How to Wrangle the President's Economists
At Statecraft, we’ve long been interested in the role economists play in policy. It’s why we interviewed the White House coordinator of CHIPS Act implementation in January, and it’s why today’s guest is Martha Gimbel, former Senior Adviser at the White House Council of Economic Advisers (CEA).
Last week, Martha launched The Budget Lab at Yale, a “non-partisan policy research center dedicated to providing in-depth analysis of federal policy proposals for the American economy.” The lab’s initial projections have prompted debate (see, for example, their analysis of what will happen if the Trump tax cuts expire vs. John Carney’s).
We sat down with Martha to unpack what the CEA is, how it works, and why you should care.
What You’ll Learn:
Why are there so many federal economic councils?
What’s the difference between economics in academia and in government?
What does the president need to know on Jobs Day?
Martha, tell us what you really care about.
I have spent most of my career both working as an economist and, for lack of a better phrase, imposing process and order on groups of economists — which is a thing I very much enjoy doing. I've done that in the private sector, I've done that in government. This is going to sound so nerdy. I was a classics major in college, and Cicero once said that the Utile and the Honestum are the same thing. The useful thing and the morally good thing, in the long run, are the same thing. As a citizen, I really care about doing the honestum, and as an economist, I really care about doing the utile. So I care about finding the places where those things intersect.
The example I always give is unemployment insurance, which I've spent a huge amount of time working on in my career. I feel that it is morally right to support people when they lose their jobs. It's also hugely economically beneficial, right? It is bad for society if people lose their jobs and then immediately just stop spending. I do think there are things that we talk about as either economic or moral imperatives, but they often cross over.
At one point you were a staff economist for the Council of Economic Advisors, and you later became a senior advisor for the Council of Economic Advisors. What is it and how does it work?
The Council of Economic Advisors (CEA) is the executive office of the president's internal economic think tank. It was established via the Employment Act of 1946.
It is the same act that established the Joint Economic Committee, which is on the Hill, where I also worked. It is the job of JEC to advise Congress on economic issues, and it is the job of CEA to advise the president on economic issues.
CEA is most effective when its work is unbiased, right? That's not to say that you're not there to support the president. It's not to say that if you're not working for a particular administration that you're not trying to help them think through their priorities. It is the job of CEA to be in the room and say, “Okay, here's this policy you're discussing, but here are all of these possible economic consequences that you haven't thought about and that we've analyzed. We've looked at the numbers and here's how you should think about it.” That is a very fun role to play.
It means that sometimes you're not everyone's favorite because you're the one coming in and saying, “Okay, we have an exciting policy proposal to give every American family a dog, but how are we going to think about paying for the increased waste disposal that comes along with that?” How do we deal with the fact that certain people's preferences are for cats? Are we providing a cash subsidy instead? Would it be more efficient to just provide a cash subsidy in general? Here's the analysis we've done on the economic implications of giving every family a dog. This will have X hit to GDP, because people will be staying home more often, because their dogs are cute. On the other hand, maybe happiness will go up.” It is your job to provide the analysis behind that.
How do you wrangle groups of economists to work together?
I'm not sure what a group of economists should be called. A chart of economists? A heteroskedasticity? Honestly it looks relatively similar. You have a group of economists together and you have to hit certain goals. In every institution that I've been a part of, the economists are there to help.
“Hi, I'm Martha Gimbel. I'm an economist and I'm here to help.” In the private sector, you were there to help the company understand the labor market. I worked at Indeed, which is a job site company, so they cared a lot about the labor market. And then also, frankly, to help them get positive PR by saying interesting things about the labor market in public so that they would get cited in the news.
When you work in government, you're trying to make sure that policymakers understand the economics that you're doing. Part of what you're doing is in trying to organize groups of economists. Economists tend to get interested in everything we see. Someone will say something and another team member will say “Oh, wait, I really want to look at that.” And you have to pull back and think, “That's a really interesting question, but is there an audience for that internally? What is the purpose that it's serving? Is this a fun side project you do on your own time? Is this something we really need to put resources behind? How does this fit in with other priorities? Is this something we need to do in the next week? Or is this something we need to be thinking about over the next six months?”
Is there a tension between the professional academic imperatives of economists, who want to pursue the most interesting projects and build their portfolio for the next 10 years, and what an organization in the private or public sector needs?
How do you convince economists to work on the thing that is immediately relevant to the organization and not their professional career?
It's not just immediately relevant, right? Academic economists in particular are used to an environment where the time to publish is years. You are spending all of that time tearing the research apart eight ways from Sunday to try to make sure that you’ve addressed any possible angle.
This is not to say that what's done in the private sector or the public sector is not rigorous, it is and it needs to be. But it's different for a couple of reasons. One is it's just faster. If the president of the United States says to you, “I have to make a decision on this policy issue tomorrow,” and you say, “Great, we'll have a paper to you on that in seven years,” that's not helpful.
Second, academic economists, with some exceptions, are looking at something that's happened in the past, and trying to figure out what we can learn from that. In the private or public sectors, you're trying to be more forward looking. If we do this thing, what are the things that might happen? What are the things that we should be thinking about?
There's absolutely some amount that is backward looking, 100%. But, a lot of the questions that are coming in are “how should we think about this moving forward? What should we be anticipating?” The instinct of many academic economists is to answer: Well, we don't know. Which, to be clear, is correct. But you have to be willing to guess, in the most rigorous but flexible way. And that is really, really hard.
Let me get a little more clarity on the non-partisan nature of CEA. I hear everything you're saying, and also, you come in for a particular administration and you're chosen by that administration. Is that right?
So you said non-partisan, which is not what I said. I said unbiased. I think that that is a really important distinction. People have partisan preferences. I'm an economist. People's preferences are their preferences. Your job in the room is to serve the person that you have chosen to work for by being unbiased and by saying to them, “You may think that X policy is a really great idea, but we've analyzed it and here are these problems that you may not be thinking about.”
Other times it is to say, “You may think this policy is a really good idea and we've analyzed it and it is a really good idea. And here's all this analysis that we've produced on why you should do it.” It's not helpful if you're in the room just telling people the version of economics that they want to hear.
How often do debates within the field of economics play out in your context of the CEA? I'm assuming that there's a heterogeneity of opinions among economists generally and at CEA itself on many questions.
I forget who said this but if you have five economists in a room, you'll have six opinions.
But that’s part of the strength of CEA. Many academics work somewhat in isolation. CEA is not a large staff — it’s somewhere between 30 and 40 people — but you’re going to have multiple people who have an opinion on an issue.
If you're the senior economist on an issue, you may write a memo saying, “I think this is the way we should think about this policy.” And then people are going to review that and say, “Well, wait a minute. Here are these six things you didn't think about. And what about this paper that you didn't take into account?” Then you go back and forth, and it makes your argument stronger.
Tell me about the institutional structure of CEA.
You have the chair who is confirmed by the Senate. Then you have two members who are the deputies. Underneath them you have senior economists — generally people with PhDs who are on leave from academic institutions or someplace in government. We've had professors from many different institutions spend a year at CEA.
Then you have the junior staff. I am a big believer in any organization that 90% of the work gets done by the junior staff. You have staff economists who are generally on leave from PhD programs. Then you have research assistants who have just graduated from college or are a year or two out.
You then have what is affectionately referred to as the front office, where I was as a senior advisor, which is in charge of imposing order and structure on the entire crew.
When you say some folks are detailed over from other federal agencies: are those typically folks who have been specifically asked for by CEA team members?
Sometimes CEA will call over to an agency and say, “Hey, we need someone who knows something about this issue. Do you have anyone?”
CEA also does something unusual; it runs an open recruiting process. If people get nothing else out of this conversation, this is the main thing I would like them to take away: CEA has an open recruiting process. You can email recruiting@cea.eop.gov and apply to be a staff economist, an RA, or a senior economist. Those applications get read. It is truly an open application process.
That's not to say that it's not hard to get through, but as someone who used to own the recruiting inbox, we do read the applications. Sometimes people would blindly apply for a senior economist role, and we would go, “Wait, this person is amazing!”
Where is the CEA office?
CEA is in the Eisenhower Executive Office Building, which is right next to the White House. Mark Twain once referred to it as the “ugliest building in D.C.” I think it's beautiful. It looks very majestic. When you think that you're working in the Executive Office of the President, you probably visualize something like the Eisenhower Executive Office Building. I'm going to get in trouble for saying this, but I used to work at the Department of Labor, and I love the Department of Labor. It's not the nicest building.
How does CEA engage with other agencies?
It is part of the policy process. Someone will say, “Hey, Russia invaded Ukraine and it’s going to have impacts on energy markets. We should figure out what to do about this.” The National Economic Council will run that process, but CEA will be in the meeting to provide whatever insights it can.
One of the things CEA does that I personally think is the most fun is reviewing inflation, jobs, and GDP numbers. CEA actually gets those numbers the night before they are public and writes a memo for the President about what he should know about the next day’s data. It's very intense because it's pre-release economic data. You have to take it very seriously, and cannot say anything about it outside of certain rooms.
For instance, on Jobs Day, you get the numbers between 12pm and 2pm the day before. A small group of you sit in a room and have a long discussion about what it means, what we’ve learned, what we should be paying attention to, and what is important for the President of the United States to know. You have to ask, “Does the President need to know about this, or do you just think it's interesting?”
What's the value to the President of reading your report before the numbers are public?
Depending on the time frame, people want to hear from the president, particularly when the data are more stressful. Some of it is so he can remove it from his brain. For instance, I wrote the memo on unemployment insurance claims when I was a staff economist at CEA from 2014-2016. Most of the time we were trying to signal with the memo, “You do not need to worry about this data, we will tell you when it comes up.”
You don't want to be in a situation where he's in a meeting with reporters and someone says something to him about unemployment insurance (UI) claims and he is unaware.
If there is a big surprise in the economic data, is there something that we need to be preparing for people in the morning? Keep in mind it’s just CEA, the President, and some very few senior people who see it. You cannot call comms or speech writing and tell them, “GDP is coming in hot tomorrow!” Part of what you’re doing is preparing at 8:31 AM to send materials that say, “You will be talking about this in an hour. Here’s where we’re going to be.”
Talk to me about the forward-looking nature of CEA's work.
There’s the direction of the economy analysis, “Where do we think this thing is heading?” There’s also, “What do we think the possible impacts of this policy could be?”
Sometimes we agree we should do some policy, but we don’t know how much we should do. For example, the supplement on UI in the CARES Act. How do we determine the right amount to add on? Nobody knows. It’s years later and we still don’t know if it should have been that amount or this amount plus or minus $50. We’ll be studying this forever. CEA is there to provide frameworks to think about this.
If you do this amount, here’s where that hits in the income distribution. Here is who that includes and leaves out. Here’s what we know about labor market impacts. CEA puts that information together so that people can make an informed decision.
It sounds like CEA does not have a lot of formal authority, but it has institutional authority as a result of being good at certain things.
There's actually very little in law that CEA has to do. We have to write the economic report of the President once a year. It’s like 700 pages, I hope everybody has read up on it!
Part of what you are trying to figure out is how to be productive and helpful in the building. Helpful in the room can look very different in academia and government. Nitpicking every little thing to death can be considered helpful in academia. In government, your job is to help people think through things. If every time someone suggests a new idea you respond with, “well, that’s just dumb,” you’re not being helpful.
More helpful is saying, “I hear your idea, and here are 6 problems. Can you talk to me about the constraints you are under so I can help you optimize for those?” We spend a lot of time figuring out how to optimize under constraints.
Could you describe CEA's comparative advantage relative to other places in the federal government?
CEA is there to be generalists. For instance, the Occupational Safety and Health Administration (OSHA) has economists, but they’re there to work on OSHA issues. People at CEA are there to provide frameworks to help think through problems, and to know who to call. If you're working on some occupational health issue, you're probably not the world expert in the economics of occupational health, just statistically. But you probably know who to call, and that's a lot of what you're doing.
What is the National Economic Council?
The National Economic Council is there to get stuff done. CEA’s job is to sit there and think about the economics of an issue. Not that NEC doesn’t do thinking or that CEA doesn’t get stuff done, but NEC's job is to make sure the policy process around a particular issue happens, and once a policy has been decided, that it actually gets implemented. CEA people often become very attached to their corresponding NEC person because you spend a lot of time on the phone with them.
Does CEA provide data or make data useful for economists writ large in academia?
I want to emphasize that the statistical agencies are independent, but one thing that CEA can do in the budget process is have conversations with academics and practitioners about what types of data are useful. Then you can call the Bureau of Labor Statistics and ask if they’ve thought about this. Sometimes they’ll say, “We’ve thought about it, we’d love to, but it would cost X amount.” Then CEA can go into the budget process and demand this thing happen.
Can you give me an example?
The contingent worker survey from BLS has not been run in quite some time. It costs money. There have been all these questions in recent years about people working gig work. Technically it’s called “contingent and alternative employment arrangements.” The survey is looking at how they are employed. Do they have control over hours? Are the jobs temporary? Etc. Gig work is partly a definition thing.
Economists would really like that survey to be run more often. The last time it was run was in May 2017. [NB: BLS proposed to re-add the Contingent Worker Supplement to the Current Population Survey conducted in July 2023.]
There are things we rely on in its place, but they are not ideal. Which is why I will get on my soapbox and say that it's incredibly important that we fund the statistical agencies in the United States government.
There's some things we try to rely on instead. Some economists have used tax data but that's hard. It's imperfect. Also, only some people get access to tax data. I can go right now and look at the contingent and alternative employment arrangements summary online. I can't just pop into the tax data. And to be clear, I don't think I should be able to for privacy reasons.
Private sector people have tried to run surveys. People have tried to work with payroll companies, things like that, but it's really hard. Sometimes you rely on data from the companies themselves. That's also not ideal. I think what you're hearing is me saying the words “not ideal” over and over again, which is why I will get back on my soapbox for a second and say again that it's incredibly important that we fund the statistical agencies in the United States government.
What are the kinds of things that would be helpful to have with greater granularity?
So, it costs money. But, to take an earlier example, there's been all these questions in recent years about, “Are more people working gig work?” things like that. Economists would really like that survey to be run more often [NB: It last ran in 2018]. It's technically called the contingent and alternative employment arrangements. And it's looking at, how are they employed? Do they have control over hours? Are the jobs temporary? This is partly a definition thing, and so this is a much longer conversation, but people will answer surveys differently depending on the type of phrasing you use, and you have to figure that out.
There was some stuff that BLS did several years ago on green jobs which they've stopped doing. Part of that is that green jobs are really hard to define. For instance, if you work in an antique store, is that a green job? You're selling things that mean that people aren't having to produce more stuff. That wouldn't necessarily fall into people's traditional definition of green jobs. It can be a little bit fuzzy sometimes.
Why is it important to fund the statistical agencies? What data do they have that a private sector company would not?
They can combine things. They have access to data from the government but they can also work with private sector companies, many of whom provide their data to the US government. They also have the capability to run large and robust surveys. Although people are generally becoming less likely to respond to surveys.
Let me go way back in the conversation. You mentioned the very stark difference in time from idea to publication between academia and the White House or CEA.
At IFP, we've looked a lot at peer review. Is the lag time in academia necessary?
To be clear, I have never been a traditional academic. I will say other disciplines seem to have a faster review and publication process than economics and still produce rigorous work. Maybe I’m wrong, but my sense is that the review process in economics is extra long.
Did your work in the private sector and at CEA suffer from a lack of an academic process?
Define “suffer.” Did we try every variation on a standard error that we possibly could before publishing something? No, we didn’t. On the other hand, we optimized for people having rigorous information in time to make a decision. There are tradeoffs between how rigorous you are and how relevant you are, and CEA is about finding that sweet spot. Academia optimizes for not being wrong, and that is important. Government optimizes for information that is useful in making a decision. Part of that is being clear about error bands and certainty.
In academia you spend a lot of time asking, “Is this the right point estimate?” In government you spend a lot more asking, “Is it positive, is it negative, is it large?” Did we try every variation on a standard error that we possibly could before publishing something or sending a memo over? No, we didn't. On the other hand, we were optimizing for people having information, rigorous information in time to make a decision. Everything's always about tradeoffs.
Is that a bit of a dodge, though? Economists do talk about Pareto-optimal solutions.
Because it's hard. I've written memos where I really thought I was right and people say, “We just don't feel that this has been worked through enough and we need you to do more.”
And I was right, and we didn't send it up. I've worked on things where people said, “Yeah, we think this is good and we sent it up and we put it out,” and then I was wrong. Academia really optimizes for not being wrong, and that is important. Government optimizes for information that is useful in making a decision. And part of that, frankly, is being clear about error bands. So part of that is being clear about, like, we are really certain about this number we're sending you. Or, here's a number, we might not be right, and we're really uncertain about it.
Would more academic fields advance quicker if they were comfortable being wrong more frequently and moved at a faster pace?
I chose to work in government, not academia. But I do think you have to be comfortable with being wrong. In March 2020, I was convinced that the economy was going to sink into an economic recession that would be incredibly painful and difficult to get out of. It was incredibly painful, but it turns out we got out of it a lot easier than we anticipated. And yes, there's been inflation. We can have all these conversations, but the unemployment rate has not stayed high. So I was wrong about that.
In the government or private sector you have to be okay with the fact that you’re going to be wrong sometimes. You have to be able to be as rigorous as you possibly can, be transparent about when you are guessing, and, when you inevitably get it wrong, sit there and figure out why. Then you dust yourself off and do it better next time.
You’re working on something new now. Tell us about it.
I am the executive director of a new center called The Budget Lab, at Yale. We are interested in trying to think through some of the big questions in the budget scoring process. I’ve started it with Natasha Sarin and Danny Yagan, who also served in government with me. Sarin was at Treasury. Yagan was at OMB. I was at CEA. For those of you who have spent time in the very niche world of budget policy, that means “Troika.” Troika is a part of the budget process where those three groups come together to generate the economic assumptions underlying the budget.
The three of us were talking and, this is in some ways very obvious, but we were really struck by the ways the budget policy process doesn’t think about the full scope of certain types of investments. Some are investments that we know pay off in the long run. For instance, investing in children is something that we know can be hugely economically beneficial, but it's generally not economically beneficial in the 10-year budget window. No five year old has ever paid for themselves in 10 years. We’re interested in thinking about the long run picture and bringing conversations around outcomes into the process.
If you go out and buy a coat, you come home and say to your partner, “I spent $50 on a coat.” You don’t come home to them and say, “I spent $50.” Right now, benefits are not always included in the budget conversation.
For instance, if you talk about the child tax credit, you'll talk about the cost of the child tax credit. You won't talk about, in the same breath, how much the child tax credit decreases poverty or how it might affect children’s earnings in the future. We're trying to link those things more closely together.
What does the product of that work look like?
Very similar to typical economic things. We’re putting out reports, tables, graphs, etc. We’re trying to give people tools they need to help policymakers make decisions and think about tradeoffs. Policymaking is thinking through tradeoffs.
What’s the benefit of doing this work externally, instead of doing it at the Congressional Budget Office (CBO), or bringing external people into CBO?
CBO does bring very effective external people in. Part of it is we can screw up. CBO is a very important government agency. When it releases estimates, it needs to get things right and make decisions based on established literature. We’re just a bunch of economists at academic institutions so we can play around and figure out what works. There's a reason we call ourselves the Budget Lab. We are here to experiment and figure out what works and is useful.
You wrote a piece for the White House blog in April 2021 about the pandemic’s effect on measured wage growth. When it comes to understanding wage growth and other subjects, debates in economics center on a few traditional metrics. You’ve argued those metrics are bad for understanding specific subjects. Could you tell me more?
I'm going to get up on my high horse here. It's not that they're bad for understanding certain things. It's that you have to understand the specifics of the data that you're looking at. All data contain information. The question is whether or not it's useful information. I find that people start playing with data and they don’t read the data handbook. What are we surveying? Who are we actually looking at? How is this data constructed?
I worked in the private sector and I think many economists, with some exceptions, just take private sector data on its face. Raw private sector data reflects the business of the company you’re looking at. You have to think about that in context before trying to extrapolate it to the broader economy. Is this representative? What are specific things that this company could be doing that could be impacting what we're looking at? If you're not thinking through the underlying issues in data construction, it’s easy to get things wrong.
When I was at Indeed, a bunch of states and municipalities were instituting $15/hour minimum wages, and they were all going to kick in around the beginning of the year. We were looking at search terms and found that, around that time, there'd been a huge bump in people searching for $15/hour jobs.
We thought this was cool. As people put these laws in place there was a big bump. In doing our due diligence we talked to the search team and said “Hey, have you guys thought about this,” and they replied, “We’re so glad you noticed! We created an auto fill for $15/hour right around that time.”
We could have easily published something and said, “Hey, here's what's going on,” when that was not what was happening. It was auto filling for people, people were more likely to search for it because it was being suggested for them.
We talked a bit about your work writing tweet threads that would go out with new labor market data. Relative to other places where economic debate happens, how important is Twitter?
I have so many emotions about this. I miss old-school Twitter. I don't know what's going to happen to the Econ Twitter community. And I'm actually very sad about this. I think that people haven't figured out where to go, at least as far as I'm aware. Maybe everyone's figured out where to go and they've left me behind.
Personally, I've met some of my favorite economists and people through Twitter and people that I would not have met otherwise. Ernie Tedeschi, who I shared an office with for two years at CEA, is someone I met through Twitter. We would never have had him if it weren't for Twitter. I had people at my wedding that I met on Twitter!
The other thing about Twitter was it was really democratizing. When I was hiring at CEA it was an amazing recruiting tool. There would always be tweets from people saying they got into various PhD programs, and if there were ones whose backgrounds looked relevant or interesting I would reach out to them and say, Congratulations! You should think about applying to be a staff economist at CEA in two years.” I wouldn't have even known that those people existed if it weren't for Twitter.
Maybe we can recreate some of the best parts of old-school Twitter here on Substack?