Odd Lots - JPMorgan's Josh Younger on Rate Derivatives and Volatility Ahead of the Election

Episode Date: October 26, 2020

For months now, traders have been positioning for a major volatility spike around the November election. But what are markets really expecting, and how are investors hedging? On this episode, we speak with Josh Younger, a rate derivatives strategist at JPMorgan to discuss how he goes about finding signal in the market's noise, how traders are positioning, and what could be a shock to the market on election day.See omnystudio.com/listener for privacy information.

On April 4th, 2023, around 2 in the morning, a man was found stabbed multiple times on a sidewalk in downtown San Francisco. Hey, who did this to you? What happened next turned the story into a political firestorm. Reports have identified the victim as Bob Lee, the founder of Cash App. From Bloomberg Podcasts, this is Foundering, the Killing of Bob Lee, beginning April 16. And welcome to another episode of the Oddlots podcast. I'm Tracy Allaway. And I'm Joe Wisenthal. Joe, we have been watching The VIX, haven't we?

We have. I mean, it's been, we've had a lot of episodes this year for pretty obvious reasons. Looking at the VIX, exploring volatility, what signal is embedded in it, what it tells you, what it doesn't. But people love hearing about the VIX. So I'm always up to talk about it. You know, as I said that sentence, I kind of realized what a weak intro it was. And I need to put more thought into how I start these things. But yes. No, it's fine.

It's fine. We talk a lot about volatility. We talk a lot about the market structure of volatility. But lately, we've been talking about volatility more in the short term, in the sense that we have this really big. potential tail risk on the horizon in the form of the U.S. elections. And if you look at the VIX, even though it's relatively low, if you look at the actual term structure or the curve of the VIX,

you can see it's quite, well, it's not as elevated as it was a few weeks or months ago, but it's still elevated compared to normal right around the time of the elections and for a few weeks after that. So that's a bunch of people pricing in the risk of something unexpected happening in the U.S. elections. Yeah, absolutely. I mean, this has been building for a while. I shouldn't note we are recording this Tuesday, October 20th.

So we are literally two weeks away from November 3rd Election Day. Although by the time you hear this, I think based on when we expect to release it, it's going to be about a week. But yeah, you're absolutely correct, which is that for several weeks or months now, the volatility term structure has been super bit up not just at the beginning of November but for a long period as the risk of a long drawn-out count is a possibility some of that has come in a little bit less concerns about that as some of the polls have widened between the two I mean I guess there's always sort of going to be some expected volatility to round election but the clear thing is because

of so much going on and all the uncertainty and economic situation there is a lot of lot of anxiety and uncertainty about the election and its ramifications. Yeah, it kind of, for me, it brings to mind three questions. So number one, are markets pricing in the results of the U.S. election correctly? So are they, you know, are the polls right this time versus what we saw in 2016? And secondly, our investors, we've discussed this before with Chrysidiole most recently, but are investors so well hedge now around the election that it's going to be really tough to actually spark a volatility event? Or third, are we mistaking that volatility premium?

And is it, in fact, market complacency? Should investors be more worried? So those are all sort of interconnected questions. But we're going to be discussing them with an odd bot's favorite on this episode. I think this is going to be his third appearance on all. thoughts. It's Josh Younger, head of U.S. interest rate derivative strategy over at J.P. Morgan Chase, and he's been writing quite a lot about this. Three episodes, does that when you get the odd lot toad tote bag? Or is that that? That's right. You get the sweatshirt. Josh, we'll send that over to you

after this episode. Welcome to the show again. Yeah, thanks. It's great to be back. So I guess, just to begin, could you maybe lay the scene for us in terms of what we're seeing in market positioning on volatility. We spoke a little bit about how at one point it looks like markets were building in quite a big risk premium around the elections and beyond it. You wrote about this that it looked like investors were bracing possibly for a contested election, but we've seen that risk premium come down recently. What does that say about where the market actually is positioned? Yeah, so maybe it's best to start by getting a sense of how we actually extract these numbers, because it's sort of easy to say, well, the market's pricing X, Y, or Z for the election,

but ultimately we need to get this from some traded instrument and the price of that instrument. And with options markets, the price of the option is proportional to or related to the potential for large moves. So that insurance is worth more if the likelihood of a large move is grand. And so you have to pay up for that insurance and vice versa if you think the likelihood of large measures is lesser. And so what we do is we take options that expire after the election and we compare the price of those options adjusting for the extra time value of them to the price of options that expire before the election.

And from that you can get a sense of how much additional risk premium there is simply because of the fact the election falls in that window. And it's tough to do this with a lot of precision in most asset classes. So you were mentioning the VIX. Those VIX futures are typically calendar months. So it's rather hard to isolate the election date itself, although options do trade at that level of precision. It's usually best done, especially if you're going to compare different asset classes with

these like benchmark type structures. So one month options on rates and equities and foreign exchange and commodities and different and equity indices and so forth. And you can compare that to say three month options on the same. And we did that back in late August, early September. And you can pull out of that the risk premium and say, isolate it to just election day. That's an assumption you can make.

And you say, how much extra risk is there on election day relative to some background level of volatility? Because obviously, if markets are volatile already, then volatility risk is going to be more expensive. And so when we did that exercise, we got something like seven to eight times the typical daily move priced for election day across really a broad range of asset classes. That's not just the U.S. equities, the VIX, the S.P 500, that was interest rates, U.S. dollar interest rates. It was foreign currencies to some extent, especially dollar C&Y exchange rates.

So that makes sense. That's a geopolitical element of the election. You can see it in gold. You could see it in oil. You can see it in credit markets. Basically everywhere that options traded, you could see some. something between six and eight times the typical daily move price for election day. And that's, again, subject to the assumption that it's all isolated to that particular day.

Now, that's not a great assumption because seven, eight times a big number. So what we could equivalently say is, well, not only is election day pricing a lot of all risk, but that risk is spread out and persists past the election itself. That's mathematically consistent. And so we interpreted that, not to say the markets expect a 25, 30 basis point moving interest rates or a 10, 15% movement equities on election day, but that the market was assuming a persistent, elevated volatility environment for weeks and potentially a month thereafter. And so that just increased the value of those options. Now, that since come down. And now that's more like three or four times.

go back over the past 25 years and why not further? It's because we really didn't have much of an options market earlier than 25, 30 years ago. So you really have a small sample there, and many of those elections weren't that close, especially in the late 90s. But if we go back over that period, we can say, look, typical event risk premium is two to three times, the background level of volatility. And we were looking at 7 to 8, and that's come down to 3 or 4, but it's still multiples of what you would expect for a typical election cycle.

That was great. On recent episodes, I feel like I've been asking really remedial questions about volatility with our recent guest. I asked why VIX curves, futures curve slope upward. I'm not going to ask about that. Now I'm going to ask it even stupider question, though. When I pull up a quote of the VIX and I see right now it's at 2880, for people who are curious about that, 2880 what? What is that number actually represent? So that's a break-even annualized move. So if you were to buy a one-year option at an implied volatility of 28%. Okay. The equity index would have to move

28% for you to make $0 relative to the insurance cost upfront. So you can change the timeframe of that. Obviously, if you're paying for a three-month option, you'd have to have less of a break-even because there's less time value. The idea with an option is you have the intrinsic value of the thing, which is how much the strike price differs from the current price of the asset. So if I bought a call on the S&P at 2,300, it's trading at 2,400. That means I can buy something worth $2,400 for $2,300. It's worth $100, right? And then the question is, if that option expires in six months to a year, if I have protection for a year, that's worth more than protection for six months. And so that price is just going to be higher, because

because of that time value, but what we want to do is put everything on equal footing apples to apples. And so we say, what is the break-even change over the expiry of this option to get sort of a neutral, you know, a neutral payoff? That's where I'm agnostic to buying the option or not. That was great. I think Joe just managed to sneak in the actual question-wise, the VIX curve upward sloping, and you very nicely answered it. So thank you for that. I have a slightly different question, which is you talked about how you've been studying the risk premium across a bunch of different asset classes. And I think you found that it looked like it was bigger or higher in the interest rates market than, for instance, in the stock market.

Why do you think that's happening? Why do you think some asset classes or some markets appear to be pricing in higher levels of expected volatility? So I think some of it makes a lot of intuitive sense. So if we think about dollar CNY, so Chinese currency exchange rates, I think it's fair to say that a Trump presidency versus Biden presidency could lead to very different outcomes there. And so the binary event has a lot of potential impact on pricing. And so you'd expect the insurance value of options that protect you against changes in that price to be very high.

So that's pretty intuitive. Interest rates is a little counterintuitive only in that. If we think of interest rates is driven by two things. One is monetary policy and the other is fiscal policy. So on the monetary policy side, how does the Fed move short-term rates to react to economic conditions? What is their framework? Who's in charge? That is important for even long-term interest rates because if I'm thinking about buying a 10-year bond, I could equivalently just roll three-month bonds for 10 years. So there's an expectations element to this. that 10-year yield should be roughly equal to the average expected short-term yield over the

next 10 years. And so whoever's in charge of the Fed and how they make decisions affects what that short-term rate is going to do. The second is on the fiscal side, which is that 10-year bond is not a rolling basket of three-month bonds. It's a 10-year instrument. The government needs to borrow money across a range of tenors, sometimes for longer, sometimes for shorter.

And so the amount they need to borrow and the way that they choose to do so, you know, So, like the different maturities they focus on, all of that tells you what the level of longer-term interest rates is going to do relative to this expectation. So this is often called term premium, right? It's the premium that is assigned as something that locks you into a position for 10 years and or the format in which that risk comes, which is a treasury bond versus corporate bond versus a rolling basket of short-term securities, et cetera. So, you know, that fiscal outlook is not that different, frankly, across.

the two candidates. The committee for responsible budget has estimates for both campaign platforms. And I think over the next 10 years or so, they have the stock of U.S. debt for Biden at 127 percent of GDP, and for Trump at 125 percent of GDP. So it's a very similar fiscal outlook. And that's consistent. The format of that debt increase maybe is different. In one case, it's more spending. And in the other, it's more tax-related, but tax expenditures and fiscal expenditures are fundamentally deficit spending anyways. And so, like, the outlook isn't that different. And then if we turn to the monetary policy side, I think it's fair to say both candidates prefer a dovish outlook. And Powell's term doesn't expire until 2022. So there's some time there. And so, and the Fed just completed a review of how

they make decisions, and they're unlikely to do a wholesale review over the short term. And so, like, The outlook isn't that different for interest rates across the two candidates. I think the reason why you'd get that knee-jerk pricing of event risk is on the one hand, everyone remembers 2016, and there was quite a bit of volatility in interest rates. And by many measures, the reaction of the interest rate market was much more chaotic to that unexpected outcome than equities and FX and other asset classes. So a lot of that post-election volatility in 2016 was really concentrated in interest rates. a lot of muscle memory there. Yeah, I remember that.

So, like, stocks, you know, mostly went up pre-Trump. They then continued going up during most of Trump's tenure, obviously. But that the tenure moved, the long end of the curve really moved violently after 2016. Probably people thought maybe Trump would deliver some sort of growth jolt due to taxes or spending that would cause a rate hike. I remember that day, November 8th, November 9th. pretty, some of the biggest moves ever. And so you think that sort of, the memory of that sort of looms large in terms of a potential rate turning point. Yeah. And I think the, one of the funny

experiences from that time, we have a model that we trained. It's a machine learning model that looks at different market signals and tries to come up with a view on treasuries for a week. So it says longer short, should I do 10% size, 50% size, 100% size. And it uses it input. I think we have 1,200 things going into it, but it's all stuff you can see. It's economic data. It's pricing. It's the yield curve. It's all levels. It's the depth of the market and different permutations of that. It doesn't know there's an election. And what I thought was really interesting about that model is that it quote unquote got the election right, meaning it was max short the rates market in the lead up to the election date itself. And so if it doesn't know there's an election,

it doesn't know what the polls are. It's not like it knew something that Nate's or didn't, it was basically just saying, like, I look at this set of market data and economic data, and I'm of the opinion as this agnostic model, that this is a rising rate environment, economic growth is accelerating, inflation is firming, and rates should be higher. And the question is then, why were they so low? And I think you can argue, I mean, it's not a particularly robust argument, but you could say, you know, the narrative that makes that reasonable is to say the election risk, the binary election risk, independent of what its impact, the impact of either outcome would have been on markets, just the fact of this big event coming up kind of held the

market back from pricing in what was otherwise like a very supportive background. And then once the election passes, you just knee-jerk price in six months of economic developments over a week or two. And so if that's the case this time around, even if these deficit outlooks, which are very similar are the same. It could also be the case this time that the event and uncertainties around the event, and even if it seems reasonably high probability of one outcome or another, everyone's saying, you know, you never know and polls could be wrong and so forth. And we might talk about that later in the episode. But just the fact of this one day that matters

and we don't really know what's going to happen can hold things back. And then in the wake of it, potentially independent of what the outcome is, you could front load all of this repricing very quickly, and that's a very volatile environment. Well, so why don't we talk about that point? Because we have the risk premium built into various markets. So you could argue that in many ways, investors are quite well positioned for something actually happening in November. But on the other hand, you could argue perhaps that the polls are wrong or there's still a chance of something completely unexpected happening. And in that sense, they might even be considered complacent,

despite that higher risk premium. So what's your take on that particular argument? So I think there's not a ton of trading that generates these price changes. So the stock of equity positions, the stock of rates positions is just enormous. And a lot of those just have to get rebalanced when there's a big change. So you can get these self-reinforcing spirals simply because like changes in the environment, especially among, you know, we often think intuitively in terms of retail investing, but, you know, the most important transfers of risk happen with institutions. They're just much larger, and they have very different incentives. So if we think about an insurance company, for example, if rates are rising, they need to shed

duration mechanically because of the way that their risks are evolving. So they're pro-cyclical with the market, and they're just very large. in size, billions and billions of dollars. And so they're less well hedged just because their risks are much more complex. On the equity side, a lot of exotic-type products can generate convexity as well. And those, you know, you can hedge that, but the risks embedded in exotic instruments, like these structured notes and other things, are very non-linear. In many cases, there are sort of binary-type risk and other things that it's just hard to keep up, frankly.

And so even if you have decent hedged positions out there, I think on the one hand, and this is more a rates thing than an equities thing, like if the environment's changing, that's going to blow through all of these protections because it's impossible. If you're fully hedged, you're not making any money. So nobody's never fully hedged. otherwise, you might as well just not be participating in investing. And so those residuals can get very big, very quickly. So I guess a sort of long-winded way to answer the question, which is I think it's sort of functionally impossible for everyone to be fully hedged in the event of a sufficiently large macro shock,

which again doesn't need to be anything other than the passing of the event itself, which in some sense was the case in 2016, you can generate these spirals that keep volatility elevated for a long time. It's important to say that presidential elections are rarely that thing. So 2016 is very much the aberration. Usually you get a couple of decently large moves in the lead-up and the aftermath of the election event itself. But even in 2000, when we didn't really know the result for weeks, like volatility was modestly elevated, but it really wasn't anything like 2016.

And then if you go back further, you know, in the 80s, these weren't really particularly contested elections. Even in the 90s, they were pretty solid results. And so now we're talking about the 70s. And I think at that point, the utility of the analogy kind of drops off quite a bit. I'm June Grosso, inviting you to join me for the Bloomberg Law podcast. Every weekday, we help you make sense of the legal stories that shape the nation and the world. Listen for complete analysis of the biggest court cases. latest actions from Congress and regulators and the legal moves driving the markets.

From corporate law to constitutional law and from state courts to the Supreme Court. At Bloomberg Law, we go beyond the day's headlines. We speak with top attorneys, judges, scholars, and policy experts to break down what the rulings really mean. We do this every weekday, then bring you the best conversations in our daily podcast. Search for Bloomberg Law on YouTube, Apple, Spotify or anywhere else you listen. On the East Coast, listen as you start your day.

And on the West Coast, catch up in the evening. That's the Bloomberg Law podcast with me, June Grosso. Subscribe today wherever you get your podcast. Talk to us about the hedging needs of investors this year, because it's been a really weird year. So, I mean, you have some investors that are probably sitting on fantastic gains that they never would have expected to, reap in a year in which we have this terrible economic outcome and a pandemic and so forth. So that's

maybe some inclination to lock in those gains. We've also had a lot of people miss this rally in some way or another, either because they were just underinvested throughout the whole thing or maybe just overly hedged throughout this whole time due to fears about some second wave or second shoe about to drop. So as people look towards the election, how has what we've seen building up to it influenced their desire to hedge against various outcomes? So I think there's two really interesting behavioral elements to this. The first is if you're sitting on very significant gains, you're very highly incentivized not to lose them.

And so if we're trying to figure out why options markets would price so much excess risk around this particular election, on the one hand, there's just a memory of 2016. But that's not a great argument because that's just saying, you know, I don't want this to happen to me again. But more importantly, I think we were talking about time value before, like the election's not that far. And so the actual dollars you need to put up to protect yourself is not that many. And so the price of that option can adjust significantly on relatively small dollar amounts of trading. And so if you're sitting on significant gains as a hedge fund or individual or asset manager or anybody, you know, that insurance cost is relatively low.

so you might as well buy it. And the converse is true too, which is why would you sell insurance premium at a relatively small dollar amount when you look insane if you get it wrong? That's a really bad look for selling those options. No one wants to have that conversation with their boss. Why did you sell a bunch of puts on the S&P like two weeks before the election? So if you have a bias towards buyers of this risk, they don't have to put up a ton of concrete dollars to get it, and the incentives are very highly skewed in favor of

protecting gains, you're going to get very expensive options optically. And this is where that adjustment, if the price of a one-year option was at levels that the price of a two-week option we're trading it, it would be a very expensive instrument. But realistically, we're talking about a handful of days of protection, and that just comes relatively inexpensive. So who's actually selling volatility protection this year then? I mean, again, it's been such an unusual year, and we've seen a lot of options, trades exploding in popularity. Who are the big sellers of protection at the moment? So in many cases, it's the dealers themselves, meaning they don't have the other side in fall. There is a pretty large and relatively persistent systematic program that's important. employed by a range of investors, but basically the idea is options tend to trade relatively rich. And so it's a good sort of risk-adusted return to keep selling them.

So harvesting that risk premium over time will be good returns relative to the volatility of that position, volatility in this case, meaning returns on that strategy of selling options. That's been around for years and years and years, 20 years, really. As of the 2005 vintage, like that was really a consequence of, at least in rates markets, Fannie and Freddie were very large. They had a trillion dollars in their retained portfolio of mortgage-backed securities. And their business model was to buy mortgage-back securities, hedge the duration risk with swaps, and then buy-back options that replicate the borrowers option to prepay their mortgage. And so they were harvesting that what people call it like the mortgage.

of the mortgage. But you can build a model that tries to use all the rate risk you can to replicate the risks embedded in a mortgage instrument. And whatever's left over is the thing that Fannie and Freddie wanted to earn. And they had 40 turns of leverage. So 0.2% times 40 is a pretty good return, basically. And so what that meant was they're willing to buy options at a relatively expensive level. And whoever was selling them those options was participating to some extent in this in this transformation and earning a bit of a transaction cost associated with the Freddie and Fannie programs. That hasn't been the case in a very long time. And yet, it's been relatively profitable to just keep selling options every day agnostic to the environment. So there have been these

programs that have built up to continue this trade, even as Fannie and Freddie have shrunk. And they've been really active at least since 2014 and grown and shrunk over time. but with the very brief exception of March and April this year, they've been pretty persistent through periods of volatility, which is supposed to be what you do with a strategy like that. If you are in a systematic strategy, you're not supposed to sit there and say, well, it doesn't feel right today, so I'm not going to do it.

And so, I mean, I think in March it was pretty easy to say it doesn't feel right today when you have the largest discrepancies in pricing and moves in basically history. We talked about that back in April. So maybe that's the exception, but you're really supposed to keep doing this through periods of stress. In fact, periods of stress are when you earn

most of your money in a strategy like that, and it's likely that that kind of activity has persisted today. Why hasn't the premium on option selling gone away? I mean, if it's been profitable for 20 years, if people seem to systematically over, overpay for protection, why isn't that just like everything else, like sort of crowded out to the point where there's no money left in? Yeah, there should be alpha decay and stuff like that.

It's one of those sort of great mysteries of interest rate derivatives markets. And some people have attributed this to central bank activity and just repression of volatility across various asset classes, especially interest rates. So basically the idea being you're not. fighting the Fed by selling options. You're going with the Fed in doing so. And the presumption that the Fed will backstop any significant period of volatility with purchases that dampen it. I don't particularly like that narrative just because it's a very strong assumption. And we don't really, there's no reason to believe that that's specifically what's going on day to day. There's also this presumption that

options are pricing in much higher risk of jumps. So when we think of an options, price, the models that people typically build, they assume some probability that there'll be these discontinuous jumps in the middle of the day. In October 15, 2014 was a great example of that where tenure yields dropped 30 basis points in 10 minutes and then came right back. And when you start pricing insignificant jump risk into the options, it bleeds through to other sort of prices associated with them. So it makes all options richer if you think there's some chance of these kinds of jumps and what would generate a jump, an election result to generate a jump that happened in 2016. You could have the Brexit outcome, things like that. You could have developments overnight

with the Chinese Yuan Deval and implications for rates markets and EM and so forth. So maybe that could generate some richness, but that's me kind of grasping at straws. Frankly, I think the usual response has been, I don't know, but it keeps working. Which is not a great reason to keep doing it, but it is a reason to keep doing it. And it's been 10 years now, 10, 15 years. And it's been a persistent, you know, reasonably good return type strategy. I wanted to go back to some of your older research. So I, gosh, I guess it was just last year, but it feels like ages ago. You created this thing called the Volfefei Index, which basically, well, you can explain it, but it attempted to quantify the impact of Donald Trump's tweeting on the rates market. And I know you've been

revisiting that index every once in a while over the past year or so. I guess my question is, how much does the rates volatility regime change if Biden wins the election and you get a more traditional president, let's say one who's less active on social media, for instance? Does change everything for the rates market? Do you have to discontinue Volpefe? Probably. So this was work that a colleague of my money or Salem did, and he came up with the name, so I shouldn't take credit for that. And basically the idea was it had felt this way to a lot of people that the president's tweets had an impact on markets. And when they veered towards certain topics, they had a greater impact on markets. And at the time, this was the height of

the trade war. And so you get these pronouncements over Twitter about the progress of talks. And I remember people being very focused on parsing the words and the likelihood of a deal or not a deal or would tariffs go up or not go up. And basically the idea was, let's try to put some numbers around this. And the reason why this works in the first place is for two elements. The first being political, which is policy announcements were made over Twitter, which is a sharp departure. I mean, Obama had a Twitter account, but he didn't announce new policies on it, at least not initially. And just the policy process has changed, and Twitter was a good or was a preferred vehicle

for making those announcements. And so we had to pay attention, whether we like it or not. The second is, it's more of a technical thing, but it turns out this is a very good problem for machine learning and natural language processing because the data we used, It was something like 30,000 tweets, which is not a massive database, but the language that Trump uses is pretty consistent. And frankly, he doesn't use that many words. And so if you want to try to digitize that in a way that can be analyzed systematically, it's a good toy problem for it.

It's something you can do in a desktop. It doesn't take up terabytes of memory. This is not like trying to teach a car how to self-drive. this is like a very concrete question, does this tweet affect markets? So we can very clearly like tag tweets as having moved or not moved markets. We don't have to try to figure out what sentiment is. Usually when we talk about natural language processing, we're talking about like good or bad. Is this a good or bad statement?

Here we're just saying did the market move? So it's easy to build a database of tweets and impact. We don't have to think about that much variation in the language because it uses pretty, clear, pretty consistent, not a ton of words. So it's something you can analyze pretty straightforwardly. And our goal was to come up with it. It was more of a detection than a forecast. I've often got the question, like, does this tweet move markets?

Or is this tweet more or less likely to move markets than that tweet? And I think the idea here was less to get a forecast for any particular Twitter announcement, but to say, what's the background of noise that this creates? Like, to what extent does the accumulated uncertainty with the potential for tweeting in general, and especially on particular topics, does that lead to elevated levels of implied volatility, meaning protection from options is more expensive because any minute now, the president can tweet about the Chinese trade negotiations. And so what we do is we took this database of tweets.

We tried to identify the words that occurred more frequently in those that moved to market, And then we built a random forest model that tried to account not only for the relatives or value of each of those words and categories in generating market moves, but also the interactions between them. So if good appeared with the word China, that had a different meaning than if good appeared with a different word. And so it's a classic NLP problem. And so what we found was when we generated this index, it had statistical significance in modeling volatility. And what does that mean? It means that if we're trying to explain the drivers of interest rate volatility and we incorporate this index into that statistical explanation, it plays a significant role. So to directly answer your question, if we move to a Biden presidency, I don't know what his Twitter habits will be.

I think it's fair to say they would be different. Maybe I should hedge a little bit and say we'd have to wait and see and try to generate enough data to build a model. but looking at the at Joe Biden Twitter feed versus the at real Donald Trump Twitter feed, I think it's fair to say his look sort of more vetted and less and much more consistent with policies that have been previously announced. And so the utility of watching Joe Biden's Twitter feed is probably less in a Biden presidency than that of watching Trump's into Trump presidency. So I don't know if you have to mothball it because who knows whether that framework

is useful in the future, but at a minimum it means the background level of uncertainty comes down because you're transitioning back to a policy process where new ideas are vetted internally, leaked out in some fashion through appearances on the Sunday shows or through articles and newspapers and going back to a more traditional policy generating process that doesn't really rely on Twitter as much. God, I can't even imagine what that world could be like. It's almost, it's like I have this vague memory that things used to operate like that. I want to ask another question about the current regime. And again, not necessarily relating to the imminent election.

But one of the things on the rate side that's really sort of characterized the past several months is essentially just this fact that the Fed has indicated very strong. that the bar to a future, the first rate hike or a future rate hike is probably higher than it's ever been before. Very ambitious goals with hitting its inflation target, full employment, far more forward guidance than we ever got during past recoveries or anything like that. And as such, even through this huge stock market rally from the end of March through now basically, we've seen almost no upward move in rates. And I'm curious, like, how does that change the game from your perspective from a rates volatility perspective, the fact that we have so much aggressive forward guided, so little left to the imagination in terms of what the Fed is going to do.

Yeah. So forward guidance is very effective at suppressing volatility. When we did an experiment recently, we took options that the expiry was a year out, two years out, three years out, five years out, ten years out. And we said, you know, these options are related to one of two things really. One, the pricing of them. One, the policy outlook. So that's forward guidance. I think we used something like the months until the next hike per the feds pronouncements. And we have a series for that. We can go back and just say, when did they think they were going to hike. And they had forward guidance in 2010 and they had forward guidance in 2012. And now they had forward guidance again. The second part was proxying these exotic flows that

that tend to drive very long-dated options, volatility. So this comes back to the Taiwanese life insurance companies, which we'll probably talk about at some point. We could get Brad on here. Red sets are on here as well. Can never get enough Taiwanese life. Exactly. And so what we found was if you go out two or three years in expiry,

so if you want protection for the next two or three years, the price of that protection is mostly related to this forward guide. So as the forward guidance becomes stronger and further out, the price of that protection comes down because, in a sense, the Fed is subsidizing it with their policy. Then when you get further out than that, you start getting into this world of, well, who really trades 10-year options on 30-year rates? And then you're really talking about much different,

not really informed by likely, not as informed by likely realized volatility over the next 10 years. You're informed by the value of that 10 years of protection to a very specific and relatively idiosyncratic subset. So, you know, to answer your question, like it suppresses, volatility at least out a ways in the term structure. And I think the Fed has made it very clear that that's their plan. So they're not purchasing assets to cap yields out to three to five years. That was something that was sort of floated or speculated at times, a form of yield

curve control, but one that's very tied to reinforcing forward guidance as opposed to say what the Bank of Japan is doing. But I think the Fed has plenty of credibility in this department. Their communications are very clear. And it just has not generally paid to fight these things. A lot of short daily news podcasts focus on just one story. But right now, you probably need more. On Up First from NPR, we bring you three of the world's top headlines every day in under 15 minutes.

Because no one's story can capture all that's happened. in this big, crazy world of ours on any given morning. Listen now to the upfirst podcast from NPR. So the Fed is suppressing volatility plus for various reasons that you've already described. You have the price of near-term volatility protection that's quite reasonable and cheap at the moment. So I guess my question is if come November we have the election and everything goes sort of not according to plan, but everything goes as expected and indicated by the polls currently. How quickly does the risk premium that's currently built into markets go away, if at all? Well, it depends a bit on the Senate.

So we talked about the presidency a lot, but that 125% of GDP target for the stock of government debt, that depends on Biden actually being able to do things. and in order to do things, he needs to have the support of both houses of Congress because the vast majority of fiscal policy is going to be an act of Congress. So the outcome of the Senate is key there. If you have a Democratic sweep, which I think betting markets have it, what, 65% or 55% and these quantitative election models like 538 have that closer to 70, 75%. So if you have that kind of outcome, then there's potential for a real shift in policy.

And that could generate volatility just simply for the reasons we were talking about earlier. Now we know what's going to happen. And so we're going to reprice the market significantly. And I don't think options markets anticipate necessarily that, meaning an extended period of repricing of interest rates. A lot of that election risk has really become more concentrated around the event itself. If you have a split power situation, you know, I think it's quite suppressive of volatility only because it's unclear. what beyond the current status quo could possibly happen. You know, the one caveat to that being, if literally nothing can get done,

then there's a macroeconomic consequence to that, which is in the absence of support from the federal government, like what is economic growth? What does GDP do? What does employment do, et cetera? And so, like, that's kind of the caveat there. But, you know, I think what's really interesting about this is it comes back, to the Fed again, which is, let's say there's a Democratic sweep and the Biden campaign platform is

implemented. So this brings up broader questions of fiscal dominance, meaning what is the role of the Fed in an environment where the debt is expanding that quickly? And the question that I think the market is grappling with is, on the one hand, the Fed is clearly not tied their purchase program to fiscal policy, nor should they, right? Independence is important. But they've also very firmly committed themselves to market functioning. And this is a lot of what was going on in March. The centrality of the Treasury market, not just as an investment, but in just the flow of money throughout the financial system

and as a provider of liquidity to the banking system. It serves a much more important purpose than simply a risk-free investment. So if market functioning is part of the mandate and part of the reaction function for their purchase program, then at some point, wider deficits could in principle generate market functioning issues, at which point the Fed has to step in. So, like, in the absence of a change to the regulatory framework that generates these risks, you know, are we setting ourselves up for fiscal dominance de facto just because of this relationship? And it's not on a short-term basis, it's not on a day-by-day basis,

but if we look out into the future and we say that the ability of dealers to intermediate the sale of treasuries and the purchase of treasuries is fixed in size or relatively fixed, but the stock of treasuries is going up substantially, then at some point the Fed has to provide an outlet for that. And so that at some point will rise to some version of fiscal dominance, even though it's not explicitly that. And so for thinking about volatility like that, that is suppressive of volatility, right? If there's a backstop on what yields can do, and there's all the time, a Fed backstop in the market, then potential for large changes is very much mitigated.

You know, I want to go back to something you're saying about looking at the polls and the 538 models in the betting markets. And of course, you know, I read a lot of sell side research where the strategists, you know, talk about DC and those charts are often in there. How much is that really being inputted in real time into models these days from clients that you deal? with where they have sort of ongoing updates of these models that then automatic, you know, take in all this stuff to betting markets and so forth and then spit out some results in terms of how they want to trade that. I don't get the sense that it's directly incorporated that often. I would say the more likely candidate for that is prediction markets. And maybe that's just our

bias as an investment types where we say, look, if there's money behind this, if it's a transactional a transactional based measure, I have a preference for it. Typically, those two things have gone together to some extent. I think there's less of that now and predict it has, what, 65% chance of a Biden victory and these quantitative election models. And I'm saying that because I want to include like the economist and Sam Wang stuff and the 538 stuff, all of them kind of converge around something like 90%. And that's frankly because there's only so many ways to do this.

and we all have the same input data. That's a pretty significant discrepancy. And I think that raises this issue that was alluded to earlier of polling errors, because ultimately these models don't assume polls are right, but they assume that they are as wrong as they've been in the past on average. And so I think the memory of 2016 again is relatively fresh, which is a little ironic in the sense that 2016 was basically a one-state, sigma polling error, it really wasn't that outsized relative to history. And if you get a one sigma

polling error in just the right way in the context of the electoral college, you can get a very unexpected outcome. But I don't think it's fair to say the polls were quote unquote wrong in 2016. So that degree of wrongness is incorporated into that 90% number. What you're seeing is markets in general, and that's reflected in option premiums. And then these prediction markets, which admittedly have very small transaction volumes, but they feel more financing than, say, pure model estimate. So, you know, those are more traded together, I think, over time. But, you know, I don't think these models are really incorporated rigorously into any investment process. I think there's just not enough time to do that. You don't have good data over a long period.

And so what you end up doing is kind of handicapping and you make one of these grids, which is like the least quantitative exercise in the world where you say like House Senate control on the Y axis and presidential control on the X axis and like what does it do to yields and what's the probability of each and that's how I come up with some target. I mean, that's kind of all you can really do. And it's not like we can choose not to participate in this election from a market's perspective. It's that you have to have a view because it matters.

but it's hard to do it in a very rigorous way. The one thing I like to highlight also and others have as well is this uncertainty cuts both ways. So I think intuitively when we talk about polling errors, we're talking about Trump winning, even when his probabilities are relatively low. But a lot of these models have a 10 plus percent Biden sweep and landslide as more likely than a Trump win. So like, the generally speaking error is symmetric, or at least reasonably symmetric. So I think there's less attention to the market on that potential outcome because a Biden victory by 10 plus percentage points is a mandate that has implications for policy that a two percentage point win would not. So it's definitely worth thinking about those scenarios.

Do you think that the fact that, you know, when you talk to clients, do you feel like that is underappreciated that fact that everyone has sort of the 2016 mental model on their head where you take Biden's lead and then you chop a few points off of it because reasons and then you get maybe this close race? And by and large, just people aren't thinking about that alternative form of error? Yeah, like 1984 never comes up as an example. So, Like the, I think the potential for a very significant Biden win that brings with it this kind of mandate. And that brings into the fold all kinds of other policies that were really not in the base case of the campaign. You know, that starts to look a little different.

And I just don't hear about that very often. That's interesting. Yeah. And it's much more along the lines of what you're describing, which is like Biden's probably going to win per the polls. but what if we're wrong for reasons unspecified? And so I need to handicap this. So my 88% turns into 65% from a betting market's perspective. And maybe from an investment side perspective,

I buy some protection on top of that because maybe I'm wrong twice. And so that's why the options get rich, right? And that's why they increase in costs because that insurance is protection against being wrong in that one. Okay. So we're going to leave it there for now. but Josh, we'll have to have you on after the elections for your fourth, awe-thought's appearance. We'll make that happen to discuss what may or may not have actually changed in the volatility regime.

Yeah, sounds good to me. Looking forward to it. Thanks, Josh. Thanks very much. So, Joe, it's always great to talk to Josh. I think he's really good at elaborating on these quite complex topics. And really, you can kind of throw anything at him. But the points he was saying about who's actually selling volatility at the moment and how it's being priced and why you might want to buy volatility protection just two weeks out ahead of the election or why you might avoid selling it.

I thought that was really interesting. We're very lucky. I feel like we've had a richness of people who are just extremely clear about explaining this time. We recently talked to Chris, Ben Eifford, who we've had on a few times. and Josh just like so clear the way he sort of describes the contours of this market. I really appreciate that conversation. Yeah, absolutely. And also the point about how no matter what happens in the elections and sort of no matter

whether it's a huge surprise relative to the current polls or not, there is going to be a need for investors to reposition. I thought that was a really important thing to mention as well. Yeah, like the idea that an event doesn't even have to be a big event, it just has to pass. And then suddenly a new regime can emerge, even if nothing really happened that fundamentally changed the outlook. Just because in the lead up to that expected event, there was so much sort of, I don't know the word I'm looking for, but hesitancy perhaps to make any big move. Yeah, absolutely. You know what else I thought was interesting is like there's like this sort of like highly quantitative flare or sort of characterization of all of this stuff.

But when you talk to Josh, like so much of what he describes in the market is sort of just like heuristics and sort of people making normal judgments that don't seem that math. It's like, do you really want to be the one that sold puts two weeks before the election when everyone remembers 2016? do you really, okay, you take this expectation, but 2016 was like this, so maybe it'll be a little closer. Like a lot of things that don't seem like that quantitative or rigorous at all and more just like gut feels about how you're supposed to play this. So it's interesting that, you know, here it's like options, derivatives, hedging volatility curves. You think it was this very sort of like mathematical approach. But a lot of it is sort of just a, you know, just people going on their gut. Yeah, well, also the example of the systematic wall sellers who are supposed to be doing that on a quantitative basis that doesn't change.

But then back in March, they sort of collectively thought, well, wait a second, it's crazy out there. Maybe we should stop doing this. Whoops. Even though it would have been profitable. Yeah, exactly. Shall we leave it there? Yeah, let's leave it there. Okay.

This has been another episode of the Aw Thoughts podcast. I'm Tracy Alloway. You can follow me on Twitter at Tracy Allo. And I'm Jill Wisenthall. You can follow me at the stalwart. Follow our producer, Laura Carlson. She's at Laura M. Carlson. Follow the Bloomberg head of podcast, Francesca Levy, at Francesca Today.

And check out all of our podcasts at Bloomberg under the handle at podcasts. Thanks for listening. Hi, I'm PJ Vote. My podcast search engine has a new two-part series for you. Of all the new technologies coming out of AI, the most transformative one might be driverless cars. They're already on the road in 10 American cities, and they're quickly coming to more. We tell the story of how we got here. The secret team at Google that spent 15 years building what might be the safest vehicle on the road, and we cover the fights brewing in blue cities,

where unions and politicians are working to keep those cars off the streets. Listen to search engine wherever you get your podcasts.