Transcript

Jessie Romero: Hi, I'm Jessie Romero, director of research publications at the Richmond Fed. Thanks for listening to Speaking of the Economy. If you like what you hear, check out our website or Apple Podcasts for more episodes or to subscribe.

I'm joined today by Richmond Fed economists Felipe Schwartzman and Chen Yeh. We're going to discuss a topic that is of perennial interest to policymakers: What causes changes in business cycles? Felipe has been studying this question from the perspective of households, while Chen has been looking at it from the perspective of firms.

Let me start with a question for Chen. What interests you about studying business cycles?

Chen Yeh: Well, it's great to be here.

What interests me about business cycles is that it's a very important topic in terms of its consequences for welfare and the growth patterns of firms, which I'm personally very interested in. We macroeconomists have relatively little understanding of it, so it's kind of a puzzle. That's what interests me to study it.

Romero: Felipe, same question for you. Why are you interested in studying this topic?

Felipe Schwartzman: A point that was made by some very important economic researchers is that there are bigger issues, right? You look at the U.S. economy, how much it grew over the last 50 years or 100 years. If you look at the difference between rich and poor people [and] rich and poor countries, these things are much larger than the five or 10 percent fluctuations in output that occur over a really large business cycle. And most business cycles are not that large.

I think the reason people care a lot about that is because business cycles have a very unequal incidence. One of the main indicators is unemployment. Unemployment is all about some people are unemployed, most people are not. Most people are still working. But the fact that you could get unemployed and the people who do get unemployed, for those people it hurts a lot. It could take a very long time for people to go back to where they were in terms of what their incomes were before they lost their jobs. I think there's a sense that policymakers ought to be able to do something about that.

There's this interplay which is very interesting between business cycles, inequality [and] politics, and how the politics then turns around and sometimes can make business cycles even worse. Understanding how these things get together is a really big deal.

Romero: Thanks. Let's talk about policy and policymakers a little bit more.

Felipe, as you note in a recent Economic Brief, policymakers often respond to recessions by trying to stimulate household consumption by giving money to people so that they'll go spend money. But there has not been a lot of attention paid to declines in household consumption as a potential cause of recessions in and of themselves. Why is that?

Schwartzman: It's not to say that people haven't looked at it, right? They have. It's just if you were to look at papers written about business cycles and the main things that people have looked at — the services that start with consumers, consumer sentiment, or consumers are low in wealth or being too levered – these things only started becoming salient after the Great Recession. When there was this big drop in housing prices, people started having a view that this was something that was directly affecting consumption decisions. Again, of course, you can go back and find papers and people who were talking about that kind of stuff before, but I think that's when it really took off.

Until then, there was this view that consumption doesn't really move a whole lot with business cycles. To the extent that it does, you can explain those movements as reactions to business cycles. People are going to smooth their consumption, they're going to try to smooth their consumptions, so consumption is going to move less than one-for-one with output for that reason. This is what you've seen in the data.

I think people have tended to look at things that move perhaps more than output, such as investment or productivity or even inventories, even though it's a relatively small part of total output. It was a little bit of the sleeping giant and then, with the Great Recession, it really became more salient in that sense.

Romero: You just mentioned the decline in housing prices as one thing that was a shock to household consumption. What are some other examples of the kinds of things that could really affect household consumption?

Schwartzman: You could imagine all sorts of things that could happen with consumer credit. There's a paper about the Great Depression where they talk about how in the depression, there was a sense of so many people didn't have access to credit. We didn't have three-year mortgages; people had to be rolling over their mortgages. So, this was a much bigger deal.

You can think about consumer sentiment. I think one thing that happened after the Great Recession in 2008 was people were not consuming as much and they were saving much more. Part of that you can attribute to people who needed to de-lever. But you could wonder whether some of that is just people became more cautious. And if people become more cautious because they're more afraid of losing a job or they've had some bad experiences or they become more pessimistic about the future, these things could also affect output.

Romero: In your research, what do you conclude about the effects of these types of shocks?

Schwartzman: I have this paper with Christian Matthes and we look at consumption shocks. Essentially, we look at instances where consumer goods are moving around disproportionately as compared to the rest of the economy. This is how we define consumer shocks. We find that they account for about 40 percent of fluctuations in the U.S. since the late '70s. That actually precedes the Great Recession as an important factor.

Romero: What are the implications of that for policymakers?

Schwartzman: There is a sense in which this helps rationalize a little bit this notion that providing insurance to people so that they can smooth their consumption is something that is going to also be helpful in stabilizing business cycles, to the extent that this is an important source of movement.

I want to talk cautiously about that. Our research wasn't really geared towards policy analysis. But that's kind of a natural thing to think about.

Romero: Great, thank you. Let's change course a little bit and talk about the firm side of things.

Chen, there has also been some kind of change in thinking or in economics literature about how firms can affect the business cycle. What was the more traditional view?

Yeh: The traditional view is that the economy could only be moved by what we call aggregate shocks. These are unexpected events that are faced by all economic agents. These could be technology shocks, like deficiency of production, or monetary policy for that matter.

In the U.S. economy we have about, let's say, five million firms. If something would happen to a handful of them, we could simply ignore it because we have five million of them. Think about the movements in a pond as business cycles. You're throwing little pebbles in there and you think of those as firms. If the pebbles are small enough, then nothing happens. They didn't cause large ripples, they didn't cause large movements.

That was the original logic behind ignoring what happens at specific sectors even specific firms. That's why we have resorted to the use of these aggregate shocks instead.

Romero: Of course, these days we have some firms that are more like boulders than little rocks. What has more recent research found?

Yeh: Exactly.

The starting point of a recent wave of research is a granular hypothesis that was put forward by a Harvard economist called Xavier Gabaix. What he was saying is that instead of pebbles, some firms are giants. Some people would call them "mega firms." So, rather than throwing pebbles, you're throwing giant rocks into this pond.

You would say, "We already knew that large firms always existed. So why did we ignore this in the past?" Well, the traditional view was saying something like, okay, large firms can exist. But the odds of these large firms occurring in economy, they're so small we can just ignore it. But what the granular hypothesis says is that, well, when we look at the data, we actually see that the incidence of these very large firms – Walmart or Apple or Amazon — is actually much higher than what we implicitly assumed. So, throwing in large rocks into this pond is actually very likely. And that's why we should take it seriously.

Romero: Great, well, thanks. That's a really great explanation.

What are you looking at in your research? How does it differ from what Gabaix was doing?

Yeh: I still adopt the granular hypothesis. But I'm, in a way, refining it a little bit.

The classical granularity theory that was started around 2010 [says] that all firms are, in a way, equal. They can be different sizes, but their volatility — let's say the growth in their employment or the growth in revenue numbers — in percentage terms that volatility is the same. But that's not exactly what we see when we look at the data.

I use confidential data from the Census Bureau that basically consists of every single employer/firm in the U.S. economy. What you see in the data is that when you compare large firms — let's say Walmart — with a typical small firm — like any mom-and-pop store — in terms of their movements, how volatile these firms are, there is actually a very stark difference. Large firms are basically much less volatile than a typical small firm.

That observation is something that you need to take into account when you think about the granular hypothesis. The logic behind this is that, well, there are many large firms in economy, there are many large rocks, so to speak. But if these rocks simply don't cause very large ripples [and] they're not volatile enough, then they won't have such a big impact in the aggregate. This idea that different firms have different volatilities was ignored. This is something that I take into account in the granular hypothesis.

Romero: So, when you take it into account, what do you find about how big those ripples are?

Yeh: Right. So, the traditional view of business cycles that we just talked about with aggregate shocks, they basically said, "Let's ignore firm-specific events because the pebbles are so small you can ignore them." You would find basically a row of firms that is exactly zero. Then you can go to the other extreme which was started by Xavier Gabaix. He was saying, okay, let's make a simplifying assumption, as I just mentioned, that all firms are equally as volatile. Then he found a number that was confirmed by some other studies, something in the vicinity of 30 to 35 percent of U.S. business cycles can be rationalized by events that happen at certain firms. When you refine this by taking into account that a large firm such as Walmart doesn't move as much as a mom-and-pop store, you basically get a number at somewhere in between. My study where I take this into account basically puts down a number of 15 percent instead.

Romero: You said you find that the granular hypothesis still accounts for some business cycle fluctuations but maybe not as large role as other researchers have found. What are the implications of that for policymakers?

Yeh: The broad interpretation of my findings and the granularity studies in general, they're basically saying how much of a role is left for shocks at the sectoral or the aggregate level, right? If I say that firm-specific shocks can account for 15 or 30 percent of business cycle fluctuations, then the residual must be taken up by other factors — shocks that could occur at the industry level or at the aggregate level. You can imagine many of those types of examples. At the sectoral level, these are shocks that occur, of course, at specific industries. Aggregate shocks can account for many types of examples — I can think about technology shocks, efficiencies shocks, or even factors such as policy. This includes monetary and fiscal policy.

My work does not quantify exactly what's in that residual of 85 percent. It's just saying there's scope for these other factors, which includes policy.

Romero: Well, Felipe and Chen, thank you both so much for talking to us today and for sharing your research. I really enjoyed our conversation.

Yeh: Thank you.

Schwartzman: Thank you very much.