Today's episode is a little different. A few months ago, I read a great essay by Sid Jha on the Chevron doctrine.
Sid had written to me saying he’d love a Statecraft interview about OIRA, the Office of Information and Regulatory Affairs: it’s a division of the Office of Management and the Budget (OMB) that reviews all major agency regulations.
I thought this was a great idea, and I asked if he'd be interested in co-hosting an episode with me. Here’s the result: an interview with John D. Graham, who was the OIRA administrator under President George W. Bush. It's a great conversation and I hope you enjoy it.
We discussed:
(00:00) Introduction
(00:43) Where OIRA comes from
(09:20) How cost-benefit analysis got better
(12:59) How OIRA kills regulations
(26:51) Which agencies hate OIRA most
(34:31) Why command and control regulation persists
(39:44) What regulations OIRA focuses on
(46:10) John D. Graham vs. Dick Cheney
(50:46) Graham and the English First movement
How much does OIRA's approach change between administrations?
OIRA sits within the Office of Management and Budget, OMB. So there's an M side of OMB and there's a B side. Questions about how federal money will be allocated are resolved in the B side, and OIRA is in the M side, so we don't have any role in how federal money is allocated.
There are two ways the federal government can make things happen. It can finance something, or it can force somebody outside the government to spend money on something. When it’s forcing someone to spend money on something, that's called a regulation. And OIRA reviews all federal regulations that come from the Cabinet departments of the federal government.
The precursor to OIRA was in the Jimmy Carter administration. Carter was a strong advocate of reviewing regulations and paperwork burdens to make sure that they were sensible. He created the office that eventually became OIRA, and he also persuaded Congress to pass the legislation that established OIRA as a body within the Office of Management and Budget.
When Reagan was elected, he took that initial organizational start and dialed it up another hundredfold. I have a new book coming out this fall called Regulatory Reform from Nixon to Biden. One of the key points I make is that the pioneering work was Carter, but it was Reagan who put it to work in an aggressive way for the first time.
Here's the big change that happened under Reagan. Suppose you're at the Food and Drug Administration, writing a regulation on food labels. You put out your proposal, you get comments from the public and the interest groups, you finalize it based upon what you think the best way to design the regulation is, and you publish it in the Federal Register. President Reagan said, “No, you don't publish it in the Federal Register. You send it to OIRA, and OIRA will review it, and they'll approve or disapprove or suggest changes.”
This was a sea change in the power structure of the federal government. Political scientists used to talk about the “fourth” branch of government. There was the Congress, there was the judiciary, there was the presidency, but then they said, “Wait a minute, how about this alphabet soup of agencies?”
They became known as the fourth branch. And one of the reasons they were called the fourth branch was that nobody was really very effective at controlling them. They were doing what they wanted to do, although Congress felt that they were a child of the Congress.
Reagan changed that with this executive order in 1981. Since then, none of these agencies can do any significant regulation unless OIRA says, “You can go forward.”
What led to Reagan making this move?
It has its roots in the 1970s. I'm going to take you back to the Nixon presidency. The chairman of the Council of Economic Advisers under Nixon was fond of saying that there was more regulation imposed on the economy in the United States of America under Nixon than under all previous presidents combined.
A lot of people don't realize that Nixon was so heavy-handed. In fact, he adopted wage and price controls throughout the entire economy after the 1973 Arab oil embargo. He went so far that he imposed a speed limit on all highways in the United States of 55 miles an hour. Can you imagine being restrained to 55 miles per hour on an interstate highway system in the United States? And you could not buy gasoline on Sundays. No gasoline sales on the Sabbath.
This detailed regulation under the Nixon administration pissed off the Democrats, it pissed off the Republicans, and it pissed off the American people. As a consequence, regulatory reform started to get a lot of play, both politically and intellectually. There was bipartisan demand in Congress to get control of these regulatory agencies. And economists were really concerned about the U.S. economy, as it had suffered a big slowdown in productivity in the 1970s compared to the 1960s. Some economists felt that the growing interference of the federal government in the day-to-day behaviors of businesses could be contributing to the slowdown.
Prior to Carter and Reagan, dozens of agencies operated in their own mode. There was no centralized oversight or review of the analytic quality that goes into the underpinnings of these regulatory proposals. Reagan mandated, somewhat naively, cost-benefit analysis.
It turned out that nobody knew how to do a cost-benefit analysis on a lot of these regulations. But that didn't bother the Reagan team, because that just meant OIRA could deny the regulation if they wanted and say, “You didn't do a cost-benefit analysis.”
It created a very adversarial relationship between OIRA and these agencies, and that got so bad that around 1986, several powerful members of the House of Representatives got a hold of the Appropriations Committee, and they just said, “OIRA's appropriation is zero.”
Fortunately for OIRA, the Senate insisted that this not happen, and OIRA survived with a cut to its appropriation. But it was a very vulnerable period, because there were many labor unions, environmental groups, and other pro-regulation forces in the political fabric that wanted these agencies to be issuing regulations, and OIRA was framed as a barrier to getting good regulations out.
How has cost-benefit analysis developed as a field since the Reagan years?
Oh, it's grown enormously and become much more sophisticated. The most obvious area would be environmental economics. When Reagan issued his order in 1981, agencies like the Environmental Protection Agency didn't have any idea how to do a cost-benefit analysis of an air quality regulation or water quality regulation.
What were they doing instead?
Well, they would do a regulation, and they would say, “This is how much we're going to reduce the concentration of this pollutant in the atmosphere. And they would provide physical evidence of how the air and water would get cleaner. “This is how many more lakes will have contaminants below a certain standard.”
And then separately, they would say, “Oh, by the way, this is going to cost the steel industry 500 million dollars. This is going to cost the utility industry a billion dollars.” There wasn't any effort to actually directly confront the question about, you know, “Is this package worth it?”
By its very nature, cost-benefit analysis nudges the analyst and the policymaker to address those questions. And EPA became a pioneer in the development of environmental economics, By the end of the Reagan administration – and certainly today – EPA was very successful at showing how the benefits of environmental quality can be translated into dollars. And you can show that a lot of these regulations are worth the costs.
One critique of OIRA is that it’s just a point in the rulemaking process where regulation that's not politically expeditious can be killed. Would you agree with that?
Well, I would say that a regulation that does not have support among the political actors of the White House can be killed easily, with or without OIRA. Now, in the old days, when the White House didn't even review these regulations, then the only way the White House would find out is after the agency published the regulation. Then they'd have to try to force the agency to withdraw it.
Or they would get an informal word: I hear EPA is developing this regulation. “John Ehrlichman, why don't you call over there and find out what's going on?” Ehrlichman was the domestic policy advisor to Richard Nixon. He was responsible for keeping an eye on EPA, but he didn't have OIRA to help him do it. These informal mechanisms would be used by the White House to clamp down.
Now let's talk about how the OIRA process works. An agency sends over a proposed regulation to OIRA, which then distributes that regulation to all White House offices and federal agencies that might have an interest in it.
Agriculture, Commerce, Housing and Urban Development, etc: all of these offices submit comments to OIRA. OIRA takes all these comments and adds its comments, because it has its own career civil servants trained in cost-benefit analysis and policy analysis. They develop a consolidated passback that goes back to the agency. “Here are our comments on your proposal.”
The way the executive order is set up, the agency is not allowed to publish that proposal for public comment until it resolves this passback with OIRA. So OIRA acts as an agent in the negotiation for all these other agencies and offices.
If a political office in the White House wants to kill one of these proposals, they might say, “OIRA, you know, why don't you just sit on this for a couple months?” There are cases where OIRA sat on regulations for more than a year and they don't want to make an announcement: “Oh, by the way, we're killing this regulation.” But just as a practical matter, they stretch it out indefinitely.
In other cases, OIRA will just issue a return letter and say, “Thank you very much, um, Department of Transportation for this proposal. We have analyzed it. We see these problems with it. We think you ought to reconsider it.” Boom. Signed, OIRA administrator. Those return letters are viewed as death knells within the agency.
What are common reasons for killing a regulation?
Here’s one of the most common strategies that agencies use. They have their preferred solution that they have gotten to through whatever analysis or consultation, and then they compare that preferred solution to only one alternative, usually only a bad alternative, usually doing nothing about the problem.
So they say, “Hey, look, our preferred solution is way better than doing nothing.” But what cost-benefit analysis likes you to do is to look at a few alternatives and see whether they might be cheaper than the preferred solution at the agency. That is the kind of process that cost-benefit analysis generates. It forces the agency to subject its preferred solution to meaningful, plausible alternatives.
Where do you feel like OIRA really adds value to the process?
Well, most regulation coming from the federal government will exist with or without OIRA. OIRA does not stop many regulations in their tracks, because usually the agency is responding to some genuine issue. Where OIRA has its biggest impact is in inducing the agency to consider smarter ways of solving the problem than the one that they originally thought about.
I'll give you a concrete example: diesel exhaust control. Industry was using these heavy trucks with diesel engines and you got exhaust spewing out of the tailpipe. EPA said, “It’s not really a good idea to have this in the breathing zones of people. We’ve got to figure out a way to really reduce the amount of the exhaust out of these tailpipes.” OIRA came back to them and said, “That is a reasonable idea, but are you aware that diesel engines are also widely used in mining, construction, and agriculture? And there are people who are living near mining operations, farms, and construction sites. Maybe you should look at whether it makes sense to also control diesel exhaust from these non-road engines.” That led to two regulations – the on-road rule and the off-road diesel rule – and they control exhaust from both sectors.
What failure modes of agencies are leading them to overlook cost-benefit analysis until OIRA forces them to pay attention?
The basic instinct of a regulator, who's often an attorney by training, will be to do what's called a command and control regulation. The agency will inform a business or an individual, by regulation, that it’s adopting a technology or engaging in a behavior that the agency is going to prohibit. “We will not allow you to sell gasoline on Sundays.” Or if you are a power plant and you're emitting pollution, “We're going to require you to put a scrubber on that stack.”
The first big innovation – it seems so obvious now, but it was a pretty smart idea when it first came up – It was this. “We're not going to tell you what to do as a regulatory agency. We're just going to tell you that you have to reduce the amount of pollution coming out of your facility by 90%. You go figure out how to do it. We're going to have monitoring equipment, and we're going to measure how you're doing, but you can do it any way you want. You can change fuels, you can use new technology, you can change process, whatever you want to do.” That's called a performance standard, instead of a design standard.
And now there's been a proliferation of new alternatives to command and control regulations. The one that's most recently gained a lot of play is the nudge, rather than the mandate.
Your successor was a big fan of this way of thinking about regulation, right?
Cass Sunstein definitely was, yes.
Was that nudge framework a part of your thinking? Was it in the water at the time you were regulator?
I don't remember us using that term when I was at OIRA. A popularization of the term occurred after I was at OIRA, but it may have already been in the academic literature before then. Certainly at OIRA we were always looking for alternatives that would make progress, but that wouldn't be as coercive or wouldn't be as costly. We didn't call it necessarily nudges, but we were always looking for something like that.
But remember, there's a problem with this strategy. The lawyers who wrote these laws often thought they were writing legislation that would force a command and control solution, because they thought that was the most direct and appropriate way to solve this problem. So we're engaging in bureaucratic behavior, in our efforts to do nudge-like solutions, at the perimeter of what's lawful. Often the courts look at that and say, “Oh, this sounds like a pretty good idea. But I'm reading the statute, and I don't think Congress ever gave you the power to do that.”
I remember vividly one that we did with the EPA on mercury pollution from coal-fired power plants. Instead of telling each power plant, “This is how much mercury you're allowed to emit from each point in the power plant,” we said, “We're just going to put a cap on the total amount of mercury that's going to come out of all the power plants in the United States. We're going to ratchet that down every 5 or 10 years. We're going to have permits, and these power plants can trade the right to emit mercury.”
We issued this rule, and the court said, “I don't know where EPA got this idea to do this pollution trading thing, but this is a fairyland of legal thinking, because there's no authorization whatsoever for this in the statute.”
Now, this doesn't necessarily stop OIRA, because if the solution is working and the stakeholders like it, it can be illegal and still be adopted and implemented. It's only if somebody sues and says, “I don't like this, I want to do something different,” that's when the courts can get in there.
And the worst they'll do is slap your hand and say, “You've got to come up with something that's within the confines of the statute.” So then we would do it again, and end up probably with some kind of command and control solution.
You have what sounds like quite a realist view of the relationship between OIRA and the courts. You try stuff, and you see if the court slaps you down or not.
Right, and remember there was a certain amount of unpredictability with the courts. It would often boil down to whether you got a good panel or a bad panel. We developed a certain kind of orientation: “We’re going to try to go for the best solution that we think is possible. And we're not going to try to guess whether the courts are going to be with us or against us, because the worst that's going to happen is they'll be against this and then we'll have to redo it again.”
Now, when I say we, I'm talking about OIRA and the regulating agency, because we have to be able to persuade the agency to actually put this in their regulation, and that's the real challenge for OIRA. How do they persuade the agency to come along with this alternative point of view?
Does the end of Chevron affect how OIRA functions?
It used to be that, from a legal point of view, the agency would be given a broad range of discretion to do what it wanted to do, as long as the statute wasn't real clear and it could make a plausible argument for what it wanted to do. Now, the courts, in theory, are supposed to make their best reading of the statute.
It's probably a little too early to tell how the post-Chevron world will play out in terms of OIRA But I think that it will cause people to appreciate OIRA even more than they do now. OIRA used to be seen as evil, particularly on the progressive left. But it has changed a lot. And you had several Democratic administrations where OIRA played a very influential, important role. You're not going to find a Democratic president who doesn’t give a substantial amount of power to OIRA — a Republican president probably even more.
Will you say a bit more about that partisan valence? How have those views of OIRA changed over time?
Particularly in the early Reagan years, the pro-regulation interest groups had learned how to work with the powerful regulatory agencies: the environmentalists with EPA, the labor unions with the Department of Labor, and so on.
Now, all of a sudden you've got this new player on the block and they're introducing a new way of thinking and kind of forcing this on the agencies. That was viewed as an intrusion into the existing power structure.
But I think what's changed is that both Democratic and Republican presidents have come to realize that they will be held accountable, politically, for what comes out of these agencies, even if they had nothing to do with it. Some harebrained regulation comes out of Washington, D.C., and the ordinary person on the street thinks, “The president's in charge of that, they should do something about that.”
And the truth of the matter is, the president does appoint the leader of all these regulatory agencies. In most cases, the president can remove them without cause. If the president thinks they can find a better person to do the job, they're going to remove him and put somebody else in. So they don't have a lot of standing to say, “Well, that was a really stupid regulation, but don't blame me.” You won't see a president take that position.
Now you've got presidents like Clinton and Obama and Biden, and they're wanting to do a lot of regulation, but they do understand that it's politically sensitive. They don't necessarily want to trust these agency people to roll the dice on their reelection. They want people close to them in whom they have confidence. So they use OIRA to channel the energies of these agencies and put them on priorities of the president, rather than issues that the president doesn't want to risk.
And the agencies don't think that way, you know. They have their own momentum. The bureaucracies have their own reasons for wanting to do things. So you’ve got to get control of that.
Democratic presidents realize that now. So they can use the OIRA for very pro-regulation purposes as well, like getting agencies on to issues that the bureaucracy might be too intimidated on their own to take on. Under Obama and under Biden, you saw a lot of instructions from the White House to agencies: “Here are the areas you will regulate, and you will give a progress report to OIRA within six months of how much progress you're making, and whether you need any additional resources and what we can do to help, but we want you to get going.” OIRA then has a mandate to enforce that.
So OIRA has become one way to help a president, Democrat or Republican, get the things done that they want to get done.
And that leads to OIRA having an adversarial relationship with the agencies?
Oh, absolutely. OIRA has had an adversarial relationship with the agencies from the beginning. I think that's built into the structure of their relationship. if there wasn't any tension at all, OIRA is probably not doing anything.
Does that relationship vary from agency to agency?
The Department of Health and Human Services is the largest Cabinet agency: it has Medicare, Medicaid, the Food and Drug Administration. It regulates everything from wheelchairs and nursing homes to prescription drugs and medical devices. I always sense a certain attitude, like, “We're dealing with medicine. This is highly technical information. What does OIRA bring to the table on these issues?”
And there was always a very strong resistance to letting OIRA get involved in health care-related regulations. It's one of the tougher agencies for OIRA to penetrate and make a difference on.
What’s the opposite – an agency that OIRA has a good system for influencing?
It's hard to generalize about OIRA as a whole. OIRA is a small agency with only about 50 career employees, and each of these employees has a desk. You might have the FDA desk, I might have the EPA desk, or an office within EPA. What happens is, these people have personalities and inclinations. Sometimes the relationship with a desk officer will be very good for an agency and they'll learn to work together. But sometimes you'll get people who just have a lot of trouble getting along.
I remember vividly the EPA folks saying to me, “John, can't you rotate some new people into the EPA desks? We're having a tough time with these people.” Partly that's personality, but partly it can be that the desk officer is pretty tough, and the agency wants to have a little easier shot at it. But they would never say it that way.
The more economic agencies like Treasury have a staffing mix closer to what OMB (OIRA’s parent office) looks like, so it’s often easier for OIRA and the agencies to get to a common solution. They have a common framing for how they think about the world.
But when you're dealing with people who are studying threatened and endangered species in the Fish and Wildlife Service, and you have an OIRA desk officer responsible for overseeing regulations about those species, that takes a lot of time. OIRA has to be very careful that it’s not engaging in issues way beyond its ability to do so competently. So it's important that they engage other agencies who have some expertise to help them review a regulation like that.
50 is a pretty small number to review all important regulation in the country. Did it feel under-resourced when you were there?
Oh, absolutely. It feels like a big city hospital emergency room on a Saturday night, and a lot of patients are rolling in there, and you barely have enough doctors and nurses to deal with all these various problems. That’s its nature.
Remember, OIRA is not intended to be the primary way we get good regulation. It's viewed as a back-end method to make sure that we don't have terrible regulations, and to make each regulation maybe a little bit better at the margin. But you can't rely on it as your primary mechanism. It's just not that potent.
What agencies fear most of all is losing their regulation in court. That is very difficult for them to get out from under, and it can take them years to redo it and then get the court to approve it. They fear the courts even more than they fear OIRA.
And remember that if you're working in the George W. Bush administration, there's a head of the regulatory agency who's an appointee of George W. Bush and a head of OIRA who's an appointee of George W. Bush, and the agency staff can have their bosses go talk to people in the White House and get this sorted out. Well, the judicial process doesn't work that way. It's very formalized, with all these briefs and these formal arguments.
And there's a trend in this direction already that could help OIRA. If it became understood that a cost-benefit kind of framework is the way the agencies are supposed to be looking at these regulations, then the court could look at that and say, “Oh, well, I didn't see any alternatives here. There was just the preferred solution, and ‘do nothing.’ That doesn't sound like a very robust cost-benefit analysis.”
From a culture perspective, what would make OIRA more effective at achieving its goals?
I'm not a big fan of expanding the size of OIRA enormously. In some ways it's best when it's small, tenacious, opportunistic, and picks the most important issues to focus on. There is a view that OIRA should be much bigger, but it's not one I would advocate.
One thing that would be helpful is if Congress were to say to the agencies and to OIRA, “You know, we've had a lot of command and control-style regulation. We've been doing that now for 50 to 100 years. There should be some presumption that you should look at something more innovative, some alternative — whether it be a nudge, whether it be a pollution tax, whether it be an information solution — where you get information to the right people so they make better decisions.”
There's currently no such presumption that an agency has to look at anything other than a command and control solution to a problem. It would help if Congress were to bust that into the statute that the agency reads, so it's not just OIRA that's singing this message.
Here’s a famous example: asbestos was used in industry, and there was no question that asbestos use needed to decline substantially. It had been linked to mesothelioma and other respiratory diseases. But there's some highly, highly valued uses of asbestos in brake pads and in other applications.
So the EPA wrote a rule that said, “Here are all the areas where asbestos is being used. We're going to regulate each of them and decide, in each case, what the right answer is.” No use of asbestos here, an exemption here, whatever. And OIRA said, “You're expressing a lot of confidence in your understanding of American industry and how this is going to change over time. Why don't you just decide that for the next five years, you want to reduce the amount of asbestos use by 50 or 70 percent, and then let these industries decide what the most valued uses are and give away other uses?” EPA refused to do that.
OIRA said, “You don't have to believe us, but there is this provision in the Toxic Substances Control Act that says you'll look for the least burdensome alternatives. Don't you think that a court's going to look at this and say, ‘You really haven't developed the least burdensome alternative?’” Sure enough, a unanimous Fifth Circuit Court of Appeals ruled in 1991 that the EPA regulation obviously had not explored alternatives that would be less burdensome.
But designing alternatives is an area that needs a lot of work. The easiest thing to do is just write a law that says we won't sell gasoline on Sundays. Not a very creative solution.
Given the growing importance of OIRA, what’s the value of trying to keep it the same size and not expand its powers?
I don't think the formal powers of OIRA have expanded substantially since the early Reagan years. In many ways, the powers have been restrained and limited. When President Clinton was elected, he issued an executive order that said, “OIRA is not going to review every regulation anymore. It's only going to review the important ones.”
That was a really useful change, because there's a tendency for OIRA desk officers to run off and spend a lot of time on things that aren't very important. Unless you have a really good OIRA administrator who's in touch with exactly how each of his desk officers is spending their time, they can waste a lot of time. Now, that's a limitation on OIRA power. I think it's a wise one, but it did reduce the amount of power that OIRA has.
Really, the informal power of OIRA is much more important than the formal or legal power. I worked for a man named Mitch Daniels, who became the OMB director. He used to say to me, “John, don't spend so much time on the legal details of where your powers are. You're going to do what I want you to do. Okay? And you're not going to go outside that territory, because we're going to be in regular communication, and I don't want you mucking around on things that aren’t worth it.”
That's an informal political constraint on the OIRA administrator. Those constraints on OIRA administrators are much more powerful than anything that's in any executive order, any formal activity. I have to work these issues out with people who President Bush appointed to these agencies. These are his people, they worked on the campaign with him. Many of them have friends in the White House. So it's a complicated job to lead OIRA.
How do you determine whether a regulation is significant or important, ex ante, before it goes through a cost-benefit analysis?
There are two levels of significance. The first level of significance is, “Is this important enough for OIRA to review it?” Then there's a second level of significance called economically significant that requires a cost-benefit analysis.
Maybe 100-200 regulations a year are economically significant. 500-700 regulations a year require OIRA reviews.
If a regulation will have $100 million in effect, that meets the cut-off for “economically significant,” as I understand it. How do you calculate that ex ante?
Biden just bumped that up to $150 million, I think. But in any event, when the agency prepares their draft analysis, they're going to look at what the impacts of the regulation are, typically on the regulated industry. That's the first cut. If you’re approaching 100 million dollars of impact, then OIRA is going to do a cost-benefit analysis. At this stage, it's a very rough call, it's not a detailed analysis.
But the harder question is, how do you decide whether we should review something in the first place, even if it's not going to meet the 100 million threshold? I'll give you a concrete example that happened under Obama. The Department of Education wanted to send out a letter to all school districts in America, saying that transgender people should be able to use whichever bathroom they want to use.
OIRA said, “Well, you know, this might be the right kind of policy and everything, but we think it’d probably be a wise idea for you to take some public comment on this idea. Get school districts ready for this. Don't just send out a letter ordering this, because you'll have parents and school boards up in arms.”
The Department of Education basically decided, “No, we’re just going to do it.” And there were a couple of embarrassing incidents in schools. It became kind of a political hot potato. That’s something that was not $100 million dollars in impact a year, but it should have had OIRA review. Some people might argue that it's not even that significant ,given all the things the federal government does, but it was highly controversial. So whether a regulation meets the “significant” threshold is a much more qualitative, subjective thing. “Economically significant” is a little more rigorous.
Will you flesh out that rigorous process for determining if something meets that $100 or $150 million threshold?
The agency is going to look at the regulated businesses affected by the regulation, and how much it would cost these entities to comply with the regulation in the near-term, not the long-term. They'll usually do a back-of-the-envelope calculation on that. If they're getting in the ballpark of a hundred million, then the agency will just declare it as economically significant and do their cost-benefit analysis.
Isn't there a perverse incentive there? Presumably the agency is incentivized to come up with an artificially low number for economic impact so that you guys don't run all over it.
Yeah. And the other thing they can do is, they can take a complicated regulation, they can divide it into pieces, and then issue three regulations, and they're all at $50 million. None of them reach $150 million.
What's happened which has diffused this a little bit is, the courts have been imposing kind of an implicit cost-benefit test. So OIRA will often argue to the agency, “You know, you don't have to do a cost-benefit analysis if you don't want: this regulation is only $50 or 75 million. But don't you think it'd be wise for you to do a cost-benefit analysis? Don't you think it'd be better to have that in the rulemaking record, being commented on? Then you can point to the cost-benefit you’ve done when you’re litigated.”
And a lot of agencies, EPA certainly, are much more sophisticated than they were 20 years ago. They're doing cost-benefit analyses of regulations well below the thresholds, because they believe it's in their interest to do it. The world has changed.
Now, Health and Human Services: totally different situation.
Say more on that.
Well, remember, a physician’s responsibility is to do the best they can for the patient. It's not “Do the best you can for your patient, subject to what the doctor can afford.” You've never heard a doctor say that. They don't like saying that.
Now, people who study health economics know there's a reality to that. It's not necessarily implemented at the level of the individual physician, but the system imposes limitations. I spent 20 years at the Harvard School of Public Health, and there’s a lot of discomfort in the healthcare culture about saying, “We’re going to put affordability limitations on what you're going to do to help people.”
The Department of Health and Human Services has that culture, and that's probably to our advantage, to a certain extent. We don't want these people accepting a lot of disease and premature death just because of a perception that it might cost too much to solve the problem. OIRA's role is to make sure that those kinds of considerations are there, but it's not natural in the bureaucracy of the Department of Health and Human Services.
That makes sense. How often did you interact with the president directly?
I probably had roughly one briefing per year, for five years, in the West Wing of the Oval Office. Not a lot, but occasionally.
So what’s the channel through which things got to the president's office?
Well, OIRA’s not going to clear or approve a regulation until it gets the go-ahead from each of the offices within the White House.
OIRA sends around a draft of the regulation seeking their comments. It wants to make sure, you know, “We're about to approve this thing. Are you sure that you don't have anything you want to say? Don't come back and tell us, ‘This is a nightmare!” You had your chance.”
Right.
So around 2003, I was working with a regulator on sport utility vehicles to raise their miles-per-gallon requirements to save fuel. This was a performance standard that might induce the car companies to add a hybrid engine or have fuel-saving tires, for example. It gave the auto industry a lot of discretion over how to do it, but they did have to gradually improve the miles-per-gallon.
I had worked on this regulation for about 18 months, and it was an expensive regulation. All the White House offices had signed off except one, the office of the vice president, at the time Dick Cheney. Mr. Cheney and his advisors did not think it was appropriate for the Bush administration to be regulating the fuel economy of cars. His view was that as gas prices were rising at the time, it was a very sensitive issue.
The president wanted fuel economy standards to make sure that people saved money as gas prices kept going up, but the vice president felt people would buy more fuel efficient cars if they wanted to. They could buy a Toyota Prius if they wanted to, they could get a smaller SUV if they wanted to.
Basically, the vice president forced this issue into the White House. And I'll never forget it, because the man who later became the chairman of the Federal Reserve Board, Ben Bernanke, was working at the time in the White House, and he was tapped by the vice president to make the case against this regulation. I had to make the case in favor of the regulation. We had to duke it out. The president really wanted the regulation, so we went forward with it. Just one concrete example of how these things go.
Are there things that you got better at over your five years?
I'm sure there were. I had worked as a college professor for 18 years, and when I landed in this job in Washington, D.C., I had no political experience. In hindsight, I very much wish that I'd spent a couple of years on Capitol Hill early in my career and just learned how the congressional side of the world works.
I had two political advisors, and I think they would say I was very oriented to what I knew, which was the cost-benefit framework, but I didn't really have a good understanding of politics. As I did the job more, I got a better sense of that.
What’s a mistake you would have made in 2001 that you didn't make in 2005?
I don't know whether I would have made the same mistake or not, but I'll give you one that happened and let your listeners decide. President Clinton signed an executive order that said that all agencies in the country that receive federal money have to offer communication services to people in their native tongue. You can't just do everything in English. You have to have it available in Spanish, etc. etc. There were a lot of people on the conservative side of the Republican base, particularly in the Midwest, who really disliked this regulation. There was even an English First movement at the time: that this is an English-speaking country.
Now, President Bush didn't feel this way. He’d been governor of Texas. He was proud of being able to speak a limited amount of Spanish when he interacted with voters in Texas. The White House took a look at this regulation and said, “It's not an ideal regulation, but we don't really want to fight this battle. We’re going to let this one go.” So then the chairman of the Appropriations Committee in the House of Representatives, a Republican, put a hold on our appropriation at OIRA and said, “You're not going to get your money for the next year until you do a cost-benefit analysis of this regulation.”
I got some advice from people within OMB, but I wasn't sure exactly how to handle this. Our staff worked on what we thought was a pretty good report. A key thing we found was, it's very important for people who are living in the country but don't speak English to learn English over time. If you’re in the medical system, for example, it’s important that you be able to understand what the physician or nurse says about your conditions.
So in this report, we were going to say that there should be an effort made by all sectors in society to try to accelerate the rate at which these people learn English and become proficient in English.
I was so naive. This report went through the White House clearance process, and the people in the White House responsible for interfacing with the Hispanic community said, “Dr. Graham, this is going to be completely offensive to that community. And we told you, this is not an issue that we want to take on. It's just not worth it for us. Get rid of that.”
How did you end up in the job, given that you’d never worked in Washington?
I was a Harvard professor for 17 years before I worked for President Bush. I had testified a lot in Congress on the problems with our regulatory system and the various kinds of solutions that might be possible, including greater use of cost-benefit analysis.
And then there was a man who was the president of the American Enterprise Institute. It's a center-right think tank in Washington.
It was my first job out of college.
Was it really? Great. Anyway, his name is Christopher Demuth, and he happened to be on the transition team for President Bush, and he knew President Bush personally. He recommended to Mitch Daniels, who became the OMB director, that I should be the OIRA administrator. That was the first way my name got in the hop.
As we wrap up, what's a place where you would want to see the next administration do more with OIRA?
There's a huge need right now to think through the issues around artificial intelligence and technology. Neither the agencies nor OIRA are really a hotbed of expertise on this issue. My concern is the government, in a well-intentioned effort to do some good, could really stumble in this territory without knowing what it’s doing.
We need to organize some way for the federal government to educate itself on the industry. Now, I know China and Europe are running out ahead of us, and they're already regulating artificial intelligence. I don't know whether they're doing it competently. I have a hard time believing that they are, but it needs to be looked at pretty carefully. Big issue on the horizon.
And then right after that, there will be Elon Musk with his self-driving cars. That's going to come pretty fast too. People are going to want to regulate the hell out of that technology. That's not going to be easy regulation to do well.
Share this post