Alp Simsek on a Risk-Centric View of Demand, Recession, and Speculation

Alp Simsek is an associate professor of economics at MIT, and joins Macro Musings to talk about the link between financial markets, uncertainty and the COVID-19 crisis. Specifically, David and Alp discuss the dual absorption problem within financial markets, how supply shocks and demand shocks have inescapably become interwoven phenomenon, and why we should look to using macroprudential policy in the future.

Read the full episode transcript:

Note: While transcripts are lightly edited, they are not rigorously proofed for accuracy. If you notice an error, please reach out to [email protected].

David Beckworth: Alp, welcome to the show.

Alp Simsek: Thanks David. I'm glad to join. I really enjoy your podcast and it's great to participate.

Beckworth: Well, thank you for listening, and thank you so much for coming on. You and Ricardo Caballero, a past guest of the podcast, have been doing some really interesting work that does a nice job of explaining why a pure supply shock like COVID-19 can actually lead to a large financial shock, and ultimately an even larger demand shock. And that's what we're going to get into today with the models and the papers you've been working on.

Beckworth: So, I think it's great work you're doing. Really helps shed light on some of these questions. Because I remember back in March, people were talking about, "Oh, this is a supply shock. The Fed should do nothing. Just step aside." But your paper really fleshes out why that view is incomplete. But your work has also been interesting to follow, Alp, because it provides some nice pushback to some of the critiques we've been discussing here on the show. And I'm guilty of them myself, so your work is going to keep my humble. Let's put it that way.

Beckworth: And that's the critiques of the Fed's approach of really just stabilizing risky asset prices, all that Fed's done is prop up the stock market, it's helped junk bonds, it's maybe propping up home prices through it's mortgage-backed security purchase, but you and Ricardo show in a very convincing manner that that's actually important. If you want to help Main Street, you actually have to think about what's happening to the price of risky assets.

Beckworth: So, I think you make some important contributions in the discussion surrounding this crisis. I'm glad to have you on the show today to talk about it. Before we get into all those things in more detail though, tell the listeners a little bit about yourself. How did you get into this area? And what motivated you to do it?

Simsek: Sure. So, I got into this area in grad school, because when I was in grad school, that is when the financial crisis of 2008 hit. I was actually a third-year graduate student at the time, just done with my courses, thinking about a thesis and research agenda, and suddenly we had a huge crisis and lots of open questions, and I got naturally drawn into it. And one thing I've noticed was that heterogeneity of beliefs were very central to understand the last crisis.

Simsek: In particular, it seemed to me that the housing boom was amplified by the heterogeneity of beliefs. Housing optimists were driving up in the boom phase house prices. In fact, they got their hands on cheap credit and they kept driving up house prices. But then the boom turned into bust and they got really hurt, and we got a huge collapse of housing bubble and kind of the fallout in the macro-economy, I started to think about these issues.

Simsek: And gradually, in terms of the macroeconomic fallout from this, my views have shifted toward demand side issues. I wasn't so much sort of focused on demand side issues early on, but the recession kind of convinced me that the demand side was important to understand why financial markets matter, and such a collapse in asset prices affect the economy.

Simsek: So, what we've been doing with Ricardo Caballero, is sort of build a framework that helps to think through these interlinkages between financial markets, things like house price, or price boom-bust cycles in financial markets, and the macro economy. And maybe I should sort of give a little bit of an overview of our framework before we get into the specific research that we have been doing around that.

Beckworth: Sure.

The Dual Absorption Problem

Simsek: So, the way we think about this framework is it's a dual absorption. It's interaction between two absorption problems. So, the first absorption problem concerns the goods market. The economy produces goods and services determined by our capacity of production, very important supply side issues, of course. But given that, there's another absorption problem, which is how do you generate enough spending by households, firms, and governments to ensure that you do operate at this capacity? A lot of kind of demand side macro is about this, right? This absorption problem.

Beckworth: Yep.

Simsek: But there is a second similar absorption problem in risk markets as well, because the economy also, economic activity generates risks. In fact, it's hard to imagine any activity that doesn't involve risks, and these risks are embedded in financial assets or financial markets, and there's an issue of how to generate enough demand for risk from investors, such as banks, institutional investors, or even households, to absorb this risk.

Simsek: And also finance asset pricing literature has been kind of focused on this absorption problem. So what we do or what we point out with Ricardo is that these two problems, although they are kind of analyzed separately, typically by separate tradition, they're actually naturally related. Because the level of asset prices affect demand in both markets, so they naturally provide a bridge by which issues in one market can travel to the other market and amplify the original problem.

Simsek: So, in particular, I think it's easy to understand why asset prices matter for risk markets. They are the prices that clear the risk markets. But they also matter for the goods market through various channels, because if, say house prices collapse, that will tend to lower spending. Just from sort of neoclassical wealth effects, right?

Beckworth: Yep.

Simsek: But also because house is a good collateral, they will tighten borrowing constraints, and force household into deleveraging. Or kind of when bond markets collapse, it will affect spending, because spreads now go up and firms will struggle to invest, so investment will go down.

Simsek: So, there's actually many, some neoclassical, some more frictional reasons for asset prices to affect demand in the goods market. So, once you make that observation you see how these two markets are very tightly linked, and how issues can go from one to the other. So, that's kind of the big picture framework that we developed and we use for a number of issues.

Beckworth: Yeah, and if I'm understanding you correctly, what you're saying is you're bringing together two literatures. And part of that is motivated by what happened in 2008, and others have been doing it as well, but in macro, kind of the workhouse New Keynesian model kind of ignores the latter channel, the risk channel, the finance channel that you just described, and we focus more on...

Beckworth: If we look at asset prices at all, we look at kind of a standard wealth effect, where it affects consumption. But you're bringing in this risk balance channel, and you, I think, outlined the core model in your paper titled *A Risk-centric Model of Demand Recessions and Speculation.* Is that where someone should start?

Simsek: Yeah, exactly. So, that paper actually is kind of where this agenda begins. We developed the framework, and like you said, it's actually a New Keynesian model, but the big difference is that we actually have aggregate risk in the model, so there's a risk market, whereas a lot of recent New Keynesian literature kind of abstracts away from this.

Simsek: They focus on a lot of other important issues, but they don't have risk markets, and therefore they don't necessarily have the effects of risk repricing on demand. So that's the thing that we bring to the table.  So, we built that framework, that bridge, but also we use it in that paper to address kind of an important issue that has been observed by the finance literature, but again, kind of not so much by the macro literature, that the risk markets feature a lot of heterogeneity, right?

Beckworth: Yep.

Simsek: So, the valuation or demand for risk by an institutional investor or by a bank is actually very different than by households. So, these institutional investors, they understand risk, they know how to price it, whereas as opposed to I have no idea how many risk markets work, right? So, a lot of these risk markets are highly complex.

Simsek: So, actually, there is very substantial heterogeneity in risk markets. The institutional differences is one source but also belief differences is another one, right? The evaluation of house price risks by a housing optimist is very different than by a housing pessimists. So for a variety of reasons, there's a lot of heterogeneity or what we call heterogeneous valuations in risk markets.

Simsek: And this naturally matters for finance, for asset prices, for risk pricing. But because of our bridge, it actually matters for the macro economy as well. And that's what we kind of analyze in this paper. In particular, we find that these belief differences make the wealth share held by the high valuation investor very important state variable for the macro economy.

Simsek: So what fraction of the wealth is controlled by relatively high valuation investors compared to the lower valuation investor? And why does this matter? Well, because when more wealth is in the hands of the high valuation investors, be it optimist or more risk tolerant agents, it is as if the economy has high valuations in general.

Simsek: So the distribution of valuations matter if you think heterogeneity is a big deal. And when there is more wealth in the hands of high valuation investors, asset prices tend to be higher. There is a natural analog of this in the recent macro literature by the way. The recent literature has emphasized MPC heterogeneity, right? So it's focusing on the goods market.

Simsek: If you are able to get more wealth in the hands of the high spenders, people who have high MPC, that will increase spending, right? So it's similar. If you're able to get more wealth into the hands of risk tolerant agents like institutional investors or housing optimists, they will drive up asset prices, and this will increase demand because high asset prices increase demand, right?

Simsek: So it's not because these guys spend more, but they raise asset prices and then they increase the demand in the economy. So this kind of makes these guys’ wealth an important state variable. And so one thing that we realized is that the state variable is highly endogenous and this means that what happens in the boom naturally matters for what happens in the recession or kind of how bad a recession we might have.

Simsek: In particular, what happens in booms, for instance let's focus on belief differences for the optimists and pessimists. What happens in the boom is that optimists naturally absorb more of the risks in the economy, right? So because they think recessions are less likely because they think kind of house prices are more likely to appreciate, they basically load on to risk.

Simsek: And in practice they usually load onto risk through leverage. So that will drive up the boom. As the boom continues these guys will get vindicated. In fact, their wealth will keep growing and that will push up asset prices. But when you now get a repricing in the housing market, let's say you realize that house prices are not as valuable as we thought.

Simsek: Now these guys will get disproportionately hit, because they were the ones loading up onto the risk in good times, so when the bad times come, their wealth share now will collapse disproportionately, and this means that now the swings on the downside will be much bigger because now a lot of the wealth is in the hands of the pessimists. So they will be the ones that are basically pricing risky assets like houses.

Simsek: So that's actually going to generate a downward swing the other direction. So this makes the asset boom bust cycle much bigger than it would otherwise be. And once you have the bridge to the real economy, it also makes the business cycle much bigger than it would otherwise be, because aggregate demand fluctuates a lot more. In good times, it sort of goes up more and in bad times it falls down more.

Simsek: And the same thing by the way applies when you focus on risk tolerance heterogeneity. So in the paper we focused on belief differences, but we also discuss how things will be very similar if you, in fact instead of optimists loading up onto risk, suppose you have the view that some agents are more risk tolerant than others. Then again, you have the same thing. Risk tolerant agents will load up onto risk because they're naturally providing insurance to the other agents.

Simsek: And again, they do it through leverage typically. When the shock comes, they're going to get disproportionately hurt and this will make the pricing swings much bigger and aggregate demand swings much bigger. So we think this kind of framework and the heterogeneity that comes with it offers a natural interpretation, or sort of a big picture but quite natural interpretation, of the last recession; but also is relevant for other recessions as well where financial market boom bust cycles are a key part of the story.

We think this kind of framework and the heterogeneity that comes with it offers a natural interpretation, or sort of a big picture but quite natural interpretation, of the last recession; but also is relevant for other recessions as well where financial market boom bust cycles are a key part of the story.

Beckworth: This is very interesting and maybe if I can summarize it, you're really bringing out in this model here the importance of beliefs and heterogeneity of beliefs with people of different beliefs. There's the optimist, there's the pessimist and this really in my mind harkens back to kind of Keynesian animal spirits. Is that fair, the importance that that role plays?

Simsek: Yeah. I mean I think Keynes recognized, I think, the importance of heterogeneity and speculation in financial markets. I mean I think there are sort of two issues and they kind of reinforce one another. One is the general level of optimism and pessimism, right? So I think in the housing boom bust cycle, there was certainly a general optimism in the boom phase that's shared by a lot of the investors.

Simsek: But on top of that, there was a heterogeneity. Actually, by the way, the recent book by Michael Lewis, *The Big Short* is all about this heterogeneity, right? How there were also pessimists in the housing market…

Beckworth: Sure.

Simsek: So once you bring this heterogeneity, what that does is it sort of amplifies the animal spirits, because now optimists exert a much bigger influence. So that, essentially, will make the boom even bigger than it would be otherwise. But when the bust comes now the pessimists become dominant and they make the bust also bigger than what would be otherwise.

What that does is it sort of amplifies the animal spirits, because now optimists exert a much bigger influence. So that, essentially, will make the boom even bigger than it would be otherwise. But when the bust comes now the pessimists become dominant and they make the bust also bigger than what would be otherwise.

Beckworth: Yeah. This is great because this also seems very consistent with what Roger Farmer has done in his work. He's talked about having a belief function and I know we'll talk about your policy implications in a minute in terms of macroprudential policy, but I think this is one of the reasons that Roger Farmer calls for targeting like an S&P 500 index fund, trying to stabilize risky assets so that you maintain a sizable share of optimists I guess in the economy. Do you see that relationship?

Simsek: Yeah, that's an interesting connection. I mean that is absolutely related because we do some policy analysis around this issue as well. I think the most direct way of targeting this issue is by macroprudential policy.

Beckworth: Sure.

Simsek: So essentially if you think the issue is optimist loading too much on risk, a very natural thing to do is then you restrict their risk taking, and you directly deal with the problem and that helps. And we show that. But importantly let me point out here that you don't do it because you think speculation is bad, that will be an added reason, but you do it for the macro economy.

Simsek: We showed that actually even if you sort of respect people's different beliefs and think gambling is actually a... you want to allow for it... you still want to interfere because you're correcting for the macroeconomic damage that these guys generate in the process of this gambling.

Simsek: But there's actually more recent work that we have done on this which sort of asks the following question. Suppose, okay, macroprudential policies, you want to do that. But suppose there is a limit to how much you can do that. And in practice there are actually potentially quite large limits to macroprudential policies because a lot of the institutions that are doing this loading on the risks fall outside of the regulatory sphere. Then there are other things you can do.

Simsek: I mean, one thing you can do is kind of try to keep asset prices in the boom phase relatively low by keeping the interest rates relatively high, even higher than what you would need for demand stabilization. So that kind of helps because that mitigates the crash. So when the crash comes, you don't fall as much because you kind of start on a lower level to begin.

Simsek: So the idea is if the Fed were to raise interest rates in the early 2000s, the house prices wouldn't boom as much, and the spreads wouldn't... the bond yields wouldn't go down as much. And this actually is a good thing because when the crash comes, then prices don't collapse as much. But, I think here is where it connects with the Farmer idea.

Simsek: So the broader idea in that paper what we call *Prudential Monetary Policy* is you try to keep the asset price swings low in ways that you can, right? So one way of doing it is in the boom, you keep the price low, but another way of doing it is in the recession, you try really hard to prop up asset prices because you again want to mitigate these swings, which damage the bank balance sheets or optimist balance sheets and hurt you, right?

Simsek: So if you are able to prop up asset prices in the recession, that's going to be good. Of course. I think keeping asset prices completely stable is extreme. We don't want to do that because there's going to be…

Beckworth: Right.

Simsek: You will let them go down a little bit, but you want to kind of-

Beckworth: Some price discovery.

Simsek: ... make them go down less than they would go in absence of these interventions if that makes sense.

Beckworth: Yeah. Let me just go back, maybe summarize some of these insights from your paper. So we want optimists, I guess at a certain fraction. You don't want to go completely pessimistic in your world, right? You want to have some optimists there to keep asset prices stabilized and keep them from falling. But then again, it goes to this critique I mentioned earlier that some folks may have, "Well, you're just propping up the rich people. You're just propping up Wall Street. You're just propping up corporations."

Beckworth: But I think the point you're making is if these asset prices are prevented from crashing, it's not just about people who hold them spend more, but they create asset valuations that cause others to spend more. The broader economy to respond. So by doing this, you're not just reaching to a certain group. You're actually reaching all the way down the Main Street. And one of the critiques of Fed policy now due to legal constraints, due to expertise constraints, so they're having a really hard time reaching Main Street.

Beckworth: They're trying this Main Street facility. There's limits to what they can do. They can only lend, they can't provide grants. And in practice what the Fed can do is reach bigger businesses, corporations. It can help stock prices indirectly. It can do all kinds of things that may not appear fair. But I think the picture you're painting is that even if it does that, it can have a positive effect indirectly on Main Street. Is that fair?

Simsek: That's exactly right David. So there are two things I want to say about that. First, yes. So this will somewhat benefit the participants in risk markets a bit more when the Fed interferes to prop up asset prices. But this will indirectly benefit also a lot of other agents in the economy. Even people, by the way, who don't participate at all because you're preventing a big demand recession. And as that happens the incomes and wealth for everyone will go up.

This will indirectly benefit also a lot of other agents in the economy. Even people, by the way, who don't participate at all because you're preventing a big demand recession. And as that happens the incomes and wealth for everyone will go up.

Simsek: Incomes, for instance, of even the hand to mouth people who don't hold any assets at all will benefit from this because as asset prices are higher, spending is higher, firms will employ more people and the incomes of these workers will go up as well. So there are actually sort of spill over benefits to a lot of the society. And that's kind of why the Fed is actually doing these things.

Simsek: But there's another thing I want to point out here. Once you are in a demand or recession situation, especially when the interest rate is constrained like we have right now, actually the prices decline more than they would in a benchmark where interest rates would be unconstrained or all outcomes would be determined by productivity and supply, right?

Simsek: So actually what the Fed is doing is not enriching these people so much as it's kind of mitigating the undue damage on these people that results from the constraint of usual monetary policy if that makes sense.

Beckworth: No. That makes complete sense to me. Another way of saying that is we got to do the right counterfactual, right? I know it's hard sometimes to think through that when you see stock prices going up, but maybe someone's home is being lost, but you're saying, "Do the right counterfactual."

Beckworth: And the right counterfactual is we're in a worse place than we should be, or maybe alternatively, if the Fed didn't intervene we would be even in a far worse situation. So we've got to think through the counterfactual and your model allows us to see that.

Simsek: Exactly. That's right. That's exactly right. That’s the main benefit of this line of models that we are doing. They are somewhat stylized but they help you really, I think, think through these issues with more discipline

Beckworth: Yeah, absolutely. I think that's just an issue of communicating this to the broader public. What if the Fed weren't intervening, where would we be today? And let me take that and run with that back to the point I made earlier. So in early March when the Fed intervened and now we've looked back, there's many stories written up. Like yesterday there was a great Wall Street journal piece on how the financial system almost froze up in mid-March if it weren't for the Fed jumping in.

Beckworth: I mean, people weren't buying and selling treasuries, which supposedly is the safest asset on the planet. That market was having problems too. So some of the discussion during that time and I was a part of this conversation myself was, "Well is this a supply shock?" Well, on one level the answer is yes, clearly it is. It's COVID-19. We're shutting down the productive capacity of the global economy to a certain degree.

David Beckworth: And so many people said, "Well that's a real business cycle shock. We just ignore it. There's nothing we can do about it. This is not a place for monetary policy to step in." And then some of us were arguing, "Well, no. Actually this could spill over into something bigger. A bigger demand shock." And I think that's a key point of your paper. Maybe I'm getting ahead of myself because that's in your second paper here.

Beckworth: But that is one takeaway is I think from this paper is it's almost impossible in the modern world to have a pure supply shock that doesn't spill over into the financial system and in turn into a demand shock. Is that fair?

Where There's a Supply Shock, There's a Demand Shock 

Simsek: That's exactly right David. I mean, this is also kind of what I've... That was our thinking. We thought initially it was a supply shock, but then it became very clear very quickly how difficult it is to get a pure supply shock. And then we realized that actually our framework offers at least one channel how this turns into a demand shock and that's kind of why we wrote this paper. It is actually a very natural continuation of the earlier paper.

Simsek: So the earlier paper was about how things that hit the financial markets, like financial crises or boom bust cycles, can travel back to the goods market and create potentially recessions. This paper actually goes from the goods market, it starts there. And so we sort of focus on shocks to the goods market and we make it a supply shock to make it very dramatic. So a reduction in our productive capacity. And that's one way of thinking about the COVID shock.

Simsek: I think there are actually elements of exogenous demand shocks as well. This wasn't a pure supply shock. But just to set the deck against us, we make it a pure supply shock in the goods market. And we find that this will naturally, again, through the bridge, travel to the risk markets, create a big problem in risk markets, a big repricing of risk, and then travel back to the goods market and turn the supply shock into a demand recession.

Simsek: Essentially here's how that works. So when the productive capacity of the economy declines, naturally asset prices decline as well just to the supply side, right? So when the economy produces less, asset prices will decline. In fact, if it didn't decline, you might have too much demand relative to capacity. So they will go... There will be downward pressure on asset prices.

Simsek: But the issue is when asset prices decline, this hurts the risk tolerant agents that I've been talking about a lot more because they were leveraged to begin with. They were more exposed to risk to begin with. And so when there's a big and surprising shock like this, their wealth declines a lot more than the risk intolerant. And what we saw in this shock is that the hedge funds got massive losses, and also institutional investors.

Simsek: And when they make big losses like this, they naturally start deleveraging, right? Because they sort of start selling risky assets, now these assets need to be held by the less risk tolerant agents, but they're going to require a higher risk premium to hold them. So essentially another way to think about what I said earlier is that you're getting a wealth shock to risk tolerant agents and this through the heterogeneity will lower the risk tolerance of the economy. So the risk premium is going to go up.

Simsek: So when the risk premium goes up, not only you get the initial supply-driven decline in asset prices, but you get an overshoot in asset prices because now we are less able to absorb risk in financial markets. So asset prices will overshoot... Now because asset prices are naturally connected to demand, you get basically also an overshoot in demand in the goods market. So this is kind of how the financial markets turn a large supply shock into a demand recession.

Beckworth: Yeah. I want to talk-

Simsek: …And then there are policy implications that come with it. 

Beckworth: And just to be clear listeners, this paper we're talking about now is titled *A Model of Asset Price Spirals and Aggregate Demand Amplification of a COVID-19 Shock.* So we'll have this and these other papers we've talked about on the show page. Now Alp, one of the stories you tell in there, the mechanisms, I think it's one of the key mechanisms and it's...

Beckworth: I liked it because you kind of told the story through a balance sheet, is that when the wealth declines in the story, it endogenously affects the balance sheets of these entities. So you said leverage rises endogenously.

Beckworth: And so the story in my mind, what I'm thinking is that you've impaired the balance. Assets have fallen relative to liability. So by definition, leverage has to go up and you're in a worst state. Your balance sheet looks worse, so you have to quickly adjust and you offload risky assets. But in doing that you're creating that demand shock.

Simsek: Exactly. That's right.

Beckworth: Okay. Now… Go ahead.

Simsek: Either they will require a very high risk premium to keep that balance sheet… you can convince them to keep very high leverage, but you need to give them a big risk premium for them to do that. Or they will, more realistically, they will just unload risky assets and now risk needs to be held by much less risk tolerant agents. Again, they will require a higher risk premium.

Beckworth: And I think in addition to natural precautionary motives and firms acting in their own interest and wealthy individuals doing the same, there's also a lot of regulations now that also amplify this.

Beckworth: So I know in the financial sector, for example, in the story of March some of these big financial intermediaries were not stepping in to the Treasury market because their balance sheets had actually gotten bigger because of some of the developments. And they literally were handcuffed by regulations from doing any kind of private sector bail-in into the market challenges during that time.

Beckworth: So I think this is a very important story that definitely can be amplified by regulations in place. What your paper has is interesting in this regard as well as heterogeneous asset valuations. And I just wanted to note in passing that this is kind of another interesting innovation to these macro models. So previously on the show, we had Ben Moll and we talked about HANK models and those are a Heterogeneous Agent, New Keynesian models, but they focus more, as you mentioned earlier, on different marginal propensities to consume where you're looking at differences in beliefs, is that right?

Simsek: That's right. Differences in beliefs or risk tolerance, or regulatory differences, like you said, create different valuations of risky assets, heterogeneity.

Beckworth: Absolutely. Okay. So there's a lot of interesting developments happening in macro models. It's hard to keep track of all of them. And heterogeneity though seems to be the name of the game [during the] past few years. Let's move back to the policy implications though. You mentioned macroprudential policy.

Beckworth: So this could be a case where monetary policy can step in. So there's regular macroprudential regulation, but you said it may not always be available or maybe it's not tapped, maybe it's hard to do so monetary policy could effectively do it on their behalf. So explain to us again how that would work.

Macroprudential Policy Implications

Simsek: Okay. I think there are two classes of policies to think about ex-ante and ex-post. The most recent paper is a lot about the ex-post interventions, like asset purchases and so on and we can come back to that. We have an earlier paper, which is about ex-ante interventions. The question that we asked in that paper is, suppose you want to do macroprudential regulation for the reasons that we've discussed. You want to restrict risk taking by the optimists or by the high valuation investors more broadly to protect the economy when the downswing comes.

Simsek: But suppose it's limited. You cannot do too much of it because some of the institutions are outside your regulatory sphere, or perhaps because this sort of the asset price boom developed very rapidly and macroprudential policy requires legislation and you won't be able to tighten leverage limits very quickly or capital requirements very quickly.

Simsek: So the question is, is there any reason to use monetary policy to raise the interest rate for prudential purposes to discipline risk taking and to mitigate the damage when the recession comes, and we find there is a reason to do that. I mean, I think there are many reasons why you may think raising interest rates can help to discipline the financial excesses.

The question is, is there any reason to use monetary policy to raise the interest rate for prudential purposes to discipline risk taking and to mitigate the damage when the recession comes, and we find there is a reason to do that.

Simsek: But one thing we emphasize, one thing that it does, is it naturally lowers... When you raise the interest rate in boom, more than what you would do in a regular macroframework, you naturally lower asset prices. And if you think the issue was that asset prices are too high and they can turn low, this helps because if you start from a lower level to begin with, when the crash comes, it becomes less damaging.

Simsek: So the way I like to think about it, I think policymakers actually have a good intuition for this. Because when asset prices are too high, they know that.. Well policymakers can not really time bubbles, but they don't need to time bubbles, right? They understand that there's a tail risk. When asset price are high, historically, they tend to fall.

Simsek: And when they fall they tend to go up. Of course, there's a lot of uncertainty, but the risk premium is very time varying, so policymakers understand that when asset prices are high, there's a risk of them collapsing. So they get a little dizzy, they worry about, "What would happen if house prices actually turn out to be a bubble and they collapse?"

Simsek: So the natural reaction that they think about, and this actually comes up in FOMC meeting minutes and so on or in discussions among policymakers, they think about raising interest rates to kind of lower asset prices. If there's a crash, there is not such a big crash. And this is the mechanism that we capture in our recent paper titled *Prudential Monetary Policy.*

Simsek: And we kind of formalize this, but we also illustrate the limitations. One limitation is of course very natural that this slows the economy in good times. You need to kind of trigger a little bit of a slowdown in the boom. So this is not free. But we find that there are some kind of conditions under which, despite this sort of cost, you may want to do a little bit of that. Not too much, but a little bit of that.

Simsek: Because if you create a little bit of a slowdown, that's not too bad, right? Going from 4% unemployment in boom to four and a half percent is not so bad because at the margin, people are kind of indifferent to work if you think it's a well-functioning economy. But once you sort of cool down asset markets, when the crash comes, you might potentially get big benefits from having done this policy in good times.

Beckworth: So is the overarching goal here in using macroprudential policy, and as it's tied back to your model, is to avoid wide swings in asset prices. So you don't want to have crashes, but you don't want to have big booms that kind of sets you up for a crash. Is that the goal?

Simsek: Exactly. Especially if you think monetary policy, when the crash comes, is going to be constrained. Like we've discussed that will make the crash even bigger because you are unable to use regular monetary policy to counter. So that makes the crash kind of bigger. So that motivates other things to do to mitigate these excessive crashes.

Beckworth: And that's why you have the name of your paper, *Prudential Monetary Policy.* So you're preventing the problem from emerging in the first place. This kind of reminds me of John Taylor's argument from 2007, 2008. I know he's making a different point because he uses a Taylor rule to say, "Hey, the Fed should have tightened sooner before and during the housing boom period." But I think your argument is least consistent with that. Have you thought about that?

Simsek: Yeah. I mean, exactly. I think one issue is that in the context of the run up to the 2000 bubble, John Taylor's argument, if I am correct, is that the policy was too loose to begin with, so… they weren't at R-star, they were actually lower than R-star, and that we can debate.

Simsek: What we are saying is that even if, the policy was at R-star, there's actually a reason to go beyond R-star on the demand side. But it is consistent with the Taylor view. Of course, if on top of that you make a mistake and swing in the other direction that will make things even worse, right?

Beckworth: Right. Right. You can tighten too much.

Beckworth: So is it possible to come up with a modified R-star or an R-star-star, some kind of a special neutral rate that accounts for this added feature in your model?

Simsek: That's a great question, David. I mean, I wish there was a simple rule like that. We are actually thinking about that because I think there's some benefits to having rules, and so on. Unfortunately, the issues here are quite subtle and they kind of depend on the probability of the tail, how likely it is the crash-

Beckworth: I see.

Simsek: ... and kind of whether you are... It's quite complex. And I think it's likely to be situation dependent. The way I think about this policy is that you don't want to use it a lot. But the way I think about this is that in practice in monetary policy discussions, usually no one really knows what R-star is, right? It’s usually like an interval where we think it might be a little higher or lower.

Simsek: So I think what this paper does is that it offers you some perspective and says that you might want to err on the side of caution of setting higher interest rates than what you would otherwise do. That's kind of... If I were to use this in practice, that's how I would use. But we are thinking about whether this could be turned into a rule, but it's not there yet.

Beckworth: Yeah. I can understand that you would probably involve some additional latent variables you don't directly observe. You can throw into that rule, make it more complicated.

Simsek: There are actually things in financial markets that you can observe because risk premium kind of…

Beckworth: That's a fair point.

Simsek: When the risk premium is low, you can see the spreads are then lower than their historical levels, price to earnings or price to rent ratios are then too high. Finance literature has actually ways of calculating the risk premium and comparing it with a benchmark. So that gives you some information, but going from there to a rule is kind of a challenge, but that's the kind of data that we want to know.

Beckworth: Yeah. This also has a little flavor of the BIS work in it. They've been very vigilant about watching credit growth. They have a series where they track credit to GDP and then look at it relative to an HP trend. So they've been on this too, but it would be interesting to see a rule that would emerge from this that would make it easy for a policymaker to think through quickly in their head when they're sitting around the table at the Board of Governors, "What should we do?" And they pull up the ALP rule and think through it.

Simsek: Exactly. I agree. It will be. And then credit definitely, I mean, amplifies this because the more leverage you have, the more fragility is in the system and then the shock becomes more damaging. So that's definitely one of the parameters that would go into the rule.

Beckworth: Okay. So we have talked about using macroprudential policy and you've listed some reasons why it doesn't often work in practice. In the US it hasn't really been applied very hard. I mean, in other countries it has been applied more so. And part of it is political, part of it is difficult to implement. I mean, part of it is a knowledge problem. What do you look at? Which indicators are the best ones?

Beckworth: So if we can't use it, you suggest using monetary policy. Let me go back, though, to the other suggestion I mentioned earlier, I mean, I think Roger Farmers' suggestion is you would target a stock market index. Now I'm not sure how you would actually do it in practice, but I could imagine a world where he has some kind of Taylor Rule where they're targeting the stock market, some kind of, I don't know, growth path where you would both tighten it and you would also ease it based on conditions. I mean, do you think you get something like that to work? Or would that be too difficult to implement?

Monetary Policy in Place of Macroprudential Policy

Simsek: Yeah, I mean, I don't know if I would look at the stock market per se, because in our paper, we have the key price, what we call the market portfolio, is essentially all the assets in the economy. Stocks, bonds, house prices and so on. So what you want, I think is target the broad financial conditions index.

Simsek: That aggregates, I think a lot of these different prices. Because essentially it could be that... I mean, I think you don't want to stabilize stock prices because there are relative pricing effects. It could be that stocks are doing poorly while houses are doing well. As long as you keep the aggregate price levels stable, that's what matters for aggregate demand. It's a combination of all these different asset classes. There's going to be a lot of relative price movements that you don't want to stabilize, but you want to target a broad level of assets prices.

Beckworth: Right. A broad asset index that includes a lot, which in itself is a whole research program developing something like it. I imagine.

Simsek: I think the Fed actually looks at it. I mean, they…

Beckworth: Oh, they do?

Simsek: ... pretty cool. Yeah. They have a kind of financial conditions index, which…

Beckworth: I see.

Simsek: ... aggregates lots of different measures. I don't know if that's the optimal one, but that's kind of going in the right direction. Sort of like the principle components of lots of prices.

Beckworth: Yes. So in their financial stability report they just put out, they provide indicators like it. That’s fair. Okay. Let me throw another suggestion out there. And now you knew this was coming because I'm a big fan of nominal GDP level targeting, but level targeting in general, whether it's a price level target or it's a nominal GDP level target. One of the arguments forward is that it stabilizes expectations.

Beckworth: And let me qualify that by saying a credible level target, and what makes credible is a whole different discussion, but let's say it's a credible level target. So households and businesses take it serious. And the argument is if it's truly credible and it's level, so it makes up for past misses then households would not cut spending in panic in the first place, same thing with businesses if they had a level target in place. So I wonder to what extent would having a level target help with the belief issue that you outlined in your model?

Could Level Targeting be a Possible Solution?

Simsek: I think this would help a lot actually. I think we should really think about making some changes in the inflation targeting framework. I mean, raising the inflation target is the simplest thing we could think about, but price level targeting will help as well. The reason these things help, I think our paper offers a little bit of a different perspective on why these things help. Essentially the issue is when you get a shock like this, especially if your interest rate becomes constrained...

Beckworth: Yeah.

Simsek: ...any sort of future interest rate reductions that you can convince people of are going to be useful. So in fact forward guidance is going to be useful. But essentially price level targeting has a forward guidance element built into it, because it sort of means that if inflation falls we're going to actually raise it and kind of stimulate the economy or stimulate after we recover.

Simsek: So there's going to be basically future low interest rate expectations, and that will get priced by financial markets. Financial markets are very forward looking. There's actually a big literature in macro about forward guidance puzzles, and so on. Actually this is a lot less puzzling once you think through financial markets, because actually the financial markets do react quite a lot to interest rate announcements.

Simsek: And I think if the Fed were to convince markets that it will do something like price level targeting, you would see a big reaction in financial markets and that directly deals with issue, right? Remember the issue was the prices overshot and they're too low. So if market prices low interest rates then it helps to stabilize asset prices and it helps to mitigate the recession.

I think if the Fed were to convince markets that it will do something like price level targeting, you would see a big reaction in financial markets and that directly deals with issue, right? So if markets price low interest rates then it helps to stabilize asset prices and it helps to mitigate the recession.

Simsek: So I'm kind of entirely on board with, I mean, I don't know whether you do GDP level targeting or price level, but I think having something like this would go a long way to mitigate these recessions.

Beckworth: Yeah. A great point. I mean, level targeting is essentially a powerful form of forward guidance-

Simsek: Exactly.

Beckworth: And your model, I can see how it compliment it. So let me ask a few follow up questions about this model in your paper on COVID-19. So what does it mean, your work with Ricardo, in terms of this popular story, one that's especially come out of the last crisis, that when you have recessions tied to equity, like the stock market crash in 2001, that tend to do better than recessions where you have high leverage, like in 2008? Does your work shed any light on that? Or is that just an altogether different question?

Equity vs High Leverage Recessions

Simsek: No, I think our framework offers a very natural way of understanding why the 2001 recession was much milder, or in general why the recessions that don't involve levered agents are much milder than others because this heterogeneous valuation that I've described becomes a lot more damaging when the agents that are hit are highly levered. Because leverage amplifies gains and the losses.

Simsek: So imagine 2001 recession. It sort of mainly had started in the stock market. So the stock market collapsed. Well, the leveraged agents in the economy, banks, institutional investors are typically not that involved in the stock market and they weren't so exposed to this collapse in the tech stocks. It was more like regular investors trading or speculating in the stock market.

Simsek: And that's kind of... That will also be bad because optimists will get somewhat wiped out, but a lot better than if banks or levered agents get a big hit, right? Because they'll get hit and that's it. Now consider the 2008 recession. So banks, institutional investors, more generally highly levered agents were directly in the middle of it. They were exposed to the house prices through these securitized products and so on.

Simsek: So when they get hit, it's actually a much bigger hit to their wealth because leverage will kind of like... Even if their assets go down by a few percent, their actual wealth can go up by a lot more in percent. So actually their wealth collapses by a lot more. So everything I said becomes a lot stronger when these shocks hit levered agents, right?

Simsek: So you're going to get much bigger repricing of risk because these banks and institutional investors are hit harder. Same with, by the way, housing investors, because housing is an asset that can be easily levered, right? It's very good collateral. So a lot of the housing investors were highly levered.

Simsek: So because the shock hit the housing market, even if it didn't hit the banks that would already be kind of a problem, because again, you're hitting the levered agents in your economy and that makes the effects much bigger. So absolutely I think our framework offers a very natural way of understanding why the 2008 recession was a lot worse than 2001, or more generally leverage makes things much worse.

Beckworth: Great. Okay. One other question to ask about your framework here. So leading up to the crisis, we saw really low yields on safe assets like treasury bonds, German government bonds as well. I mean, in fact, in Europe, many long-term government bonds like in Switzerland, Germany, they were in negative. 10 year yields were negative. Nominal yields were negative. And in the US they've gotten really low as well.

Beckworth: So that that kind of screams a safe asset shortage story, which your colleague Ricardo Caballero has written a lot about. But at the same time stock market was just soaring. It was taking off and some people have this disconnect they say, "Hey Beckworth. How can there be a safe asset shortage when you have these rising stock prices?" And yes, they acknowledged that interest rates are really low.

Beckworth: I think you can ask that same question even today, right? So yields have gone down even more during the crisis. Now the stock market did come down, it's come back up a little bit. But do you have a way to explain that at least the pre-crisis period, why you could have this disconnect between, or what appears to be a disconnect between stock prices and safe asset yields?

The Disconnect Between Stock Prices and Safe Asset Yields

Simsek: Okay. That's a fascinating issue that we are thinking about with Ricardo. So I think there are two aspects to this. I think sort of there's a chronic shortage of safe assets, and then there's the crisis phase. I'm going to come to that. Let's think about the chronic. So I think Ricardo is absolutely right that there has been a huge and increasing demand for safe assets and that pushed interest rates down and we see even kind of negative rates around the world.

Simsek: And that came with sort of high stock valuations. But it's natural because low interest rates will actually increase stock valuations as well. And I mean there's a question of what the risk premium on stocks were before, whether they were low or high, and I don't want to get into that debate, but they weren't particularly low because the interest rate, the alternative to stock investments, were also extremely low.

Simsek: So actually it's somewhat natural that the returns on stocks will go down. It doesn't mean their premium is necessarily low, especially considering that volatility indices like VIX were kind of low in the run up to the crisis. So essentially stocks were kind of commanding a relatively high premium in a low risk environment, which suggests risk tolerance can be fairly low and that's reflected in the low interest rates.

Simsek: Now there's a crisis aspect of this. So when we got to the crisis, I think basically now the safe asset shortage became even elevated. And that's consistent with the mechanism I was describing, because what happens when you get a big shock like this, people who are holding the risk get hurt a lot more right, the risk tolerant. And the safe assets seekers are hurt a lot less.

Simsek: So now they command a lot more of the wealth in the economy and there's actually even bigger demand for safe assets. And that pushes down the interest rates even more, and that put us into the zero lower bound in the US. So you might then ask, "Okay, so what about the recent recovery in the stock market?" How do you explain it? 

Simsek: I think there are two things I want to say about that. First, the stocks have recovered, but they’re still much below than their levels early this year. So I think if you just focus on the recent trends, you sort of lose the perspective that they declined so much that even after they recovered…

Beckworth: That's a fair point.

Simsek: ... they are fairly low. But the more important point is that I think the stock prices, when you put them in perspective with other assets, help you understand how actually the risk premium can be still quite high. In particular, if you look at safer assets, in fact, if you look at treasuries, you see they didn't go down at all, they actually went up a lot, right?

Beckworth: Right.

Simsek: Because all the interests rate expectations have fallen. If you look at investment grade bonds, you'll see they've gone down but much less than stocks. If you now move to high yields bonds, you see actually they've gone done more. So basically if you look across asset classes, according to their risk, you see the risk assets are hit a lot harder even after the recovery compared to the safe assets.

Simsek: And the fact that safe asset prices went up means that some of the recovery in stocks is basically the low interest rates, right? Because low interest rates will raise price of everything. It's very easy to see in the case of treasuries, but in stock prices, essentially what's going on is stock prices have declined because of low risk tolerance, but they kind of recovered because of low interest rates and we see the level that they have. But it is consistent with them still being priced with low risk tolerance and high premiums.

The fact that safe asset prices went up means that some of the recovery in stocks is basically the low interest rates, right? Because low interest rates will raise price of everything. It's very easy to see in the case of treasuries, but in stock prices, essentially what's going on is stock prices have declined because of low risk tolerance, but they kind of recovered because of low interest rates.

Beckworth: Okay. That makes a lot of sense. Well, the time we have left, I want to ask a few more questions about your second paper, the one that was titled *A Model of Asset Price Spirals and Aggregate Demand Amplification of a COVID-19 Shock.* So tying it back into the current crisis we're experiencing, and you actually have some policy prescriptions in that. So talk us through what should the Fed be doing? And do you think the Fed has been doing them?

How Should the Fed be Addressing the COVID Shock?

Simsek: Okay. The first thing the Fed should do is of course cut the interest rates and the Fed has done that very early on and then hit the lower bound. Our model actually shows that when you hit the lower bound, if you don't do anything else, these shocks can become a lot bigger.

Simsek: And here's how that works. And that I think motivates the other policies I'm going to describe in a minute. Because when you hit the lower bound and sort of the price pressure is high. Well, you'll get a big decline in risky asset prices, kind of like we saw in March. But when risk prices now start going down, even beyond the initial shock, everything that I told you earlier becomes amplified, right?

Simsek: Because the risk tolerant agents now get hurt, not just because of the supply shock, but because of the additional repricing. So their wealth declines even more, and this actually leads to an even bigger increase in risk premium and even bigger reduction of risk prices and so on. So we can get into a very bad spiral, kind of like what we saw in March, where you get a huge dramatic...

Simsek: Essentially what happened in March is that some financial indicators, like the volatility index has reached to levels that we have last seen in the last financial crash, even though this wasn't something that originated in risk markets. So that makes kind of alternative policies very important, right? Because you cannot directly deal with regular monetary policy. What we show is that there are still things we can do, and we've seen the central banks do this, you want to do policies that support risk markets more directly.

What we show is that there are still things we can do, and we've seen the central banks do this, you want to do policies that support risk markets more directly.

Simsek: So risky asset purchases or credit markets supporting facilities, things that the Fed announced in March, work very well in our model. And the reason they work is basically, remember the problem was that market suddenly becomes risk intolerant. So there is a supply of risk, but you're unable to absorb it because you are very risk intolerant. What these risky asset purchases, or the facilities that kind of promise these types of purchases, do is that... 

Simsek: They take the risk away from the markets, right? They basically take away some of the risk, and directly deal with the problem. They bring that risk to the consolidated government's balance sheet. And now the market, even though it's less risk tolerant, is able to absorb the risk. So the risk premium doesn't go up as much. And that kind of raises asset prices. And once you are able to raise asset prices, now all these knock-on effects work in your favor, right?

Simsek: So you sort of undo the spirals that I've described, and you can get a very big bang out of the buck with these policies. Our model clarifies how that works. But I think basically it's also very consistent with what we've seen in practice because once the Fed announced these policies, shortly after these announcements, in fact, the risk market stabilized and spreads came down and VIX came down and stock markets recovered and so on.

Simsek: So I think our paper offers a natural explanation and also suggests that the central banks did the right thing and they should be ready to do more if necessary, because the crisis is not over. So if things turn south, you will need more of these facilities.

I think our paper offers a natural explanation and also suggests that the central banks did the right thing and they should be ready to do more if necessary, because the crisis is not over. So if things turn south, you will need more of these facilities.

Beckworth: So for all the people out there who are crying, "Oh, the Fed is doing credit policy. Should just stick to monetary policy," your response is, but look, by doing this, it is actually helping the economy. Spreads have come down. It's enabled some risk taking back in the economy. In fact, some of the announcements in fact themselves are powerful.

Beckworth: I know that when the secondary market was announced, the high yield ETF market stabilized on corporate bonds. And your point is overall, you see a similar picture, so we should be grateful for that. And I think, secondly, you would say consider the counterfactual, if the Fed hadn't done that it would be far, far worse.

Simsek: Exactly. In fact very recently we are adding an extension where we also focus on debt overhang type issues in the corporate sector. Let's say some firms are kind of in debt and they might go bankrupt. Actually, these policies help a lot with that as well, because like you said, they lower the spreads. They enable the firms to do kind of bridge financing to stay afloat.

Simsek: And of course you should do other policies as well. We are never saying that you shouldn't be doing anything else. But these are very natural complements to all the other policies that the Treasury and the government is doing. They very directly actually help real firms and real households and then help the economy. So exactly, I agree.

Beckworth: So this speaks to another issue we've talked about on the show before and that is, what should a central bank buy in terms of assets? And one of the, I don't know, guiding principles you often hear is, "Well, the central bank, the Fed, the ECB should try to have a neutral footprint." But I think being neutral would be kind of doing what you've just said, kind of restoring the pre-crisis equilibrium, but here are the two stories I've heard.

Beckworth: And I've mentioned this on the show many times, but kind of the US-UK view is that central banks should buy only government securities because from a consolidated balance sheet perspective, it leaves a small footprint that the Feds buying up all this debt and therefore it's not intervening in a major way into private credit markets. That is one definition of neutral.

Beckworth: Another definition is the German view, which is, I think more practiced at the ECB and definitely has a more sympathetic hearing over there. And that is the central bank should try to replicate or mimic something like an average portfolio of assets held by some average investor, if there's such a thing. Which I know it's aggregation could create a lot of problems in this context, but the idea is you're not going to be neutral by just buying government securities.

Beckworth: In fact, you're going to be weighing your portfolio heavily towards government securities and ignoring all these other assets. And so it makes sense to buy a little bit of everything, representative of maybe what the market is doing. And I think your paper, if I can kind of weave that into it, suggests the German view makes a little more sense. So what are your thoughts?

Simsek: No, this is a great question on some open issues that I hope we hope to get into. So it really depends on the purpose of monetary policy. I think if you view the primary purpose of monetary policy is increasing the money supply and cutting the interest rate, which is the usual view, right? Or the usual mechanism, cutting interest rates.

Simsek: Then I think buying safe assets makes some sense because it's not about taking risks. It's about increasing the liability of the Fed but in sort of a minimalist way. But in situations like this, when your regular monetary policy is constrained and you're sort of trying to interfere in risk markets and take the risk away the market, with the backing of the Treasury of course, then I think the German view starts to make more sense. Because when you're buying risk, there is a danger of picking and choosing and creating kind of winners and losers. And you don't want to do that.

Simsek: You want to kind of... In fact, in our model, you buy the market portfolio, you buy a little bit of everything depending on their market value. That might be a bit extreme, that might be reasons why you may want to tilt a little more toward bonds relative to stocks, because there's additional channels not in our model, by which spreads might matter more and so on.

Simsek: But the starting point should be kind of you want to do something non-targeted. If your goal is to absorb risk, you want to do something non-targeted and try to absorb risk the same way, kind of a representative investor will absorb.

Beckworth: Okay. So maybe a tentative compromise would be, we act like the US-UK view during normal times, we take on the German view during crisis times.

Simsek: That's fair.

Beckworth: Okay. Well, that's been great. Well, our time is up and our guest today has been Alp Simsek. Alp thanks so much for coming on the show.

Simsek: Thank you, David. It was a pleasure. Thank you.

Photo by Spencer Platt via Getty Images