From the blog of The Socialist Party of Great Britain, via Markus Mobius:
What Shapley and Roth had in fact worked on was how to allocate resources to needs in a non-market context. As the Times went on to say, they worked out in theory (Shapley) and practice (Roth) how to match ‘doctors to hospitals, students to dorm rooms and organs to transplant patients,’ adding ‘such matching arrangements are essential in most Western countries where organ-selling is illegal, and the free market cannot do the normal work of resource allocation’ (like allocating organs to those who can pay the most).
And this:
So, we really are talking about a non-market way of allocating resources. As socialism will be a non-market society where the price mechanism won’t apply to anything, the winners’ research will be able to be used for certain purposes even after the end of capitalism; which is not something that can be said of the work of most winners of the Nobel Prize for Economics.
No doubt it would continue to be used to allocate organs to transplant patients and students to rooms. In fact, this last could be extended to allocating housing to people living in a particular area. While they may not get their first choice, people would get something for which they had expressed some preference and that corresponded to their needs and circumstances. It might even help answer Bernard Shaw’s question, ‘Who will live on Richmond Hill in socialism?’ Since socialism will be a non-market society the answer can’t be, as it is under capitalism today, ‘those who want to and who can afford to.’ This would not only be ‘repugnant’ but impossible.
Cheap Talk 
6 comments
Comments feed for this article
December 9, 2012 at 11:06 pm
Tayfun Sonmez
Hi Jeff, I hope this doesn’t turn into an amplified version of cooperative game theory vs. non-cooperative game theory debates that was less than constructive when I started working on these problems on late 90s 🙂
That said, at least much of the work on school choice, kidney exchange, house allocation, etc. is certainly intended to be compatible with a wide spectrum of economic systems.
December 10, 2012 at 5:43 am
Pedro F.
I imagine that soon we could have a huge, open source, computerized allocation system for the whole economy. It could be like a wiki, where people could add (peer-reviewed) changes and improvements, to always keep it up with new innovations, new products and fashions.
Already today large retailers, like Walmart, with automatized selling systems can know immediately how much people are demanding of each good and how much each individual seller has on stock. If you unify these systems, you could have such knowledge for (at least a large portion of) the entire economy. Huge lines and lack of some goods, which are imagines often associated with communism, wouldn’t happen. You could very quickly notice changes in consumer behavior and adapt production for it, or divert resources for where it is needed the most.
Or imagine this scenario: every week you can put on your tablet or notebook a ranked list of what you want to eat, what you want to buy and what you watch and do this week. Then the above mentioned algorithm gets together these preferences and defines how much of each each person gets (maybe he provides each person with some token money and makes a simultaneous ascending auction?).
Anyway, the possibilities are plenty, and I feel very enthusiastic that economics is helping making it all possible. This seems to me like it is (or at least should be) the dream of 21st century communists.
December 10, 2012 at 2:14 pm
Mounir
I’ve seen this point being made a few times now. It takes the correct view that the allocation problem can be solved efficiently and assumes that this proves that markets are no longer necessary. This forgets that the price mechanism which equates supply and demand is lost without markets. If there is a steep increase in the number of doctors, the Shapley mechanism will not incentivize new hospitals to open up, or existing hospitals to hire more doctors – if the wage (i.e. the price) is fixed.
December 10, 2012 at 5:34 pm
kerokan
I am a political scientist and it seems strange to me to extrapolate from the current applications of mechanism design (MD) to other aspects of managing the society. The current applications, for example kidney exchange, seem very appropriate for MD, because everyone involved (ie. kidney patients and their close relatives) would benefit from a more efficient trading system. In other cases, for example land distribution, Pareto efficiency is not possible, because some actors benefit a lot from the current system even though it does not efficiently use the resources. Moreover, in the current applications (ie. assigning medical residencies or kidney exchange) we can sign a contract and be pretty confident that it will be enforced. As Acemoglu and Robinson argue, however, many goods in the society provide power to their owners. For example, once we transfer land to the poor, there is a high risk that they will use their new economic power to renege on the payments they promised to make when the system was designed. I am skeptical that such problems can be solved by MD.
December 11, 2012 at 10:42 am
twicker
Seems the socialists are ignoring key components of the situations where Roth-Shapley applied:
1) Many of these situations are ones where:
(a) A strong majority of the population believes that the situation is not one where money should determine which part of supply goes to which demander/consumer, and
(b) We still need allocation of the resource.
For example, most Americans would think it incredibly unfair if the rich were preferred recipients of organs. People are completely fine with the idea that the rich are the only ones who buy personal yachts or planes or multiple vacation houses, etc.; “being alive” requires none of those things, so there’s no taboo against letting the monetary market decide (same with buying filet mignon v. chicken breast: sure, one is more expensive, but you’re not going to starve if you only get the cheaper one). The socialists want to apply it to everything – including those things where the vast majority of people would think that there’s no “right” to the thing (a “right” to filet mignon?), that there’s no moral reason why everyone should have a thing.
There are good moral, ethical reasons why we should act (and, in fact, act more than we are) to make sure that everyone has a safe place to live; and food; and a minimum amount of healthcare, including preventive care; we can pretty much all agree on this (except the Randian Objectivists and Social Darwinists, who seem to think the poor should just hurry up and die). Few people think that we all have a “right” to, say, expensive wine, or a particular kind of car, etc.; those remain as incentives for working harder. Socialists don’t seem to understand that.
2) Other situations are one where a monetary or other market already acted; the question now is, how does one allocate what remains to the people there, and do it in an equitable way (e.g., matching doctors with hospitals, matching dorms with students)? The people being matched have already expended the resource to be part of that market (graduating from med school, being accepted into the universities); what remains is how to divvy up a particular part of the market resource amongst the people who have already “bought”/earned their way into the market.
—————————-
In short, the socialists are taking great ideas and extending them illogically. They might need to be reminded that we’ve already seen societies that tried to implement socialism to the extent described in the quote. Evidently, the socialists, or at least the SPGB, need to be reminded that we refer to those societies in the past tense for a reason.
December 12, 2012 at 8:55 pm
ggclass
I do not know much about the specifics of the mechanisms proposed but it seems that there is a minimum level of heterogeneity for the matching to be meaningful in any way (allows the best matches to be made or makes the matches stable). This would imply that there are settings where these mechanisms are not relevant or do not work well. In their proposed example about housing it would make sense if we’re choosing between neighborhoods with different desirable characteristics over which consumers’ preferences vary–say, good for families vs good for hipsters–but not if the ranking is the same for everyone–say, suburb vs ghetto. In the latter case the discovery that the mechanism makes is meaningless and the allocation of housing will fall back to some other mechanism like lottery or discretion. So no socialist panacea here.
I believe Roth would prefer a functioning price system for liver exchanges rather than his mechanism which takes the absence of prices as a constraint.