Nov 19, 2018

Adam Ozimek on Inflation, Migration, and Productivity

A Macro Musings Transcript
David Beckworth Senior Research Fellow , Adam Ozimek

The following transcript is from Macro Musings, a podcast series hosted by Mercatus scholar David Beckworth that explores the past, present, and future of macroeconomic policy. Subscribe on Apple Podcasts or your favorite podcast app.

David Beckworth: Welcome to Macro Musings. The podcast series where each week we pull back the curtain and take a closer look at the important macroeconomic issues of the past, present and future. I'm your host David Beckworth of the Mercatus Center. We are glad you've decided to join us.

David Beckworth: Our guest today is Adam Ozimek. Adam is a senior economist at Moody's Analytics where he covers US labor markets and demographics. Adam also actively blogs and tweets on economic issues and can be found under the Twitter handle Modeled Behavior.

David Beckworth: Adam joins us today to discuss fed policy and demographics. Adam, welcome to the show.

Adam Ozimek: Thanks for having me. Glad to be here.

David Beckworth: Well, it's great to have you on. We have followed each other for some time on Twitter, blogging before that. In fact you're a part of Modeled Behavior. I mentioned that in the intro. You're actively engaged in commentary on websites, again on these mediums. I want to know, how did you get into all of this? What was your journey into the policy world?

Adam Ozimek: So, my journey in economics begins in high school. I had a government policy and economics class, or something like that. That really sort of piqued my interest. So, I went into undergrad as an econ major. I stayed through the whole time. It really, I found that it combined a bunch of things that interest me. Like business, I've always, especially in high school I really enjoyed business classes and social science. Then once you get into the college level it starts to involve math and philosophy a little bit more, which I really enjoy.

Adam Ozimek: So, that sort of was the hook for me that pulled me in and just spoke to a lot of my interests. Also, current events. It's very relevant to what's happening in the newspaper both in policy and in economics news, and business news.

Adam Ozimek: So, I majored in that in undergrad and then when I left school, started kind of something that would become a theme through my career is that I like to do things there are fun and interesting to me. I make a lot of decisions based on what's going to keep me entertained and interested.

David Beckworth: That's fair.

Adam Ozimek: Yeah. I mean, probably put a little bit extra weight on that than the average person.

Adam Ozimek: So when I left undergrad I realized that a lot of the jobs that I was qualified for, and that were kind of out in the market, didn't seem like they were going to be that fun. I owe it partly to not getting enough empirical skills in undergrad. So, I thought I'll go get a terminal masters that'll really beef up my empirical skills and let me do something more fun.

Adam Ozimek: So, that's what I did. Before I even finished the masters I did get a more fun job in economics consulting. So that was really applied micro kind of stuff. It was very real estate economics focused. As I was doing that I realized that once again the more fun stuff was done by PhDs in consulting. Having the PhD, I could see the people at my company that had PhDs got to do the more fun things. If you have a masters or less you don't do as many fun things long run. So, once again to maximize my enjoyment and entertainment I sort of kept going through the end of the masters to pursue my PhD while working full-time in consulting.

Adam Ozimek: That's one way to do it. I don't necessarily recommend that to everybody.

 David Beckworth: Right. Sounds pretty tough.

Adam Ozimek: It was a lot of work. It was a lot of work. Consulting a PhD at the same time. But I made it through and it was around this time that I started doing the blogging. I was writing for Modeled Behavior, which Carl Smith started, and started doing Twitter as well. Getting really into general economics stuff. That was really fun for me. I like that a lot more than the kind of applied micro real estate stuff that I had been doing in my day job. It was a lot of commercial real estate valuation, local house price indices, damage analysis for litigation in real estate.

Adam Ozimek: That stuff is kind of fun, but general economics commentary really, I enjoyed that a lot. So I tried to make a switch in my career to become more of a generalist. That's when I made the jump to Moody's Analytics. That's what I've been doing ever since.

David Beckworth: Yes. So you are a big user of social media like Twitter. I'm just wondering, what role does that play in your day-to-day job. Does it inform what you do, or does it go back the other direction?

Adam Ozimek: Yeah. So it goes both ways really. One of my roles at Moody's Analytics is I write for our blog, which is free. A lot of our stuff is behind the pay wall. But I write for our blog and being on social media helps me understand what people are talking about, what people are interested in. It's a great way to find leads for new hypotheses. There's constantly, constantly people being wrong about economics on there and that's the fuel for my job is finding people saying wrong things and disproving them with data, ideally.

Adam Ozimek: So, it's a great way to find stories. It's also a great way to promote the work that I'm doing. Sometimes I'm writing for eight people and those eight people are all financial economics journalists and I just want them to see it then sort of write about what I'm writing about. It's a great place to connect with those people.

David Beckworth: Absolutely. Again, our listeners will want to check out Adam @ModeledBehavior. That's his Twitter handle.

David Beckworth: Well, one of the things you have written about quite a bit and you've done a lot of commentary, but you have a nice piece that kind of focuses in on this particular issue, is Fed policy. We just had the September FOMC meeting, they raised their target interest rate another 25 basis points. It's now in the 2-2.25 range. They've signaled that there's much more on the way. So they're going to continue hiking throughout next year, maybe even once more this year. You have an interesting paper that's titled 'The Fed’s Mistake'. Tell us about that.

Adam Ozimek: So, I think that there hasn't really been a look back at the rate path that the Fed has taken so far. There hasn't been enough questioning of when they decided to raise rates. That's what I tried to do in my paper. One of the most important pieces of data to look at is the feds changing beliefs about labor market slack.

Adam Ozimek: So, when they started raising rates in the end of 2015 they figured out through their projections, you can tell here's what they thought long run unemployment was and here's what unemployment was at the time. Through those two numbers you can gauge what the fed thought the amount of labor market slack was. Over time they've revised downward their estimate of long run unemployment basically accepting that the labor market had a lot further to heal than they did.

Adam Ozimek: So basically I look at those downward revisions and I say, "This represents the Fed’s admission of how much it has underestimated labor market slack." I look at how that changes over time and what that implies about monetary policy.

David Beckworth: Yeah. You have a great quote in here from Neel Kashkari, he's President of the Minneapolis Federal Reserve Bank, a past guest on this show. He says, "Members of the FOMC, including me, now believe that both the neutral real interest rate and the natural rate of unemployment are lower than we had realized in prior years. The implication of these revisions, admittedly with the benefit of hindsight, is that monetary policy was less accommodative than we previously thought." You also go on to note that even Fed Chair Jay Powell kind of acknowledges this point as well. Because he says, "Policy was less accommodative than thought at the beginning of the normalization."

David Beckworth: So, what you're saying is the estimate the FOMC has of how low unemployment can go before inflation takes off is constantly being revised. It has been revised a lot since 2015, and even before that.

Adam Ozimek: Yeah. Sort of like taking their thoughts and finishing them. So, Kashkari and Powell, they both say, "Our monetary policy, turns out it was less accommodative than we thought." But then my argument is, "Well, complete that. Complete that thought." That invites that it was less accommodative than it should have been. I mean this is controversial and this isn't it even ... It's not the consensus among economists. It's not even the consensus among economists at Moody's Analytics. It's my opinion, it's not the firms opinion. But it really seems like such a small leap to me. Like, the Fed admits they were less accommodative than they thought, are they not then admitting they were less accommodative than they should have been?

David Beckworth: Yeah. Absolutely. I mean, another way of framing it, and I'll be more explicit, they were effectively too tight. If the unemployment gap was bigger than they thought it was, then raising interest rates was a mistake. Right? They tightened too quickly. So I think it's a very powerful insight.

David Beckworth: What you do in the paper is you kind of take their framework they would use. You think of a Phillips Curve thing. A Phillips Curve implies there's some link between the amount of slack, how much unemployment the economy can tolerate before inflation kicks in. You take that and you kind of kick out these estimates of where the federal funds rate should have been. You also use a Taylor Rule. So you use kind of the tools that they would use and you look at the path that the federal funds rate should have gone down versus the one that it did. Then don't you also take that and then tie it into estimates of the economy itself, what does it mean for employment-

Adam Ozimek: Yeah. So I basically, like you said, I take a couple different paths to get to what the optimal rate path really was. One way that I do that is I just say, "Alright, here's what the Fed tried to do." So to take a very specific example when they started raising rates in December of 2015 they thought that long run unemployment was 4.9 percent, and actual unemployment was five percent at the time. So, basically they thought that the amount of slack in labor market measured by the unemployment rate was .1 percentage point.

Adam Ozimek: Now, they've revised their estimate of long run unemployment. So I say, okay if the Fed thought rates should start to rise when the unemployment gap was .1 percentage points. Then just using their, what I call their revealed preferences for the rate path, when should they actually have started raising? Definitely Spring 2017 and as late as 2018 depending on where you think actually long run unemployment is. So, using that approach and a Taylor Rule approach I come up with alternative paths that are all more dovish than what they actually did.

Adam Ozimek: Then I run them through our macro-econometric model, the Moody's Analytics model. Basically, I use this to determine what was the effect on the economy of departing from this optimal rate path of raising rates sooner than they should have been. What the model suggests is that had the Fed followed the rate path that I say ex-ante they should have followed unemployment would be between .3 and .6 percentage points lower. There would be between 500,000 and a million more jobs would have been created.

Adam Ozimek: So, obviously there's really wide bands around this kind of exercise. There's a huge amount of uncertainty. But, I think order of magnitude, that this is what we're talking about. We're talking about a pretty substantial effect on the economy over this time period.

David Beckworth: Well, I totally bite. It fits my priors, full disclosure.

David Beckworth: But, I do think it's a good point. It's effectively the Fed put a governor on the engine of the economy and never let it kind of take off, heat up, until really Trump’s policies come in and really kind of add the winds behind the sails that’s really pushed the economy forward a lot faster recently. But, I think maybe we would have seen some of this sooner had the fed not had raised rates as soon. That's what your analysis shows. I mean, that's quite a big number, 500,000 to one million more net new jobs.

Adam Ozimek: Yeah. I think that there's some ... One defense of the Fed is to say, "Well, this is just how fast the economy should have recovered. This is a responsible recovery speed and anything faster than that would have been pushing it." But, I find that to be a pretty puzzling response in the aftermath of what's the slowest economic recovery from a recession in modern US history to say that, no, actually that's the right speed. We wouldn't have wanted to go any faster. In any other context talking about this recession I don't think anyone accepts defending the Fed’s rate path. I don't think anyone would try to make that claim.

David Beckworth: Yeah. I mean, normally after a steep recession, or any recession, there's a bounce back, kind of a V shape returning to kind of the level trend path any of these measures were on. Instead we've had more of an L-shaped recovery. A kind of a unit root in that every economic theory is that it falls down and stays at a lower path. It's growing but it never returned to its previous path.

David Beckworth: Kind of the analogy I like to use, and I talked about this with Jared Bernstein on a recent episode, is that imagine that the economy is a truck and the Fed’s the driver of the truck. They have to be at a certain place at a certain time. The speed limit is 65 miles an hour, and that's where they've got to get to. Their destination, their final place. If they hit a traffic jam, there's a wreck, that wreck could be the recession. They come to a stop or they really, really slow down, what has to happen afterwards is that truck has to actually speed faster than 65 miles an hour. It needs to pick up to maybe 75, 85 for a while until it gets back on its projected trajectory at a certain time. So, speeding up is important. That's what we normally see after a recession. Usually the economy, inflation, things tend to grow a little bit faster than normal. Then they settle back down. But we did not experience that after the great recession.

David Beckworth: What you're showing is that this in part, not entirely, but in part due to fed policy.

Adam Ozimek: Yeah. I definitely want to stress that the Fed raising rates too soon is a part of the story, it's not the story. So, there are obviously a lot of really big important headwinds pushing the economy down. I just, I find it very strange that we have this debate in the early days of the recession, like will this be a V-shaped recovery or will it be an L-shaped recovery? You had this debate and no one, at the time, would have said, "It's going to be an L-shaped recovery. And by the way, that's a good thing." It seems obvious to me that if the Fed could have done something to make it more of a V-shaped recovery to make the recovery happen faster that is what they should have done. It's strange to sort of now look back and say, "Well, that's impossible." Either it's impossible or it would have been bad, it would have been too risky.

Adam Ozimek: I'm not saying the Fed could have dug us out of that hole in a year or two, but getting us out all I bit faster seems wholly within the realm of possible. Especially when you consider that ... In my analysis I focus on the actual rate path. But that's just part of it. I don't talk about forward guidance at all, but that's another area where their consistent underestimating of labor market slack had consequences. I think if you ... My 500,000 to one million jobs would be even larger if you could estimate the forward guidance counter factual where the Fed understood early on how much labor slack there was and sort of communicated their preferred rate path that was consistent with that.

David Beckworth: So this speaks to two issues I see. Number one, there simply was a challenge in getting macroeconomic policy to where it needed to be at for the great recession given kind of the norms and the values of what's appropriate policy. Those policies worked well during the Great Moderation, you tolerate a little bit of inflation, you don't ever overshoot. But after a steep recession I think more was needed. I think we kind of boxed ourselves into a corner where that kind of policy is not possible.

David Beckworth: But there's a second point, I think, that comes out of your paper and this conversation. A more general point. That is this view that what happened during the Great Recession, and what's happened afterwards, is just to be accepted. It was a structural story. We're talking about the Fed here. But there were a lot of other people, you've highlighted this in your other work, in your other comments. A lot of people said, "Look, this is structural unemployment. There's nothing more we can do about it whether it's monetary policy, fiscal policy. It's structural, just accept it." That's kind of where we are today. This L-shaped recovery is just kind of taken for granted. This is life.

David Beckworth: As opposed to maybe there's some hysteresis, maybe there is some things that could have driven more rapid growth. I mean, I think back to the great depression, right, if one had said in the late 1930s this is just the way it is you would have had a story to what we've had today. Now, of course, they had World War II and some unusual circumstances that pulled them out. But I do get the sense that people have just settled, become complacent, with what we've got.

Adam Ozimek: Yeah. I don't know what it is about structural stories that there is something appealing to them to people. I think this, it's appealing to economists, it's appealing to journalists, it's certainly appealing to readers because I watch stories about structural causes of unemployment just go viral so much faster than stories about just demand. I saw this throughout the recovery and it was one thing after another. One structural story after another and you just, you knock down one with data and people just drift onto the next one.

Adam Ozimek: I don't know what the psychological, cultural explanation for it is, but as someone who has argued for demand side causes for the last few years and someone who has spent a lot of time arguing against the structural causes, the amount of support that the structural causes get, it's just disproportionate.

Adam Ozimek: I don't know what it is, but … I'm hopeful that when the dust has settled from this recovery it's something that we can look back on and maybe do better at next time, because there is an absolute bias there. I'm a supply sider. You know I mean this isn't, it's not my ... I'm not like a liberal economist, I'm not, if you had to place me in the scale, I'm in the center right, so I wasn't talking about demand, demand, demand before the Great Recession, it's where the data led me and so it's kind of been frustrating to watch the consistent biases towards the structural explanations and it's been a problem.

David Beckworth: But this view, your view has been validated by what we've seen under this Trump experiment with added fiscal stimulus, right? To some extent at least, we're seeing people come back into the labor force who otherwise were being classified as disabled, or some of the elderly are joining the labor force. So when the economy gets hot enough, and again this is up to a point. I think we both agree there's a limit to how much demand can do, but relative to where people say potential real GDP is today relative to where it could be, there's still some room for growth I think is what both of us would say. But clearly there's a ceiling. You can't change the whole world based on demand. But there is maybe some more added capacity that could be created through running the economy hot.

Adam Ozimek: Yeah. And I'm not arguing for, and I don't think you are either, some extremely heterodox economic view. I feel like this should be the mainstream conclusion. And I'm hopeful that it will become so as history settles. And I do think that ... Certainly in the aggregate these stories have been debunked and that the recovery has simply gone on much farther than they thought it would. The unemployment rate has gone farther down than they thought it would. The employment depopulation rates recovered farther than they thought. Inflation has remained subdued. Wage growth is not ... It's not that it's just continued to improve, but it's continued to improve very linearly, at a modest gradual acceleration. It didn't explode overnight like some of these structural theories predicted. All the pieces of economic history are laying into place to prove these demand side stories correct in the long run. And I just hope that we have this look back, and make a judgment about here are the theories that did a good job of explaining the recovery, and here are the theories that did a bad job. And let’s sort of raise the status of the good theories, and lower the status of the bad theories, so that when we have another recession, which we will, that we don't make this mistake as badly again.

David Beckworth: Well, if we used Milton Friedman and Anna J. Schwartz’s ‘Monetary History’ as a guide, it took them about thirty years to get at least a monetary version of demand story for the Great Depression back to the mainstream level. So, that took awhile. I guess the hope is today people like you and others are doing research currently that is hopefully changing the view. Maybe we'll have a sooner recognition the role demand played in the Great Recession and the slow recovery.

Adam Ozimek: That's a great comparison. And I hope it doesn't take that long, but I do think that there is an unwillingness to look back and judge these theories kind of critically. I don't see as much looking back as makes sense, and so I do think it will take time. I think people need to forget what their own bad forecasts were so that they feel comfortable looking back and being critical of the Fed’s bad forecast.

David Beckworth: Yeah. And it's easy to fall back into your camp, your ideology, your view of how the world works once you’re past the worst part of the recovery and the recession. But there is something, again, I don't want belabor this point, but in the 1930's this is an illustration. In '36, '37 the recession was in part due to the Fed tightening policy and the treasury sterilizing gold inflows. And the reason they did this is because they were concerned about inflation. 1936, '37, they still hadn't returned to full capacity. The economy still wasn't firing in all parts of the engine. And yet they were worried about inflation. And I don't know if this is something hard wired into our brain. You know it spans generations that you can't have flexible inflation.

David Beckworth: Now over the long run we want a stable nominal anchor. We want inflation to be anchored over the long run, but in the short run we got to have the flexibility to allow V-shaped recoveries, and that's my theory is that we simply don't have it at a big broad institutional level. Whether you want to view it through fiscal or monetary policy, I just get the sense the norms institutions in place aren't flexible enough to tolerate it. So it's not a question of maybe who's running the Fed, it's the constraints on the Fed, and the fiscal policy that they're under.

Adam Ozimek: Yeah. It's funny. You know you can look back at '36, '37 and it seems so puzzling now with hindsight that they would be fixated on inflation. But I have to admit that having lived through the Great Recession and the obsessions with inflation in the aftermath of that, it seems less surprising than it used to, that people ... Looking around at the political and economic commentary today, I understand how '36 and '37 happened.

David Beckworth: Yeah. And to be fair to some of the old timers out there, I should be careful, but they also kind of cut their teeth on the great inflation of the 1970's, so they have that in mind as well. On top of maybe any inherent bias or heuristic that's hard wired into our brain to not like inflation, period.

David Beckworth: Well, let's move on. That was a great paper. Is that paper available to the public or is that one behind the paywall?

Adam Ozimek: That one is behind the paywall. But if someone wants to read it they can shoot me an email and I can send them-

David Beckworth: Okay. There you go listeners. You have inside access to a great piece, so directly reach out to Adam. But on the topic of inflation, you've done some other work. And one of the puzzles over the past decade is that inflation has been lower than the Fed actually wanted, at least the Fed's target. And I imagine part of the story is the Fed’s policy as you mentioned in your previous paper we just finished, but part of the story you also attribute to population growth. So you have another paper, I'm surmising is behind the Moody’s paywall, but it's called ‘Population Growth and Inflation’. And you do some interesting work in this paper. You show through a series of models that population growth is strongly related to inflation. So you do a cross country panel. You look also at a panel of US metro areas from 1971 to 2016. And then you do a long run model of inflation and population. So tell us first, what is the theory between population growth and inflation? Because when I first saw this I had to do a double take, because I've got this monetarist blood flowing through me, and it just kind of didn't resonate at first until I had to read your paper carefully. So what is the theoretical link between the two? And then we can get into the evidence.

Adam Ozimek: Yeah. I mean I'm right there with you. This was a fun and surprising paper for me to do. And oh, by the way, this one you can get if you go to adamozimek.com. One of my papers that I have listed up there. This one is outside the paywall.

David Beckworth: Fantastic.

Adam Ozimek: So if you think about supply and demand in the aggregate, it's not clear right away why a growing population should be inflationary because population growth is supply and demand. And like you say, the monetary authorities should be determining the money supply and that should set the inflation rate. But while population growth brings more labor and demand, one of the inputs that's fixed in the production function is land. And so my argument is that population growth, because of the inelastic supply of land, this doesn't adjust to changes in population growth. And so basically what happens is it's through the housing market channel. Where population growth accelerates, housing demand doesn't respond right away, and land is in relatively fixed proportions. And so you have this slow and maybe even incomplete increase in housing supply, and so that translates to higher house prices and then higher inflation. And on the other end when population declines, then you have this housing stock that doesn't get torn down overnight and only depreciates slowly, and so that's going to depress house prices and that's going to be a headwind for inflation.

David Beckworth: So if understand correctly and correct me if I'm wrong, this is kind of a negative supply shock story that there's an insufficient amount of housing relative to demand, so supply effectively is too short which causes prices to go up, and since housing’s an important part of the price index it gets reflected more broadly. Is that the right story?

Adam Ozimek: Yes. On the way up and on the way down.

David Beckworth: And on the way down. And now again, the macro person might say, "Well, but supply shocks are like a one-time hit to the price level." But what I think you're arguing is that this is like a sequence of negative supply shocks, right? As a population's growing, this has been a persistent housing shortfall for many years, so it's been a persistent ... It's not just a one-time problem. This has been a continual problem.

Adam Ozimek: Yeah. And I mean you can see it in the cross sections of the data. So the rational expectations economist might say, "Well you know, homeowners, you'll get hit once." But then home builders will anticipate and they'll increase supply in expectations of population growth. But empirically that doesn't seem to be the case. And when you're talking about declining population or population growth that's very slow, like I said, it's not like home builders tear down homes. So that's not really anything that forward looking agents can do anything about it except to build less. And there's only so much that that can go.

David Beckworth: Let me ask you this. If you took a price level measure that excluded housing, what would you see then? Would there be less of a connection between population growth and the inflation rate?

Adam Ozimek: It would probably be less of a connection, but I think housing is an input into everyone's production function. There's nothing that you make that doesn't involve land prices, whether it's services, retail. You know land prices translate to wages and they translate to input prices. So it's not just that housing is something that we measure in the CPI, but it's that housing is an input, and land is an input to all these other prices. So I don't think it would go out completely.

David Beckworth: So in other words, my scenario’s too farfetched. You can't simply remove the effect of housing because it's so pervasive in so many parts of what we do in our life.

Adam Ozimek: Right. It's an input. It's up there with land and ... Land is up there with labor and capital as a core input into the production function for basically all households and firms.

David Beckworth: And to be clear you're not saying macro forces ... I would say medium to long run inflation rates determined by the quantity of money relative to demand. You're saying that's still important, but this is an important piece of the puzzle we've ignored.

Adam Ozimek: Right.

David Beckworth: It's not the only piece of the puzzle.

Adam Ozimek: It's a headwind, and it's tailwind. And it wouldn't necessarily be important if monetary authorities were fully aware of it, and fully anticipated it, and fully offset it. But I don't think that that's the case because you just don't see population growth being talked about that much. So my study is the first to show this for the US. But there's a study before that did it with Japanese pre fixtures. And you certainly don't hear this story in Japan where they've struggled to generate inflation, and you don't hear the story basically being one about housing markets generating consistent deflation. And so if monetary authorities don't recognize this problem, then they can't offset it. And if the problem is so big that it requires them to really change their reaction functions, then it's easy to see how they wouldn't fully offset it either.

David Beckworth: Yeah, so Adam this is a very interesting study, but I'll admit there's one part of it that really disappointed me because it destroyed one of my priors, one of my hobby horses. And that is my argument going into this is, yeah. There can be a link between these two things, but I was viewing it in terms age. And I know your answer already is that it's not age, it's old people, it's prime age. So I'm going to throw out what my theory was, and you can tell me what you found that contradicts this. But my theory, because I'd seen some of this evidence from Japan as well. So my thinking was okay, in the case of Japan the reason demographics, population growth matters there is because it's a slowing population, it's an aging population, and the older folks, they live on fixed income a lot. In fact they hold a lot of these government bonds and the last thing that they want is to see higher inflation. So it's kind of a political economy story there. They're not going to tolerate, since they're a big part of the population, since they really are sensitive to inflation as well, they're not going to politically tolerate anyone who's really going to gin up the inflation rate. And so there's kind of that story.

David Beckworth: And then I'd say back here in the US we have all the Baby Boomers who lived through the great inflation we talked about earlier, so you get all these policymakers still in power who still think about the 1970's, and they freak out when people like Adam Ozimek and David Beckworth bring it up as something not to worry about. But what you show is that's not really a story, right? You find the link is somewhere else.

Adam Ozimek: Yeah. I was equally as open to that theory as well when I started looking at the data because you do see some studies make this claim, whether it's through the political economy effect, or a savings rate effect, or something else about the aging of the population. But I found it to be ... You could find a result in the data where aging affects inflation, but it was very sensitive to specification, and depending on how you measured the aging of the population, sometimes it went one way, sometimes that went the other. And it was a very sensitive result that wasn't robust, that I don't really believe. And so it doesn't surprise me given this sensitivity that you do have studies out there that claim to show it, because it can be found in the data. But it's different than population growth. Like population growth in the data, in the US metro data was really, really, robust. Like you can't shake it.

Adam Ozimek: One of the fun things that I found was that for a handful of metro areas, you actually have CPI data going back like 100 years, which I don't think you even see that data used very often. So I was able to look at it in the very long run. I was able to look at it in the short run, and then international data. And I used an instrumental variable and you just can't shake the result of population growth. But aging is, it's sensitive to specification and I just don't really believe it. And if you think about it, the theory is kind of second or third order stuff. It's like aging affects this, which affects this, which affects inflation. Whereas the housing channel's like super direct.

David Beckworth: Right. So this is an interesting finding for sure, and it's something that you suggest policymakers need to be more cognizant of. They need to be thinking about it more carefully. And I wonder what would you tell Jay Powell if you saw him. He said, "Hey Adam, come into my office. Sit down and explain this finding to me." How would you give him that elevator pitch? What to do? What to think about?

Adam Ozimek: So I would just put it on the list of reasons to probably not worry about inflation as much. It's just another thing for the chalkboard, another reason why compared to the ‘60s and ‘70s inflationary pressures are not going to be as strong today as they used to be. And it's going to be even less going forward as the US population growth slows. So I would just put it on the list. I wouldn't put it at the top of the list. I think cyclical supply is still the most important thing. It's something I think they should think of as a headwind. In the forecasting, in macro forecasting we talk a lot about headwinds and tailwinds, and things that are pushing up or down on certain things. And I would just put it on the list of inflationary headwinds.

David Beckworth: Yeah. And one more point to tease out from this study before we move on, is that what you found was an asymmetric effect on population growth that if it's a positive growth it's still there, but if the magnitude is much smaller than if there's actually negative growth. Is that right?

Adam Ozimek: Yeah. Which makes sense because housing supply, elasticity when you're talking about falling population is going to be pretty weak. If you look at a place like Detroit where you lose population and lot of these places in the Midwest, it takes a long time for the quantity of housing to respond to the decline in population, which means the prices do a lot of the work. And I'm not the first person to show that asymmetric effect either. So when you're talking about declining population or you're talking about parts of the country where you have declining population, the effect is going to be larger than for a faster growing population.

David Beckworth: Very interesting. Also very intuitive, when you think about elasticity of housing supply. All right. So I have to put this plug in one time for nominal GDP targeting here, but this seems to me to be a good argument for, or another argument for nominal GDP targeting, because there's so many things that could distort or distract what inflation's trying to tell you, right? And this seems to be another one of them.

Adam Ozimek: Yeah. I think that that makes sense to me. Nominal GDP targeting has a lot of appeal to it, and it's something that I think the Fed should definitely consider and ...

Adam Ozimek: That I think decided to definitely consider and take seriously, along with the simpler things like raising the inflation target in the aftermath of the Great Recession. I would put it on the list of things to consider and I'm a fan of nominal GDP targeting, but not as settled on it as the one true policy reform going forward.

David Beckworth: Yeah. It's just my argument, listeners will know this, is it's a whole lot easier just to focus on stabilizing demand than spending, than trying to figure out month to month, what's driving these changes in inflation. When you have so many things, so many supply side shocks, or distortions that can feed in. This is another, in my mind at least, another reason to be cautious about getting worked up about the latest inflation numbers.

Adam Ozimek: Yeah. I agree with that, but on the other hand, I also don't feel like understanding the economy correctly in the aftermath of the Great Recession was impossible based on the data that we did have. It's almost a, if we had simply had better policymakers, they should have been able to do a better job and it's not a fundamental impossibility based on imperfect measurement. It gets a little frustrating sometimes with people who, when we go out there and we say, "Well, The Fed should have been more dovish" and then people say, "Well yeah, but that's impossible to have known."

Adam Ozimek: It's like, "I was arguing with you three years ago and telling you why it was possible to be known. I showed with the data" and it's like, it's a frustrating argument to go from someone saying, "No, you're wrong" and then three years later being like, "Well yeah, but it would have been impossible to know at the time."

David Beckworth: Yeah, no. That's a fair point. I want to be clear I'm not trying to articulate the Fed couldn't have done something different. I'm just saying month to month, there's a lot of noise, but you can start looking at trends pretty easily. Again, over medium term, you should be able to infer something is wrong with monetary policy if for the last ten years, you're not hitting two percent inflation.

David Beckworth: Anyway, let's move on to another paper and this also ties into population growth. This paper is about how population growth may affect the lack of dynamism in the US economy. The title is “Firm Startups, Population Growth, and Domestic Migration.” Tell us about this paper and what are some of the insights you've found.

Adam Ozimek: So, we have this decline in entrepreneurship over the last 30 years, and I'm not going to say it's a mystery, but a lot of it is unexplained and there have been a lot of great economists trying to figure out exactly what has caused the decline in entrepreneurship, and the best papers that have been written on this, their conclusion is like, "Well, here are 15 things that it isn't."

Adam Ozimek: They very persuasively established all the things that it isn't, and one of the most I think persuasive things I've seen is to look at demographics. So, the basic argument is there's two kinds of channels. [Hathaway] and Litan had a great study where they just did a cross-sectional metro look at like, "Hey, population growth slowing down and startup rates slowing down kind of go hand-in-hand at the metro level."

Adam Ozimek: That was an inspiration for me and my co-author, Martin Worm, to dig into that a little bit more, and to focus a little bit more on identification and what we found is that when population growth slows down, you also have a slowdown in startup rates. We use an instrumental variable approach, we use a panel model approach, and really what we did is we took the [Hathaway] and Litan results and we tried to add a few extra elements of rigor to it.

Adam Ozimek: The results stand up, and it really suggests that we should think of that slowdown in population growth in the US, especially in a lot of parts of the country, as a significant driver of lower levels of entrepreneurship.

David Beckworth: Yeah, so this is putting another piece of the puzzle together and our understanding of why we have this slowdown in business dynamism and firm startups, and other measures of the health of the US economy. You have a quote there in the paper from Ryan Decker et al. Ryan Decker is a friend of ours on Twitter and they say in their paper, 2015 paper, "We do not fully understand the cause of decline and indicators of business dynamism and entrepreneurship."

David Beckworth: What you do is you provide a little bit more clarity, right? You've taken the muddy water, cleared it up a bit, so we know now that population growth is one of the contributors, or the lack of population growth.

Adam Ozimek: Yeah. There's two channels it can operate through. One is the demand side, one is the supply side. Basically the story is that when there's new demand in the local area, more customers, people want to buy more things, who responds to that demand? What research has found is that it's new firms that respond to it. New firms are much more likely to respond to increases in demand than existing firms.

Adam Ozimek: This provides a channel through which mere population growth, by creating new customers, can increase the number of new firms which is then the startup rate. There's a counterintuitiveness to this, which is when economists think about population, they think about it as very rote in the production function. Like, you have more people, that's more supply, that's more demand.

Adam Ozimek: There's no connection between that and entrepreneurship and productivity growth that would matter, but if you think of new demand as favoring new firms, then that's a way for population growth to drive entrepreneurship. The other way is new population means new workers and the startups may also be the ones who respond to those labor supply shocks, too.

Adam Ozimek: If you have a demand shock, or a labor supply shock, who's going to respond to that? Who's going to create the new firms? Who's going to expand to meet a demand? And the evidence suggests it's new firms. This is how you get from mere population growth to increases in startup rates.

David Beckworth: Okay. Part of your paper also deals with domestic migration, so what's the tie in there?

Adam Ozimek: Domestic migration is a nice instrument. Basically, we take Greg Howard had a paper where he uses metro to metro migration flows to create an exogenous source of metro population growth, and we thought that was a really cool instrument, and what we wanted to do was take that and connect it to this question of entrepreneurship and population growth.

Adam Ozimek: Basically, we take Howard's instrument and we use ... If you're looking at one metro, basically it's like what is the migration to this metro that you would predict based on the migration patterns between all the other metros? Is the simple way to look at it, and it makes for a nice instrument for metro population growth, and using that plausibly, exogenous source of population growth, we can see that it is related to startup rates. That's how we bring some more rigor to this question.

David Beckworth: Okay. Well, let's move on to another study that you do that's tied to demographics and this one's tied to age. We've talked about population growth, but you have a paper that focuses in on aging, where it is consequential. This paper's title is “Aging and the Productivity Puzzle.” Can you tell us about that and what did you find in it?

Adam Ozimek: Alongside the declining startup rate, one of the other big macro puzzles out there is why has productivity growth been so weak lately? Actually some people at my company came to me a while ago and said, "Do you have any interest in working on productivity growth and trying to figure out what's going on there?" I said no, because I thought I'm not going to crack this puzzle. Am I going to figure out the thing that Robert Gordon and Chad Syverson haven't figured out, and Larry Summers?

Adam Ozimek: It seemed like such an active and contested area of research. I was really skeptical we were going to find anything. But then I was just playing around with some data and I noticed a really interesting and strong cross-sectional relationship between the share of workers in a state and industry who are 65 and up, and productivity growth in that state and industry.

Adam Ozimek: Basically, if you think of like 20 or so NAICS code major industry categories, so different kinds of manufacturing, leisure, hospitality, think about those industries and you think about the 50 states, you end up with 50 by 20 observations, and you look at the share of workers in those industries that are old and productivity growth, there's really strong relationship.

Adam Ozimek: It stands up to industry and state fixed effects. So, if you look at workers in California, health care, after controlling for California fixed effects, and after controlling for US health care fixed effects, the productivity growth and aging relationship is really strong. That's what pulled me into this research, and I started working on it with my colleagues, Dante D'Antonio and Mark Zandi, who's the chief economist here.

Adam Ozimek: We wanted to take a closer look at this theory using microdata, and so one of the things we do at Moody's Analytics is we produce the monthly employment report for ADP, the payroll software company. We have access for research purposes to ADP's microdata.

David Beckworth: Wow.

Adam Ozimek: That's really-

David Beckworth: That's nice.

Adam Ozimek: ... nice thing to have. When we come up with ideas, we take 'em to ADP, get the permission and ADP has researchers who work on their data through questions and research that are interesting to them. It really allowed us to take a detailed look at what happens at the firm level, to wages, as a result of aging of the workforce. It really allowed us to make sure that we weren't ... This isn't a composition effects thing. This isn't your firm gets older and those older workers somehow make less money and that brings down average wages.

Adam Ozimek: This is, if you're looking at an individual worker, controlling for their own age, controlling for their own industry, what happens when they work with more people who are 65 and up? We find through a lot of different specifications and tests, we find that older workers bring down pay for other workers.

David Beckworth: Okay, so you find a strong connection between aging and declining productivity. What is the theoretical link between the two?

Adam Ozimek: We came at this with two theories actually. One was the wiseman theory, and one was the albatross theory. The wiseman theory was that an aging workforce was associated with a lot of retirements and older workers were the most skilled and the best, and they had all this firm specific human capital. When they left, that was like bad for the firm. That was the wiseman theory, that was my theory, which was wrong.

Adam Ozimek: The alternative theory which Mark Zandi came up with, was the albatross theory. That was that older workers are bad for the firm and they're holding back productivity. This is through like blocking technological improvements, blocking changes in the way things are done, that they hold back other workers. That's the theory that we find the most evidence for, that when firms see ways to do things differently, they see new technologies to embrace, that older workers block this somehow.

Adam Ozimek: Either it's rational for the firm to not do it when you have a lot of older workers, or it's rational for the workers to not do it. We don't know the mechanism there, but it appears to be associated with more high-skilled occupations, this older workforce effect. We think it has something to do with technology blocking.

David Beckworth: You find this evidence at the firm level. Has it aggregated up? I mean, is this something that can explain part of the aggregate decline of productivity in the US? Let me reframe it. The decline in the productivity growth rate.

Adam Ozimek: Yes. There's obviously a lot of uncertainty here, but I think maybe between 0.3 percentage points to 0.75 percentage points, somewhere in that range per year. It's ongoing, so this is a pessimistic result. The share of workers aged 65 and up has basically doubled over the last decade or so, and slightly more than a decade, but this is going to continue going on. It's big enough of an effect to explain not all of, but a significant chunk of the slowdown of productivity growth.

David Beckworth: Okay. So, there's some policy implications from this paper, aging and productivity, as well as from your firm startup population growth paper, at least to me there seems to be some policy implications. One is we need to have more babies, higher fertility rate, but unfortunately that's actually going down. The second one would be we need to have more legal immigration, particularly in the high-skilled area where you find these results. Is that a fair interpretation of the policy implications?

Adam Ozimek: Yeah, I think that you look at this evidence collectively as a whole, the work that I've done on demographics, it suggests that there are important spillovers to a young and growing population that economists are underestimating. That's been the wider theme of this research. I do think that there's a pro-immigration conclusion to take from all of this.

David Beckworth: Yeah, if want to make America great again, we need more legal immigrants, right?

Adam Ozimek: Yeah, and also if you want to make a lot of places in America specifically great again. It's not just about getting the total number here, but I think it's also about helping immigrants be distributed throughout the country more broadly, to places that aren’t getting enough of them now and to understand immigration is able to bring more prosperity to struggling parts of the country.

David Beckworth: Yes. One other thought I had after reading these papers is so immigration or a younger workforce through higher fertility rates is one solution. The other solution would be to have enough technological gains to maybe offset these aging folks who don't adopt it. I'm thinking back to 2002 to 2004 where we did see at least a temporary spike in productivity growth. My understanding, correct me if I'm wrong, that was tied to the IT, the leftover residual IT investments during the dotcom period. There were some initial gains.

David Beckworth: I'm doing a leap here, of faith, what if we had more robots, more technological innovation? Could that, you think maybe close the gap or make up for the decline that's going to be implicit in the continuing aging of our population?

Adam Ozimek: Yes. I think that that's definitely true, and again, as a forecaster, I like to think of headwinds and tailwinds. I think one of the mistakes that people make when they talk about productivity growth is they identify one current headwind or one current tailwind and they conclude therefore productivity will be low going forward or it'll be high going forward.

Adam Ozimek: But I think the real takeaway from Robert Gordon's really impressive work of economic history and productivity growth, is that every era has its measurement problems. Every era has its headwinds and tailwinds, and you can't just focus on having one particular measurement problem now, or one particular headwind or tailwind now, and conclude productivity growth is either going to be a huge success or that it's doomed.

Adam Ozimek: There are always factors buffeting productivity growth up and down and just because aging is going to be holding back productivity growth for the next 15 years or so doesn't mean that we can still have great productivity growth. It's just one factor pushing in one direction that is going to require a lift in the other direction.

David Beckworth: Yep. That's good to remember because productivity is the ultimate thing we care about in terms of driving up living standards, higher per capita income growth. I had Jared Bernstein, I mentioned earlier, on the show, and we talked about how this was an important part of the frictions that we now see in politics, is the decline in productivity growth, that would solve a lot of problems if we could get more rapid productivity growth. It's an important area to continue looking into.

David Beckworth: All right, with that, our time is up. Our guest today has been Adam Ozimek. Adam, thank you so much for coming on the show.

Adam Ozimek: Thanks for having me. Glad to be here.

David Beckworth: Macro Musings is produced by the Mercatus Center at George Mason University. If you haven't already, please subscribe via iTunes or your favorite podcast app, and while you're there, please consider rating us and leaving a review. This helps other thoughtful people like you find the podcast. Thanks for listening.

Photo credit: Joshua Roberts/Getty Images

Support Mercatus

Your support allows us to continue bridging the gap between academic ideas and real-world policy solutions.Donate