On my blog, there has been little relationship between the foundational importance of an argument, and its coverage. I tend to cover what I find new and interesting, and those do not tend to be very big shifts in my basic views. They tend to be the merlons and embrasures on the castle of thought, and not the castle foundation itself. So today, I would like to defend a very basic, but important, thing.
I think capitalism – or more specifically, private property rights – are good. I believe it leads to people living improved lives, with more of the things which they want. When resources can be employed in many different ways, capitalism allows us to elicit people’s private information and allocate them to the right uses. It does this by making people bear the costs and benefits of their decisions. In other words, capitalism is an incentive compatible mechanism. If a central planner were implementing a mechanism to efficiently allocate resources, the mechanism they choose would include prices.
Let me give an example, which I take largely from Prendergast (2017). Feeding America is the largest food bank system in America. It receives donations of food from companies, and has to allocate them among their food banks. Until 2005, the system worked as follows. Each food bank would be placed in a queue, with their placement determined by the number of people below and near the poverty line in their area (with above poverty line recipients being appropriately weighted), compared to the amount of food the food bank had received. Whenever Feeding America was offered a truckload of food (say, Walmart discovers that they have 20,000 pounds of loose bananas that they would really like to get rid of) they would contact the first food bank in the line. That food bank would have 4 to 6 hours to accept or reject, and then their position is reshuffled based upon the new amount of food received.
There are some problems with this system. First, Feeding America lacks information about what food the food banks have on hand, as food banks have multiple sources of food. This means that food banks have private values for different types of food, and randomly offering the truckloads leads to places not getting the types of food they want. What does Idaho do with more potatoes? What does Wisconsin do with more dairy? What does North Carolina do with more bacon? Second, offering it to only one food bank at a time meant that picking up the truckloads would be long delayed. Food could spoil on the loading docks, or at the very least the donor would be greatly displeased that they have had 20,000 pounds of bananas sitting in their warehouse for three days and it’s beginning to smell really bad! In the case of goods like bananas, or fresh produce generally, they have no good way to get food banks to take the goods to at least get them off the hands of the donors. (Doing this maintains good relations, and means that they will continue to offer actually useful goods in the future).
Fundamentally, the problem is that items lack prices. We are unable to say how much different goods matter relative to one another, or how those values might differ between different food banks. There is wide heterogeneity in preferences between food banks, and prices are simply the incentive compatible way to elicit those preferences.
In 2005, with the advice of some UChicago economists, Feeding America replaced this with a market, called the “Choice System”. They distributed a fake currency, which could be used to bid on auctions for truckloads of food. These auctions could now occur twice per day and simultaneously, with the fake currency – “shares” – being distributed back to food banks by formula. The system was designed to be extremely simple to use, even if some gains are possible if food banks had unlimited calculating power. The auctions are sealed bid and first price, and each truckload is independent, as food banks cannot place combinatorial bids.
There is considerable empirical evidence for privatization and capitalism and the price system and so on, but I don’t want to focus on such reduced form evidence. It is agnostic about the channel through which capitalism improves welfare. Rather, it is possible to precisely estimate how much the new mechanism raised welfare by, once we have estimated structural parameters.
Altmann (2023) estimates that implementing the Choice System was equivalent to raising food donations by 32.7%. He is able to do this by estimating the food banks preferences using their current bids and past winnings, and then simulating various “games” to estimate counterfactuals. So, for instance, he can show clearly that food banks have storage costs, as their bids for items which they had won recently decline. Food goes to food banks which are closer, cutting down on transport costs. The auctions being simultaneous is responsible for the bulk of the welfare gains, as food banks can plan around what is being offered that day. Food wastage is reduced by a quarter. Storage costs per pound of food are down. The vast majority of individual food banks are definitely better off, and those that aren’t are simply the ones who accept every load of food offered under the old system. None of this is possible without prices.
Now, some of you might say that eliciting the preferences for different types of food, in the case of food banks, can’t be that hard. We could require that they report their full stock of food at any given moment, and then calculate the optimal allocation of food. Doing this, however, simply transfers who pays to report the allocation of food. If Feeding America hired agents to go around and check, they would simply be replacing the information already elicited through the price mechanism – it can do no better, and probably worse. (We can prove this formally, so long as there are no externalities and no market power, which this market clearly satisfies – see Maskin (2015) on the subject).
And what should we do when these aren’t food banks, but people? No one can go around and tally up the desires that you have. They have to ask you. When you are asked, you must face a cost to receive the item that means you will only receive it if you place a value on it sufficient to cover the cost of making it – else, we will produce items which make us worse off.
This is why private property rights are good as well. To give someone property means that they have the right to the full product of it, for both good and bad. If the cost of investment, or the returns to the property, are shared among other people, then there must necessarily be some instances where there was wasteful investment in the former, or insufficient investment in the latter.
What we very consistently find is that new economic systems, which try to avoid property rights and money and markets, find themselves with obvious flaws which are fixed only by making themselves more and more like market firms. Take labor cooperatives; in particular those of Yugoslavia under Tito. These were set up as democratic organizations – all workers would vote to decide at least upon the directors, who would then be given a mandate to maximize output. This led to four main problems. (I follow Milton Friedman here). First, the cooperatives maximize average output per worker, not total output. They restrict employment, just like how a monopsonist would, in order to maximize their wages. Second, as a consequence of this, resources would be poorly allocated across sectors. People would be unable to move to the capital intensive sectors. Third, you could not take your employment with you. As you got to the end of your career, you will choose higher pay now, instead of investment for future generations. Finally, and this is now from Kremer (1997), if the distribution of skill is skewed then the cooperatives will have excessively egalitarian wage schedules. The lower skilled members will redistribute from the higher skilled members to themselves.
The solution to each and every one of these problems is property rights and prices. In the first case, someone owns the company, and pays each employee their marginal product. Since they don’t have to pay everyone the same wage, they can then hire the number of employees which maximizes social product. This also solves the sectoral misallocation problems, as now there is now incentive to maximize the labor share – only total output. The third is solved by making the cooperative into a corporation, with the employees getting shares which they can then sell, or receive indefinite payments from. And the fourth is due to voting not being a price mechanism – there is no way to indicate how much you want something, only yes or no.
This is something I believe you should have burnt into your soul. Markets are not something to be avoided. There is no purity found in avoiding them. Instead, they are a tool. In many cases, they are the best one. When I discuss market failures, like in the market for innovation, and what we can do about them, this is not a call for socialism. It is a call for improvements which are built on the foundation of markets and prices.
Where do you think the intuitive distaste for markets comes from (if you grant that it exists)? Is it that markets improve the legibility of our status hierarchies, and we would prefer to not have our relative value to others be quantifiable? That unfortunately might make the benefits of markets inseparable from negative moral intuitions about them.
Although I don't know how I would square this argument with the even more rigid and quantifiable status hierarchies that existed in largely marketless feudal societies, even amongst the top stratum of individuals that had some sort of power.