Skip to Main Content
Speaking of the Economy
exploding lightbulb
Speaking of the Economy
July 12, 2023

This Paper Blew My Mind

Audiences: Economists, Bankers, General Public

On the 100th episode of the podcast, three leaders of the Richmond Fed 's Research department — Kartik Athreya, Huberto Ennis and Ann Macheras — share the research paper that had a major impact on their careers and the economics profession in general. Paper topics range from bank runs to insuring against risks to knowledge spillovers.


Tim Sablik: Hello, I'm Tim Sablik, a senior economics writer at the Richmond Fed.

Welcome to a very special 100th episode of Speaking of the Economy. To celebrate, I'm delighted to have on as guests three leaders of the Richmond Fed's Research department: Kartik Athreya, executive vice president and director of research; Huberto Ennis, group vice president for macro, micro and financial economics; and Ann Macheras, group vice president for regional analysis and outreach and Research communications.

Thank you all for being with me today.

Huberto Ennis and Ann Macheras (overlapping): Thank you for having us.

Kartik Athreya: Thank you, Tim.

Sablik: As listeners will have already seen from the title of this episode, we've got a fun topic of discussion for today. You're each going to be sharing an economic paper that changed your life, or at least left a big impression on each of you.

Why don't we dive right in? We'll start with you, Huberto. Tell us a little bit about your paper, what it's about, and what the main takeaways are.

Ennis: I selected a paper that is a classic by Diamond and Dybvig. These two scholars received the Nobel Prize last year for their contributions to financial economics.

Doug Diamond has been associated with our Research department for over 20 years. It's been a great pleasure for me and, of course, an honor to interact with him during the visits he [has] made to our department over the years. Doug is really a celebrity in financial economics research.

The title of the paper is "Bank Runs, Deposit Insurance, and Liquidity." It was published first in the JPE [Journal of Political Economy] in 1983. Interestingly, after the stresses in regional banks early this year, the paper is as topical now as it ever was.

The broad objectives of the paper are to study a rational theory of banking panics and identify a role for deposit insurance within that framework. It was pioneer work at the time, not just for the foundations that it provided of one of the main government interventions in financial markets, that is, deposit insurance. In its efforts to understand panics, it set up a theory of banking more generally that was incredibly useful for thinking about many other financial topics.

Sablik: When did you first encounter this paper?

Ennis: I read some version of the model back in Argentina in my undergraduate years, say 2003, 2004. If you grew up in Argentina, you have to be interested in financial stability. You live with it all the time.

I think I became most familiar with the paper while working on my Ph.D. dissertation at Cornell. I gave a presentation of the paper in a theory class for Professor Karl Shell. It was an advanced class and we met on strange times and days. I think my presentation was [on] a winter night at Cornell in upstate New York, way past sunset, for sure.

One thing I remember is that after the presentation, Karl Shell asked me for my notes. Professors at Cornell, and in particular Karl, always thought of his students almost as colleagues. In Argentina where I was coming from, that was much less common.

Sablik: When you were thinking about which paper to choose, did this paper change your mind or open your eyes to something that you hadn't previously considered?

Ennis: The part of the paper about deposit insurance, which is in the title, is actually not my favorite. At the time the paper was written, there was a sense in the academic community that the bank run problem had been solved. The question was, how was it solved? Diamond and Dybvig wrote this paper that formalized the most common explanation — which was deposit insurance — in a very elegant and eloquent way.

After the 2008 financial crisis and also the recent stresses at regional banks, it seems that the bank run question is alive and well. Things are more complicated than we initially thought and more work on this topic is currently being done, a lot of it based on the Diamond-Dybvig framework.

Leaving aside the deposit insurance aspect, there's a remarkable insight that comes from the Diamond-Dybvig model. It's a more subtle insight, but a very important one.

When we see banks getting in trouble and the government is intervening, people get frustrated because these are interventions that put at risk taxpayers' money. Many people end up with an extreme reaction of advocating for what's called narrow banking. The idea is to ask banks taking deposits to hold fully liquid assets that they can quickly use to repay deposits on demand in case a panic happens.

That sounds reasonable in principle. But, as it turns out, the Diamond-Dybvig model is an amazing tool for understanding the social costs of narrow banking. In that framework, the bank is part of an insurance arrangement. Most of the time, depositors do not need their deposits immediately. The bank can invest in more productive, possibly illiquid activities, using a portion of those resources and still be able to satisfy those depositors who happen to have an immediate need to withdraw their funds. There's a social value on allowing the bank to do this — to invest in illiquid productive activities while keeping some funds available in case depositors need them.

Now, of course, when depositors get panicky, the bank has a problem. Some of the resources are invested in these productive, illiquid enterprises, so they're not immediately available. Yet the Diamond-Dybvig model makes clear that if those panics are not very likely, then narrow banking is not the answer to the optimal financial arrangement.

Narrow banking is a simple solution, but it is not the best solution. It has costs, social costs. And it's important to understand those costs to make appropriate policy decisions. Diamond and Dybvig gave a framework that does just that.

Sablik: Yeah, that's super interesting, and not a point that I would immediately associate with that paper.

As you mentioned, this paper has obviously been hugely influential on the economics profession, with thousands of citations. Ann and Kartik, do you have any thoughts about how this paper has impacted you on your professional or educational journeys?

Athreya: Sure, Tim.

Every time I think about financial stability — especially as it relates to banking-like arrangements, not just those that happen in a formal bank but any time you have an institution that's borrowing in short-term instruments of one kind or another and then using the proceeds to invest in something that is longer term — all of those arrangements to me have the flavor of the setting that Diamond and Dybvig studied. So in one sense, it's broadened my view of the sources of financial instability to well beyond something that may call itself a bank.

Ennis: Also, you see it in the reaction to financial crises. A lot of policymakers familiar with economics literature reference that framework to explain the way they were reacting to events during that time. It's sort of a bedrock of a lot of the thinking about banking and bank stability, and instability, at the Fed and other central banks around the world.

Athreya: I don't know if, Huberto, you would agree with this characterization. But to me, it puts economists in the proper kind of humbleness. If we see something out there, we don't automatically presume that it's somehow a deficient arrangement or a silly arrangement. Rather, we ask the question, when might this arrangement actually make a ton of sense? That gives some respect to the people in the private sector that are thinking all the time about how to contract with each other.

Sablik: Alright. Well, Kartik, why don't you tell us about the paper that you selected and how you first encountered it.

Athreya: The paper that I selected was "The Risk-Free Rate in [Heterogeneous-Agent] Incomplete- Insurance Economies" by a guy named Mark Huggett. This is a paper that was published back in 1993, 30 years ago, in a journal called the Journal of Economic Dynamics and Control.

This is a paper that's influenced me and the entire profession really profoundly. The paper was part of an early group of papers that tried to grapple with what happens when people have to deal with risks that they can't buy easy, clean insurance contracts to deal with. Historically, examples include unemployment insurance, disability insurance, workplace injury insurance.

Mark's paper was motivated by an observation in the data: the interest rate that risk-free forms of saving, and in particular the U.S. Treasury bill rate, why was it so low all the time? He had about 100 years of data that he looked at and noticed that, in real terms, that rate was really close to zero.

The way he wanted to think about that was if we live in a world where people face risk to their earnings, but they don't have well defined insurance arrangements to deal with that, they're going to try to do something. And the something that they're going to try to do is build up buffers when times are good for them that they can then draw down when times are not so good for them. This is a very natural idea for any of us when we build up a savings balance because we think, man, things down the road might be tougher than they are right now, so I want to make sure I'm provisioned for it. Or, I'm going to retire one day, so I'm going to try to be provisioned for it.

What his insight was is that if people as a group wanted to do this, that would put downward pressure on the interest rate. The interest rate that the market would then deliver wouldn't have to be as high simply because everybody was so eager to find somewhere — anywhere — to park savings. It was a beautiful, empirically motivated paper that used a tight model of behavior and decision-making to explain a fact.

Sablik: When you first read this paper, was it eye-opening to you or did it change your mind about certain things?

Athreya: Yeah, it was eye-opening to me. The time that I encountered it, I had been scratching my head a little bit to think about how to interpret data on consumers not repaying debts.

When I started thinking about that, it occurred to me — and I think would occur to anybody thinking about it — that people borrow, expecting things to be better in their futures than they are right now. That's really the main reason to borrow. For some people, it works out and they repay their debt. For other people, it doesn't work out and they have trouble repaying their debt. When they have so much trouble repaying their debt that it makes sense to pay the costs associated with defaulting, some people do it.

Well, if you had rules governing how costly it was to default, but then you use Mark's and the papers related to Mark's framing of the issue where people face these risks that they had to then borrow and deal with at times and saving in good times and so on, it offered a ready-made vehicle that I had to just tweak a little bit in order to get answers to the question I was interested in.

Sablik: Yeah, you alluded to the fact that this paper certainly had an impact on your career and has been influential on the profession as a whole. So, what has been the impact for the economics profession as a whole?

Athreya: This paper was part of an effort that allowed economists to think about the implications of uninsurable risks on a host of issues. In any setting where people face the kinds of risks that we've been talking about — where you could get sick, you could get unemployed — and you build up a savings buffer or you draw it down or borrow, that's going to mean that, at any point in time, even people that were kind of the same to begin with in terms of, let's say, their education and age, as they go through their lives they're going to wind up in different positions in terms of what they've got in the bank. Some people are going to have better luck than others and they're going to wind up in different places in terms of the savings positions that they have.

So, these models very naturally open the door to being models of wealth inequality. They take earnings-related uncertainty and tell you what that might imply for wealth-related inequality. That's obviously a topic that has been of tremendous interest to the profession.

Macheras: Yeah, I'll jump in.

As I was listening to Kartik talk about the paper, my mind was going to wealth inequality. Since one of the groups that I oversee at the [Richmond] Fed is our community development function, the focus there is on low- and moderate-income populations and we focus a lot on why some groups lag behind other groups in terms of their accumulation of wealth. What are their investable opportunities, if they can save at all?

Athreya: That's a great point.

Macheras: I just think … I just thought that was interesting. That's where my mind was going.

Ennis: I was also thinking when Kartik was talking [that] you see the impact of computers on academic research. We've always been concerned about this distributional issue, these disparate experiences of agents. People in 1960s were thinking about these issues, but it was really hard to get interesting, clear implications from economic thinking. Mark is one of the early people that exploited that newly available computational power.

Athreya: For me, a standout component of that paper is the care that Mark takes to explain to a reader how he did things. If you read the paper, it's just a model of clarity in its exposition. He explains how to compute the solution, to Huberto's point. I can read this, soup to nuts, and figure out how I can just …

Ennis: It's a recipe.

Athreya: Yeah, it's a recipe.

The second point that I want to quickly make is this [paper] occurred in the mid to late '80s, early '90s. It was a period in which some of the pioneering work was actually done by women economists. Our profession has, I think, struggled over time to have the kind of representation, especially in macroeconomics, of women. It should be pointed out that one of the earliest papers in the literature that I'm talking about was written by somebody named Ayse Imrohoroglu. That paper predated Mark's by a little bit. I picked Mark's for the set of reasons that I did, but I wanted to point out how important her paper was also to launching what came after it.

Sablik: Super interesting. Alright, Ann, why don't you tell us about the paper that you picked?

Macheras: Sure. I went in a little bit of a different direction. My paper isn't as connected to the ones that Huberto and Kartik discussed.

My paper is called "Growth [in] Cities" by Edward Glaeser, Hedi Kallal, José Scheinkman, and Andrei Shleifer. It was published in 1992 in the Journal of Political Economy.

It's kind of interesting that we all picked papers that came out around the same time.

This paper focuses on theories of economic growth, specifically growth that's generated by knowledge spillovers that come about when firms are clustered together. In practice, these knowledge spillovers result in one firm's productivity gains generating productivity gains for other firms, at no cost to them. Since firms are most likely to be geographically concentrated in cities, this explains the fact that cities continue to grow, even though costs are high there — the cost of living, doing business there is higher compared to non-urban areas.

The paper tests different views on the dynamics of growth from knowledge spillovers. Do spillovers occur between firms in the same industry? If so, this means that the geographic specialization of industries is important for growth. Are knowledge spillovers more prevalent across industries? If this is true, then we'd expect a greater diversity of industries within a city to matter more for growth.

The authors conclude that the latter proposition is supported, at least by the data they had available at the time. Adding to this notion that we were just talking about in terms of technology, data availability has changed a lot since 1992 when this paper came out.

Sablik: When did you first encounter the paper?

Macheras: I didn't exactly encounter the paper when I should have. I say that because I should have been more aware of it when I was a young professor teaching economic development. I did bring some of the concepts into my class through the work of Michael Porter, who I'd say arguably started popularizing the concepts and made the idea of industry clusters, which is very much dependent on this concept of knowledge spillovers.

I didn't really come to this paper until about a decade after it was written. At that time, I was working in the field of economic development. I was working at a state economic development agency. By then, I had transitioned into being more of an applied regional economist.

Sablik: When you did read [the paper], did it open your eyes to anything or change your mind about something?

Macheras: It did. This paper, with its focus on the broader concept of what we call agglomeration economies, provided a really useful framework for thinking about why some places grow faster than others. This is something I was asked all the time as an economist at an economic development agency. Regions want to know, "How can we grow faster?"

Those agglomeration economies, or also we call them positive externalities, occur because of the knowledge spillovers that Glaeser and his colleagues studied, but also because firms that are in close proximity to each other can share sources of inputs. They can also share a common pool of skilled labor. All of these things lower cost to the firms.

Working within an economic development agency as an economist, I needed some sound economic principles to guide my thinking.

Sablik: That touches on my next question. How did this paper influence your professional career?

Macheras: As a regional economist, it was a small jump to put those concepts into practice. I worked on a project to identify industry clusters for various regions in Virginia and it was an interesting opportunity to work with our community college system. I was focused on industry clusters and a colleague was focused on occupation clusters to help identify opportunities for targeting industries to attract to a region.

As an economist, the idea of targeting industries might seem odd. Shouldn't the market take care of that? Why do we have to worry about what industries to target? That might be true. But I was really comfortable with this type of analysis because it provided a lot of information to the economic development professionals in the region about what they already have and what it makes sense to think about for future development.

Sablik: Right. It's definitely clear to see how these ideas have had an impact on applied economics. Do you have any sense of how the paper has been impactful on the economics profession as a whole?

Macheras: I'd say that the paper was a solid contribution to the empirical analysis of knowledge spillovers. There has been a lot more work since the empirical analysis, again with better technology, better data, more computing power. Edward Glaeser continues to explore the growth of cities and his work is cited by many economists.

In fact, he was the economist that was interviewed for our Fourth Quarter 2021 issue of Econ Focus, so you can read more about his ideas there.

Sablik: That's right. We'll definitely put a link to that in the transcript and show notes.

Kartik and Huberto, did you have any thoughts to add?

Ennis: Ann was linking these ideas of agglomerations to growth. Around that time in the economics profession, there was a boom of thinking about economic growth. [Robert] Lucas had said, around that time, once you start thinking about economic growth, you can't think about anything else.

Athreya: And to what Huberto said I would add, if you look across the three papers that we've been talking about, or the three kinds of papers we've been talking about, they're all motivated by looking around and seeing something happen — whether that's banking, whether that's an interest rate that looks awfully low, whether that's the fact that industries and human beings tend to clump together in space. These are all empirical observations, broadly speaking. In all cases, they have spawned some of the best reasoning that economists have come up with to explain or understand those phenomena.

I'm just struck by the practicality of a lot of what we've been talking about today. I think that's something that sometimes gets overlooked in discussions of what economists do.

Sablik: I think that's a great way to sum everything up. Kartik, Huberto and Ann, thank you so much for coming on today to make this 100th episode really special.

Ennis: Thank you.

Macheras: Thank you.

Athreya: Thanks, Tim.

Sablik: If you enjoyed this episode, please consider leaving us a rating and review on your favorite podcast app.