:
You're listening to At Any Rate, JP Morgan's global research podcast, where we take a look at the story behind some of the biggest trends and themes in fixed income, currency, and commodity markets today. Welcome to this special edition, where we take a deeper dive into swap spreads in the US markets and how that connects to a kind of term premium that we call term funding premium. I'm Ipeko Zil, Senior Interest Rate Derivative Strategist with JP Morgan. And today I am joined by my colleague, Shunir Ramaswamy. Co-head of US rate strategy. We're recording this on Thursday, May 2nd, 2024. And our comments today are based on our publication on this topic from earlier this week, which is available to institutional clients of JPMorgan on our website, JPMorganMarkets. So Shrini, let's take this all the way back to the beginning or actually even one step before the beginning. Can you explain to our listeners what problem we're trying to solve? Yeah, let's do that. So let's actually begin by defining what SwapsPress actually are. In the US markets, the convention is to define a swap spread as the swap yield minus the US treasury. So for example, a five-year swap spread would be the five-year swap yield minus the five-year on-the-run treasury note yield. Now, back in the days when LIBOR was the benchmark for the swap market, the swap yield had some credit component to it, and therefore the swap spread had some credit component to it. And the swap spread had some macro risk exposures that people generally understood well. But ever since SOFR replaced Libor as the short rate benchmark in the swaps market, that has changed the very essence of what swap spreads are. It has become much more nuanced, and it represents actually a kind of term premium, as we will discuss later here. We have actually written about this connection, you know, quite a while back. Several years back, we made the connection between swap spreads and term premium. But now the benefit is we have a few more years of historical data under our bells. And we can actually flesh this out a bit further. And that's what we've done in our work. So long story short, term funding premium, as we will call it and define it. It affects swap spreads at every single maturity. It actually affects long end spreads more than it does short end spreads. which means by focusing on the term structure, you can extract information about term premium. And conversely, if you focus on term premium or term funding premium, you can adopt what we call a term structure first approach to thinking about swap spreads and modeling swap spreads. And we think by doing that, you should be able to do a better job with respect to forecasting that. So that takes us to the beginning. So maybe let's start getting into some of the details. Can you explain what we mean by term funding premium? Yeah, actually the place I would like to start is actually to look at the treasury floating rate notes or the FRN market. So treasury FRNs are basically two year maturity instruments and they pay returns that are linked to three month table. So they actually pay three month table returns every three months plus a spread. And basically over the life of this instrument, if you held it, your returns will be very similar to owning bills and rolling them every quarter for a two-year period, but you'll actually earn an additional spread on top of that. Now, why do you earn this additional spread? It's because the investor is lending money to the US Treasury for two years upfront instead of lending it for three months at a time and then rolling it every quarter. So the additional return, the market price is in. for committing principal amounts to a longer duration. That's what we refer to as term funding premium. Now this is directly observable in the case of FRNs, but it's a component of yields everywhere, including in nominal treasuries. And what we have done is to actually look at that a little bit further and extract that information from swap spreads or the term structure of swap spreads. Okay, so let's go there next. Can you explain how swap spreads contain this term funding premium? Yeah, let's start by reflecting on what a 10-year swap yield and a 10-year treasury yield actually are. So SOFR is a risk-free overnight rate, which means the 10-year swap yield basically reflects the market's expected return from rolling overnight risk-free loans every single day, rolling it over a 10-year horizon. A 10-year treasury yield, on the other hand, reflects also returns from a 10-year term risk-free loan. But here you're committing the principal upfront for 10 years to term. One key difference between these two strategies is this component of terming out your principal and therefore the swap spread basically, to a considerable extent is a reflection of that term funding premium. Yeah, so that makes a lot of sense. And so, so far we've defined term funding premium and how swap spreads contain this term funding premium. So can you also explain the connection this term funding premium and the term structure of swap spreads? Yeah. As we have discussed so far, we do know that swap spreads to a considerable extent in any given maturity point reflects term funding premium. But that's not all that goes into swap spreads. Swap spreads are also impacted by other factors, liquidity premium, top-term flows, hedging demands from some sector of the market. swapping of issuance, et cetera, et cetera. There's a lot of other stuff that can also impact swap spreads. So the question is, how do you extract term funding premium information knowing that these other influences can also exist? And the way we do that is actually we look at the entire term structure. Since term funding premium impacts swap spreads at every maturity, but in a manner that's proportional to its modified duration, you can actually look at the entire term structure and extract this component that is common to all the different maturities. So specifically, imagine taking a snapshot today of swap spreads at every point. You pick every benchmark treasured point. So you pick the twos, threes, fives, sevens, tens, twenties, and thirties, and you plot the maturity match swap spread versus the modified duration of each point. Now imagine drawing the best fit line through this plot. Well, the slope of this line is basically a reflection of term funding premium. The higher the term funding premium, the more steeply inverted the swap spread curve will be. So that long end spreads will become much more negative relative to front end spreads. So based on that observation, we actually use this link to define term funding premium. Specifically, we define it as basically the negative of the slope of that fitted line. which means our measure of term funding premium, we extract that from not any given swap spread, but from the collection of swap spreads across different maturities. So conversely, you could also say that if we think term funding premium is going higher or lower, that will have an impact on swap spreads at every maturity. So, Ipek, so far we've described the term funding premium and its link to swap spreads. Why don't you give our listeners an outline of what comes next? Sure. So basically I would describe our recent work as having four major elements. So the first one, we've already talked about it. We define a way to measure term funding premium that's priced into the market on any given day. And we infer this from the term structure of swap spreads. But the quantifying it is of no use to us in markets unless we can forecast it. So the second major element in our work is that we have developed an empirical model that links term funding premium to supply and demand factors, which allows us to project this quantity into the future. So, when I say demand side factors like Fed balance sheet developments are on the supply side US Treasury supply dynamics. So we combine all of these and translate that into a view on term funding premium. Third step. We talk about how we can use these term funding projections to make a first approximation baseline projection for swap spreads at any desired maturity sector. And finally, the last step, we develop secondary models that help to refine our swap spreads forecast further by incorporating sector-specific influences. So I guess you can say that this is the roadmap of our recent work in this area. Okay, let's actually dive a little bit more into the details here. Can you maybe describe the factors that we use to project a funding premium on a forward looking basis? Sure. So as I said that we look at it on a supply and demand, we look at supply and demand factors. So on the supply side, it's relatively easy. It turns out that the best quantification of US Treasury supply for this purpose is to measure Treasury issues in duration weighted units. for example, how many billions of 10 year note equivalents are issued in any given month. So this is the important supply side factor in our model. And then we have important two demand side factors. The first one is the size of the Fed's balance sheet. So as we know, when the Fed conducts quantitative easing, it does so by purchasing US treasuries. So the balance sheet growth should pressure term funding premium lower and vice versa. And then we also use aggregate AUM at major bond funds as a second demand side factor. So the fourth factor, it turns out that the size of the overnight RRP facility is also a relevant factor. And the channel for this is a bit nuanced and interesting. And we think the channel is actually like a substitution mechanism. So since we're already controlling for the size of the Fed's balance sheet in our model, a higher RRP balance reflects drain of liquidity which should pressure term funding premium higher and that is in fact what we see in our model. Okay, let's try for the finish line here. Why don't you walk us through the remainder of the work? Sure. So we use this model together with our projections for the underlying drivers to take a view on term funding premium going forward. We think this premium should rise a little in the near term, which means that swap spread curves will be slightly more inverted going forward, all are SQL. But that is not all. As it turns out, swap spreads are also impacted by numerous other influences, which can also be actually different in different sectors. So we developed secondary models for these adjustments, if you will, one for each sector. That may be too much detail for this podcast, but our clients can access this in the publication we referenced earlier. So using these models, we can also refine our swap spread forecast a little further. And when we put it all together, what we see is that we think swap spreads will likely be biased somewhat wider across much of the curve in the coming months. And the exception to this view is the long end. So in the 30 year sector, we think competing influences make it a wash and we remain neutral on third-year swap spreads. Okay, Bekkah, thank you so much for joining me today. I think that's more than enough that you want for one podcast. So to all our listeners, hopefully you find our perspectives and views interesting and useful. You can find more details in our published work, which is available on jpmorganmarkets.com to institutional clients. Stay tuned for more episodes of At Any Rate, JP Morgan's global research podcast series. This communication is provided for information purposes only. Please read the JPMorgan research reports related to its contents for more information, including important disclosures. Copyright 2024, JPMorgan Chase and Company, all rights reserved. This episode was recorded on May 2nd, 2024.