Money Markets, Collateral And Monetary Policy

>> John Taylor: Let's get started, we're all here and happy to have Harold Ulick talk to us about an amazing topic. It's hard to get through, I have to say, Money markets, Collateral and Monetary policy, yeah, different titles.

>> Harald Uhlig: Yeah, no, so.

>> John Taylor: Well, yeah, from the University of Chicago, PhD, Minnesota, I just looked through a gigantic two volume book, what, two volumes.

We wrote together called the Handbook of Macroeconomics 2016, not that long ago, we'll have to do another one, okay? Harald, welcome, you're really welcome, many times over, so thank you for being here.

>> Harald Uhlig: Yeah, no, thanks for having me, visit Uber, it's actually, I'm having a great time here, and so thanks for all the support.

And thanks to the two drawings, I mean we wrote a paper together in 1990 already, so that's a long time ago, so thanks for being here. So this is a joint work with Fiorela De Fior was at the bang for international settlement, Marie was at the European Central bank, that's at work.

And Karen Rogers, who used to be a PhD student here, and now it's at ASHI Paris, Paris, and paper has been a long time in the making. We were interested in thinking about the Eurozone crisis was the richer motivating factor. And when you think about the Eurozone crisis the interbank market was in, let's say upheaval.

There was a word there that said it's, it can be impaired, but I don't know what that means, but that's the word that central bank documents would say. So when you look at the eurozone crisis from 2010 to 2015, a bunch of things happened. But in particular, as far as interbank markets were concerned, the share of unsecured interbank borrowing declined throughout European Monetary Union.

The bank borrowing from European Central bank in the south increased Eightfold. The market haircuts on southern government bonds increased substantially, nonetheless, the wholesale deposits at banks remain stable. There's a bit of u-shaped pattern for that and I'm gonna talk about it, but they remain pretty much stable. So all these things have affected the interbank money markets, and these interbank money markets are important for liquidity management.

It's something that central bankers certainly worry about a lot, and so central banks now have developed all these additional tools to try to deal with that. Often we write down, I don't know, three equations, Keynesian model. There's a normal interest rate, it's set according to the chain rule or some other principle.

But central banks do a lot more than that, and actually, they worry about financial markets a lot. So here we are going, want to go beyond the question of how to set interest rates in terms of inflation and so forth, output gaps. And quantitative easing want, and think about how does the central bank interacts with that and what are the macro impacts.

So what we do is we construct the quantitative general equilibrium model to understand this. And what we're gonna have, you're gonna have heterogeneous banks because we have this north-south distinction. We have this distinction between banks that can borrow unsecured and unsecured. so there's a lot of heterogeneity across banks here, we're gonna have one central bank at the mainland against collateral and when it does so it imposes a haircut.

So these prices, private sector and central bank sector haircuts play a crucial role here. And then in the end we want to compare, I mean, we had a bunch of policies to compare. But in the end, we settled on two, so we compared the benchmark policy to an alternative policy.

So what's a benchmark policy? The benchmark policy is what the ECB has done, namely lend against bonds that are favorable to the haircut. So they didn't impose draconian haircuts at private market, they had light haircuts. I mean, you could think maybe that's a bad idea or good idea, we don't have anything to say, that's what they did.

And we compare that to an alternative where the bank actually doesn't land at all. I mean, you can't go to the central bank and borrow from there, you have to get reserves if you need them.

>> Harald Uhlig: Pardon?

>> Speaker 3: Badgett. That's right. Walter Badgett.

>> Harald Uhlig: Yeah, I mean,

>> Speaker 3: so they should lend at a penalty rate.

>> Harald Uhlig: Right, they should lend against good collateral at penalty rate. So there is an emergency lending program that the ECB has and has been used in Greece, technically it's actually done by the National Central Bank. So it was the Central bank of Greece that did it, I mean it played a great role during the Greek sovereign debt crisis, but overall not so much.

So lots of banks, yeah, I mean you could say it's an implementation of the bargeload principle, if you like, yeah, mm-hm.

>> John Cochrane: It wasn't like a constant policy in this period, there was a bunch of innovations,

>> Harald Uhlig: Right.

>> John Cochrane: Was for the banks, the ECB said, hey, we will take all sorts of stuff as collateral that previously we wouldn't take at all.

>> Harald Uhlig: Right.

>> John Cochrane: So for example, this move from interbank market to ECB market. Well, the ECB says come and get the money and you can post whatever drunkest collateral you want. And we'll give you a great haircut for it, well, surprise, surprise that they went and did that.

Similarly, there was other things like Mario Draghi said, whatever it takes.h

>> Harald Uhlig: No, I mean there's a ton of things going on and we thought about them, they came up, the model in some ways, I mean, I don't know, is it simple as complicated? Anyways, the model that we have has, that's gonna be my defense, has already 91 equations.

And it took us actually, more time than I care to remember to cut through those. And so whenever you're gonna say, well, why don't you have that feature in the model? I'm gonna point at that line and say, look at it, so,

>> Speaker 5: You have a parcel on us, 91 equation.

>> Harald Uhlig: Exactly, it's as parsimonious as you could write it down in essence. So the policies differ concerning the rise of the private sector haircuts, that's really where these two policies have a difference. And the foreign output, the initial foreign output would have been twice as high under the alternative rather than on the benchmark.

So that's gonna be what we're gonna show, so here's a picture of that, So I'm gonna explain the lines here. But the yellow line is when the private sector haircuts started dramatically rising. And when the banks would approach the central bank to borrow from there. And you can see this benchmark policy in blue and the red policy is the alternative.

Now, interestingly enough, there's a rebound that's later, right? And so you could say there was a bigger whipsaw, would have been a big whipsaw under the alternative policy, okay? So let me get into this, there's lots of literature, so here's my literature review, we really only mainly rely on two papers.

One is what's a theory of bank leverage, here we rely on Karate Kiyotaki. So they have a model where you can maybe run away with a share of the assets. And that imposes a leverage constraint, thus become very popular, and so we use that. But then you also needed, The reason for liquidity management.

And there's a nice paper by Bianca pigeon econometric in 2022 that we did on. So essentially we're combining these two into one model. Okay, so let me first show you some data. Here's where the data comes from. Not that interesting, but we distinguish between two regions. North for us is Germany and France and South is Italy, Spain and Portugal.

Now you could ask why not also Greece, why not other countries? But that had to do with the limitations of the data that we had. So anyways, so let me show you. So here's the first observation that we saw this decline in the unsecured shares. So you see the solid line represents the unsecured borrowing where banks just borrow from each other without having to post collateral.

And you see the decline both in the north and the south. So that's a very symmetric movement with a secured borrowing rising. You first used to have only this slide that just shows the shares. But in contrast to the United States, the interbank market in Europe is alive and well.

It shows that now you have to post collateral tomorrow. I mean in the US that interbank market has largely collapsed because they're all borrowing from the Fed.

>> Speaker 5: The unsecure amount rise or.

>> Harald Uhlig: Yeah, so the secured amount rises, right? It's substituting for the unsecured. The overall amount-

>> Speaker 5: The overall amount is unsecured. Where's the unsecured levels doing?

>> Harald Uhlig: Well, so the secured rise of some and the unsecured fall. So there's a substitution. I don't know what the total is. And if that's the equation.

>> Speaker 5: Yeah, radically access.

>> Harald Uhlig: Yeah. I think the share was initially 50 50. So..

>> John Cochrane: all borrowing from the ECB is secured.

>> Harald Uhlig: Yeah, so that's not part of this year. That's the interbank market. Yes, right. Yes.

>> John Cochrane: Interbank is unsecured.

>> Harald Uhlig: Yes, can be unsecured. Some of it.

>> John Cochrane: Some of this is a choice to go to the ECB.

>> Harald Uhlig: Exactly.

>> John Cochrane: Choice, the interbank market to start asking for collateral.

>> Harald Uhlig: Right, so we're getting into a portfolio problem here right away, right? Because some banks can borrow unsecured, some banks, if they borrow secured, they have to post collateral and then they can decide do I go to the ECB where I have to post collateral there or they'll go into bank market.

 

So that these portfolio tradeoffs that are inherent to what we do.

>> John Taylor: It's just a wholesale funding, right?

>> Harald Uhlig: Yeah, yes.

>> John Cochrane: US they borrow from the agencies and pension funds.

>> Harald Uhlig: Right, yeah. It's I think in the United States at least federal funds make it, right? I mean it kind of disappeared.

It's just the foreign banks and home loan agency essentially.

>> Speaker 6: Can I ask you about market structure? Thank you, in practice. So this is an interloan. This is a market crossed by markets for just banks in new countries.

>> Harald Uhlig: Yeah.

>> Speaker 6: And in your denominated.

>> Harald Uhlig: Right.

>> Speaker 6: Assets or bank from a non-EU country.

I was thinking about is there any currency considerations if the market access.

>> Harald Uhlig: Yeah, so I'm pretty sure there's the interbank market of banks borrowing and lending in Europe, right? I mean and foreign lending, reserves to each other.

>> Speaker 6: Country that is physical in Europe but does not have the Euro would have a bank that can participate in this market.

>> Harald Uhlig: Well, it's not here we don't measure it, right? Because there, I mean these five countries, that's a good question.

>> Speaker 5: Whether they have access to the ECB balance sheet? I mean probably some of them, but I don't.

>> John Cochrane: How much? Does it matter?

>> Speaker 5: You only have these five countries.

>> Harald Uhlig: Well, it's these are the five large countries so the other ones are all smaller. So I think it's a-

>> Speaker 3: Same picture?

>> John Taylor: But we don't have the data for all the other countries and so there's a data limitation.

>> Speaker 5: Okay. So one thing that was going on in the middle of this period was the Fed had swap lines.

>> Harald Uhlig: Yeah.

>> Speaker 5: For the other central bank. What role does that play and what's helping to facilitate if any.

>> Harald Uhlig: Yeah, so the only thing that we consider here is collateralized borrowing. Borrowing from the central bank. And I mean there many more instruments. I mean, so 91 equations.

>> Speaker 5: Were they more willing to do this cuz the swap lines of the Fed.

>> Harald Uhlig: Yeah, would be interesting to compare. I mean it's not a comparison to the United States, right? I mean it's specific to Europe. It's

>> Speaker 5: What I'm just wondering-

>> Harald Uhlig: The central banks introduce sort of a whole plethora of instruments, right?

And announcements and what have you, and this is only focusing on one part. Okay, so observation tool. So that's the solid line the boring from the ECB didn't change much in the in the north, right? It stayed pretty flat but increase increased eightfold in the south. And so there's a clear distinction here between northern banks and southern banks.

Now the same picture also is just a question.

>> Speaker 7: Is it the distinction between the banks or the assets they have to pledge?

>> Harald Uhlig: It's going to be. Well, I mean in the model we're gonna say it's the same, right? Are we just hardwired to be the same?

>> Speaker 7: Because presumably all of the lending that they're doing and all the borrowing they're doing in the secured market, the southern banks, where they're doing it in the private market, they're posting collateral, which is bonds and other high quality bonds. And it's because they have a lot of the low quality bonds that have a huge haircut.

That's what they're taking here.

>> Harald Uhlig: No, exactly. So I mean, this is, it's a really strange feature. There's a strong home country bias in terms of the bonds that banks hold. So Spanish banks overwhelming the old Spanish bonds, German banks overwhelming the old German bonds. It's really weird.

It should be the opposite, right? I mean, Spanish banks should hold German bonds in the other way around in order to unfreeze yourself from being tied to the fate of the sovereigns.

>> Speaker 3: It's not a surprise. That's not an irrational behavioral bias.

>> Harald Uhlig: Yeah. I mean, so here we hardwire it.

>> John Cochrane: Hold your own country bonds, your regulators. Luigi, you're not at the bond auction. Should we come look at your books?

>> Harald Uhlig: Financial repression, I have another explanation in another paper. I'm happy to talk about it. Which has to do with the, if a German bank called Spanish bonds and Spain goes under, right?

Then in principle the fiscal authority in Germany has to kind of back the German bank and the ECB is made whole. But if the Spanish bank sold Spanish bonds and Spain goes under and the bank goes under, the ECB is stuck with the loss. So I wrote this up as a working paper when I was an ECB visitor, and I had the paper on American methods and they said, I can issue the American methods working paper, you can't issue that one, as I said in working paper.

So, I mean, I don't know whether that's the right explanation.

>> John Cochrane: This is really simple. So everybody wanted to get rid of Greek, Italian bonds, especially French and German banks wanted to dump them. The ECB wanted to prop up this market, there's a limit to how much it was gonna buy directly on its own balance sheet, so it simply told the Italian, Greek, Spanish, we don't have Greece here.

But he told the southern bank, look, you guys buy this stuff up, we'll lend you the money to do it and we can post the stuff as collateral. It's kind of an indirect way for the ECB to buy up all its debt.

>> Harald Uhlig: That's essentially, if you like, the story, right, so here they keep the haircut favorable in contrast to private markets, that's our benchmark.

>> John Cochrane: That's like an arbitrage for banks.

>> Harald Uhlig: Well, not quite, right, because they are, I mean they're getting the bond back, right, I mean they're not selling the bond directly, just so.

>> John Cochrane: Sense of arbitrage is there was no capital charge, so there's never been a capital charge on sovereign debt, which is just scandalous.

>> Harald Uhlig: No.

>> John Cochrane: As long as default doesn't happen.

>> Harald Uhlig: Right.

>> John Cochrane: Yeah, it's an arbitrage except for default, but if there's no capital charge, no limit to this whatsoever.

>> Speaker 5: If the Eurozone collapses, you have to target two balances that have to.

>> John Cochrane: The German can come to the rescue.

>> Harald Uhlig: Yeah, so that's another huge topic, so I'm gonna leave that away, so this is just sort of taking the segment. This is the same picture because there's also dotted line here and these are the haircuts. And you can see, it's sort of in line with the discussion that the haircuts practically for northern bonds didn't change, whereas the haircuts for southern bonds changed dramatically.

I mean, they went up a lot, and so that explains the borrowing from the ECB that favorable haircuts. And then finally, I mean, here's also the borrowing from the ECB, but you can also see the borrowing from the households or the deposits as a share of the assets.

And what you can see, I'd like to emphasize that it remained fairly stable, but you can also see sort of a dip there that moves in the opposite direction from the borrowing from the ECB. And so we're gonna have that dip also in the quantitative implication, so we like to think of it as a substitution.

>> Speaker 6: Do you think of it as a substitute?

>> Harald Uhlig: Yeah, it's substitution effect that the banks went borrowed from the ECB rather than from depositors and then eventually went back and bought from depositors rather than the ECB, that there was.

>> Speaker 7: You have an account for why deposits were so stable in this area of time.

At a first order, it's not rational, if you're a household, if you're to keep your deposit in an Italian bank, it seems like upside downside. I'm just wondering, I think that is has been a striking feature of the Eurozone experience, but I wonder if you have a theory of that.

>> Harald Uhlig: Right, so in the model we kind of hardwired it because as you're gonna see, households don't have much of a choice of assets so they can hold deposits of money. So they're stuck, but I think the answer may just be in the deposit insurance scheme that Europe has embraced, and there was this moment during the 2008 financial crisis that I vividly remember.

There was sort of an interbank bank that was ready to collapse in Germany, there was talk among the finance ministers at a conference to issue blanket deposit insurance. And Germany said no, that's up to the banks to do it, we don't do this, I think this was November 2008 if I remember during the financial crisis.

And then the bank collapsed and then the German government, Merkel issued this statement, your deposits are safe and was publicized in a widely circulating newspaper in the Bild Zeitung. And it's not clear what that really meant, it didn't say we guarantee your deposits, but I think, everybody understood that this was an implicit bailout guarantee by the government.

And then very quickly after she issued that statement, money flew into Germany like crazy from all the neighboring countries, so all the other countries then had to follow suit and also issue these implicit deposit guarantees. And so my guess is it's these deposits guarantees.

>> Speaker 7: Although the second part of your story deepens the mystery because why wouldn't that continue, because the promise from the Italian government is not the same credibility as the German government.

>> Harald Uhlig: Maybe, no, I mean, fair enough, right.

>> Speaker 6: Just understand in Europe, unless you are very large investors for an Italian household, to perform these operations, there's free mobility of individuals, but not of capital among your country, so.

>> Speaker 7: But corporations could do, I mean presumably.

>> Speaker 6: This is the household, no, it's borrowing from the private houses.

>> Harald Uhlig: Yeah, the deposits here is borrowing from the house, I mean deposit guarantees are private, the federal guarantees in Europe are implicit, at least in Germany, and so it's also they are different from the United States. Okay, so let me show you the model, it's a general model with really two key features, so we have this leverage constraint and we have the Bianchi-Bigio liquidity feature.

So let me talk it through, so discrete periods in the morning, the households choose between deposits and money, right? So that partly answered that question, and they choose how much to work and how much to consume and production takes place, production is kind of boring here. Foreigners choose bond, we need that somebody else, rather than just banks sold and central banks sold bonds, households don't do it.

So we have foreigners having elastic bond demand, and then what do the banks do? Well, they collect returns on the assets, they pay dividends as a fraction of the net worth, they're just hardwired. They get randomly assigned at the beginning of the period to north or south. I mean that ideally would just banks always be northern banks, always southern banks, but that introduced additional state variables, and so he just did it like that.

So every period the Deutsch bank doesn't know whether it wakes up being a Spanish bank or German bank that's a feature of the model. And then it can be connected or unconnected, and that's what connected means, you can borrow without having to post collateral from other banks, unconnected means you cannot.

And we had in mind some kind of network theory and that was the remainder of that. So we don't say why you have to post collateral, we just hardwired that some banks can, some banks cannot. So then they have to make a portfolio choice, they have to invest in capital, but they also can hold reserves which they may need.

They can invest in bonds, two different bonds, right, with different haircuts, and they can borrow from the central banks and they also issue deposits. Okay, so there's a Gortler-Karadi-Kiyotaki collateral constraint, and there's also constraint visa-vis the central bank. I mean, I think there should be common there. Anyways, and then in the afternoon, yeah, they're government bonds, right, they're only government bonds and no private bonds.

So they're private.

>> Harald Uhlig: So the banks own the capital stock.

>> Speaker 5: Okay. Yeah, yeah. Separate corporate bond, okay, good.

>> Harald Uhlig: Okay, good, yeah, there are capital producing firms and they sell, and so the banks hold the capital stock here. Right, yeah, sorry, yeah, so the only bonds that are here are these government bonds.

And in the afternoon, that's the trick by Bianchi-Bigio there are these random withdrawal of deposits. I mean people make payments, but of course when I make a payment to Bob and I send money, that my bank account goes. Goes down but his bank account goes up. So it's really reshuffling between the banks.

And so there are some banks, I don't know, flush in reserves and other banks that are not. And the ones that are not may have to borrow from the ones that are flush. Some if you're connected, you can borrow unsecured. But if you borrow these private markets you have to borrow against your government bonds and they're subject to the country specific care cuts.

So j denotes here the country or the region could be North or South. But you can also use reserves, if you have reserves, you can just use those to meet these withdrawals. So you're gonna see how that works. And at the end of the afternoon all these flows of payments are reversed.

So Bob repays me at the end of the afternoon and the banks go out with the same portfolio that they had at the beginning of the afternoon. So this is just really as a way of modeling the liquidity need of banks.

>> Speaker 5: Do I redeposit in the same bank or took it out?

>> Harald Uhlig: Yeah, so the banks end up at the same amount of deposits.

>> Speaker 5: What's the bank's problem during the day?

>> Harald Uhlig: Well, they have to meet these withdrawals, right? So, I come to my bank and I sent, I want 1,000 euros or the bank has to produce-

>> Speaker 6: Cash out during day.

>> Harald Uhlig: Yeah, yeah, that's right, that's right. I take cash out of my cash machine, the bank has to meet that cash withdrawal. I go to Bob and give him, he deposits a 1,000 euros so his bank has an additional 1,000 euros. If you think hard about it, there's something odd about the timing here, right?

Because by the time my bank needs to borrow the 1,000 euros, maybe Bob hasn't deposited 1,000 euros. So we ignore that, I mean essentially-

>> John Cochrane: That's in the morning or in the afternoon?

>> Harald Uhlig: Yeah, that's in the afternoon, right?

>> Speaker 7: That's all happening in the afternoon.

>> Harald Uhlig: Pardon?

>> Steven Davis: No risk, her risk is the timing.

>> Harald Uhlig: Yeah, yeah, right, exactly. I mean.

>> Speaker 3: Whether you're connected or unconnected is random, it's exogenous.

>> Harald Uhlig: It's totally exogenous reality, yeah.

>> Speaker 3: I guess the reality, whether you can borrow secured or unsecured, there could be a lot of adverse selection there, right?

>> Harald Uhlig: Absolutely, yeah, absolutely, right. It's 91 equations, it's my.

>> Harald Uhlig: It's true.

>> John Cochrane: True worrying about these liquidity risks, I thought the whole problem here was solvency of the government bonds.

>> Harald Uhlig: Yeah, so-

>> John Cochrane: Taxes and paying coupons these bonds in the model.

>> Harald Uhlig: Yeah, so we didn't model the sovereign default risk either, right?

You could conceptually embed it in the sovereign default model, and.

>> John Cochrane: I thought fear of sovereign default was like the central issue of this whole episode.

>> Harald Uhlig: Yeah, so we're gonna have private sector haircut rising for reasons. I mean, they're just gonna be exogenous in the model. Of course the reason that they're rising is because of sovereign default fears, which we don't explicitly model.

We just have the haircuts. I mean that would be, again.

>> John Cochrane: Okay, so, but I'm getting pictures.

>> Harald Uhlig: Yeah.

>> John Cochrane: You want to see, how does the sovereign thing spill into the real economy and the banking sector and its ability to make real loans?

>> Speaker 6: Yeah.

>> Harald Uhlig: Right, exactly.

>> John Cochrane: The haircut going up now I've got less ability to borrow in order to make these cash needs. And now it's not really about the sovereign default.

>> Harald Uhlig: Right, yes, exactly, i's repercussions.

>> Harald Uhlig: Exactly, why nobody defaults, right? And of course, if a country defaults, then I don't know, you're often not a branch here.

Okay, so here's another graphical way of showing that, right? So there's a central bank and there's this whole banking system, right? They're connected north, unconnected north, connected south, unconnected south. The southern banks only hold southern bonds, the northern banks only have northern bonds. So that's hardwired. The central bank can fund these banks against collateral, right?

So it's collateralized funding, the households deposit. So the Central bank issues M0 in two ways, right? It issues money to households as well as reserve funding to the banks. The banks fund firms, holding their capital. And there these governments that issue bonds, northern bonds and southern bonds subject to different haircuts and there for different prices and the foreigners also buy them.

And then you have the start line, there's the afternoon, there's beginning after, I mean two parts really. So households start withdrawing and they deposit, right? And it all, you think of that as simultaneous. And, so there will be liquidity rich banks that are at the bottom and liquidity poor banks at the top.

And maybe the liquidity poor banks have to now borrow if they're unsecured, if they're connected, that's great. They can borrow unsecured. They just call up their friends and say, please, I need a million. And they send it over, but only for the other ones they have to post collateral.

Okay, so now there's a lot of stuff going on and, keeping track, understanding the impulse responses and so forth.

>> Speaker 6: What is the household's problem? So they receive profits from the bank, what do they save in? I'm thinking about the household problem.

>> Harald Uhlig: Yeah, they also maximize utility subject to their budget-

>> Speaker 6: No labor, but they. I wonder how the firm's profits are rebated back to the household.

>> Harald Uhlig: So the firms are in perfect competition.

>> Speaker 6: Yeah.

>> Harald Uhlig: And it's, it's just pricing.

>> Speaker 6: Okay.

>> Harald Uhlig: Right, I mean, they're just price capital, their price wages. There's an adjustment cost, I mean.

>> Speaker 6: Okay, yeah, yeah.

>> Harald Uhlig: I mean, there's a technology to transform old capital to new capital.

>> Speaker 6: So what do they save then, households?

>> Harald Uhlig: They use deposits on money.

>> Speaker 6: Okay.

>> Harald Uhlig: So they can't hold bonds directly.

>> Speaker 6: Okay.

>> Harald Uhlig: Yeah, yeah. So these models already have a bunch of features that show up in the impulse responses, but there are a few that are, that I think of new here.

There are four main forces or four main effects. One is this capital crowding out effect. If you have to post more collateral, that means you can invest less in capital and that will dampen investment output. There's a bond reserve substitution effect. So banks have to solve this portfolio, the unconnected banks.

Should you hold more bonds or should you hold more reserves, right? You can get reserves quickly, there's a total amount of reserves. And maybe as haircuts rise, maybe you want to hold reserves rather than bonds. Because bonds are no longer as attractive to meet your liquidity needs in the afternoon.

There's a north-south liquidity spillover effect, as the southern banks shift into reserves, this affects the collateral premium, collateral value, even for northern banks. So even though the haircut is gonna rise in the south, the northern banks are affected because of this pecuniary spillover. And then finally, since the central bank offers a different haircut than the private sector, then as John mentioned, that makes banks move into funding from the central bank, pardon.

>> Speaker 5: Tapping the interest rate on reserves.

>> Harald Uhlig: So we have an interest rate on central bank funding that we keep constant. So the interest rate is kept constant here. There are all kinds of interest rates floating around, right? The wholesale stuff, they don't borrow directly from the central bank, they have deposits.

So something happens with the interest rate here. We don't have sticky prices in there. So there's an entirely Fisherian model inflation.

>> Speaker 5: North south liquidity spillover effect.

>> Harald Uhlig: Yeah.

>> Steven Davis: Arises, as I understand it, from the difference in the interest rate on bonds and reserves getting affected by the reserve scarcity originating in southern bank.

Is that right?

>> Harald Uhlig: Yeah, so the Southern bank suddenly they need more reserves or more collateral and but the Northern unsecured also need the collateral. And so they're competing for the same source of collateral and that makes the collateral more expensive for the Northern banks. And as a result, even though the haircut only rise in the South, that's what we're gonna assume.

The Northern banks are affected because they're competing for the same pool of collateral, be it reserves or be it, yeah, I'm really just-

>> Speaker 5: Interest rates not a sufficient measure of credit.

>> Harald Uhlig: Exactly, yeah, so they-

>> Speaker 5: It's all going in the price of cloud.

>> Harald Uhlig: Yeah, exactly, right, so the whole question of what interest rates on the bonds change, interest rates on deposit change, but we on purpose get the interest rate on central bank funding constant.

>> Speaker 5: I don't know what we assumed, I think 1%, 2% or something,

>> Harald Uhlig: cuz that's where the action could be but we just shut that one down. So then here's a model, I don't know how much should I say about this, maybe not much.

>> John Cochrane: Hey.

>> John Cochrane: That's the model.

>> Harald Uhlig: That's the model, right, okay, that's actually should be the shortest slide, so lemme go through this, yeah, that's fine. So there's a whole so maximizing utility, so you sum this over an infinite horizon, of course. And they can hold deposits at a risk free deposit rate and they can hold cash, and they have money in utility function there are firms, your fine good firms that produce output using capital and labor.

They're capital producing firms, they transform old capital and new capital, and the capital is held by banks, so banks are always the holders of capital. Fiscal policy are entirely mechanical and you could think of the two regions but representative household. So we don't have Italians and Germans in the model, we only have Europeans and they get taxed at the same rate.

So really ultimately what only matters is essentially the consolidated government budget constraint, that's the second line there. And you can think of implicitly transfers between the North and the South going on that could be of interest but that interested. So yeah, so there's a tax rate on labor income, they issue these bonds and then they have a fiscal rule that's there in the debt chain.

So if you're far away from where you want to be in terms of your debt position, then you just reduce the debt and you just raise taxes, so that's going to happen.

>> Speaker 6: The household problem because of the substitution effect that you were hinted in the graph. If I think about the European Union, I think about the taxation regimes as being country specific, but the transfers as equalizing.

>> Harald Uhlig: Say this again.

>> Speaker 6: Right, so I was fixated on the household problem because of the graph you showed us.

>> Harald Uhlig: Right, yeah.

>> Speaker 6: Borrowing, and then I was thinking, so it matters how I say, so it matters how much I gain by, I mean, labor income wise.

And I thought about the EU, I mean as a picture, right, as a collection of country with the country specific taxation regimes. But with an authority that fiscally transfer and tries to equalize consumption exposed via transfers. So if you had asked me, I would have expected the Tau to be country specific, but the capital T to do a lot of the exposed adjustment, so to speak.

Does it matter? The Tau is common across countries.

>> Harald Uhlig: The households are common.

>> Speaker 6: A Tau.

>> Harald Uhlig: TAU, yeah, it probably does, I mean again, that was not another kind of warmth if you wanted to open, right? I mean, so it's a European household, European capital market, right, I mean again, you could think that you're investing, investment in the South rather than investment in the North.

It's common capital market, common household, common tax rate, so all of that is common-

>> Speaker 6: And the conjecture as well, but it affects your savings, so the labor income tax affect your saving effectively. And so your ability to save must have an impact on the ability of banks to borrow.

>> Harald Uhlig: Well, I mean there's a European market for deposits, right, and so again, it doesn't matter to the household whether you deposit a Spanish bank or German bank. And there's a supply of deposits coming from the wholesale, depending on the interest rate. There's demand for deposits coming from the banking system, right, and they have to sum up so there's a market clearing condition.

>> Speaker 6: Okay.

>> Harald Uhlig: I don't know why that answers this.

>> John Cochrane: Call this a new Keynesian model, it looks like it's entirely real. Yeah.

>> Harald Uhlig: And it's not a new Keynesian model.

>> Harald Uhlig: Sorry, maybe I mumbled this, right?

>> John Cochrane: You said it, that's what you said.

>> Harald Uhlig: Yeah.

>> John Cochrane: It's flexible price.

>> Harald Uhlig: It's totally flexible price, and thanks to your blogs, I'm more convinced than ever that we don't know really what drives inflation, so here,

>> Harald Uhlig: We just have a Fisherian economy, which means that nominal rates and inflation move in lockstep. And so this model, on the face of it, will have strange implications for inflation, but that's not what this model is about.

Okay, here's the central bank, it can hold Northern bonds, it can hold Southern bonds, it can issue these loans to banks. And they're one period loans, so this is really just a collateral on the asset side, these are just the collaterals that they hold against these loans. And then liabilities are reserved plus currency in circulation, and then there's finally bank equity.

So the collateralized loans, there's a haircut, so let's say ADO T is 0.97, that means for €100 posted, you get a loan of €97. Normally the haircut is described as 1 minus ADO T, in fact the ADO T is kept constant, sorry, that there should be no T on there.

And that's regardless of region, and that's a benefit to the Southern bonds in particular.

>> Speaker 7: Now, given our remark that this is a real model, why do you bother with having all these nominal variables?

>> Harald Uhlig: Yeah, because prices do move and the bonds are nominally safe bonds, so we-

>> Speaker 7: That's an important feature of the bond.

>> Harald Uhlig: Well, I mean,

>> Speaker 7: Reduce 91 to some smaller number by

>> Harald Uhlig: So I mean, prices do move and therefore I mean, I don't know, you can look at interest rates.

>> Speaker 7: Well, that's why there's a difference between.

>> Harald Uhlig: We could have in the end divided everything might be fair, but we felt more comfortable writing a German nominal because bonds have nominally fixed payouts.

In fact we have bonds in the paper, our bonds have a six year maturity, and so then there's all kinds of dynamics as a result of that. But I agree, I mean conceptually we could probably write it at least at this level down in real terms.

>> Speaker 7: I'm on your side of 91, a smaller number.

>> Harald Uhlig: Yeah, so they get this shock, this IID connectedness shock, a network shock knew whether they can force Secured, unsecured, and they get this ID shock, whether they are holders of northern or southern bonds, maybe that's really the right way of thinking about it. And then given that they learn this in the morning, they pay their dividends, they get the shock, and the, right.

Banks would always like to keep net worth inside the bank, but we force them to pay dividends and then they make the portfolio choice. Capital bonds of the particular region they're in and depends on the network type whether they want to hold any, how much reserves money do they hold.

We call reserves money here, same thing, and then liabilities, deposits and loans, so it's a portfolio problem that took us time to think through. And I'm gonna to bother you with all the details of that, unless you really want to know, okay? And then the afternoon, there's a liquidity shock that realizes some omega that's smaller equal than some max omega and you just have to make good on these payments.

So we assume that, they're not allowed to go bankrupt here, so they know this might be coming and they have to make sure that they can get enough money or reserves, to pay demanding depositors in this case. So here's a bank balance sheet, so that's a bank balance sheet, they hold capital, the old bonds of the country, they have to pay dividends, they hold reserves and there's liabilities.

They have deposits at the central bank loans and they have their own net worth. The central bank loans are collateralized, so that's the ADA there is the central bank haircut parameter or 1 minus ADA is the haircut imposed by the central bank regardless of region. And then in the afternoon, so there's a leverage constraint here.

Okay, so the bottom is the Go Karate GEOT leverage constraint that essentially says that how much you can, how large a balance sheet can be relates to the franchise value of a bank, right? So you can one away, that's the idea, you can run away with a lump of share of your assets.

So these are your bonds and your reserves and live on an island happily ever after, or you can keep running the bank and running the bank must be better. So that's on the right hand side, that's the franchise value of new J and that constrains collateral. It also means that how high the leverage can be is a very forward looking constraint with the franchise and that shows up later on if the franchise value of the bank declines.

For some reason, because the future is bad, that means you can level up less today, so there's immediate impact and that will show up in some of the impulse responses.

>> Speaker 6: Sorry, dopey question what is big F?

>> Harald Uhlig: Big F is the, is the borrowing you do from the central bank, so funding from the central bank.

>> Speaker 6: Okay, sorry.

>> Harald Uhlig: Yeah, and then the afternoon, so this year's the Bianchi Bidgio, that's maybe the less known portion, right, this Bianchi Bidgio liquidity management. So they are the deposits of your bank of network type new and region J and you must be prepared to pay off omega max times that.

So Omega max recalibrate to point 1 you must be able to pay off 10% of depositors on the spot, and you do this either with reserves, so that's the MUJ that you may hold, or with borrowing from other banks, at least for the unsecured. Sorry, for the unconnected banks, right, that's the constraint that they have to worry about.

The connected banks don't have to worry about it because if the depositors come, they just borrow from their friends, and so they're not worried about that constraint. They can get the funding, it's the unconnected banks that have to respect this constraint, and you see, for the bonds, they can only pledge the bonds that haven't already pledged to the central bank.

So that's why this is different, the value of the bonds is QJ, so QJ is a market price for these bonds, and then they're only getting Ada tilde on that. So if Ada tilde, if haircuts rise, so the Ada tilde collapses, that means they're getting very little for the bonds that they're holding, and that creates, that creates this portfolio problem.

Okay, so that's the model, then we calibrate it, let me maybe not say much more to that and let me remind you of the forces. So they're gonna show up in these impulse responses, capital crowding out, that means if you buy more collateral, you can hold less capital.

The banks have to make this choice between reserves and funds, there's a spillover from south to north, as the south demands more collateral, that makes collateral also more expensive for the north. And then as the gap opens up between the haircut and the central bank to the private market, banks shift into a central bank funding this.

>> Speaker 6: NN sorry, Ada T or without T Ada tilde JT are backed out.

>> Harald Uhlig: From the data, yeah.

>> Speaker 6: Internally based on the pictures.

>> Harald Uhlig: Yeah, so this is the assumption, actually that's exactly the answer to this so we looked at north and south were the same, so as far as connectedness goes, right.

So we just get the site, I mean it's sort of a little jagged in the data but that's a curve that we fit.

>> John Cochrane: Could you say a little about the calibration of previous slide.

>> Harald Uhlig: Yeah, so some of them are. Let's see what you want to know.

There's a longer slide, I mean many more parameters.

>> John Taylor: Well I know say a little bit of where these are coming from.

>> Harald Uhlig: Yeah, so there's sort of high quality liquid assets and based on that we said the omega max is 10%, I mean that's we had to come up with a number that was did something in the paper, but.

>> John Taylor: You got five assets.

>> Harald Uhlig: Five assets, yeah, so right. Maybe five, six, I forgot I didn't know of them right. And omega max is how many, what's a fraction of deposits that you have to be able to cover on the spot in case all these depositors come on to withdraw.

And we're saying the max of size is 10% of your deposit base is something that you have to come up with, and we looked at high quality liquid assets as a way of measuring what batteries would do. Capa that's maturity, so we want our bonds to be of 6 years maturity, in the presentation I focus on one period bonds but once last couple of years.

This row is the inverse elasticity of the semi elasticity of foreign bond demand, and we got that from a pay by Cogen et al. The haircuts originally are set at 3% so that explains why all these adults are the same, the RF that's the rate that we always keep constant, that's a central bank rate.

If you borrow from the central bank you have to pay 0 point to 5% per quarter or 1% per year, so that's fairly low, and point to its observed share in 2007 of unconnected banks. And then there's some other parameters here, what's the dividend share that you have to payout.

What's the Gertler Kiyotaki Lamta, what's the money demand parameter, what's government spending, what's average debt? Actually it's known real, right, so that's why. Twist and turn these parameters until they match these.

>> John Taylor: So you have estimated all this?

>> Harald Uhlig: Not estimated, I mean-

>> Harald Uhlig: Estimation would be a glory statement, right?

I mean we fiddled with them until it fit the I mean I guess it's a sort of let's say estimation.

>> John Cochrane: You said we don't have standard analysis.

>> Speaker 9: Harold the model is already super rich. But I guess one thing that's not directly in the model is the sort of housing market.

And I was wondering the extent to which that was important in Europe in over this period. I mean the United States, subprime lending was obviously a big part of the 2008 financial crisis.

>> Harald Uhlig: Yeah, I think-

>> Speaker 9: Spain as well.

>> Harald Uhlig: Right.

>> Speaker 6: Spain.

>> Harald Uhlig: In Spain particularly the housing market collapsed and there were these-

>> Speaker 9: Because that could affect the impulse responses.

>> Harald Uhlig: Right.

>> Speaker 9: It could be a direct wealth channel.

>> Harald Uhlig: Right, I think in Europe they never had this morph back security crisis, though, that was in the United States. I remember when that hit the first banks that went under were actually German banks because the German Landesbank invested heavily in those and had to be bailed out by the German government.

But they weren't-

>> Speaker 7: What is had to be?

>> Harald Uhlig: Hm?

>> Speaker 7: the German government elected to.

>> Harald Uhlig: Well you could say that. But there were federally owned bank. William, I mean technically but it's tricky. I mean they're sort of public private partnerships perhaps. Anyway, but these mortgage backed securities, they were never a big deal in Europe, all the mortgage lending is pretty much done by the banks directly.

But certainly the Spanish housing market suffered greatly.

>> John Taylor: Now you'll tell us the results.

>> Harald Uhlig: Yeah, exactly. So okay, so first the four facts model versus data. So, the secured share, that's on the left. I mean that's pretty much by assumption, you just say we match the data.

Here's the Central Bank borrowing and you can see that it's gonna spike both these shocks in there. So we have two shocks here, that's the other thing that I wanted to say. So we have the disconnection shock and then a little later the haircut shock. So the haircut starts rising or the ADA tilde starts falling quite a bit.

After the disconnected happens, you see that it's sort of constant first and it's three years later and that's gonna matter to some degree. And that's where the yellow line is. That's where the haircut-

>> John Cochrane: Great, tell us what's going on here. I'm used to impulse response functions where you start at 0.

There's a shock and something happens that's not. This is something else.

>> Harald Uhlig: Well, so in 2007, out of a sudden, and to the surprise of everybody, XI is having the red evolution. Everybody understands in 2007 that XI is no longer gonna be constant, but that is gradually dropping to much lower levels.

So psi is the fraction of banks that can borrow without having to post collateral.

>> John Cochrane: Is this an exogenous?

>> Harald Uhlig: It's an exogenous event. So this is exogenous. Yeah, this is exogenous.

>> John Cochrane: 0.42 until 2007.

>> Harald Uhlig: Right.

>> John Cochrane: And then you're feeding in, it's the red line.

>> Harald Uhlig: Right, we're feeding in this red line.

>> John Cochrane: Okay.

>> Harald Uhlig: So the shock is learning that the red line is going to be, and you learn this in 2007. And then three years later, actually in the data it's four years later and we kept that three years anyways, three years later so that will be 2010.

They also learn that haircuts in the south will be dramatic. And we follow this red line, right? So, it's constant until then. I mean, that's not quite true maybe. Well, I don't know. Maybe it's from the data. And then it falls off fairly rapidly. And we had these yellow dots on the blue line that's really the data, right?

And so the blue line it's an artifact of putting a line through the yellow dots. And so the red line, we then tried to put a, I think it was an AR2 through that, or something that explained that shape. But it's pretty much-

>> Harald Uhlig: Private sector haircuts, yeah.

>> John Cochrane: Haircuts or something.

>> Harald Uhlig: They're constant, yeah. The ECB haircuts remain constant throughout.

>> John Cochrane: It is the haircut. Remind me what psi was again?

>> Harald Uhlig: Psi was the fraction of banks that could borrow without having to post collateral from the private sector. As far as the ECB is concerned, you always have to post collateral.

And it's collapsing from 0.422 to 0.19 or s-

>> Speaker 5: Chose to not the number that couldn't.

>> Harald Uhlig: So you have to post haircut, sorry, you have to post collateral if you're an unconnected bank. And psi is a fraction of banks that don't have to post, they don't have to post collateral.

They can be on the afternoon market, their customers say we want 1 million euros, and they just go to their friendly network neighbor bank and say we need 1 million euros and they ship it over

>> Harald Uhlig: These are the true shocks yeah. So the stochastic, yes, exactly. So The Stochastic Experience, 2007, everybody wakes up, they realize it's a red line.

And in 2010, everybody additionally wakes up and says, my god. Also, the private sector haircuts move and they move in this particular shape that-

>> Speaker 6: Agents expect that they will continue past 2014, 2018, they anticipate entire path forever.

>> Harald Uhlig: Yeah, they anticipate entire paths and the haircuts are going to be permanently lower, for example, and the size is gonna be permanently lower.

>> John Taylor: So it approaches, I think 0.15 or 0.1, I can't quite read it here.

>> Harald Uhlig: Yeah, you could draw it out, but it goes to lower limit in the end. So the steady state moves, right? Yeah, yeah, right, the steady state moves. And okay, so that's the haircuts move, that's all.

The third line is by assumption. And the deposit share, you get a little bit of movement, it's fairly constant. But we get this little funky dip that you also see in the data. That's kind of the converse from the Central Bank boring. So that's just saying that the four observations that we had got matched.

>> Speaker 9: Here is a question.

>> Harald Uhlig: Sorry, Athanasios?

>> Speaker 10: Let me figure out how to unmute. I need to figure out how to unmute here.

>> Harald Uhlig: No, you're muted. Yes.

>> Speaker 10: Very good.

>> Harald Uhlig: Okay, so thank you.

>> Speaker 10: So, Harald, this is really a fascinating model. I'm glad you're doing this, but I don't agree with the way you are modeling the baseline because I was there in 2007-2011, which is the period you are modeling.

It would have been wonderful if the ACB had maintained a 3% haircuts. As a matter of fact, the experiment I think that would be worthwhile doing is the reverse. What we messed up in 2010 is that the ECB's framework started threatening countries, including Portugal and Italy that you have in the sample, that the haircut would rise to 100% if they were downgraded.

And because of that threat, markets actually reacted, the spreads went up and the whole thing unwound. So, you really need to capture the thread of 100% haircut creating the crisis instead of saying, yeah, yeah, it was 3% and it was so nice for stabilizing things. And you have the tools to do this.

I think you have a wonderful model where you can actually ask the question. The Euro crisis, which in my view was totally unnecessary, was 90% of that due to this peak headed ECB policy or 50% or 60% or 30%. I think you can quantitatively answer that question, but you are not doing that because your baseline is kind of way off.

>> Harald Uhlig: Yeah, that's a great Suggestion, part of the problem is of course the private sector haircuts here are exogenous, right? So you are telling the story why the private sector haircuts endorse respond to what the ECB does.

>> Speaker 10: Because the problem is that the collateral policy can create multiple equilibria and debt markets.

And by doing what we were doing, we effectively moved bond markets to bad equilibrium. And then everybody reacted to that. But that was a choice by the central bank. This is the Fed telling us tomorrow that US debt dynamics are unsustainable. So as of next week they will no longer be accepting a government that is collateral, think about that, that's the sort of thing that the ECB did.

>> Harald Uhlig: I'm very sympathetic to that argument, I mean it's a great argument. It's just that, it would require another extension of the model in some ways we have that actually because our alternative policy has the ECB not accepting haircuts. Don't worry about the 51 slides, half of them are just additional slides that are brought along.

So I'm gonna be actually getting close at the end here, so this is not. Yeah it's my fault, I'm gonna end on slide 29 or something. So anyway, so this is the two shocks, the output movement here together, there's a sort of this funky, whiplash effect on output.

And that has to do with this good Kiyotaki Karate, net worth forward looking feature, let me not dwell on this. But you see this output drop that happens here once the haircut shock happens. Okay, so impulse Responses, so remember the sign moves only very gradually, right? In 2007 the banks, everybody wakes up and says gradually, the number of banks that can borrow without paying collateral is falling.

Nonetheless you get immediately drop an output, and that's because of this forward looking, the nature of the GKK constraint. Everybody realizes that the franchise value of banks has fallen, and that immediately dampens the investment that happens. So you get an immediate effect on capital markets and so forth, and that's what is shown here.

But then if you add also the adhatoda shotgun we saw this response and output, and it's driven largely by the response and investment that also has this whiplash. The thing that I want to emphasize here on the slide, is the two panels on the bottom, the capital and the south and the capital and the north.

So when the haircut rises in the south, that makes collateral more expensive, and you see this in the collateral premium for the south and for the bonds, right? It makes it, it makes it more expensive for them, and so naturally they invest less. You get this shifting out of this first force, you shift out of capital and posting collateral.

But because the north and the south, they both compete for these reserves, that the central bank is somewhat reluctant to put in here, the north doesn't, counter that effect. The capital that the north holds doesn't rise, right? You would think that if the south invests less in capital, the north will now invest more in capital and just replace them.

But the collateral premium has risen overall in the euro area, and that's this north, south pillow effect that happens here.

>> John Cochrane: Basic output story, so we demand more collateral from banks, they can't give you intraday money as much as they used to be able to, now what happens?

Somehow you invest less or.

>> Harald Uhlig: Yeah, so the banks shift out of investing in capital into investing in bonds, so there's less investment because banks are holding all the capital. So the banks have to hold more bonds, and so as a result hold the balance sheet constant, there's so many things that can move.

But if the balance sheet hold is constant and the only thing that happens is the mix between capital and bonds, as they have to hold more bonds for the afternoon, they have to hold less capital.

>> John Cochrane: The people that they bought the bonds from, what do they end up not allowed to hold capital, what do they do?

>> Harald Uhlig: They're foreigners, right? So it could be a gift to the foreigners at that point, because that's true, the bond price falls, right? No, the bond price rises here, so because banks are now willing to pay more for the bonds. It's a bit more complicated because return to capital ought to fall, actually, the return to capital rises, right?

>> John Cochrane: And those deposits used to go to buy investment goods, I guess, or what happened to cash, but now the deposits go abroad to buy bonds.

>> Harald Uhlig: Well, the deposits always only help the households, so households hold deposits and money. So partly the banks compete against the money that the wholesales have, right?

So the central bank has Issued a certain amount of reserves, and it's now needed by the banking sector because they need these reserves now in the afternoon to meet sudden withdrawals. And they're moving into that because, bonds have become relatively expensive so they're moving out of. I mean they have to hold more collateral, that's one thing, but they're also shifting from bond collateral into money, so all these things are happening at the same time.

>> John Cochrane: A classic fallacy is that, banks are making paper investments not real investments so the economy is falling. So I got to the classic fallacy but I'm trying to tie up the general equilibrium of what people did with the money.

>> Harald Uhlig: That's why question, so I'll leave it there.

Maybe the best answer is not just about windfalls, it's also about the net worth evolution, given that in the future you have to hold more collateral and it's harder to borrow. The franchise value of banks have fallen, and that means the leverage constraint has become tighter. You might get a windfall gain on the bonds, but nonetheless the amount of leverage that you can have may become lower, because the franchise value of the banking system has gone down.

It's one of these weird looking good like Kiyotaki.

>> Speaker 6: I would have thought that deposits in the south, and capital in the south, they would have mirrored each other and they do up to.

>> Harald Uhlig: So that's right, so the deposits through this whiplash. So capital falls, return on capital rises and that means the ADA here has this evolution, it also recovers after a while, right?

And so that means the initial shock happens but then after a little while actually the future looks better than it just looked yesterday. And that means the forward looking banks say okay, now the franchise value for the banks goes up so they can leverage up a little bit more, but they have to hold more collateral so they can't invest in the capital.

>> Speaker 6: What I said is module, the path of the whatever side, I never know.

>> Harald Uhlig: Right. And then And then you can compare the benchmark, I have to probably finish soon, can compare the benchmark for the alternative. And really the main story is here that, in the alternative, you can't borrow from the central bank, right, and so if you look at the middle row, all the way to the right, that's really where the difference is.

The funding from the central bank, it's all now in deal, that's why we use small letters, the funding from the central bank and the benchmark goes up. But in the alternative policy, that's sort of the Athanasius version, you can't borrow from the central bank at all. And because you can't borrow from the central bank, that worsens the collateral situation, collateral premium is higher.

Because collateral premium is higher, banks now can invest less, you also see what happens with the capital actually goes down both in the south and north. And the alternative policy, because it's spillover, north, south, spillover is so strong, so chart up with the.

>> Steven Davis: Far left corner up the top is the same as the trip.

>> Harald Uhlig: Yeah, that's the same, yeah, it's exactly, that's the same chart here, and I used to have charts with all 93 participants, but I'm gonna explain it out. So there's a lot, a lot more going on and all the other variables, which is, which is really what you have to look at to sort out, is it the net worth, is it, know, paper gains or what's going on?

Exactly what's happening to the foreigners and so forth, so it's not going on. Anyways, so the summary here is interbank markets important, I documented these four observations. Central banks deal with these disruptions or these frictions or these movements on these financial markets, and so this model here was designed to deal with that.

It's not about interest rate movements and inflation to say, but about these balance sheet policies and the haircuts, in this case, just the haircut, fitness center, bank charges. And I agree, you could think about many of the other there's a myriad of policies that the central banks introduce that are outside standard Taylor rule, interest rate setting considerations.

And I think models of that type in principle could help answer what they did.

>> Speaker 5: So, Harold, how would this compare to Giannikopoulos's collateral cycle?

>> Harald Uhlig: I have to reread that paper, I read that long time ago.

>> Speaker 5: Absolutely

>> Speaker 5: Okay, and something that's probably worth thinking about.

>> Harald Uhlig: Yeah,

>> Speaker 5: because he has this endogeneity that has major, real effects.

>> Harald Uhlig: Yeah, definitely.

>> Speaker 5: Well, thanks for pointing that out.

>> John Taylor: What do you come away from as an overall conclusion?

>> Harald Uhlig: Well, that these, that these policies by the central bank are really important, make a big difference for output, right.

They're not just affecting health of the banking system.

>> John Taylor: How do we see that in terms.

>> Harald Uhlig: Well, that's the, that's this chart here, right, the alternative policy was benchmark policy that the drop in output would have been twice as large. I mean it's half a percent is that large, right, it would have been bit more than 1% plus we get the, I mean if you do the discounted value, maybe they're the same here.

You might, if you look at that, which is, which is kind of weird but that's, that's, it has to do with this, with this good, like Kiyotaki karate dynamics of the, of the, of the leverage constraint and the return to capital. So I mean that's what it is, I mean it surprises us too, now was it a good idea for the central bank to do what it did, right.

I mean that's a different question because you don't have, I mean it's not wealth analysis, it's just saying this is what is quantitatively what we think has happened and what would have happened if they hadn't done this, and Roger may be right.

>> John Taylor: What assumptions are you putting here that you're most concerned about, what assumptions are you most concerned?

>> Harald Uhlig: Well, one is the flexible prices that bites us because something happens to inflation here that's kind of weird, let's just put it this way. And by this, due to this Fisherian assumption and probably want to replace it with something else, but we didn't. And the other part is the bank switching their status over the north source, having that randomly switch every period and also the connected unconnected switching every period, that's not attractive.

It would be nicer to really have the regions and have the hardware the regions, but it just complicates the model even further.

>> Speaker 10: Harald, if I could add on the takeaways, the main takeaway is this is something that no other country has done it, no other central bank has done it.

So this paper is beautiful in that it gives us a model that can be used to explain the Euro crisis, it wasn't really about inflation monetary policy and it was the collateral policy of the central bank that created the problem. And this is really unique to the euro area, so this is beautifully done, but you need to do that experiment because it's like actual policy was closer to 100% haircut, so it would have been wonderful if it was 3%.

>> Harald Uhlig: Right, yeah, no, I like that idea, I guess model with trash actually sovereign default built in, and it's in which the private sector haircuts are endogenous, right. Love to make them endogenous, but maybe by cutting down other parts of the model and reintroducing Mavic or maybe go after that.

>> Speaker 10: And this is also a very important lesson that was picked up during the pandemic when following effectively the Fed's intervention to incorporate debt market, the ECB suspended its collateral framework. And this is why we didn't have any problems during the pandemic, they suspended the, the threat of, of cutting off collateral eligibility for two years.

>> Harald Uhlig: So of course the different views in the euro system of whether you know what's good and what's bad there, but it's, it's politically highly contentious, but I agree, it's important, right. And maybe that's what the paper points to that, and one can look at the macro effects of doing that

>> Speaker 7: just to amplify the point. The pattern of haircut that you treat as exogenous.

>> Harald Uhlig: Right.

>> Speaker 7: That also the turning point is also endogenous to policy.

>> Harald Uhlig: Yeah.

>> Speaker 7: The Bumblebee speech, whatever it takes, which is the ECB saying that they will provide liquidity. Right, they're gonna provide liquidity by buying the bonds or financing bonds, etc, so it's not only in the origins of the crisis, but this other variable that's treated as exogenous is very much endogen to apologize.

>> Speaker 10: A number of policies like that followed because effectively my reading of the history that basically created a problem and then it reduced the cost of the problem with the SMT, then with the OMT, with the Bumblebee. So those kinds of efforts to contain the problem that was already created in 2010.

>> Harald Uhlig: Yeah, I'm very sympathetic to that perspective, so, and you can see when was the Draghi speech, was it 2012?

>> Speaker 10: 2012 2012, July, August.

>> Harald Uhlig: Starting around here, right, I mean you could potentially have three shocks here, right, you could have, you could have the Draghi speed shock built in and that would be another surprise, the private haircuts go back, I mean, I don't know.

>> Bob Hall: Could you remind us of all which agents, government and non government are treated exogenously, which ones?

>> Harald Uhlig: Yeah, all the government agents are treated as exogenous, so they're not optimizing, they're just very mechanical.

>> Bob Hall: What does this responsive mean if you're freezing?

>> Harald Uhlig: Well, it's.

>> Bob Hall: What do we get out of the experiment?

Well, the impulse response function when you're freezing, a first order important decision, or you think that monetary policy is not very important?

>> Harald Uhlig: No, I mean it is the haircut policy is very important, right, I mean, that's the point of this picture, I mean, I think of this, I'd like to think of this as giving central bankers a menu.

They can look at this and say scenarios, what happens if you do this, what happens if you do this? No, but the other agents, government and otherwise, are assumed not to do anything, is there some natural behavior of the other important agents, I'm confused. I mean, there's fiscal policy that's also mechanical, so they have a certain rule for the evolution of the bonds when they have government spending.

And then they have a certain debt policy, they want to return to certain, that's your GDP ratio essentially, but these, these are the.

>> Bob Hall: In your impulse response functions, even though these are important, the impulse response functions assume the absence of any response of those agents.

>> Harald Uhlig: Well, it's mechanical response, right, so the government, if they see that the debt is higher than where they wanted to be, they will raise taxes, I mean, that's taxes in the very.

>> Bob Hall: I'm still, I think we should have a conversation, I think we should have a private conversation, but happy to.