(Bloomberg) -- The rate banks pay on savings accounts hit the headlines earlier this year, when an outflow of deposits contributed to the collapse of Silicon Valley Bank and other lenders. Suddenly, the mechanics of how banks attract deposits — and what they actually do with them — became a hot topic. And even before then, there'd been a lot of discussion over why many banks hadn't passed on the surge in benchmark rates to their customers by raising rates on savings accounts. So what exactly do banks use deposits for? How do those deposits behave? And can that behavior change in different interest rate environments? In this episode we speak with Josh Younger, senior adviser at the Federal Reserve Bank of New York and formerly at JPMorgan, about his recent research looking at how banks pass on higher interest rates and what it means for their own exposure to interest rates. This transcript was lightly edited for clarity. 

Key insight from the pod: How do banks typically manage rate risk? — 6:06How do banks pass along higher rates? — 8:41The importance of the depositor mix — 12:25How deposit betas change across the rate cycle — 17:12How deposit stress changes the asset side of a bank — 21:24What are bank risk controls? — 23:39What does this mean for the transmission of monetary policy? — 28:15

-------------------- 

Tracy Alloway (00:09):Hello and welcome to another episode of the Odd Lots podcast. I'm Tracy Alloway.

Joe Weisenthal (00:14):And I'm Joe Weisenthal.

Tracy (00:15):Joe, what do you know about the dark arts of asset liability management? 

Joe (00:22):Only things that have been told to us about it on the Odd Lots podcast. 

Tracy (00:26):That's actually quite a lot.

Joe (00:27):No, that's true. It seems tough. Deposits can be flighty sometimes they can be stable. Sometimes when rates go up, they get passed on to deposits [but] sometimes they don't. Deposit betas change over time. That's about it. 

Tracy (00:45):No, that's pretty good. You're absolutely right about your characterization of deposits. And I think one of the really interesting things about banking is it is kind of built on this tension, which is that in theory, I can put my money in a bank as a deposit, and then I am theoretically entitled to pull it out at any time.

And I, for one, can't really imagine a business where at any single point I could see a majority of my customers suddenly vanish. But it is also true in practice that people who put their money in banks, they don't tend to move it around that much. And that is the thing, like the magic behavior that enables banks to actually lend money or buy assets like bonds and things like that.

Joe (01:38):Obviously part of the reason anyone is having this conversation still is that we all remember this past March with SVB, but I recall at the time and some of our episodes around then thinking about how, I don't know if paradox is the word, funny is the word or weird it is that on some models, bank deposits are extremely sticky and long duration.

And I think there's that stat, maybe it's fake, but people are more likely to get divorced than leave a bank is one of those things that people say. And if you have a customer, that's great. On the other hand, sometimes the deposits do leave and can leave in five seconds and it doesn't take very much. Do deposits exist on some continuum of sticky to long duration? Or is it more like Schrodinger's deposits? Either they're there or they're not there, but that there's no sort of in-between stay.

Tracy (02:38):We did that great episode with Joe Abate over at Barclay's about bank deposits. And we've been joking on the podcast since that in order to improve the transmission of monetary policy, everyone needs to go out, do their market research, find a high paying bank account somewhere and actually move their money. I am joking, but also kind of not, there is an interaction between banks and deposit rates and the effective transmission of monetary policy. I mean, banks are supposed to be the primary mechanism through which a lot of monetary policy is transmitted. And so the question of how deposits are actually working, so whether or not they're sticky, whether or not that stickiness changes in an environment where rates are going up or down, and also how those deposits sort of feed into monetary policy. That is a really big topic and I think we should do more on it.

Joe (03:35):The thing that really strikes me in this is in a bank run is the perfect example of non-linearity, right? And so something is one state and then something is suddenly a very different state. We've learned in recent times how tricky that is to manage and I think understanding the actual management of that non-linearity in practice is something we can dive into further.

Tracy (04:01):Right, and I did call it the dark art of asset liability management at the opening. And to some extent it is kind of opaque. Like no one gets a lot of transparency into how an individual bank is actually managing things like deposit and interest rate risk. So I'm glad we can do this episode and more than that, we are really doing it with the perfect guest. We have someone who's been on Odd Lots many times now. 

Joe (04:25):He's now climbing the ranks. Don’t you think? 

Tracy (04:30): He must be. 

Joe (04:31):He must be. I mean, he's up there.

Tracy (04:31):So we're going to be speaking to one of our perennial favorite guests, Josh Younger, formerly of JP Morgan, now a senior advisor over at the Federal Reserve Bank of New York. Josh, thank you for coming back on Odd Lots for like the fifth, sixth time. I don't know.

Josh Younger (04:48):I've lost count. But, it's great to be back. I really appreciate the opportunity. It's great to talk about this.

Tracy (04:53):And I should just mention that we're recording this on November 9th. And Josh, you have something to say to our audience now?

Josh (05:02):Yes, I almost memorized it, but we'll see if I get all the parts right. But these views are my own, and those are my co-authors from the paper that we wrote and they don't necessarily reflect the views of the Federal Reserve system, the Federal Reserve Bank of New York, or the Federal Reserve Bank of Dallas.

Tracy (05:17):I should mention that. I mean, I kind of figured that Josh would be an expert on this anyway, but one of the reasons we wanted to talk to him is because he did, he did just publish a paper at the Federal Reserve Bank of Dallas called “Deposit Convexity, Monetary Policy and Financial Stability” and it gets into the nuts and bolts of all the themes that Joe and I were just discussing in the intro.

So speaking of the paper, I mean, one of the things that you mention in the research is that you and your authors have sort of personal experience with the way banks actually deal with interest rate risk. Can you talk to us a little bit about how banks typically manage this? Like how are they thinking about things like deposits and interest rate exposure?

Josh (06:06):It's very tempting. I mean, Joe, you were saying earlier that deposits are there until they're not but they're also an overnight demandable liability, meaning you can get your money back whenever you want your money back and you tend not to. So it's a very behavioral process.

And so a lot of the work of managing the interest rate risk of a bank is really understanding that behavior because deposits are typically by far the largest source of funding for most banks. Depending on the kind of banking you're doing, you might have more or less of it, but it's typically a huge chunk. And it's not something you can, I mean, we had made an attempt to write it down on paper, but it's all based on the premise that people tend to leave their money at banks for long periods of time even if that interest rate is below what the broader market would consider risk free. 

Now there's lots of reasons for that. You certainly need bank deposits to live your life, right? So people usually call this a convenience yield because you use bank deposits for actual transactions as opposed to like money market fund shares, which you don't. But if as a bank you have one set of liabilities, you have to understand. And then on the asset side, sometimes there's some complexity, like mortgages are relatively complex. Lots of loans have embedded quote unquote options, meaning floors and caps and things like that. But those are sort of easier to model from a first principles like financial math perspective. And so you have a pretty good understanding of the interest rate risk on the asset side of the portfolio. And a lot of the legwork is spent understanding the liability side, specifically deposits. 

Joe (07:35):In light of that, Tracy mentioned that we did an episode with Joe Abate over at Barclays, and that was just a few weeks before SVB. There was no intention, but the question then was, well, rate hikes have gone up, or sorry, rates have gone up by then at that point the Fed had done a ton of hiking and yet you look around and deposit rates or the rates that banks were paying out on people's deposits hadn't gone up that more. So we explored this idea of deposit betas. In your research, what can we say just based on your research about the nature of how banks time the passing along of higher rates?

Josh (08:18):So it brings up the mechanism, which is we talk about this beta. What does it represent? Why does it move? Because there's a temptation to think there's one big book of deposits and a giant lever somewhere in the bank.

Tracy (08:30):We should just say, so beta is like the degree to which these two things are moving in tandem. So benchmark interest rates going up, are rates paid out on retail bank deposits or other types of bank deposits actually going up? 

Josh (08:41):So if a quote unquote “market” yields the yield you would get from a money market fund, for example, or close to that goes up by a hundred basis points, how much of that is represented in your deposit rate paid as a bank? You think as a bank thinks about paying your interest on deposits so we call it rate paid. But obviously the depositor is getting paid that interest, that's the rate you earn.

And a beta of, take, 25% would mean for every hundred basis points or one percentage move in market yield, 25 basis points or a quarter of a percentage point is actually reflected in the deposit rate. The temptation is to say, well, there's just this big book of deposits. There's a giant dial somewhere in the bank, and you kind of move the dial less than the market yield moved. And you do that to achieve some business outcome and there, there's sort of two things to think about with that.

The first is that the actual beta, the changes in beta or the level of the beta is to a great extent the consequence of the deposit mix, meaning how much of these things are in time deposit certificates of deposit, things with term exposure, and how much of that is, let's say zero interest checking. 

Tracy (09:45):Demand deposits, they're called, right? Where you can withdraw them whenever you want, like on demand. 

Josh (09:49):Anytime. So you could say, I want a demand deposit account, I want to get my money whenever I want. Or you say, I'll get my money in three months and I want to earn a slightly higher rate of interest for that. Typically those time deposits are priced at some spread or difference to the market yield and demand deposits typically are at a very low interest rate.

And so as customers move from demand deposits to time deposits, there's naturally more reactivity just because they're terming out their deposits. So that's one source of behavior and then there is to some extent a dial that reflects savings and checking account rates, particularly on the institutional side. So the rates paid to corporations and investment managers and things like that who tend to be much more sensitive to these economic considerations, they leave a lot more money on the table, basically when, when deposit rates are incrementally higher or lower.

And in that case, the question is, do I want to grow my business or shrink my business or keep it the same? So there's kind of a debate when you're talking about a beta, are you talking about the market neutral beta, meaning my market share, say, is constant? Are you talking about the beta required to grow my business? Are you talking about beta required to shrink my business?

There's been a lot of discussion about how size can sometimes be more expensive, not more profitable. And so you might want to shrink a business, for example. And so when we think about topics around interest rate risk management, I always like to think in terms of the sort of market neutral beta, which is what is the market beta I need to pay to keep all the deposits I currently have, or at least my share of the deposits in the system. Because that's more reflective of the actual underlying interest rate risk of deposit liabilities as opposed to the business management and planning decisions of an individual institution.

Tracy (11:44):Getting back to the idea that in theory people can pull their money out of banks whenever they want unless they're in something like a certificate of deposit. But in practice they tend not to and I remember this is something we spoke to Abate about, like there are a lot of people who have an account at a specific bank because they're getting something else out of it. There's a relationship there, a business relationship, they know each other, maybe they're getting a loan, a business loan or something like that. How sticky are deposits in practice and does it vary by deposit type beyond whether it's a demand deposit versus a termed out deposit? 

Josh (12:25):The type of depositor is really important. Maybe taking a big step back, so think about the system as a whole. Why do we think deposits are long lived? If you go back to the state banking era of the 1830s and go all the way to today, for most of that period, deposit balances are growing at or faster than gross domestic product, which tells you that the deposits in the system as a system are there for the most part. Now, there are more outlets now than there were back then to leave the banking system and go to say, money market funds. There certainly weren't money market funds in the 1850s, but like generally speaking, deposits tend to grow with or faster than GDP. So even if you take your deposit out of one bank, you're probably putting it at another bank.

So deposits in the system are very long-lived. But then there's the question of the type of depositor. And so on the one hand you have what we call retail depositors. I'm just using liquidity coverage ratio terminology. This is what the regulators have used to establish different types of deposits. So there's a common language for this now. Retail or stable retail deposits are individuals like people at this table. We tend to be relatively insensitive to rates on the whole, because we use our deposits for the most part, for lots of stuff. So we go buy coffee with it and we go to the movies or whatever. Like you pay your gas bill, even if it's a direct deposit from your bank account, it's still from a bank account, it's not from a money market fund.

And so the convenience yield, so to speak, of having bank deposits is relatively high. And even if most people don't keep a lot of money at the bank, there are a lot of people. And so that adds up to a big deposit base. So most of the useful money in the economy is in deposit format used by individuals. We call that a low beta deposit or a sticky deposit and that's again, a statement about the rate paid, not necessarily the propensity for that deposit to leave the bank. We can talk about that separately, but this is assuming I pay whatever rate is necessary to keep that deposit. So that's a relatively low beta, low rate deposit, not much reactivity to market yield. And then if you go into what the regulations term, the wholesale deposit space, which is really institutions, asset managers who have excess cash, corporations who have excess cash, they kind of split it into two buckets.

One is what's called operational deposits. Operational [deposits] are used to meet payroll and various expenses. If you import and export, you have to pay somebody with that deposit. And so those still need to be in the banking system somewhere. Now you can move them around between banks but the rate paid on those tends to be a little lower than market yields, for example, just because they still need to be a bank deposit at the end of the day. So there's still a convenience yield, so to speak to have that. And then at the high beta end of the spectrum, you have non-operational deposits, which is excess cash that's being left in a deposit account for some unknown reasons, it could easily go to a money market fund. It's money that corporations don't need to meet regular payroll and expenses so they're kind of keeping it because it's convenient to have some excess savings just in case maybe or something like that. But it's not being used on a daily basis. So in principle, it could be in anything. It's more of an investment in deposit accounts. So to keep those funds, you need to pay a relatively high beta, pretty close to one because otherwise they'll just go somewhere else or to a money market fund.

Joe (15:40):What happens when the Fed embarks on rates? And we've had this hiking cycle, maybe it's over but past hiking cycles. What does your research show really happens to deposit betas over time as a hiking cycle picks up?

Josh (15:56):So here we're thinking about the consolidated beta of the whole institution. So this is, if you take whatever fraction you have in retail, wholesale, operational, wholesale, non-operational, and even wholesale funding and term funding, you just think of the funding rate the bank has to pay to fund itself. That beta, just the ratio of that rate to the risk-free overnight rate tends to increase as interest rates rise.

So this is related to the propensity of corporations in particular, but also individuals to move their excess savings into money market funds when interest rates are above zero. And the opportunity cost of keeping them in a bank increases. That's not true for everybody. But on average and over the long sweep of history, it's tended to be the case that the beta goes up when rates go up.

Tracy (16:42):Wait, so this is different to how people normally think of the way bank deposit rates work, which is people tend to think of them as like a linear thing. Like rates go up a hundred basis points and then deposits will go up a hundred basis points, maybe not a hundred exactly. For obvious reasons that we've discussed on the podcast, but by like a set amount each time. But you're kind of arguing that actually it's a non-linear relationship and as interest rates go up the beta, the relationship actually gets stronger.

Josh (17:12):Yeah, exactly. So linear means a static beta. A static beta means 25% of every interest rate increase goes through to deposits and for example, if that beta increases, the non-linear relationship starts turning up and getting some convexity to it and that means that the last a hundred basis points are not the same as the first a hundred basis points, for example, from the depositors' perspective.

Tracy (17:33):So on that note, you mentioned money market funds just then. Like how much of this dynamic is driven by competition with money market funds? Because If I think about things that have changed in the financial system it feels like MMFs are a bigger presence, although maybe they're not compared to 2008, I can't remember the numbers, but it feels like nowadays there is this range of options for where you want to put your excess cash. And we have seen inflows to money market funds pick up, we have seen that play a role in things like SVB and Silver Gate obviously, or sorry, arguably. So I'm just wondering, how do banks compete with MMFs and how does that feed into that non-linear relationship that you just described?

Josh (18:17):

They're competing primarily with each other, I would argue. A bank has expenses that a money market fund does not, namely capital and liquidity. Money market funds are essentially a pass through instrument and so they don't have branches, they don't have payment processing services, they don't have all of the overhead a bank deals with. So it's kind of a misnomer to think that a bank could quote unquote, compete with a money fund purely on the yield that they offer, because It's just a very different business. 

Tracy (18:42):

But this is what I mean, like, banks are always going to have expenses, so they're always going to have to pay out less than a money market fund. Like they won't be able to pass on that full interest rate hike, right?

Josh (18:51):

Generally speaking, yes. And now they can compete with each other and they can try to source deposits that are associated with profitable activities. So we haven't really talked about what you're using these deposits for. If you use them to just hold excess cash, it's probably less attractive. If you have businesses like wealth management and market making trading that generate non-interest revenue, it might be more interesting. But you have to have something to do with the funding. That's why in some cases, banks were very happy to see these deposits roll off and when you have non-operational deposits supporting cash holdings, there's not much spread in that, you still have all of the expenses. So you have a relatively high interest deposit funding to support non-spread earning cash assets and that's really not what they're in the business of doing. And so under those circumstances, higher rates and runoff to money market funds is reasonably fine. But then again, if you have very profitable businesses that need funding, you want to bid to retain that funding because otherwise you have to shrink the business. And so I would think of this as more of, to some extent competing with money funds, but also banks competing with each other for those marginal deposits that support profitable activity.

Tracy (20:03):Joe, do you remember, I think it must have been just a couple years ago when banks were complaining about excess deposits.

Joe (20:09):Absolutely. It took me actually many Odd Lots episodes to even understand the coherence of that statement. Now in November 2023, I can understand why having too many deposits would be a problem, but it took me a while to wrap my head around that concept. Obviously in the case of SVB, that one was very straightforward because there were not many natural things they could do on the asset side of the business with all of that VC cash that they were getting in and so forth. They were in a traditional lending bank.

But that actually segues to my next question, which is, in a time of moving rates, when rates are stable, whatever it seems, banking doesn't seem that hard. But in a time of when rates are on the move, and as you mentioned the market neutral beta that banks want to keep their business the way it is, move around the deposit, move around the rate to hold on to your deposits. Moving around the assets does not seem very easy and they're illiquid and in many cases elongating. Talk to us about how the deposit mix changes or they have to pay out what kind of strain that pay places on the asset side for the banks.

Josh (21:24):Well, I think the key is that the bank is exposed to interest rate risk from these deposits. So earlier the question is how can you have long-term interest rate risk from overnight demandable liabilities, demand deposits, I can get my money back anytime, this is an overnight liability, right? And the answer is, if people keep their money there for a long time, that beta below one, meaning not all of those interest rate increases get passed through, generates a ton of interest rate risk for that kind of liability.

And you can think of it by splitting it up into what we call a replicating portfolio. What's equivalent to a 50% beta deposit liability? It's $50 worth of Fed funds linked floaters, perfectly floating rate liabilities and $50 worth of zero coupon debt. That generates an interest rate expense that's 50% of the Fed funds rate or the overnight funding rate. 

And so if you think about it in those terms, if I give you $50 worth of zero coupon funding, you would say, well, that's going to be worth a lot more when rates are higher than lower. So if I make money when rates go up and I lose money when rates go down, well that sounds a whole lot like short duration. And so that's why deposits as a general matter, have duration risk associated with them. And that duration risk is tied to the beta, not necessarily the runoff although that's important as well. Even if you're realizing exactly the expected runoff that you would anticipate given the level of rates and all the things we described, or if you have a static balance sheet even, as that beta moves around, your interest rate risk changes.

And so the job of a bank ALM department is to try to anticipate on the one hand that behavior and on the other, the mix of assets that satisfies all the other criteria and mitigates the risks associated with the interest rate exposure of the deposits on the liability side. So you have to pick the right assets, and to do that, you really have to understand your deposits. And so part of the point we're making in the paper is that the fact that the beta is variable, while not particularly controversial, people have kind of known this intuitively for a long time, has important consequences for what happens on the asset side of the balance sheet, especially when rates go up quickly.

Joe (23:33):When you say important consequences on the asset side, expand on that a little bit further.

Josh (23:39):So the bank is subject generally speaking to risk controls. So you can't just take all the risk you want. There are risk limits on the duration of your equity. There's a rule called the economic value sensitivity rule that's a basal requirement that looks at the sort of fair value sensitivity of the entire balance sheet under interest rate changes. And so the key is that, generally speaking, banks have to hedge their interest rate risk.

Now they can take positions relative to it to some extent, but they are due to regulatory and risk management requirements going to hedge most of it. And so when betas change relative to expectations, you have to do something on the asset side of the portfolio to adjust for it. And it's important to bear in mind that, and we'll talk about policy transmissions for things like bank lending channel type things, which we have in the paper. The securities holdings of a bank, the treasuries they buy, the mortgage backed securities they buy are not the only source of duration when they make a mortgage and keep it as a loan or when they make a long-term loan to a corporation. That's also interest rate risk.

Tracy (24:39):Right. Which is also moving around when rates are going up for obvious reasons. 

Josh (24:43):So banks have a lot of negative convexity. Negative convexity means that as rates go up, the exposure of the bank increases to that risk. So when rates go higher, they get longer duration, when rates go lower, they get shorter duration. This is just like a shortened option. And so banks have options through a variety of channels. On the asset side, that's through mortgages and other things and on the liability side, which is a huge chunk of it, probably coequal, if not larger source of negative convexity is the sensitivity of those deposits and the beta in particular to interest rates.

Tracy (25:18):Wait, could I just ask on this topic, the relationship between the deposit mix and the assets that a bank is actually investing in. In practice does a treasurer at a bank talk to the risk manager or whoever's in charge of the deposit mix? Because I always had it in my head as like, someone sat there and they have like X million or billions of dollars to play around with, and they can kind of within a certain framework do what they want with it. But do you think they would actually consider like, oh, actually this is what my deposit mix looks like. Here's a rough estimate of my duration. How much of this is sticky and isn't, and this is going to limit or shape my asset or my lending decisions? 

Josh (26:03):That comes to the risk limits. So I can only speak in generalities but when each institution has to satisfy risk limits, that means they can't take all the risks that they want. Now, where those particular things live within the institution, I can't really say. But, as a general matter, like if your deposits are losing duration or gaining duration, that affects all of the other activities that you're taking, which can contribute or detract from the duration of the asset side. It's all about the consolidated exposure of the whole balance sheet because you think about it as one giant interest rate exposure for the entire institution. You roll it all up, all the lines of business, all the lending businesses, all the deposit taking businesses.

Josh (26:43):At the end of the day, you have one number and that number has to be held within limits. That number is how many dollars you gain or lose when rates go up or down. That's the thing and you can compare that to the equity of the institution you can compare that to the overall risk of the institution. You can compare that to a bunch of different things and there's all kinds of nuance underlying that. But, generally speaking, the whole balance sheet has to be structured such that it does not have too much exposure in aggregate to changes in interest rates.

Joe (27:31):So I noticed reading your paper that one of the authors you cite regularly was actually a recent guest on the show, Itamar Dreschler and we talked about the fact that we talked about sort of this in the context of the 1970s and the impairment of monetary policy back then due to, it’s sort of different, but the nature of deposits and deposits moving out and that curtailing lending and so forth. But talk to us, what are the consequences from your research and maybe bring it forward to today? When you think about it, people try to understand the transmission of monetary policy, which as Tracy mentioned earlier, in many ways, goes through the banking system, like it kind of has to. What does your research imply for that? 

Josh (28:15):So a lot of our work is built off of what they've done and they basically make the argument that the fact that the beta is less than one means the deposit franchise has value meaning your funding advantage from deposits that cost you less than the risk-free rate to support is worth more when rates are higher and less, when rates are lower. Whenever you make money when rates move around, that's interest rate risk and so the deposit franchise generates short duration risk, which again, just means you make money when rates go up. That balances the long duration exposure on the asset side of the portfolio. You can think of this somewhat differently and say, the value of the bank is related to its funding cost. If its funding cost is lower, its value is higher.

And so the right quote-unquote “discount rate” to apply to the assets of a bank is its funding curve, not necessarily sort of the risk-free rate in general. It's an equivalent statement. We're just adjusting what they're saying and arguing that if the beta is variable, their beta is generally static because they're making a pedagogical point about how interest rate risk is generated in a bank portfolio. And we're just saying that variability and convexity can affect the decisions that banks make in a way that matters for a variety of participants in the market. So, you mentioned monetary policy transmission. The academic literature talks about the bank lending channel. The original version of the bank lending channel is more related to reserve requirements and the creation and destruction of reserves to facilitate monetary policy setting.

And therefore that has an impact on not just the pricing of lending, but the quantity of lending. The key is that banks can make fewer loans as a consequence of certain monetary policy decisions or more loans as a consequence of certain policy decisions. And so Dreschler and his collaborators make an argument that deposit rates are a secondary channel for this, or now that there are no reserve requirements, a primary channel for this, where if the deposit rate goes up, your expected funding cost for certain loans goes up, that gets Fed through into pricing and quantities and therefore there's a channel through which policy rate decisions can affect bank credit allocation decisions. Our argument in this paper is one, the nonlinearity of deposit rates amplifies that channel to some extent just because the change in expected funding cost for a given loan is higher if you think betas are going up when rates are going up.

And the secondary effect is that if the bank is losing capacity to support duration in its lending, which is another way to say this, if the deposits are getting shorter, that means the assets have to get shorter. And that means there's less space to make long-term fixed rate loans. Mortgages, for example, are long-term and fixed rate. Some commercial term lending is long-term and fixed rate. A 10 year loan commercial real estate lending is sometimes long-term fixed rate. That in addition to all the securities holdings that they have are all sources of duration exposure. When deposits get shorter, the ability to generate long-term assets included within that long-term fixed rate loans is simply lesser as a general matter.

Tracy (31:26):If deposits are getting shorter in a rising rate environment, like how does that interact with the macroeconomic outlook as well? Because I take the point about funding costs and things like that, but also if interest rates are going up, it's probably because the economy is doing relatively well. And so in that environment, maybe banks would want to lend more, but you have this sort of natural constraint that's going on through the deposit's impact on the asset side.

Josh (31:59):It's a duration capacity constraint and I should be clear when I say deposits are shorter, I don't mean they're more prone to runoff. I mean that the beta is higher. Therefore, in this example where I have zero coupon debt and floating rate debt, that mixture changes such that I have less long-term zero coupon debt and more floating rate debt when the beta is higher. So it's just a higher degree to which the cost floats. That means less interest rate risk in my liabilities, we can call that shorter because it has less risk. And when my liabilities get shorter, my balance sheet gets longer because I'm short my liabilities, they get less duration, then the whole balance sheet gets more duration. And so there's less capacity to support the long-term fixed rate lending I was referring to. So this is where it's about quantities not pricing.

Tracy (32:42):Oh, I see. 

Josh (32:43):It's within the constraints on the bank's ability to take interest rate risk, there's simply less room for long-term fixed rate lending.

Tracy (32:50):And then just on the general impact of having non-linear deposit beta as rates are going up, does that mean like as rates increase that the effect of monetary policy tightening gets amplified through the bank channel?

Josh (33:07):Well, there's a lot of academic literature about it. I wouldn't necessarily say amplified, but there’s definitely a lot of academic work that argues that this is one of the ways that monetary policy affects the economy because it's not necessarily an amplification mechanism, it's just a mechanism. Basically our argument is in addition to the traditional channel there's this sort of duration channel that affects a very specific subset of lending but still affects it.

Tracy (33:37):Should we all be researching higher paying bank accounts? Is that the suggestion? Should we do this?

Josh (33:44):There are companies that do that. I mean, I wouldn't necessarily say that's the solution to this particular problem. I've always found it interesting that this is generally true. Like this deposit beta effect is true this cycle, it was true last cycle. Bank accounts have a value that's reflected in these betas and it is frustrating for people who want kind of like a no arbitrage type model where no money is left on the table. At the end of the day, the behavior of depositors is clear and the value they see, which is not necessarily reflected in the interest rate, is real. This is the risk that the bank has, which makes it very hard to manage. It’s behavioral. 

Joe (34:31):I still think we should pay banks for holding our money. 

Tracy (34:35):Joe, don’t. You’re going to ruin it. 

Joe (34:37):No, I’m kidding. I know in theory we're lending them money, but come on, they're providing a service, I get a card.

Tracy (34:42):This is your negative interest rate thing all over again. 

Joe (34:46):They're providing a service, they run a nice website, I can pay my bills. I really think we should be paying them. I don't understand how they pay us. I can't take interest. 

Tracy (34:57):Producers, please cut this bit of the podcast out.

Joe (34:59):No, keep it in.

Tracy (35:01):Well Josh, really appreciate you coming on the show for another appearance. Again, I've lost count. That was a really great insight into how this kind of opaque aspect of the banking business, but really the heart of the banking business actually works. So thank you so much.

Josh (35:18):

Well thanks for having me.

Joe (35:19):I think Josh gets a mug now because he's been on so many times. 

Tracy (35:21):Have we given you a mug before? 

Josh (35:24):I have a mug. I'll take another mug, but I have a mug.

Joe (35:27):Good to know. 

Tracy (35:28):You can take a mug and next time you see Jerome Powell or something [give him one]. Joe, have I convinced you to open up a certificate of deposit or like a different bank account yet? 

Joe (35:53):No, I'm going to email my bank tonight and ask if there's any way that I can pay them a little bit for the great service they provide me.

Tracy (36:01):I think you've missed the entire point of this conversation. No, but it is really interesting. In some respects it's always surprising to me how much new information there is to learn about the way that finance and economics actually works. Because you think that throughout history people have modeled deposit betas as sort of linear or having a stable relationship with benchmark interest rates, but then a paper like this comes along and actually says, no, it doesn't really work like that. 

Joe (36:31):

It really is fascinating how young it all seems, how new it all seems. . There's a line in Josh's paper saying first deposit convexity amplifies the bank lending channel, monetary transmission, eg. and then he cites a paper from Bernanke and Alan Blinder in 1988. Like, that's not even that long ago. The idea that some of these really interesting ideas or sort of core ideas, some of them, the key paper it feels like it's day one in our understanding of this. Sometimes it feels like that.

Tracy (37:03):No, absolutely and we're also still trying to understand it in a very different environment. And we didn't even really touch on this, maybe Josh mentioned it once, but we've had QE. Interest rate hikes are happening at the same time that the Fed is reducing its balance sheet and that impacts assets as well. We also have things like the reverse repo facility, the RRP, which we didn't have before. And so we're still learning things about how all of this works against a very different financial and economic backdrop. So there really are so many moving parts.

Joe (37:39):And then just the idea of the pressure that this puts on the asset side and the constraints that come up and the idea that one thing moves and then you can only be so limited in how quickly you can hedge or alter or change your asset mix for the bank. So lots of interesting stuff there.

Tracy (37:57):I wouldn't want to be a risk manager at a bank, but I also wouldn't want to pay them for managing my money so I'm a hypocrite. Alright. Shall we leave it there? 

Joe (38:05):Let’s leave it there. 

©2023 Bloomberg L.P.