Isabella Weber On a New Way to Think About Inflation

I'm like supply chains and disruptions for this so that reason, and all of the various times we've used the term perfect storm to describe certain things in certain industries, perfect storm of perfect storms. My guess right now, you know, in January, is that episodes will be a little less dominated by these topics. That would be my guest. I'm guessing that this year we do a few fewer perfect storm episode. I think that's right. But I think, you know, we we spoke a lot about what it was that

people didn't see coming when it comes to inflation. Why did a lot of economists get it wrong? Why was the inflation that was supposed to be transitory? You know, maybe it was transitory in the sense that it was narrow, you know, not a big sort of like macro unleashed inflation, but it definitely stuck around longer than a lot of

people expected. And so every time we have these big questions like why aren't we better at forecasting inflation, it provides an opportunity to maybe learn something and start thinking about it in a slightly different way. Well, yeah, absolutely, And I would say there's really two things that I feel are unanswered by all of the conversations that we've

had in the last year. So one is still like, is inflation like a macro or a micro thing didn't happen because a few categories really had some disruptions and then spilled elsewhere, And therefore it's not really about fiscal

or monetary policy specifically. And be okay, we do have very high inflation right now, even if there's evidence coming down, what tools, like, if it is the case that a lot of it is related to disruptions and chip shortages and freezes in Texas, etcetera, what are the tools that are best to address that, because it is important to

get inflation down. But on the other hand, there's a pretty good argument that if the issue is some sort of disruption at the ports or whatever, that sort of like strict blunt instruments like raising rates rates aren't necessarily the best approach to dealing with that kind of where raising rates won't grow more trees to turn into lumber and like that, or more births at the ports or

anything like that. So I'm so glad you said that because today we are going to be speaking with one of our Odd Lots favorites, and she has just written a new paper which she says is inspired by some of the conversations that we've had on Odd Lots, but it's also just really interesting because it presents a new sort of third way potentially of thinking about inflation, not transitory, not persistent, a new more interesting third option, and that

maybe could help us think about ways of accepting, Yes, inflation is real, it is a problem, but that some of these blunt instruments that just treat inflation as a function of there's too much money in the economy, we need there to be less. Maybe there are better approaches than just this sort of like blunt monetary approaches to addressing them. Absolutely, So without further ado, we are going

to be speaking today to Isabella Vabor. She has, of course a economics professor over at the University of Massachusetts Amherst, and you might remember her from from some previous episodes. So Isabella, thank you so much for coming back on odd Lots. Thank you so much for having me. It's a pleasure. Thank you. So the paper is called Inflation in Times of overlapping emergencies systemically significant prices from an

input output perspective. But I just want to get you to say on camera that this is inspired by all it is. I mean, I have been listening to your podcast or through the pandemic um and as you were just saying, obviously you have been tracing all these price shocks that have been rippling through the economy. So um, this paper is trying to come up with the framework to trace these shocks and ripple effects in a somewhat more aggregate and possibly less fine print, but maybe a

little bit more like formal kind of fashion. I love that the most self serving first question we've ever asked on an interview, What is you know, what what is an input output approach mean? Because my understanding is that this is actually like a very old idea in economics.

But that is some kind of has actually been forgotten, is my understanding, And that the sort of like various versions of monitorist thinking which sort of treat if prices are hive, we want to understand prices, well, just look at how much money or how much credit is in the economy. Rein that in and you've seen this paper seems to be like going back to like an older tradition in economics. Can you talk a little bit about

what this is. Yeah, So, as we said, we tend to think about inflation as a micro phenomenon, right, where

it's basically just about the movement of aggregate measures. Whereas what we're trying to do here is to think of prices as kind of an interconnected network where since one sector's output is another sector's input um, and therefore one sector's output prices at the cost of another sector UM, you can't kind of trace um the price movements across the whole production network, which input out put tables allow you to do. So. These tables basically like Rchester the

relationships of input and outputs across the whole economy. Historically, UM input output tables really had a breakthrough during the War time UM. Where the question was, how can we hit the enemy's UM economy in ways that we kind of like, with the minimum number of bombs UM, create the maximal damage to really UM, I mean undermine UM the enemy economy's ability to even fight a war. I

mean concretely. Of course, this is mainly about the German economy UM, and so therefore it really is a method of identifying points UM that are of particular systemic significance for the economy as a whole. Back then, the idea was to identify these points of vulnerability to UM, I mean,

as I said, create destruction UM. The idea of our papers to say, if we can identify these points of vulnerabilities, then we can actually UM kind of UM know what what the potential sources of UM of these report effects that can create UM macro outcomes UM could be so if some prices matter more than others, we want to know what these prices are, and input output is one

method of trying to identify these systemically significant sectors. So Tracy might take away from that is that in a war it makes more sense to say bomb and oil refinery, even a candy factory like it, right, Like if you're thinking about, well, what are these sectors that will have the biggest ripple effects across the economy, Then that would be the implication, Well, maybe we should talk about morale

in that context. But no, okay, there's another there's another analogy um that you use in the paper, which is, you know, if you're trying to identify systemically important sources of inflation and maybe address them before they start actually contributing to price increases, it's kind of like trying to identify systemically important banks and then making sure that they, you know, hold more regulatory capital or maybe go under

preempt of stress tests or things like that. Can you talk about, maybe, you know, before we get into policy solutions, can you talk about how that approach maybe differs to traditional ways of thinking about inflation, because you know, in my mind, there's really there's the money terrorist view it's all about the money supply, and then there's a sort of new Canesian view where it's more about you know, supply side and capacity and demand and things like that.

Can you place this new approach in the context of those two older um ways of thinking about it so as different as like kind of monetarism and new caneskiness can be. They share the understanding that inflation is always driven by macroeconomic factors. Right now, what we are doing here, I mean bond case, it's the distance from from aggregate capacity utilization. In the other case, it's more that like classic story of too much money chasing too few goods.

But still it's like trying to locate the the origins of inflation on the aggregate level. What we're trying to do here is to say, well, if there are micro origins of inflation, if shocks to specific sectors can matter in ways that they can unleash processes that actually unsettled the stability of prices overall, that we want to understand

what these sectors are. We want to know where these points of vulnerability are so that we can react to these shocks before they kind of ripple throughout the whole system. And as you said, um interest rates are already being recognized as a systemically significant kind of price, right That's why we have central banks, which of course historically at some point was also not the case. So it was a historical evolution to recognize the systemic significance of interest rates.

In some sense, what we are arguing here is to um is to say that there are more prices than the price of boring money that can acquire systemic significance in bays, that can have a very large implications for monitor stability. Just a shout, I'll be drawn here on the workous of salad Amarova, who has been working on systemically significant price what someone else We definitely have to have on the podcast at some point. It's so funny because you know, of course, and we talked about this

the last time you're on late last year. You took a lot of heat for saying, well, maybe there's a time for having some discussion about price controls, and everyone freaked out about that. And yet they're like, Okay, now let's control the price of money, as if that isn't a form of price control. And yet of course central

banking ultimate price control, the ultimate price control. But you know, so I joked, but I guess it's not really joked that, Like, there are some areas where it's kind of obvious that some sort of some functions in the economy are more crucial to other industries than others. So an oil refinery is going to be more crucial to other industries than a candy factory. But that's obvious. How do you go about systematically identifying beyond the sort of really crude examples.

What is this? This sort of like a rigorous or empirical approach to actually identifying what parts of the economy are in fact the most likely to have ripple effects elsewhere. So, but what we have done in this paper is that we have simulated shocks to every industry in the input out the table, and just for orientation, seventy one industry. So it's not super disaggregated, I mean also not super aggregate compared to microeconomic variables, but it's the fairly broad right.

So we run a shock on each of these sectors, and then we simulate how this shock runs through the whole economy, reciting in an indirect impact on on on the CPI, right, Because if the price of oil goes up, the price of plastic goes up, the price of plastic toys goes up. So therefore in the CPI you do not only have the direct effect of people of people consuming fewer or gas, but you also have indirect effects of plastic and plastic toys. And then in all sorts

of packaging and so on. Right, So we are tracing this direct and indirect effect that results from a price shock in any one individual sector, and we run this UM this simulation UM for for every separate sector, so that we then get distinct magnitudes that show us whether as shocked to UM to to one sector matters more in comparison to another sector. In other words, we can create a ranking of what we called the total inflation

impact from a shock in these sectors. Now, UM, with the simulation, we basically have three determinants that can render a sector systemically significant. The first determinant is the bait in the CPI UM, and housing is a great example here. Housing is not something that is very upstream and that creates a lot of propert effects in other industries, but it has a very large bay in the CPI right. So therefore UM, if there is a price change in housing, it has a pretty large impact on on the CPI UM.

Something like UM like oil and gas is actually pretty upstream, but not as upstream as something like wholesale trade because of the ways in which the upstreamness measures are constructed. But for oil and gas UM you have very large price movements. So the magnitude of the shocks that we use UM are either using average volatilities UM in the in the two decades before the pandemic, or using the actual price change in the pandemic and in the context

of the Ukraine War. So in oil and gas we actually had very large price movements already before the pandemic and then again during the pandemic UM and in the context of the war. So here the drivers would be kind of all three components the importance UM in terms of indirect effects, creating a relatively large total the eight in the CPI UM the large price movements and relatively upstream even though not as upstream as hole sate trade, where it's for holes trade, it's really basically because of

the upstream nous of that sector. And then in the pandemic, of course we also had fairly large price movements there. But before the pandemic the price movements in hole say trade would have been much smaller than in in something like oil and gas extraction. Right, So it's it's these three dimensions that we are capturing in in UM in creating this ranking. So just on this point, can I just press you when it comes to identifying the systemically

important industries or I think you call them ubiquitous industries. Like, how do you just just aggregate their weight in the inflation indusseries versus the extent to which they matter for other prices, Because I'm sure there will be some people who listen to this and say, like, well, obviously, energy and you know, maybe some consumer goods and things like that have a higher weight in the c p I.

And so that's why you're getting these results. If you look at the paper, which I'm not expecting anyone to do, we can we can distinguish between a direct and an indirect effect. Right, So what our direct inflation impact is is just the weight in the CPI, Right, this is just giving you this is what the CPI shows us UM is the weight of UM of the change in save petroleum in core products UM for the change in

the CPI UM. But then there's this addition is share, which we call the indirect effect, which comes from tracing the indirect price UM movements that reside from this initiative shock in say petroleum in core products. Now, of course we have to make assumptions on how industries UM hand over um UH cost increases, right, and be in the paper, we make two distinct assumptions. One is that it's just a pastor of one, so firms just have a cost

increase and just passed us on to their customers. The second assumption that we make is to say, um, what if firms actually don't just pass on the cost, but they actually want to protect their profit margins. Now, if their cost goes up, go up, and they were to increase prices by just the amount of the increasing costs, their profit margin would go down, right, So what if what if they actually protect their profit margins are therefore increase prices by more than the increase in costs um.

So this then gives us different magnitudes of the total effect, but we find that the rankings are relatively stable independent of these different assumptions that we are making. And the CPI that we are using here is a synthactic CPI, because of course we have to break it down to

UM these seventy one industries that we half UM. So it's it's it's it's not the CPI that you're downard from the b A UM if you just look for CPI, but it's a CPI that you get from the b A if you look into it put out for Tavist. You know, something I'm interested in, and I don't know if it's something you've specifically looked at, but it sort of reminds me of this, Like, you know, we talked a lot about or economists have talked a lot in

the last year about goods versus services. Inflation is if these are like two distinct categories of types of things that people buy, that you can draw a bright line and say, Okay, goods have gone down, but service is still up. But it seems to me that any good that we buy is also implicitly a bundle of services that need to go into the You know, if I buy a refrigerator, well, there's some sort of service person

who helped delivery deliver the furniture. Um, and there are services for you know, the truck driver whatever it is. Does your approach the sort of input output approach sort of Um, I'm trying to think exactly the way to phrase it, But in your view, does it offer a more useful way of thinking about categories of goods beyond just or sort of would seem to me like these arbitrary distinctions between between the types of things that get

bought in the economy. Good question, Yeah, great question. UM. I mean I would of course say yes, thank you, thank you for saying that. A great ka. Sorry, keep going, UM. To be sure, services are part of the input out put tables, and we actually find that they are pretty upstream because of the fact that you just described, right, because it's like some form of even like small administrative

service UM involved in pretty much everything. So if we talk about it administ sorry, if we talk about upstream sectors, we tend to think about physical stuff like oil and gas or matters or candicurstans on, right. But what we actually see when we do the analysis that is that some that some services are very upstream. Now, the price movements in services are relatively small on average over time because they are very much UM tied to wages, right, and wages tend to move much less than let's say,

commodity prices. Right. So therefore, when we run these shocks, the service sectoris even though um they are pretty upstream, end up not being very important for the general movement of prices in in this model because just the initial shock um is so small. If we model the shock um based on magnitudes of past price movements and price

movements in the in the in the the COVID nineteen inflation. Now, I do think that this kind of dust give us a different way of distinguishing um categories of of of of sectors if you want so, because the idea here really is to say, Okay, we don't care if it's services, or if it's commodities, or if it's processed goods, or if it's manufacturing of whatever it might be. But all that we care about is um the importance of this sector.

It's if you want so, it's centrality in relation to all other sectors, and in relation to people, it's um um uh consumption patterns. Right. So what we find then is that the sectors that we identify systemically significant basically in three groups. So it's like basic necessities stuff like housing, food forms that of course produce a lot of food

utilities um, and of course also energy. And then basic production inputs stuff like the fosil fields that we haven't really talked about but also can a good products um. And then kind of like basic circulation infrastructure, so things like always say trade right, which is critical for commerce. It's a kind of a basic commercial infrastructure. So I do think that this does give us a different way of kind of UM distinguishing the nature of different sectors.

So once you've identified these systemically important industries, you know, these ubiquitous industries for inflation, things like basic necessities housing, farms, food and utilities, and energy, how does that inform the policy response. So the idea here is that because these UM centers are so important that if there are large price movements in these sectors that have this has implications way beyond these specific sectors. We should be paying more

attention to what is happening in these sectors. So the first occasion is to say we need more monitoring capacity. Why do we need more monitoring capacity? Because we are living in some sort of a age of overlapping emergencies right where UM we have, of course a pandemic that is not over looking for example, at what's happening in China and how this UM impact global production networks, but also looking at climate change and you're a great episode on the Mississippi River and how this is a kind

of just making a whole UM sector. In this case, of course grain UM and other commodities m gryan to a hold UM. But we also have these massive geopolitical tensions that can have huge implications for the ways in which production is organized globally. So in other words, it seems very likely from my perspective that more shocks maybe in the pipeline. Of course, no one wants these shocks, and everybody is hoping that things will be calm and stable. UM.

But um. But like from the perspective of the dynamics of overlapping emergencies, even if inflation is now easing and it seems like in the next couple of years these

kind of shocks UM are likely to keep coming. So if that is the case, you kin't want to have capacity on the side of the state to be able to monitor these sectors that are so important in UM in in ways that allow you to react to these shocks before they kind of create these huge cascading effects throughout the whole economy, and you then actually get some sort of potentially more generalized kind of inflation UM. Beyond

monitoring capacity, of course, not enough to watch. You want to be able to kind of step in and stabilize, right and UM here then UM, I think the big shift in in policy think that emergence from this paper is that once we go on the sector level, we kind of leave the world of more or less homogeneous aggregates where we can talk about interest rates up at one percent or down by one percent or point five

or point seven five or whatever. But it's like pretty one dimension, right, and pretty clear that there's a one dimension that we can measure in very clear ways quantitatively in percentage points, very straightforward. If we now think about the prices of chemicals or um the stability of the flow of goods and holes trade and therefore the prices attached to hols trade or the prices of commodities, we

enter the word of qualitative differences. Right, we enter the word of the last uh two years of odd lots episodes right where you have been unpacking this incredible amount of detail on the qualitative differences that have huge quantitative implications for pricing. But that required quiet um an extraordinary extent of understanding of the specific specifics of these sectors. So to be able to react to shocks in these sectors, I think one would really need quite a bit of

capacity that is quite tailored to these sectors. So there's no like kind of um uh one solution that does at all if you think about housing versus um oil refineries, you would obviously need a very different kind of policy approach, Right, So this then means that kind of these and I mean, there is a lot of capacity out there, but it needs to be connected back to the question of macroeconomic

and monetary stability. It's always so funny to me that there exists a data point on the terminals that the FED monitors called capacity utilization. Is if there's is if the concept of industrial capacity could ever be homogenized in a single index of like here's your finding capacity, here's apartment capacity, here's capacity to make cars. Like it just like sort of blows my mind that that's like a that that could ever be boiled down to a single number.

Let me ask you a random question. Is there a sector or a part of the economy that in your research surprised you as having more ripple effects across other prices than you might have expected that maybe people I mean oils obvious, right, we all know that everything needs energy or whatever, but other sectors that maybe people don't

think of that have outside effects. Generally speaking, I would say that the results are not terribly surprising, which might make you say like, Yeah, then why I bothered modeling, to which I would answer, well, it is nice to kind of be able to capture these aggregate effects and trace them in a systematic way throughout the economy. One of the sectors that I think is quite interesting is chemical products, like, which is just in everything, right, It's

like as ubiquitous almost as UM as as fossive feuds. UM, and UM is incredibly important and apparently it's also important not only like from a quantity perspective of composition of production, but also from a from a price perspective. So this

is one that I personally hadn't hadn't thought about as much. UM. I think holds trade kind of came very much to UM to the top of our minds in the pandemic, But our simulations for before the pandemic UM also show that holes trade was already pretty important, which again is something that I think, like from the pre COVID mindset, would not have been something that I would necessarily have associated with thinking about inflation. Sorry, really, what do you what?

Hole shield trade? Man? You said, what do what specifically you're talking about? Yeah, so holds the trade again, Like this is actually one of the points where probably the level of aggregation can become a problem. I mean generally speaking, it's stuff like logistics, but also whole say traders, I mean any kind of company that that provide that that basically does Whoway say merchandising, right, which yeah, so Isabelle it can I ask one thing you mentioned in your paper.

You talk about the possibility of minimum inventory requirements. So if you know that a specific industry or thing is important from an inflation perspective, maybe we should build in additional inventory, some resilience into the system. And this is you know, inventories. Um. The idea of business is moving to just in time and maybe being a little bit more vulnerable to big shocks and demand. This is almost

classic odd thoughts territory. And it seems like the difficulty there is how do you encourage companies to build up that extra capacity in their system when maybe their incentives are more skewed towards you know, just making money and profits and short term things. How do you actually go about doing that? How realistic is it? Yeah? Absolutely great question, um?

And I think that um. I mean, if it is about making money and some of the companies that have experienced bottomnecks actually have experience that they have managed to increase their prices and ways that rendered them even more profitive with them before the pandemic. Right, so then your incentive of of increasing your inventory it might actually be pretty low because you think, like in normal times, UM, I don't wanna have inventories because I want to be

as efficient as I can be. And then if shocks hit, if everybody is kind of um running this same model, like all competitors in one in one segment running this same model, so then they all don't have a lot of inventories, then there's this exactor white UM supply change shock which all allows them to high prices and base in which they could not hide prices at normal times because now they kind of had this mutual knowledge of

UM of of of shortage UM. So and then they end up being actually in a pretty good position, which we have seen in some of the sectors that have experienced extreme UH impacts on their supply change during the pandemic. Right, So, therefore, we somehow need a way to get out of this. And I think because of what I just laid out, it's not clear that companies by themselves would necessarily increase

the inventories space. Um that that that sufficiently prevent this situation, at least not to the extent as it would be like kind of um socially desirable or desirable from a more like macroeconomic kind of standpoint. How to do it practically? Again, Like, I really see this paper as providing a framework and kind of starting a conversation, and I think to get to the question of how to do it practical, one

really has to start. I'm talking to people who understand inventory management and companies rather than me like kind of as the arm to economists coming up with some some sort of fix all the inventories of US corporations and

one type of approach. We're almost out of time. I have one very short question, but you know, we did an episode recently and it was pointed out by one of our guests of the way a lot of economists think is that if the price of gas goes down, for example, that doesn't improve inflation because the sort of general equilibrium, Well, that's more money in people's podcasts, and they're just going to spend more on haircuts now, or

they're just gonna spend more on cars. Why is a limit to how many haircuts, yeah, right, or maybe they'll spend more on going out to eat, okay, so uh and then it doesn't really get us anywhere. Like what do you say? I'm just curious your response to that. That's like, okay, you target a sector, great, you target energy, great, But then everything is cheaper, people have more money and they spend elsewhere, and you don't get anywhere. Why should

that not? Is that not a fatal flaw of your approach? I mean this is the famous I mean one of the famous fretument frequent quotes also where he's saying exactly

what you just said. We were back to monitor, Yeah, yeah, I mean that's the kind of brings you back to the question of how firms are setting prices, right, um, And if we are in a situation where we have very highly concentrated um corporate structures, which I think is a fairly fair descriptions of large parts of the American economy, then we can actually see that that the demand response, sorry, that the price response to demand is surprisingly small. In

many cases, the prices are actually quite surprisingly stable. I mean, if you think back to the two decades decades before covid Um, where of course there have been periods of more demand and less demandents on, but prices were surprisingly stable, right, and everybody was a kind of surprising. Why are price

is so stable? Well because in a very concentrated sectors, firms tend to UM to compete over market share, UM, compete over conquering new segments of markets, UM, compete over cutting costs, and so on, less than UM than using any kind of small increase in demand by immediately raising prices. Right, Because if you raise prices in your competitor doesn't raise prices because both of you are price makers and not

price takers, then UM, that can actually harm you. So therefore UM in in in the kind of institution setting that we find ourselves in, UM, I don't think it's clear that UM, if people spent less on gas, then immediately the prices of everything else that they are UM assuming are going up. And I also don't think that this is something that we see empirically, that the price of gas andoy going down, the inflation in other parts

of the economy suddenly like going up by any large margins. Isabella, We're going to have to leave it there, but thank you so much for coming on. Odd lots You're definitely one of our favorites. And I'm not just saying that because you've translated the past two years into actual academic research. UM.

Fascinated discussion. Thank you, Thanks Isabella. That conversation was great, UM, and the paper is definitely worth a read, although I know Isabella said she didn't think anyone was actually going

to read it. One thing I was thinking is it does kind of go back to remember some of the conversations we had with Stephanie Kelton on modern monetary theory, and you know, her solution was, well, we need instead of reducing spending, like maybe we identify where the bottlenecks are happening and we increased capacity or try to increase capacity. And my criticism of that was it's difficult to do

it in real time. But I think studies like this maybe go some way towards identifying where to look, right. I think this use of input output tables and Isabella said it is actually a very old thing that you never hear mainstream economists talk about. I don't you know, it could be is like very useful idea and like, okay, it's like difficult. Sure, it's a lot more difficult to identify critical sectors, and it is to just raise rates when CPI comes in higher than expected. But it's doable.

And I did read the paper. But when I say I read the paper, what I mean is I read the first four pages, skipped over the forty pages of equations. You have to see the intro and then the conclusions. So I read the I read the intro, I skipped over like all the equations and Greek symbols or whatever, and then the conclusion, and I thought it was really interesting.

And I think, like I suspect and maybe as a result of all this, the pandemic and everything, there might be renewed interest in this sort of like pretty vigorous approach to identifying critical sectors and how they distribute prices across the ecomptment exactly this. I would be really disappointed if we came out of the past two or three years without a sort of like new way of thinking about inflation, or at least maybe an additional dimension. Yeah,

all right, shall we leave it there. Let's leave it there. This has been another episode of the ad Thoughts podcast. I'm Tracy Alloway. You can follow me on Twitter at Tracy Alloway and I'm Joe, why isn't Though? You can follow me on Twitter at the Stalwart. Follow our guest Isabella Vaber. She's at Isabella and Vaber, and check out her paper, Inflation in times of overlapping emergencies systemically significant prices from an input output perspective. Follow our producers Carmen

Rodriguez at Carmen Armand and Dash Bennett at Dashbot. And check out all of our podcasts as Bloomberg under the handle at podcasts and for more odd Lots content, go to Bloomberg dot com slash odd Lots, where we post transcripts. Tracy and I blog right away. Weekly newsletter every Friday. Go there, subscribe to it, get it in your inboxes. Thanks for listening.