Against Business Schools

Ethics in economics, power in the labor markets, and the social responsibility of business

Milton Friedman had an extremely influential article arguing that the only social responsibility of business is to increase its profits. Managers, to Friedman, have no business claiming that they have responsibility to society. If the shareholders of a corporation wished for the manager to contribute to social causes, they are free to direct the manager to do so, or to form a corporation for that purpose. Otherwise, the manager would be misdirecting the money of other people. Instead, the manager should be focused on producing the best products at the best cost. Though his behavior may be selfish, it is well-aligned toward making the world better.

Increasing one’s profits need not, however, be socially beneficial. Firms often have market power which, for some reason, they do not exploit. They may not know that they have it, or they may suspect it but believe that exploiting it is unfair or unethical. My argument in this essay is that beliefs about fairness are indeed efficient, and they keep us from behaving in socially destructive ways. Business schools, in teaching businessmen how to raise profits, are upsetting a cooperative equilibrium, and their effect on society is considerably worse than may appear.

But first, an example.

Dube and Misra are two professors at the University of Chicago’s Booth School of Business. Ziprecruiter is an online hiring website which matches jobseekers to employers. Ziprecruiter makes its money by charging businesses for the right to post a job advertisement. Ziprecruiter wanted to know how much they should price their products at, and approached Dube and Misra for assistance. In particular, they wanted to know two things: first, if they have a uniform price, if they should raise it; and if they would increase profits by offering an individualized price to each customer based on observed characteristics.

Dube and Misra implemented an experiment, where new customers (after filling out some basic information about their firm) are given randomized prices everywhere from $19 to $399. Before the experiment, the price Ziprecruiter charged to small firms using their service for the first time was $99 a month. Dube and Misra, found that Ziprecruiter was significantly under-pricing their product – of the prices they tested, $399 was the most profitable. With some assumptions about how consumer demand is distributed, they estimate that $327 is the profit maximizing uniform price. If they give a personalized price for everyone, the average price falls to $277, although prices offered range from $126 to $6,292.

Ziprecruiter took all of this in mind, and raised their price to $249 a month.

Doing this lowered welfare. With personalized pricing, there is some argument for it increasing welfare – if it allows more people to be served – but not for a price hike. The marginal cost of allowing an additional firm to post a job advertisement is plainly close to zero – there is perhaps some cost of congestion for looking through more postings, but this could hardly be close to $150. Dube and Misra knew this going in, and knew that their would would make the world worse.

When this paper was presented in class, I asked the professor if he felt he had any ethical responsibility not to assist if a company were to ask him to help raise their profits in such a manner. He replied that he did not feel he had such an ethical obligation, and along the way made some remarks about business schools. In economics, he said, we study market power in order to see how far we are from the optimum. In business schools, they study market power in order to learn how to do it better and harder!

And there is evidence that MBAs affect what businessmen choose to do. Acemoglu, He, and le Maire (2023) use detailed information on managers in Denmark and the United States, combined with manager deaths and retirements as an instrumental variable, to estimate how much appointing a manager trained at a business school reduces worker wages and the labor share by. They find that wages decline by 6% in 5 years, and that this is not driven by changes in the composition of workers. When plausibly exogenous positive shocks due to increased foreign demand come, non-business managers are more likely to share the newly gained rents with workers. The working paper version has suggestive evidence that going to business school is in fact causal, although the fact it was dropped makes me concerned.

It’s not all bad. Being a better manager is not merely waking up to the fact that you could exploit people, but are not. There is extensive evidence – I refer to “What Drives Differences in Management” as an up-to-date reference – that better management practices increase the productivity of a firm. Nobody need lose out when someone starts recording inventories or why machines broke down, and these simple practices will increase productivity. Nor do we think that it is unreasonable to reward competence, take attendance, and assign people to different tasks.

Nor is firing people necessarily evidence of zero-sum competition. So often, efficiency improvements don’t take the form of increasing the output of a particular project, but of letting go the people who aren’t contributing anything. If you wanted me to advise New York City how to construct subway tunnels at less cost, my first piece of advice would be to simply fire all the people contributing no value! They are free to go do literally anything else, and in doing so, raise output.

Firing people could be the firm waking up to its monopsony power, though. You really can’t tell without separately estimating the markup on materials or the markdown on labor. I am extremely hesitant to cite them, though, because there’s a lot of shaky stuff going on in the macroeconomic estimation of markups. I do not want this essay to turn into an “everything wrong with De Loecker-Eeckhout-Unger (2020)”, as I have one in the works. I will instead point you to this article from Brian Albrecht to cover everything up to 2024 (and then to the new James Traina paper, which hasn’t come out yet but is good).

For starters, you have papers like “Labor Market Concentration” by Azar, Marinescu, and Steinbaum (2019), which is just running regressions on measures of market concentration (as defined by HHI, which is the share of a market controlled by the top 4 firms), instrumented by a modified Autor-Dorn-Hanson style shift-share. In other words, they use national trends for some job type to control for local shocks. Just as with ADH, though, this means that exogeneity is broken by national trends. ADH could convincingly argue that China was getting more productive, and not that the United States was getting less productive – I don’t think that they can do the same! And anyway, isn’t running regressions on market concentration the sort of structure-conduct-performance stuff we were trying to get away from? Azar, Berry, and Marinescu (2022) take an approach that can actually answer the question, and do find a substantial gap between marginal product and wage. It is still dependent, admittedly, on somewhat shaky instruments (in this case, the tendency toward uniform wages for all employees of a given job type).

I think the best paper, as far as I can tell, is Yeh, Macaluso, and Hershbein (2022). They avoid many of the criticisms of DLEU by only using Census data, which does limit things to manufacturing alone. Markdowns have increased considerably since the early 2000s, although they did decrease from the 70s till then. Something of a wash on the story that business schools are encouraging the exploitation of the unexploited.

But all this has gotten too far away from the original point of the article. We have a sense of some things being fair and unfair. It doesn’t seem right to us for someone to extract everything they can from their workers or from their relations. I think this is the same sort of thing which allows us to cooperate in prisoners’ dilemmas, and not collapse to behavior that nobody wants. Business schools teach an approach to profit maximization which is fools’ gold, taking into account only your own profits, and not your profits if society as a whole should adopt it.

Despite what Friedman wrote, we economists actually do have a social responsibility. We cannot use anyone else’s money for it, but we can choose not to assist projects which are likely to have negative effects. I would think that one should not be a management consultant for Auschwitz.

A brief appendix. A reasonable counterargument for a firm extracting monopoly profits is that, by doing so, it provides an incentive for other people to innovate and provide better products. My lack of mathematical training is a hindrance here – I cannot construct an example where it raises welfare, but neither can I prove it does not. My best argument would be that if the generous monopolist were producing until marginal cost, we are at the first best optimum; and if they were underproducing given their price, another firm could enter.

Suppose firms pay a fixed cost of .1 to enter. Demand is [0,1] – [1,0] and the marginal cost of the firm is 0. The profit maximizing firm would pay to enter, then price at .5 and make a profit of .15 after the fixed cost. Now suppose the firm forewent the profit, and just priced at .2, producing .5 units. Another firm could enter, pay the fixed cost, and produce more units, whether or not the first firm maximized their profits. All that matters it the opportunity for profits, not whether it was exercised, so long as production is at or above marginal cost.