundbreaking research?

v_lisivka9• 8 years ago

You should compare number of implemented innovations and market disruptions instead.

baursak7• 8 years ago

Innovation works best in competitive capitalistic environments

Well, a huge chunk of innovation in capitalist countries happen on university campuses or within government funded organizations like DARPA or NASA, or within communist-like arrangements like Open Source (Linux, Apache, etc). Which are then appropriated by private interests to be monetized.

SII_of_SII• 8 years ago

One thing I got out of the article is that an alternate-reality communist economy can be more efficient than a capitalist economy, given a good objective function. The way to do it is to simulate a market economy on your central-planning computer: Simulate a network where each node models a consumer or supply chain unit. This runs the same computation as a market economy, but propagates price signals much faster. It's faster and cheaper to do an RDMA access than to wander around a grocery store comparing the price of bread to rice.

The trouble is in formulating a good objective function. Perhaps you can solve the problem of consumers expressing preferences that Cosma discusses in "The Given Assortment, and Planner's Preferences" by imitating a market economy. Each person receives a fixed salary which they spend at state stores. Since this is how the computation performed by a market inputs consumer preferences, this is the input used by the central computer in its market simulation. But the simulation reaches equilibrium much faster than the real market and its objective function is more fair than a real market (since everyone gets the same salary).

I suppose there are bigger problems with formulating the objective function than consumer preferences. Creative kinds of economic activity are important and leaders that want to hold control may not allow you to formulate the objective even if you could.

leghifla2• 8 years ago

"I suppose there are bigger problems with formulating the objective function than consumer preferences."

Maybe producer preferences? The fact that one prefers to be a mechanic or a nurse does not influence the planning algorithm (AFAIK): people are just pawns on a chess board, like any other resources.

guiriduro• 8 years ago

Is that necessarily so? You could argue that producer preferences are some kind of soft constraint that can be maximised (albeit subsidiary to the hard constraint solution)

nine_k2• 8 years ago

Second-order effects kill it: https://en.m.wikipedia.org/wiki/Economic_calculation_problem At least, this has always occurred in the reality we have got. Visit Venezuela for the latest example.

ASipos• 8 years ago

Also, Scott Alexander on Red Plenty:

http://slatestarcodex.com/2014/09/24/book-review-red-plenty/

7sidedmarble5• 8 years ago

Stopped reading at:

In certain cases, Russians were very well-incentivized by things like “We will kill you unless you meet the production target”.

Come on. Outside of POWs during the war (and that's not what were talking about) no factory workers are getting dragged out back and shot. The story of collectivization on farms is more complicated but it still didn't involve on the spot murder.

It's like people watched that movie where the Russians are at stalingrad, and the commisar gives every two men a rifle and shouts "first man shoots the rifle, if the first man dies, the second man picks up and shoots the rifle." -- And then everyone collectively thought "oh, not only is this true but this absolves me of any responsibility to actually learn about the USSR."

Read some books, talk to some older Russians. There was some extremely dark times in the history of the USSR. But for a lot of it's history, it was just a normal western-ish country. It wasn't like war time Nazi Germany 24/7. You weren't getting shot for not making enough hinges one day.

Fins2• 8 years ago

Umm, have you ever lived in USSR? Being born in the 80's doesn't count either.

There absolutely were times when being a few minutes late to work (not during the war either) would get you a long prison sentence at the very least.

Also, it never were "a normal western-sh country". Russia hasn't been, and USSR in toto had no reasons to be either.

Works of marxist "economists" shouldn't even be dignified with a discussion, but is is greatly concerning that there is such a thing as whitewashing of the soviet regime, and that it is considered to be far more acceptable that whitewashing nazis.

unknownkadath• 8 years ago

So if P=NP, Communism will triumph?

julienreszka• 8 years ago

No. There will be a definite proof of it's failure.

In Soviet Union, Optimization Problem Solves You (2012)

Related Stories