Friedman’s Legacy: From Doctrine to Theorem

Friedman was more right than his detractors claim and more wrong than his supporters would like us to believe. However, after 50 years of lax enforcement and a digital revolution that increased network externalities and created digital monopolies, we cannot continue to assume we live in a society where there are no monopolies and where corporations do not have much influence on legislation, as Friedman did.

Editor’s note: To mark the 50-year anniversary of Milton Friedman’s influential NYT piece on the social responsibility of business, we are launching a series of articles on the shareholder-stakeholder debate. Read previous installments here.

Love it or hate it, the piece that Milton Friedman wrote in the New York Times 50 years ago should be considered one of the most influential op-eds of the 20^th century. The fact that 50 years later the best minds in economics and the law have been willing to debate its consequences on ProMarket is a testimony to the enduring legacy of Friedman’s contribution.

In the popular press, Friedman’s principle is treated as a matter of religion; while in the more formal academic publications, it is ignored (after all, it was not written as a formal theorem). As a result, even half a century later, it is difficult for an educated reader to grasp what Friedman’s enduring lesson is. The 27 articles ProMarket published provide an excellent overview of the diversity of opinions still existing on this topic. In my concluding piece, it is impossible to do justice to this variety. I will try my best to weave them together in a common thread, highlighting where Friedman was right, where he was wrong, and where the matter has not been settled yet. 

First of all, there are two ways to read Friedman’s contribution. Is Friedman writing about what is optimal for shareholders or what is optimal for society? The title seems to suggest the second interpretation, but in the text, Friedman focuses mostly on the first. The first dimension, emphasized by Steven Kaplan and Lucian Bebchuk and Roberto Tallarita, is less controversial. While Friedman gets the law wrong (corporate managers are not “employees of the owners of the business,” but of the corporations), he gets the substance right. If shareholders desire “to make as much money as possible,” corporate managers, who have a fiduciary responsibility towards shareholders, need to focus on maximizing profits.

The second dimension—i.e., whether it is socially desirable that firms maximize profits—is much more controversial. In his piece, Alex Edmans correctly compares Friedman’s idea to the celebrated Modigliani and Miller Theorem, a benchmark that every finance student confronts. The Modigliani and Miller theorem is not an article of faith; very few people believe that in practice the cost of capital is totally unaffected by the financing mix, but it is an extremely useful instrument to understand why financing matters. The problem was that in the New York Times piece, Friedman’s idea was not stated as theorem, but literally as a “doctrine,” triggering a religious reaction, rather than a more balanced academic response.

Five decades later, it is important to fix this mistake and restate Friedman as a theorem. Under what conditions is it socially efficient for managers to focus only on maximizing shareholder value? First, companies should operate in a competitive environment, which I will define as firms being both price and rules takers. Second, there should not be externalities (or the government should be able to address perfectly these externalities through regulation and taxation). Third, contracts are complete, in the sense that we can specify in a contract all relevant contingencies at no cost. 

If these conditions are satisfied, Friedman’s result, which I will label the Friedman Separation Theorem, holds. In the 1930s, Irving Fisher established that a firm can determine how much to invest solely on the basis of the market interest rate, regardless of the temporal preferences of its shareholders. This result is known in the literature as the Fisher Separation Theorem. Friedman significantly extends Fisher’s results establishing when corporate managers can ignore not just the temporal preferences of their shareholders, but all of their preferences and the preferences of their workers, suppliers, etc. Friedman Separation Theorem makes the life of a corporate manager much, much simpler—this is one of its many attractions.

Any well-trained economist will recognize that my formulation of the Friedman Separation Theorem is nothing more than a restatement of the celebrated First Welfare Theorem, establishing the social optimality of competitive equilibria. Proven in its generality by Arrow and Debreu in 1951, this idea was initially relegated to the abstractions of mathematical economics. Friedman has the merit to see the tremendous practical implications of this theorem and to push it into the real world.

Addressing a general audience, Friedman also understands that the mathematical subtleties of the Arrow and Debreu proofs will not cut it. Instead, he chooses to present an argument that appeals directly to the core American values of freedom and independence. In a free economy, Friedman notices, stakeholders voluntarily get together. By and large, they choose to provide everybody a fixed-payoff contract, except the shareholders. As a result, shareholders absorb all the risk and reap all the rewards of the company’s performance. In this context, maximizing profits (i.e., the residual after all the fixed contracts are paid) is tantamount to maximizing the size of the entire pie.

Furthermore, under those assumptions, any additional social responsibility imposed on a company is a cost born by shareholders (and only shareholders). Thus, imposing this responsibility against their consent is tantamount to taxation without representation. Thus, Friedman unleashes the shareholder revolution appealing directly to the American Revolution. 

Given these assumptions, the Friedman Separation Theorem holds. But do these assumptions hold, at least approximately, in the real world? If they do not, should the Friedman Separation Theorem be thrown away altogether? Let me analyze the assumptions in reverse order.

Are contracts complete? The answer is a resounding “no.” When contracts are incomplete, all the stakeholders (not just the shareholders) are affected by corporate decisions. When a company fails, the suppliers lose their credit, the workers experience a  drop in wages, and entire communities are devastated. In a world of complete contracts, shareholders could have provided full insurance against this risk. In a world of incomplete contracts, they cannot. Anticipating these partial expropriations, workers, suppliers, and customers will be reluctant to make a firm-specific investment, reducing the value of the firm. In other terms, when contracts are incomplete, maximizing the firm’s value and maximizing shareholder value are not one and the same, and Raghuram Rajan is right: managers should maximize the firm’s value. If that is the objective, why should we leave this decision solely to shareholders?

This is where Friedman’s “proof” of the theorem becomes handy. Let’s look at what people choose voluntarily. If we rule out externalities (e.g., tort claimants) and we focus on contractual stakeholders (suppliers, customers, employees, and financiers), we should ask ourselves why the for-profit corporate form, where shareholders are given the exclusive rights to appoint directors, is so widespread. It is possible for businesses to organize as cooperative, as non-profit, and in recent years even as benefit corporations, where managers are required by law to consider all stakeholders in their decisions. Why do businesses keep choosing the traditional corporate form in overwhelming numbers? When Leo Strine was still a Delaware Justice, he stated that power is purpose. If companies keep giving shareholders (and shareholders only) the power to elect corporate directors, this implies that they want those directors to do what shareholders want: to maximize the value of their claims.          

I am old enough to know that the evolutionary argument in economics is often abused. Even in the biological world, evolution leads to the survival of the most prolific, not necessarily of the fittest organism. Why should it be different in the economic world? It is easy to concoct examples where inefficient organizational forms prevail because of wealth constraints (especially of workers) or bounded rationality. Yet, the burden of proof is on the stakeholders’ advocates. Even in the most progressive Silicon Valley start-ups, workers receive stock options, but not voting rights. Companies only offer voting rights to workers where they are required to by law, as in Germany. They do not do so voluntarily. If it is so efficient to allocate votes to workers, why don’t they do it voluntarily?      

The empirical implausibility of assumption three does not, in my view, undermine the practical implications of the Friedman Separation Theorem, but it should temper its most cynical interpretations. Companies should maximize shareholders value, but not in an opportunistic way, or at the expense of other stakeholders. As Friedman himself recognizes, corporations should “make as much money as possible while conforming to their basic rules of the society, both those embodied in law and those embodied in ethical custom.” It is not Friedman’s fault, but society’s fault (and the fault of us teachers of business executives) if the ethical customs of business have deteriorated so much that they do not impose any limitation to opportunistic behavior other than those imposed by reputation.

When it comes to the second assumption, nobody in their right mind will defend the idea that we live in a world without externalities. Thus, to believe that assumption two holds (at least approximately) we need to believe that the government is relatively good at addressing these externalities via taxation or regulation (including the creation of “rights to pollute”).

Before we address the plausibility of this assumption, we need to confront the fact that corporations are born with an original sin: the ability to externalize some of their costs. While limited liability vis-à-vis contractual claimants can be achieved via ordinary market contracting, limited liability vis-à-vis tort claimants is a privilege granted by the State. Historically, as Stefano Zamagni reminds us, this privilege was granted only when a social purpose (like building a road or a bridge) was involved. Over time, the explicit social purpose was dropped, because the economic growth triggered by freedom of incorporation was considered a sufficient social benefit.

Yet, it is important to remember that corporations are not just purely contractual entities; they are granted by the State an extraordinary privilege and, thus, the State itself can demand something in exchange for this privilege. The British Academy goes too far in this direction, asking to reintroduce an explicit mandatory social purpose for corporations. This is a very dangerous position because it delegates to the state the right to decide not only what is good and what is bad, but also which business can be pursued and which one cannot, undermining the freedom of enterprise. But some milder mandates (more on this later) could be justified.   

Having understood that when it comes to corporations the production of externalities is a feature, not a bug, we are left assessing how good the government is at addressing these externalities. As economists, our preferred way to deal with externalities is taxation (referred to as Pigouvian taxes). As I explain in my book, A Capitalism for the People, political economy considerations make the approval of Pigouvian taxes very difficult. Just look at the carbon tax, overwhelmingly supported by economists, but rejected even in the most liberal states of the union. Thus, claiming that Pigouvian taxes will address all the externalities is untenable.

When it comes to regulation, the argument is more subtle. There is certainly a lot of regulation aimed at addressing the externalities produced by corporations. For relatively small companies, it is debatable whether this regulation is enough. But when we look at very large corporations, the ones Anat Admati defines as too big to jail, regulation becomes ineffective. Roy Shapira and I have studied how, in 1984, DuPont decided to dump lethal toxic waste in the Ohio River and how it got away with it for more than thirty years. Not only was DuPont able to capture the regulators with a smart combination of revolving doors and aggressive lobbying, but it was able to hide the very dangers of the substance it was producing and freely disposing of in the environment. When discovered, DuPont was able to draft the regulation to its own needs to avoid liability and was successful in doing so while its CEO was celebrated as an environmental leader.

It was only a combination of a clever lawyer and many strokes of luck that eventually led to a conviction of DuPont more than thirty years later. All the corporate scandals listed by Admati suggest that DuPont is not the exception, but the rule: large companies are subject to regulation only de jure, not de facto.

If the government is unable to fully address the externalities, should managers maximize profits? From a societal point of view, the answer is clearly “no.” But the answer is “no” even from a shareholder point of view if shareholders have some social objectives. Friedman recognizes this possibility by saying that the desire of investors “generally will be to make as much money as possible.” Yet, he is quick to suggest that even when this is not the case (as when shareholders are interested in charity), it is better for companies to maximize profits, distribute them, and let shareholders donate those profits to their favorite charities. 

This result had enormous consequences for the asset management industry. If the Friedman Separation Theorem holds, asset managers do not have to care about the social preferences of investors, they just have to maximize financial returns. For example, one of the goals of the Bill & Melinda Gates Foundation is to reduce inequities in health by developing new tools and strategies to reduce the burden of infectious disease and the leading causes of child mortality in developing countries. However, their asset managers do not have to decide whether the firms they invest in are pursuing or undermining the goals of the foundation. Managers can invest purely based on financial considerations, to maximize the money the Gates Foundation has to pursue its goals. Is this the best way for the foundation to reach its objectives? No.

Oliver Hart and I show that Friedman’s result only holds in the knife-edge case when a company does not have any comparative advantage vis-à-vis shareholders in pursuing a social objective. This is certainly the case for corporate charity. One dollar donated by Microsoft is equally as effective as a dollar donated by me. Yet, it is not the case for most other social objectives. It is cheaper not to pollute than to pollute and clean up. It is cheaper not to sell guns to the Saudi than to do so and then help the Yemen refugees, and so on and so forth. Why should companies pursue the maximization of monetary returns at the expense of the social goals of their investors? If I am a 100 percent owner of a company, I will be running it to maximize my utility, not my wealth. Why should it be different when many of us own it? The modified Friedman principle should be that managers should maximize shareholders’ welfare, not value.

As Eugene F. Fama and John G. Matsusaka correctly point out, the difference between a firm owned by an individual and one owned by a multitude of individuals is the difficulty in reconciling different opinions. When there are multiple owners, they will disagree on what social objectives to pursue. Some very conservative investors would want a pharmaceutical company not to pursue an abortion pill even when such a pursuit is profitable, while some very liberal investors would like to sell such a pill below cost to help poor pregnant women. How to reconcile these preferences? We know from Arrow’s impossibility theorem that no voting system can convert the preferences of individuals into a community-wide ranking, while also satisfying a set of desirable properties. Yet, Arrow’s impossibility theorem has not led us to give up political democracy, so why should we give up shareholder democracy?

The difference between maximizing shareholder welfare rather than value might appear to be mere semantics, but it is big. It really destroys the separation result that made Friedman’s theorem so appealing to the asset management industry. Taken seriously the Hart and Zingales result will make corporate democracy necessary, revolutionizing the way companies are run. This democracy, as all democracies, will impose information costs. In spite of these costs, shareholder pressure seems able to reduce toxic chemical releases and inmate suicides in private prisons. Whether the costs are worth the benefits is still open for debate. 

Information costs aside, can these externalities be resolved by the private sector? In a recent paper with Eleonora Broccardo and Oliver Hart, we show that if the (wealth weighted) majority of investors are even slightly altruistic and if they are well diversified, then they will vote as a benevolent planner would, so they will lead companies to internalize externalities. Thus, there is hope that shareholder democracy could fix even this problem, but only if we let democracy work. Unfortunately, the Department of Labor is currently trying to prevent asset managers from considering any factor other than financial returns in casting their corporate ballots. It is ironic that alleged libertarians resort to the coercive power of the state to prevent shareholders from expressing their opinions about the companies they own. If we eliminate these restrictions, however, the modified Friedman principle (i.e., maximization of shareholder welfare) leads to desirable social outcomes.   

The really problematic assumption is assumption number one. Friedman himself recognizes that a monopolist maximizing shareholder value is not good for society. Yet, writing in 1970, at the peak of the US antitrust enforcement, Friedman was not overly concerned about monopolies. After 50 years of lax enforcement and a digital revolution that increased network externalities and created many digital monopolies, we cannot be so cavalier. Consider Google’s choice of how to rank news about political candidates. Experimental evidence shows that searches leading to a higher ranking of bad news about a political candidate result in fewer votes for that candidate. A board that maximizes profits would tweak the search results to disadvantage those candidates who want to tax Google or to break it up. In a competitive market, reputation prevents these abuses, but Google is a de facto monopolist. It will take a lot of abuses for a customer to switch to Bing. Do we want Google to manipulate elections?  

While Friedman was well aware that the “price taker” assumption could be violated, he was not equally aware of the “rule taker” assumption. In his essay, he states that corporations should “make as much money as possible while conforming to their basic rules of the society” as if these rules were exogenously given. Only six months after Friedman wrote his NYT piece, George Stigler published his regulatory capture piece in an academic journal, explaining how regulated companies distort regulation in their favor against the public interest.

Stigler and Friedman were not only colleagues at the University of Chicago, but also close friends, who had lunch together almost every day. It is hard to imagine that Friedman had not been exposed to Stigler’s idea even before he wrote his NYT piece. Certainly, he was exposed to it afterward. Yet, there is no record (at least none that I am aware of) of any attempt to integrate the two points of view. 

If lifting environmental regulation has the effect of killing thousands of people, but increases the profits of chemical companies, should a CEO pursue this objective through any legal means at her disposal? I refuse to think that Milton Friedman would have endorsed this idea. After all, in his 1970 piece, he wrote that maximization of profits should be pursued “so long as it stays within the rules of the game, which is to say, engages in open and free competition without deception or fraud.” Thus, the only society in which Friedman’s principle applies is a society where there are no monopolies and where corporations do not have much influence on legislation. This might have been the world Friedman lived in then, it is certainly not the world we live in today, as Martin Wolf clearly describes.  

Thus, in 2020, how should we interpret the practical implications of Friedman’s idea? If you are a small to medium-sized company, let’s say Chuck E. Cheese, a company with no market power and no real power to influence regulation or elections, maximizing shareholder welfare is the right goal to follow. Especially if this goal is pursued with attention not only to legal rules but also ethical customs, like Friedman advocated, but most companies ignored.

When it comes to super corporations, corporations that have market power, like Google and Facebook, or political power, like BlackRock or JP Morgan, or regulatory power, like DuPont or Monsanto, a single-minded pursuit of shareholder value maximization can be extremely bad for society. This is the reason why Oliver Hart and I have advocated requiring boards of monopolies, like Google, or of firms too big to regulate, like Blackrock, to maximize social welfare, the utility of society as a whole, not shareholder welfare. 

How do we convince a board elected by the shareholders to maximize social welfare? The simple answer is that we impose an additional fiduciary duty toward society, in addition to the existing one toward shareholders. This fiduciary duty would make directors personally liable (for a multiple of the director’s fee they receive), if the company inefficiently pollutes, influences legislation, or abuses its monopoly power. 

The proposal advanced in this debate by Colin Mayer, Leo Strine, and Jaap Winter goes in a similar direction. They require companies with sales above $1bn to adopt the Delaware public benefit corporation statute. This statute requires directors to “balance the pecuniary interests of the stockholders, the best interests of those materially affected by the corporation’s conduct, and the specific public benefit or public benefits identified in its certificate of incorporation.” This is tantamount to a license for directors to do whatever they want. It is difficult to imagine that directors, elected by shareholders, will change their behavior based on this prescription. By contrast, the fiduciary duty comes with an already existing enforcement mechanism.

The Mayer et al. mandate is triggered by a fixed level of sales ($1bn). We can quibble where it is too low (even Chuck E. Cheese might trigger it). We think that a flexible criterion, which looks at the effective political and regulatory power of the super corporations, might be more effective. This criterion would be similar to the Systemically Important Financial Institution (SIFI) label imposed by Dodd-Frank. In this way, the mandate will have an effect of discouraging opportunistic behavior even by firms below the threshold, which fear to fall under the mandate. 

In sum, Friedman was more right than his detractors claim and more wrong than his supporters would like us to believe. His “theorem” has greatly contributed to determining when maximizing shareholder value is good for society and when it is not. The discipline imposed by Friedman’s theorem also forces greater accountability on managers. In the world of 2020, the biggest shareholder in most corporations is all of us, who have their pension money invested in stocks. We are the real silent majority. Corporate managers finance political candidates, lobby for self-serving legislation, and capture regulation. They have the power to use our money to fight against our own interest. While Friedman did not anticipate these degenerations, he warned us against the risk of unaccountable managers. This warning will remain his most enduring contribution.