The Beveridge Curve and Implications for a Soft Landing
This 17-minute podcast was released Feb. 22, 2023, as a part of the Timely Topics podcast series.
“I would say that, in the present conditions, a soft landing is possible given that it’s likely that poaching vacancies go down by more than unemployment vacancies,” says Paulina Restrepo-Echavarria, a senior economist at the Federal Reserve Bank of St. Louis. “We’re kind of hopeful.” Restrepo-Echavarria sat down to discuss the Beveridge Curve with Laura Taylor of the St. Louis Fed.
Paulina Restrepo-Echavarria: I would say that, in the present conditions, a soft landing is possible given that it’s likely that poaching vacancies go down by more than unemployment vacancies.
Laura Taylor: Hello and welcome to the St. Louis Fed’s Timely Topics podcast. I am Laura Taylor, your host. Today I’m here with Paulina Restrepo-Echavarria and we’re here to discuss our favorite beverages. And, of course, since we work at the Federal Reserve, that’s the Beveridge Curve.
Thanks, Paulina, for joining me today and thanks for putting up with the bad jokes.
Restrepo-Echavarria: [Laughs] You’re welcome.
Taylor: Paulina is a senior economist at the St. Louis Fed and focuses on international macroeconomics and search theory. Now, today we’re here to learn more about her and her co-author’s work taking a fresh look at the 50-year-old concept known as the Beveridge Curve.
So Paulina, let’s do some level setting to get us started. I imagine most folks listening to this podcast have heard of the Beveridge Curve, but let’s back up a bit for those who might not be as familiar.
An easy place to start is, of course, unemployment. So most everyone hears about the unemployment rate on the nightly news or reads about it in newspapers. It’s basically a measure of those who do not have a job but are able to work and actively are seeking employment. Another basic concept is job openings. This is measured by the number of vacancies available on the last day of each month. So they could be full-time, part-time, or even seasonal, but the open position has to be there and available.
OK. So I think that’s a good setup for where we’re going next. Now, Paulina, tell us how the Beveridge Curve kind of fits into all of this.
Restrepo-Echavarria: Yes. So basically, the Beveridge Curve is the relationship between vacancies and unemployment. So it tells us that when we lower the number of available vacancies, by how much would unemployment increase. And vice-versa, basically, if we lower unemployment, how big of an increase in vacancies we would need. So it’s just this relationship, which we tend to think that it’s negatively sloped in the sense that the higher the number of vacancies, the lower unemployment, and the lower the number of vacancies, the higher unemployment.
Taylor: So measuring a relationship.
Taylor: Great. So the Beveridge Curve is, basically, like putting together a recipe. We add some ingredients like the unemployment rate and job openings and we get a product. So if we use more or less of the ingredients—say unemployment goes up or openings decrease—then, of course, we get a different result. So now, what’s interesting about the research you’ve been conducting is you said, “Wait, what if we change those ingredients? You know, what would the final product look like then?” So tell us a little bit about your approach.
Restrepo-Echavarria: Yeah. So basically, it’s this idea that not every vacancy is made equal in the sense that we know that firms can hire people from unemployment or they can hire people from employment. So they can go to another firm and poach an employee and bring them to work with them. And this will make a definite difference in the Beveridge Curve because if a firm, let’s say, opens a vacancy that is tailored for an employed person, then that person would make a job-to-job transition, they would never go through unemployment, and will just have a different job. But we wouldn’t see that in the Beveridge Curve. So to that extent, the Beveridge Curve only applies to people that are unemployed and moving to employment because, you know, they are part of that unemployment pool that you mentioned earlier, which are those people that are willing to work and looking actively for a job and they find a position somewhere.
So when we think about vacancies divided into two different types of vacancies—let’s say one for employed people and another one for unemployed people—then those vacancies that are open for employed people would not affect the Beveridge Curve. So, in some sense, we’re thinking about those vacancies together with those that are open for unemployed people, but they are not really affecting that relationship between unemployment and vacancies…
Restrepo-Echavarria: …because you kind of skip that step, to some extent.
Taylor: It’s almost a more robust measure.
Restrepo-Echavarria: Yes, exactly, because we think that the hiring process when it comes to employed people might be very different than the hiring process for unemployed people.
Taylor: You know, that’s very interesting. So by separating the types of job openings, it can have an effect on the outcome or the shape of the curve. So after you did this, what did you find?
Restrepo-Echavarria: So what we observed in the recent years is that the Beveridge Curve has had this very particular shape that we have not observed before. So basically, as I described before, you usually observe this negative relationship between vacancies and unemployment, how far from the origin in the sense that if you plug that in a graph, you know, like, that can shift outwards, can shift inwards. But usually, it has that negative relationship.
However, in these last years we’ve seen that, first, it’s flatter; that curve has become really, really flat in the sense that, for some period of time, the sensitivity of vacancies to unemployment seems to be very low, but for another period of time, it kind of turned vertical, meaning the opposite. So basically, you could shift around vacancies quite a bit and unemployment wouldn’t move. So it’s been having this very strange relationship that everyone has been kind of scratching their heads about and wondering what’s going on with labor markets that the Beveridge Curve has been looking so weird.
And so, thinking about this, we thought, “Well, you know, there has to be a reason or an explanation.” And when you think about vacancies divided into these two different types of vacancies, you’re able to explain this because once you remove the vacancies that belong to employed people, then you get back your usual Beveridge Curve that is just a negative relationship between vacancies and unemployment.
Taylor: Interesting. You know, if we go back to this idea of the recipe metaphor, it’s almost like some ingredient wasn’t behaving the way it was supposed to or it has traditionally behaved. And so, you know, we talk about job poaching. You know, it’s a really interesting way to look at this 50-year-old concept. You know, you said, “Hey, there’s more to this. Looking at this from the surface level doesn’t look like the way it should, so it’s not telling the whole story.” So what originally made you and your co-author say, you know, “We need to look into this?”
Restrepo-Echavarria: So Anton started looking at this first—he works for the Dallas Fed—and he got very interested in thinking about the shape of the Beveridge Curve. And so he started thinking about that and then, you know, like, discussing things, it kind of seemed obvious that one alternative was we knew that job-to-job transitions have increased over time. And so this should matter. Then, the natural thing would be to think, “Well, if job-to-job transitions have been increasing, something that we have not looked at, or no one has really looked at, is the vacancy side.” We tend to ignore the vacancy side. We put kind of all of our efforts in trying to understand unemployment and search effort and how much time people put into looking for a job. But then, you know, we’ve had vacancies kind of sitting on the side and, you know, not paying too much attention to them.
And so once you think that these job-to-job transitions are important, well, why not think about, “Well, maybe vacancies are different?” Because we know that this is true; we know that firms hire from employment, we know they hire from unemployment. These could be due to many different reasons. Probably the skill, the level of skill they’re looking for in a worker, could be one of them and could dictate whether they look from one pool or the other. So that’s kind of what led to thinking about these issues.
Taylor: That makes a lot of sense. We certainly heard in the headlines a lot during the pandemic that folks were switching jobs at a very high rate. And when job openings are targeting existing workers and the increase in hiring is really just job switching, then it’s obviously not taking workers out of that unemployment pool, so it wouldn’t decrease that unemployment rate.
This is quickly becoming a complicated recipe. So [laughs], of course, there must be potential implications to your initial findings. The Beveridge Curve and its traditional construct is a useful indicator to understand the labor force. So using this new way to measure these new findings, what does that mean for looking at or understanding the labor force?
Restrepo-Echavarria: So I think that it’s important when we think about policy implications because these would have an important meaning for how sensitive labor markets are to macroeconomic conditions. So in other words, it makes a big difference when we think about vacancies going down if the vacancies that go down are vacancies for employed people or vacancies for the unemployed. Because, of course, if vacancies for the unemployed people go down, then this generates unemployment and we don’t want that. You know, it’s one of our mandates as, you know, for the Fed to keep full employment. But if we know that the number of vacancies that is going down is the number of vacancies for employed, well, those people already have a job. So in that sense, it’s fine.
So it really matters to think and to better understand, how do firms decide whether they poach someone or whether they hire from unemployment? What dictates that decision? And also, is it the case that policy or changes in policy affect the firm’s decision of which vacancies to alter the most or if it’s the same.
Taylor: Absolutely. And you touched a little bit on the Fed’s dual mandate, which, of course, is stable prices and full employment. So it’s obviously very important that while we’re measuring employment or unemployment that we’re looking at it accurately and really getting a true understanding, a deep understanding, of what makes up those numbers, what makes up those ingredients.
And another, you know, sort of timely topic that’s been in the news is this idea of a soft landing. So a soft landing is the idea that the Fed can implement tighter monetary policy or raise rates and it won’t lead to a large increase in unemployment. So typically, when rates are increased, companies may slow borrowing, may slow spending, and ultimately, that can slow or stop hiring, which leads to higher unemployment. But what you’re seeing is that this time, it may look a little different because of the mix of vacancies for, you know, folks who are currently employed or vacancies for folks who are unemployed.
Restrepo-Echavarria: Yes. So I think that this is probably the most important implication for our work, which is, we’re very concerned, of course, always, when we’re increasing rates that we’re going to slow down economic activity. This implies that firms slow down spending. They decrease credit demand. And so usually, they decrease the number of vacancies, they hire less people, there’s more unemployment. However, it would make a very big difference if those vacancies that are being “sacrificed,” quote, unquote, when there’s an economic slowdown, are the vacancies for the employed. Because if the vacancies for the employed, as I said before, are the ones that are sacrificed, then there’s no unemployment, there’s no reaction from the unemployment side.
And so something that we find in our work is that, over time, when we estimate what’s the fraction of vacancies open for poaching and what’s the fraction of vacancies open for the unemployed, that fraction of vacancies open for poaching has been increasing, especially since 2015. It’s been larger over time. And in 2020, actually, it peaks and goes up drastically such that the overall share of vacancies open for poaching is two-thirds out of the whole pie, basically. And what we’ve seen with previous recessions and monetary tightenings is that in the face of a monetary tightening, it seems like both poaching vacancies and unemployment vacancies have decreased by the same amount. So by the share that they represent in total vacancies.
And so this kind of indicates that right now, there wouldn’t be any reason to think that with a monetary tightening, you know, vacancies open for the unemployed would be sacrificed in a larger amount than the vacancies for poaching. So, you know, if what we’ve historically observed keeps on going, with the tightening of the Fed, basically, vacancies for the employed should go down by more than vacancies for unemployed people, so we really shouldn’t observe as much unemployment due to the tightening of monetary policy.
And since we wrote the paper in September of 2022, all the numbers that have come out after that have been supporting this. We have not been observing increases in unemployment since we kept increasing rates. So we’re kind of hopeful, you know, that these might be an indicator that, hopefully, a soft landing is possible; we won’t see as much unemployment in the face of the increasing rates that we have been making. So, you know, there’s some hope out there [laughs], I think.
Taylor: That’s such a great way to put it. And we’re here talking in January of 2023, so we’ve got a little bit of time under our belt since you first initially looked at this and to be able to say, “Let’s look at this in a new way,” because, like you mentioned, it’s not behaving the way it historically has, so we need to pull apart these pieces and see what we’re really looking at. Like you said, it gives us hope for that soft landing, as folks call it.
So this is such a good reminder, too, that this is why the Fed conducts this type of research. It’s important to peel back the layers and see if the same old recipes are worth tweaking a bit here and there.
So as I mentioned, Paulina, both of us work here at the St. Louis Fed, which means we can’t get away with talking about any research without also talking about the caveats. So, tell us, you know, as you’re doing this research, you know, what are some of the caveats here? What are some of the things you might want to explain about this research?
Restrepo-Echavarria: So I think the main caveat is that we don’t really know how firms decide whether, you know, they poach someone or they hire from unemployment. We know very little about how vacancies are formed and how they are decided, in some sense. Because we know that, legally, firms cannot put out a vacancy and say, “Hey, we’re only hiring employed people,” you know, because of Equal Opportunity rights, et cetera, et cetera, this cannot happen. So finding evidence about this division between, you know, poaching vacancies versus unemployed vacancies, I think it’s tough, you know. When you talk to headhunters and so on, they all agree on the fact that, yes, most of the time, they know where they’re hiring from. But there’s no solid evidence about this.
So that’s one thing. And that’s very connected to the fact that when there’s some monetary tightening, we don’t really have a way of knowing which type of vacancy’s going to go down by more. So as I said, historically, we’ve seen that they kind of go down proportionally, but this doesn’t have to be the case in every situation, you know. Like, we know that, over time, the economy looks different; firms make decisions that are dependent on economic conditions at the time and so on. So this tells me that we really need to increase our research in terms of understanding vacancies better and how firms decide how to hire people, from where, and how open are they at substituting, you know, individuals from one pool of workers to another, and so on. Because this can really tell us, you know, data elasticity and how likely really that soft landing is because if, all of a sudden, in this situation that we’re living, they decide to just reduce the vacancies for unemployed people rather than the poaching vacancies, then…
Taylor: What a difference.
Restrepo-Echavarria: …we won’t have a soft landing and this will generate unemployment. And we really don’t know what these events are. So this is an important caveat, I think. We need to better understand how this happens.
Taylor: Thanks for explaining that. It’s always good to understand.
Paulina, thank you so much for joining me on Timely Topics. You’ve got such interesting research to share with our listeners.
And if our listeners are interested in learning more about Paulina’s work or the work of others here at the St. Louis Fed, visit stlouisfed.org today. Thank you.