Jennifer Nou, "The Nature of the Agency"

In The Nature of the Firm, Ronald Coase argued that private firms arise because of market transaction costs. His theory yielded predictions about the scope of firms and their internal design. What power, if any, do these insights have for understanding administrative agencies? After all, social scientists have long conceived of political actors as engaged in market-like transactions. The lack of profit-maximization and price signals, however, has also drawn skepticism about the parallels. These questions will be explored amidst contemporary debates about the administrative state.

Jennifer Nou is Professor of Law and Ronald H. Coase Teaching Scholar.

This Coase Lecture was delivered on January 22, 2019.

Transcript

 [APPLAUSE]

JENNIFER NOU: So thank you, Dean Miles, for that extremely generous introduction. It's an honor to be giving the annual Coase Lecture this year. And indeed, as Dean Miles mentioned, in 1991, Ronald Coase won the Nobel Memorial Prize in Economics.

And as Coase tells it in the lecture accompanying the award, he first had the idea that will form the core insight explored in this talk-- indeed, the theory of the firm-- at the tender age of 21. He remarked, is a strange experience to be praised in my 80s for work that I did in my 20s. Indeed, his little paper, "The nature of the firm"-- it's little. It's 20 pages long-- was basically ignored for three decades before eventually serving as the foundation for entire subfields of research today.

Now, let me tell you, this is a very heartening thing to hear both for me and for you. Because for me, this suggests that there's still hope that somebody maybe decades in the future will actually read my work. For you, this means that most of you-- gosh, in your 20s now-- here at this law school in the next few years could have an idea that overturns a conventional wisdom.

And I think that prospect is highest and likeliest-- no offense, 2Ls and 3Ls-- when you are in your first year, when you still have the outsider instincts. You're not yet fully acculturated into the law. You still have the instincts to ask, wait, did what Professor Levmore just say about the efficiency of tort law really make any sense? What is this account missing knowing what I know about the real world?

So what was the conventional wisdom that the young Coase questioned and challenged in the late 1930s? To state it very simply, it was that prices coordinate market exchanges of goods in such a way that consumers that value the goods highest will get them.

So the producers of these goods, economic firms themselves, are usually price takers. And they will use inputs to produce these valued outputs by using capital and labor to maximize profit, thus economic nirvana achieved, unlocked. We've got efficiency.

So sitting there in his graduate seminar, the young Coase asked-- there was a puzzle that arose for him. He said, you know, if that's right-- that prices do all the coordination necessary for production-- then what was the point of these guys? What was the point of these guys? And the lack of reaction suggests to me that I'm already culturally out of touch.

This is David Brent from the UK version of The Office, the better version of The Office. And he is the main manager of Wernham Hogg, a company that produces paper. And Coase asks, why are there inputs to production like managers in firms whose function it is to also coordinate these exchanges of goods and services?

As Coase put the question vividly, quoting DH Robertson, what explains these islands of conscious power in this ocean of unconscious cooperation-- the invisible hand of the market-- like lumps of butter coagulated in a pail of buttermilk? What was Coase's answer? How did he solve this puzzle, if you will?

Well, he realized that there are costs to using the price mechanism. In his words, what the prices are have to be discovered. There are negotiations to be undertaken. Contracts have to be drawn up. Inspections have to be made. Arrangements have to be made to settle disputes and so on. And these, of course, are known as "market transaction costs."

By contrast, transactions within a firm usually occur through what you might call "authoritarian fiat." So let me give you an example. So I searched the web. And I found an economic firm called the Conestoga Restaurant located in Lancaster, Pennsylvania, which might explain why the pizza looks as it does.

And the restaurant tells us that the pizza dough is made fresh daily in house. No need for the Conestoga Restaurant to contract with Pillsbury pizza crust, as I'm sure other restaurants in Lancaster probably do, to get the pizza crust. You don't have to haggle over what the price is, what's going to happen when, gosh, there's some natural disaster and pizza crust can't be supplied.

Instead, the manager of Conestoga Restaurant can simply turn to his employee and say, employee, this is how you're going to make the pizza crust. And if circumstances change and suddenly consumers want gluten-free pizza, well, gosh, that's easy to do, too. You can just tell by command your employee how to make the pizza. There's no costly haggling or bargaining going on.

So the punchline then for Coase was-- this is a nice little piece of abstract art that I made for you to hopefully represent this. The punchline was that the boundary of the firm will arise just at the point on the margin when the marginal transaction costs for making some exchange on the market is higher than in the firm. In other words, you're going to want to bring a transaction in house when the transaction costs are lower in house than contracting it out onto the market.

So I've deliberately, by the way, left labels off this because I'm going to come back to it. These lines, if you will, and these institutions can mean various things. But this insight implies that a firm size decreases with the increasing marginal costs of organizing more transactions within the firm.

So for example, if there are decreasing returns on managerial ability-- David Brent is suddenly making the pizza-- well, gosh, then there's more incentive to contract that out and have the pizza dough made on the market. As I've suggested, many have since broadened the problem ultimately as one of comparative institutional arrangements. It's not just a question of the market versus the firm.

Rather, it's a broader question of how do different institutions and organizations deliver goods and services. And so in other words, it's not just about the boundary between the firm and the market. But you can also look at boundaries within a firm and understand these boundaries along a spectrum all the way from the market to the firm to within the firm and so on.

Now, what does any of this have to do with the main topic of my remarks today regarding an institution that has become just as salient as the private companies of Coase's youth? This talk is called The Nature of the Agency. It's a riff, obviously, on Coase's paper. And I'm talking, of course, about administrative agencies and their vast variation in form.

Now, this next slide is going to give heartburn to the libertarians in the room horrified at how big our government has become. But this is a list of the dozens and dozens of administrative agencies that look extremely different in form and subject area. So you've got your familiar Environmental Protection Agency. You've got another familiar Food and Drug Administration. You've got less familiar ones, like the Federal Library and Information Center Committee, the Foreign Claims Settlement Commission.

So suffice it to say that there is a lot of variation and a lot of administrative agencies. And the question is, can their emergence and this variation be explained or eliminated at all by the ideas like transaction cost? And what's to be gained by attempting to do so?

Now, this question, I should hasten to say, has been pursued in earnest by political scientists for decades. I'm not asking it for the first time. There's seminal work by Terry Moe, Barry Weingast, among many others.

And so in the remainder of my remarks today, I want to address first the initial objections that you have to applying economic models to the political sphere. And I'm going to address and talk about how your objections have been answered or attempted to been answered in the extant literature.

Next, I'm going to discuss what I perceive to be important blind spots in that literature when thinking about the bureaucracy. And I'm going to propose what I hope are some relatively new directions for that work, addressed both to social scientists and also to lawyers. And then I'm going to end with just some brief methodological reflections on law and economics as a whole.

So let's begin with your initial objections to the analogy between the economic and the political spheres. Political behavior, of course, does not occur on a free market with prices, nor is there an incentive for profit maximization as such.

But one response to that might be to say that this objection goes to skepticism about whether political markets, as it were, maximize social welfare. I mean, who would think that politics maximizes social welfare, right? It does not foreclose the possibility that political actors, like economic ones, are utility maximizing. Perhaps we can understand that utility function in terms of maximizing the chances of re-election, maybe maximizing the probability that one's policy preferences will be achieved.

Another possible response to this objection is mainly to resist the false dichotomy between economics and politics and instead to suggest that political institutions are just another kind of firm in the market-to-firm spectrum. Like private firms, the government often participates in the market both as a buyer and seller. And the boundaries between private firms and governmental institutions are famously blurred.

So take, for example, the United States Postal Service. You can find it at www.usps.com. It thinks of itself as a commercial enterprise. You're skeptical that that's any evidence. But believe me, the bottom line does matter at the USPS. Some people praise it as one of the most efficient government institutions out there, which might suggest the baseline is low.

Also in this category you've got Freddie and Fannie Mac, government-sponsored enterprises that have different features of market dynamics and governmental power. And on the flip side, economists like Luigi Zingales here at our very own University of Chicago have argued very recently that firms, particularly large corporations like the Google and Facebooks of the world, are themselves political actors, thus the need for what he calls a "political theory of the firm."

And I think this recent move in the theory of the firm literature is potentially fruitful because not only does it help mirror developments in the real world. But it also takes us even further afield from simplistic textbook ideas about firms as mirror production functions. After all, not all firms maximize profit. We've got nonprofit corporations, as well as for-profit corporations, who now profess to pursue socially responsible ends instead of shareholder value. And Henry Hansmann and many others have applied the transaction cost of economics even to these non-for-profit-maximizing firms.

So if social scientists are right that transaction cost economics can be meaningfully applied to politics, then how and why do these political actors create administrative agencies? Now, for the 1Ls in the room, it's first important to understand that administrative agencies are creatures of statute, statutes passed by Congress. Agencies have no authority to act. Their actions are [INAUDIBLE] if they lack delegated power from Congress. And this fact may help to explain what has been called the congressional dominance of political science models on the bureaucracy to illustrate this strand.

I just want to tell you about one very famous argument advanced by David Epstein and Sharyn O'Halloran which explicitly draws on the transaction cost framework. So in their account-- and you're not meant to see the box. It's just meant to suggest how complicated the legislative process is. All these boxes tell you is about all the steps. It's not Schoolhouse Rock's a bill becomes a law.

The real process of legislating is really costly and expensive. And they suggest that one way to think about what Congress does is they face a make-or-buy decision. That is to say, they can make policy by passing a statute through this incredibly time- and resource-intensive committee process by specifying a statute very specifically about what that policy is going to be.

So let me give you an example. So let's say Congress is debating how to, I don't know, gosh, reduce air pollutants. And gosh, one option on the table is to pass a statute that mandates a particular kind of technology.

And these generalist legislators-- not always the brightest bulbs in the room-- are going to have to be debating things like, geez, should we use wet scrubbers, fabric filters, electrostatic precipitators, activated carbon injection afterburners, other options to pass and choose a statute that says something like, all major sources of air pollution shall install electrostatic precipitators. That's going to entail a lot of transaction costs.

By contrast, it can pass a statute that instead buys a policy from an agency, buys in the sense that it's still costly to run the bill through Congress. But gosh, it's going to be a lot less costly. It's going to require a lot less information to pass a different kind of statute that instead says, all major sources of air pollution shall be reasonably regulated by the EPA. That's delegating the authority to Congress.

And they want to convince us that that is like-- they are just analogies, but is like-- facing the make-or-buy decision that is to make the pizza dough, the statute, in house or to contract it out through incomplete contract to an agency. And the implication, of course, is Coase's implication that when the transaction costs are lower in Congress, they're more likely to make it in house rather than to contract it out.

And when is this most likely to occur? Well, first, the condition is that the more uncertain and complex the policy arena-- what political scientists call the "uncertainty principle." The more uncertain and complex the area is, they hypothesize, the more likely Congress is going to contract that out to an agency. Why? Because gosh, debating electrostatic whatever, precipitators, air filters, and all of that is really costly. And it requires a lot of information.

Second, the closer the agency's policy preferences to the median preferences on the legislative floor, the more likely Congress is to delegate. And that makes sense because it's much cheaper for them, at least in the short run, to get a policy made by the EPA that it wants when it knows that EPA is going to adopt the same policy that median legislator also wants.

And these are very intuitive hypotheses. They test them with a lot of empirical data. And it suggests that indeed maybe this theory has some traction, that indeed this occurs. Again, you've got the uncertainty principle and the allied principle. You're more likely to see delegation. Congress is more likely to buy the policy.

And it is because of these principles, if you will, that political scientists-- that is, uncertainty and lack of preference divergence-- that political scientists have earnestly turned to what is called the "politics of structure," the ways in which political actors attempt to lock in their interests and insulate those interests from change by political adversaries in the future. Because the more uncertain something is, the more likely future Congress are going to overturn what you do or a future agency is going to overturn what you do, the more you want to lock that in through structure.

And what do these structures look like? Well, they're always contained in statutes. And political scientists often describe them as increasing political transaction costs.

So some examples of this are, for example, designing an agency with multiple members at the head. So these are, for example, all the boards and commissions that you will learn about in administrative law if haven't taken them. But it's like the Securities Exchange Commission, the Federal Election Commission, the National Labor Relations Board. These are multi-member commissions.

And gosh, it's much harder for an executive to change policy from the status quo when you have multiple members rather than a single-headed agency like the EPA because, first of all, it will take the president more time to get a majority of members on the board that accords with his preferences, especially with partisan balancing requirements. And then on top of that, even if it's the chair implementing the policy, it's much harder to convince a group than just having to get one single person to go along with you.

A second example of this structure is removal restrictions. This is also very common in literature. And this is just the idea that Congress can say that a particular agency head can only be removable for cause. That is, you need to give some kind of a reason before firing that agency head rather than just firing them at will.

And of course, as you imagine, the president has much more control when you can fire somebody at will. It's much easier to change the policy as a result, much easier for the president to change the policy when an agency has a removable at will.

And so the resulting predictions are that in cases of political hold up-- that is, when a future agency is unlikely to act in ways desired by Congress, for example, when there is divided government-- the more likely legislature is to choose structural design choices of the kind that I've just described. In other words-- prediction-- when you have divided government, the more likely you are to see multi-member commissions or removal restrictions.

So in this manner, the dominant way of thinking about agency structure and design has been characterized by what I perceive-- and many others have noted this as well-- a myopic focus on Congress and the statutory structure of administrative agencies. Part of the reason for this is surely because the US code statutes are just easier to code and evaluate than the messiness of what the president does in the federal register.

Executive orders, legal memoranda, presidential declarations all have uncertain varying legal status. It's really hard for political scientists to get their head around and also to measure all the activities of the president. And I think this focus has also come at the expense of much richer thinking about the determinants of the bureaucracy.

And in our time of legislative gridlock and polarization, I submit more important determinants of bureaucratic behavior. Most obviously, presidents actually play a critical role in establishing and structuring agencies themselves through executive order, departmental order, reorganization acts. So in other words, contrary to this picture that Congress is the only one setting up and structuring these agencies, the president does so as well.

So William [? Howell and David ?] [? Lewis ?] have convincingly shown that since the end of World War II, presidents have established 50% of the new administrative agencies that were on my list. These include agencies such as the National Security Agency and the Peace Corps. Even more significantly for my purposes, since 1946, departmental orders are responsible for creating fully 40% of those new agencies.

Now, I said "departmental orders" to contrast it from executive orders from the president. Departmental orders are issued by department heads-- the secretary of Health and Human Services, the administrator of the EPA.

And given that presidents rarely get personally involved in reorganization matters, I think this fact suggests that heads of departments and other agency heads are likely the principal architects of these schemes. They know a lot more about their agency. So even if the president does have some role, they're going to defer a lot probably to how the agency head thinks an agency should be restructured.

So public law and economics, I want to argue, needs to shift its focus from Congress-- and more recently, it's shifted to the president-- to thinking more critically about the role of agency heads in the administrative state. Again, these are the secretaries and the administrators authorized by law to make what I argue are actually the most consequential decisions impacting the agency on the ground.

Why should this be the next frontier in the wave of research? Well, I want to say first, understanding agency heads as principals in their own right appreciates the preference divergences between the president and agency heads that are very real.

So a historical example includes Christine Todd Whitman. This was George W. Bush's appointee to head the EPA. She resigned over substantive policy disagreements with the administration. There was a real clash in environmental policy between the agency head and the president.

And second, I would argue that this shift to agency heads also renews attention to the internal structure and process within administrative agencies, how political appointees attempt to lock in their preferences from attack by future administrators or from attack even from the president or Congress. For example, agency heads make important delegation decisions of their own. That is, they take authority delegated to them by Congress and give that authority to some lower-level official in the agency, often with finality. And I think these design choices, again, are just as, if not more, consequential to specific policy outputs.

Let me just give you a quick example to illustrate this abstract idea. Again, you're not meant to read that. I'm just going to point out the subdelegation. So this is the subdelegation in a famous case called Accardi versus Shaughnessy, which I'll come back to. And what happened in the case is that top box is the attorney general. And he issued a rule. And the CFR refers to the Code of Federal Regulations.

This is a codified rule that says, in considering and determining appeals, the Board of Immigration Appeals shall exercise such discretion and power conferred upon the attorney general by law as is appropriate and necessary for the disposition of the case. There's more. The attorney general can choose to review it. But if he doesn't choose to review it, then the board's decision is a final one.

Some delegations like this, of authority from the agency head to internal actors, have become particularly important under the current administration given the high number of vacancies in high-level positions. Many of these appointees don't even know subdelegations exist.

There are these really funny quotes where historically, at least, some SEC commissioners have said that they've seen something in the newspaper about the SEC taking some action. And they turn around they say, I don't remember voting on that or taking that action. And lo and behold, it's because of some delegated authority to some staff member in the agency. And they made that decision. And they didn't discover that until years into their terms.

So as a result, I think one could imagine asking all the questions that political scientists have asked for decades about congressional delegation to agency subdelegation. And when these subdelegations are given to actors outside of the agency, which is often what happens when-- and the EPA often delegates its authority to state administrative agencies. Then make-or-buy questions arguably arise once again.

Relatedly, agency heads also make important choices about how to organize agencies, offices, and bureaus. Where agencies place particular offices in the internal hierarchy can also impact the substance of agency decision making, as well as the likelihood that an agency decision is going to be upheld in court.

So focusing attention here would help to remedy what I perceive to be another relative blind spot in this literature in agency structure. And that is the link between agency structure and agency effectiveness. After all, mechanisms of political control may help to contain discretion. But they also contribute to an agency's failure to achieve particular goals. In other words, the more you have to go through notice and comment, it's much harder to actually achieve something. So there's a lot of internal gridlock within an agency.

And I think one likely reason for this blind spot has been the increasing segmentation of those studying the bureaucracy into law schools, political science, and economics departments. And critically, into public policy schools, there a lot of classes on public administration. And as a result, I think the literature that public administration scholars have been exploring for decades about organizational effectiveness has really been ignored to the detriment of law and economic scholars in this tradition.

And I think we public law types would do better to incorporate the work. One of the best conferences I went to last year, I met these public administration scholars for the first time and was just blown away by all the really deep, qualitative, thick work that they were doing. It was extremely illuminating.

So thinking critically about the actors, the agency heads that manage and run agencies day to day, arguably brings us full circle back to Coasean transaction cost economics. Subsequent efforts following Coase to explain not only the boundary between the market and the firm but also the boundaries within the firm have tried to take up the mantle. So let me just give you a quick illustration.

So there was a famous study. I'm going to simplify. This is just to give you a little bit of history about what the General Motors corporation used to look like. And this is a simplification of that hierarchy to illustrate two models that economists have explored. This is Oliver Williamson's work in particular.

He notes that in the 1920s, General Motors was organized according to what he calls and others have called the "unitary form." Because this goes into business schools, they have to name it kind of snappy. So they call it the U form of a business.

And what the U form of business looks like is the business is organized functionally. So you've got General Motors at the top of the hierarchy. Then you've got functional divisions, like marketing, human resources, and finance.

And then in the 1930s, General Motors reorganized itself differently into what has then been called the multidivisional form, or the M form, of how to run a business. And instead of having these functional offices, what General Motors did is it established the Chevrolet division and the Oldsmobile division that were much more product focused.

And as a result, Oliver Williamson and others contend that a transaction cost story can be told about these two forms. And it essentially looks like this. In the U form of doing business, it's really difficult to monitor and measure and discern and see the output of these functional divisions because all of their work is being just channeled to the headquarters without any clear sense of where the profit centers are within the firm.

And as a result, he theorized that the heads of these departments-- the marketing department, human resources, finance department-- their budget maximizers are just going to bargain for what the budget should be without a clear sense of the returns to the firm. And he hypothesized that it would be much more inefficient.

And as the scale of the firm increases, he argues that we should have more M-form organizations, where now, because you can tell what profit the Chevrolet division is making or the Oldsmobile division is making, it's much easier. There are lower transaction costs to having activity within the firm that contributes to a firm's profit maximization.

Now, surely by now you again are going to renew your objection that this analogy does not apply to government agencies who don't engage in clear profit maximization and thus lack the incentives to vindicate the potential advantages of the M firm. And I want to ask a question. There's serious objections to this, too.

But I wonder if the objection loses some force when we consider that agencies that engage in cost-benefit analysis may possess more data about their contributions to social welfare that allow for these kinds of evaluations that are more analogous-- not completely so, but more analogous-- to profit maximization in firms. And as courts, Congress, and the president are increasingly using cost-benefit analysis to evaluate agency efforts, perhaps adopting the practice could help to increase agency organizational efficiency and maximizing social welfare.

Quick, quick, very quick illustration. Let me just-- I'll describe this. So this is a hierarchy of the EPA. It has product-based, subject-based offices. There is an office for air, an office for solid waste, an office for water regulation.

And it used to be the case that economists were in each of these little offices. And after a Reagan order mandating cost-benefit analysis in the executive branch, the EPA reorganized itself to centralize functionally all the economists into something called the National Center for Environmental Economics.

And one hypothesis you might have given what I've just told you about the U and M forms is that perhaps this shift from a more M-form organization to a U-form organization where-- again, the social welfare benefits of each of these offices is much more dispersed. Did that change lead to less agency effectiveness within the agency?

I suspect some of you are going to say, can that be empirically tested? Well, we'll talk about it. I think that there are ways that potentially you could get some traction on that. But the point just for now is perhaps debating these decisions against the backdrop of debates about unitary and multidimensional organizations can help impose some much-needed discipline on what now seems very arbitrary and path-dependent decisions about how to reorganize agencies.

Taking a step back, some of the 1Ls in this room and others may be asking, gosh, what does any of this have to do with law? Aren't we in a law school? Well, let me suggest just two underdeveloped legal areas that might benefit from the perspective I'm suggesting today. First is the law of subdelegation.

So one of Coase's insights was that courts are available to enforce market contracts but not to enforce intra-organizational arrangements. So in other words, the fact that GM is a U form, an M form-- no court is going to step in and enforce the org chart that comes down from headquarters.

And if you think that there is some analogy in the administrative agency context for an agency head's choice to subdelegate to an internal actor and thus by a policy-- or even from an external actor by a policy, which is probably the closer analogy-- then the more courts make these subdelegations judicially enforceable, the more one would expect some delegation to occur. Because again, this is just Coase's insight that this subdelegation-- let me show you.

Recall the subdelegation that we had. It's an incomplete contract between the subdelegate and the delegator-- the agency head, in this case. The more we can make that traditionally enforceable, the less likely there is for hold up in terms of the relationship-specific investments that the subdelegate and the delegate's going to be investing in the decision.

And therefore, the agency head is more likely to subdelegate when it knows that a court will reduce the risk for that political holdup because a court will step in and say, no, no, no, no, no. This is what this incomplete contract said.

So the way to get at this in the administrative law context is fittingly known as the Accardi Doctrine. What happened in Accardi? Well, in Accardi, the petitioner challenged the validity of his denial of an application for a suspension of deportation by the Board of Immigration Appeals.

Why? Well, he somehow found out that the attorney general had circulated a list of unsavory characters, of which he was one, to this Board of Immigration Appeals, saying whenever this application comes before you, deny it. And he claimed that this violated this subdelegation because this contract says that the Board of Immigration Appeals shall exercise discretion and power, not the attorney general.

And the court bought his argument. The court said, indeed, in agency in particular contexts, which I will describe, if it has rules of procedure, including rules like this, they shall abide by those rules. The problem is currently that the doctrine introduces a lot of uncertainty in terms of when these contracts are going to be enforced or not.

And there are a couple sources of uncertainty. One is the doctrine concurrently says that these subdelegations will be enforced when they take the form of what's known as "legislative rules."

When you take administrative law, you will quickly understand that the line between legislative rules and non-legislative rules is famously blurry. All you have to know is that courts haven't figured out how to solve the problem, the distinction, in a principled way. And so an agency head promulgating to subdelegations is going to have a lot of uncertainty. Is this going to be enforced or not? Is this legislative or non-legislative?

The second source of uncertainty is going to be in how the court interprets that subdelegation because this is an interpretation of a regulation. It's regulatory interpretation. And courts are all over the map. Forget statutory interpretation. Courts are kind of a disaster there, but they're getting better.

In regulatory interpretation, nobody knows what to do. Are we textualists there? Are we purposivists? Are we looking more broadly? What materials do we use? And as a result, there, too, there's another source of uncertainty.

So one implication of all of this is that perhaps Coase's work helps us to better appreciate various calls for courts to adopt more bright-line rules in the Accardi context. It doesn't matter what the rules are. Just adopt something that's simple.

So for example, you could have a rule that says, well, as long as this subdelegation has gone through notice and comment, you get Accardi. Regulatory interpretation-- we're just going to look at the clear statement of the text. Lots of ambiguities to be sure, but it's certainly less uncertain than it is now by adopting a rule like that.

OK. A second set of legal implications concerns the traditional story that our law students will hear about, the constitutional separation of powers. So in the cartoon version of this story, we know that Congress passes a statute through bicameralism and presentment, which is then enforced by the president, subject to judicial review by the courts. And these actors have various means through which they can check and balance each other.

Social scientists have imposed the transaction cost framework to examine the interplay between president, courts, and Congress to draw normative conclusions about these institutions. So take for example [? Jonathan ?] Macy. He's argued that how the structural constitution currently fixes transaction costs is non-optimal for allowing interest groups to gain favorable legislation and therefore a non-optimal amount of rent seeking.

What is less appreciated is that administrative agencies have internal separation of powers problems on their own. And is the last substantive idea that I will introduce before concluding. And this is something that's still in code. This is a current project that I'm working on. It's simply to observe that particularly on multi-member boards and commissions that I've told you about, there's often disagreements about what the executive power entails.

So take this statute-- I'm going to call it a "constitution" to import all our ideas about what constitutions do-- of the Chemical Safety and Hazard Investigation Board. This is a five-member commission. The chairperson of the commission one day woke up and said, gosh, now that I am the chief executive of this board, I alone am going to set the budget for the entire agency. That is my executive or administrative function.

Well, budget setting has huge consequences. As you can imagine, the other members of the board pushed back. And this question was finally settled by the Office of Legal Counsel, OLC, about what is the scope of the executive power within this agency.

And OLC, interestingly, used to the words of the Chief Executive Officer, the CEO, of the board to say, gosh, well, if you look at corporate law, CEOs' decisions are subject to the approval of the board. That is, when there is a dispute about the scope of authority, the board has the final say. And it says we'll apply that principle here. Sorry, chairperson. You're out of luck.

And interestingly, I think, in the corporate law context, many have observed that that form of government looks parliamentarian in form-- that is to say, when the executive is subservient to the assembly. And so I think that there is a lot of potential to import the ideas from comparative constitutional law, presidential versus parliamentarianism, parliamentary governments, where comparative constitutional law scholars and political scientists have long asked which form of government is more effective for achieving particular ends to the administrative context.

So wrapping up in our time together thus far, I first attempted to introduce one of Coase's key insights on the role of transaction costs in thinking about comparative institutional arrangements. I've asked whether the voluminous work done on the topic in the private firm context should be extended to governmental actors. I briefly surveyed some of the work that political scientists have done in answering in the affirmative.

And considering this literature, I proposed two areas that I perceive to be in particular need of further theoretical and empirical development, of which I hope to have contributed a bit. One, understanding the role of agency heads, political appointees granted delegated power, in structuring and designing the administrative agency. Two, in thinking more about the ways in which these agency structures may be linked to agency performance.

So now I'm finally going to conclude with a broader observation about the development of law and economics as applied to the bureaucracy. And in doing so, I'm going to invoke an observation that my colleague, Eric Posner, has made about the broader field.

He notes what he perceives to be a troubling trend of law and economics into two subdisciplines-- one, what he calls an "economics law and economics," and two, a "law law and economics." The first is mathematical and descriptive in orientation, while the second is more verbal and normative in orientation. And the fact that these two groups don't talk to each other as much as they could raises the risk that intellectual history is going to repeat itself.

Because when you read Coase's reflections on his own work, they reveal a strong aversion to the mathematical abstractions that fill the political science journals of today, which is, by the way, somewhat ironic for someone who has a so-called theorem named after him. And Coase's aversion to this mathematical abstraction was not due to a lack of capacity, but rather a worry about the potential for such abstraction to lead the fields astray, just like price theory did many, many decades ago.

And so the need for-- to use the buzzword of the day-- microfoundations and qualitative work of the kind often pursued in cognate fields like public administration is, I think, critical because I don't think "empirical" only means larger data sets. "Empirical" simply means we need to understand more about the real world. Work and talk with the individuals whose lives we purport to study. And what a tribute to Ronald Coase that the Coase-Sandor Institute of Law and Economics here at our very own law school fosters these kinds of conversations every year.

Now, I'm biased to be sure. But I think lawyers in particular can bring much to a law and economics conversation about the bureaucracy, for lawyers are perhaps best at what might be called "mid-level theorizing," beginning with empirical observations to deduce more verifiable general statements.

And I think maybe one reason for that is because we ourselves learn law through individual cases to deduce legal rules. Maybe that's where we get our orientation. Maybe it's because lawyers have clients who often seek to gain something from or defend the government. And so maybe we more than social scientists need to appreciate where the real mechanisms of power lie in government.

Whatever the case may be, since I began with Coase's Nobel lecture, it seems fitting to end it how he did. To paraphrase his words, I'm very much aware that many economists and political scientists whom I respect and admire will not agree with the opinions that I have expressed. And some even may be offended by them.

But a scholar must be content with the knowledge that what is false in what she says will soon be exposed. And as for what is true, maybe she can count on ultimately seeing it accepted if she lives long enough. Thanks.

[APPLAUSE]

So I think I have some time for questions, hopefully from students in particular.

AUDIENCE: [INAUDIBLE]

JENNIFER NOU: Gosh, OK. So there has been a lot of-- so the question if you didn't hear it was, is the Nondelegation Doctrine efficient? OK. So back up. What is the Nondelegation Doctrine? The Nondelegation Doctrine essentially says that Congress cannot delegate its legislative power. Footnote-- the Nondelegation Doctrine is in desuetude. No courts actually enforce it.

And first I'll say [? Jonathan ?] Macy argues that it is. He actually argues that we should have this Nondelegation Doctrine be enforced more strictly. And I'm going to think in a minute critically about it to see whether I buy it.

But his argument is essentially that by delegating so much to the agency, one way to put it is that the accountability of legislators is now compromised. And so if you think about the political market robustly, political actors and voters don't really have a way to hold Congress accountable. And therefore, efficiency is reduced.

I wonder if this is one of those places where I feel like qualitative work would be more useful in answering that question because I think we have to know, first of all, how much delegation-- empirically, we have to know how much delegation is actually given to agencies. And what do agencies do with that?

And then I think we need a more qualitative account of the extent to which rent seeking actually does and doesn't occur because there are robust stories that that is not actually what occurs on the ground. So there is an argument that it is. There's arguments against it. And ultimately, easy for me to say we need more empirical work on it. Yeah.

AUDIENCE: I wonder what you think of agency heads and political entrepreneurs? And to take kind of a fanciful example, suppose someone wanted to reform education and made a choice between being head of the Department of Education or talking Google into setting up an education division that would make school systems compete with public schools.

So you can see one choice where the political entrepreneur wants to be at the top of a U form, which is the regulatory state regulating what private and public institutions do, versus what you could call an M form, where there are private corporations like the [INAUDIBLE] corporations like Google actively and powerfully joining to create public goods and taking on an avowed political position partly in competition with the regulatory state. Some of it's a question of the role of a huge, politically [? empowered ?] company.

JENNIFER NOU: So I am very sympathetic to the idea of local entrepreneurs for two reasons. One, I think the concept highlights this conception of bureaucratic autonomy that I think is very important. So Dan Carpenter and others actually explicitly refer to all the entrepreneurial ways that bureaucrats-- all the methods that they use to increase the scope of their activities. So I like that it highlights the autonomy that agency heads have.

And then second, what I also like is that it maps onto this observation that for whatever reason, particularly in this current administration, there is a real celebration of agency heads that have had previous private corporate experience, sometimes with not such great ends. So take Rex Tillerson in the Department of State. People celebrated the fact that he used to head Exxon. And gosh, he was going to bring all of this managerial efficiency and success to the State Department.

Now, there was an incredibly complicated story about why he didn't succeed. But I think that it maps onto, again, this observation that we have agency heads that think of themselves in private entrepreneurial terms.

And the rhetoric, I think, is becoming surrounding private market ideas into government. So witness, for example, Jared Kushner's establishment of a White House Office of Innovation to bring business leaders to tell government how to do things. I think that's gotten a lot of traction. And so I think that it's a useful image. Yeah.

AUDIENCE: I had a question. I was wondering if you had a view about what authoritative fiat is and how it works. So I guess to illustrate, I was kind of thinking if you're in the Army and your officer tells you to get up over the trenches and run, there's two views about maybe why you do it. One is you're just balancing the cost of shame versus the cost of getting shot in the head. And you decide you'd rather get shot in the head, I guess.

And the other view might be that there's just something about being in the Army. You've got an authority figure. [INAUDIBLE]. You had an authority to tell you to do. And you just do it, and that's the way it goes. I was wondering if you have [INAUDIBLE] what authoritative fiat is and how it works.

JENNIFER NOU: Yeah. Well, the first observation is that I think there is a lot of variation in what authoritative fiat looks like and how it operates in various organizations. So the military is paradigmatic. I think you chose it for a reason, right? It's the most top-down organization that there is. And so there is much less room for pushback, if you will.

Incidentally, the current people I've been thinking about is about civil servant resistance. And I'm just going to note that even in the military context-- I've been looking at the military for a comparison. And in the military, there is this notion that even those in the military have a duty to disobey a clearly illegal order. And there's a whole body of law about what that looks like, which is just to say that even there, there are pockets of legitimate resistance.

But as compared to other organizations, we have a spectrum to more traditional agencies and then to private corporations. Gosh, I think authoritative fiat looks very different in a private corporation, especially when you're at will. You could be fired. It's much easier just to be like, this is how you're going to make my pizza dough.

And in the bureaucracy, where you have a civil service that has removal protections and has been there longer for political appointees, it would be a stretch to say that the bureaucracy operates by authoritative fiat. There is a huge history of resistance by the civil service to the political appointees. We see that a lot in the Trump administration right now.

So last comment on that. I mean, you could answer that in two levels. One is to say about the norms versus the practice. The norm, I think, is that of fiat in bureaucracies. That is to say, if you're in the civil service, the norm is that you do whatever your political superiors tell you. But again, for various reasons, the practice departs from that. Do we have time for more? OK. One more. Yeah.

AUDIENCE: [INAUDIBLE]. But I wonder if there's a possibility that [INAUDIBLE] where your congressman can't get involved. There might be situations where he doesn't want a political appointee to maintain control over the legislative process.

JENNIFER NOU: Yeah. OK, great. So the question was, gosh, if courts enforce these subdelegations more, well, isn't there going to be a greater likelihood that Congress gets involved? Yes. Because amongst many other things, litigation is going to foster transparency. It's simply because these are going to become more high profile and, in some ways, kind of brings these fire alarms.

Busy people in Congress, even busy agency heads, did not even know these subdelegations existed. And so normally, I want to say that would be great. But anything that Congress did to get more involved in the administrative state would be great. But gosh, if it's going to get involved in subdelegation, that has so many accountability benefits and all these other benefits that I think would ultimately be a net positive for the administrative standard.

PRESENTER: So thank you, Professor Nou, for an insightful 2019 Coase Lecture. You have validated my position to delegate to you.

[LAUGHTER]

Please join me in thanking Professor Nou.

[APPLAUSE]