[Music] Hello and welcome to this podcast from the Oxford Institute for Energy Studies. [Music] Welcome to the latest edition of the OIS Podcast series. My name is Bassan Fattu, and I'm the Director of the Oxford Institute for Energy Studies. I'm delighted to be joined today by Dr. Ilya Busheev. Ilya is the managing partner at Pental and Investment, Senior Research Fellow at OIS, and an Anjunt Professor at New York University. He's also the author of the book "Virtual Burles", which was published in 2023, and recently launched YouTube channel under the same name. Few years ago OIS in collaboration with Ilya started a new series, "The Energy Continent" tells. In which he analyzes some of the trends that are shaping oil markets from the financial side. I will be discussing with Ilya his most recent papers from this series. The first one centers around the revival of the volatility risk premium. The second and third papers center around algorithmic trading and investment strategies of financial players and alcoms. Hopefully in today's podcasts we'll be able to unpack some of the themes in this papers in more detail. To stamp the date, we are recording this podcast on Wednesday, February 80. Ilya, it's great to have you back on our podcast. Thank you, by thumb, and my pleasure to be back. Ilya, before we dig deeply into these papers, a few people ask me when the new I was hosting you. What does Ilya think about the role of fundamentals in price formation? Do you think fundamentals play a limited role in oil price formation and at which time horizon? Or is your point that analysts don't pay much attention to the role of financials? And actually this has become more dominant in influencing prices as the size of what you call virtual and balance actually overwhelms the size of physical balance. Interesting question, very timely of course. Obviously we keep talking about the fundamental blood, but the market is up a year to date about 10-12 percent. The market is comfortably backward. So what's going on? I guess I was even joking yesterday on Twitter. I said my forward summary of the current market condition is that everybody is bearish and everybody is long. So I guess that's the state of the world. So back to your question. So let me answer it in two different ways. Of course, of course, fundamentals do matter, but I'll give you an answer from two different perspectives. You know, Basan that I like the commodity history. I travel a lot. In the ancient days when commodity traders were shipping products, they used to navigate by stars and the sun. So to me, fundamentals, that's like your North Star. That's generally the direction that you kind of want to go, but the best route may not necessarily be the direct because there are storms, there are pirates, the wind may not blow the right way, and that's why you may need to deviate from that. So at the end of the day, fundamentals is like a long term gravitational force. It's pushing you to a certain direction, but fundamentals don't change on a daily basis and prices do. So let me kind of illustrate it with a second angle. A bit more, a bit more mathematic, but I think it conveys the point. So we kind of look in the quantum world, we say that oil prices follow basically as the caustic process. There are two terms. There is a drift. So we keep drifting in a certain direction and there is a noise. So generally, the drift is proportional to time and the noise is proportional to what we call square root of time. But if time is very, very small, then the second term is much bigger than the first. For example, if the time is one day, the second stochastic, the variance term is 16 times larger than the drift term. So the fundamental reflect the drift. So we keep drifting in a certain direction, but the noise term for one day is 16 times larger. So that's my take on fundamentals. Yes, it's a gravitational force. So we keep drifting towards that. But in the meantime, on a day-to-day basis, the prices are driven by a lot of other things. In finishing what we need to do, the client demand directly do not buy future smart yet. People or algorithms do. In finishing what we need to do, we need to model more how humans or machines, how they interpret fundamentals. We need to model their behavior. But hope this kind of two examples illustrate my take on that again. There is a long-term goal that keeps drifting towards that. But in the meantime, whether it's a wind or pirates or stochastic term, we have to, we have to deviate. Well, it's fascinating, this one's a very interesting way to look at this whole debate of physical versus financials. And I'm sure today we'll be able to unpack this more. Let's now move to some of your recent papers. So in the paper on the volatility risk premiums, you discuss how option risk premiums are not constant. And they depend on the balance between those who want to hedge and take insurance and those who sell the insurance. You also mentioned that as recently, the number of players who are willing to provide insurance has reduced them
atically. Why is this the case? And what have been the implications on oil markets? I think if anything in today's market, yeah, there is an imbalance in the options market between buyers and sellers for sure. And that imbalance is growing. I think more recently, I think it's coming more from increasing demand for protection, rather than decreasing supply. I mean, there is a steady group of suppliers that provide the insurance, but the demand is now coming much more from retail traders. And it's happening not in oil specifically. You see what's happening in this gold option, silver, equity market. There is a huge interest on behalf of retail traders. And that incremental demand for protection, there is not enough suppliers for that. So the suppliers are there, but the pool of suppliers of volatility is not increasing because the dealers are there, but the sophisticated financial investors can't really help to sell volatility because they generally look at commodities or oil as a hedge. And it would be very odd for them to hedge by selling options. So if anything, they can't really do that. And the other thing I think I probably should mention it to a bit to a lesser degree, but I should mention there is less reducer hedging in the market now comparing to five or ten years ago. And producers used to sell coal options, used to buy food options and sell coal options to the dealer, but this selling of coal optionality provided huge buffer to the market. Now that buffer is much weaker. It's still there, but it's much smaller because the entire share industry or US oil industry has consolidated a lot. A lot of smaller producers that used to hedge quite aggressively, they got absorbed, acquired by majors and majors generally don't hedge. So you see, you clearly see a decrease suppliers, this coal volatility, and it's getting a little bit more nervous now, whereas the upside could be capped by these producers. So it's a combination of both combined from retail and a little bit less selling of coal options from producers. So, Elia, sticking with the same paper and I think related to what you said last, you argue that end user option buyers are willing to pay higher premiums to get protection. Does this mean that option sellers are having a free lunch or are there some extreme risks that they should deal with and therefore should be compensated for? And in case of sharp price movements, do these sellers accelerate or do they counteract the price movements? Yeah, there is no free lunch in trading. We all know that. So when you sell option, you're taking massive risks. You're selling an insurance product and it has a highly, highly asymmetric negative payoff. As in all insurance contracts, you generally make a little bit of money on the regular basis and then periodically you lose a lot when you have to pay off on your insurance contract. So you have really no choice but to be extremely disciplined, hedging your risk. And the way you do that as a hedger, so you basically have to sell futures very aggressively when you sell the insurance against downside move in the form of food options. And then you have to buy futures very aggressively when you sell a call option, basically you sell in
insurance against the upside move. In either case, you sell the lows and buy the highs, you clearly exacerbate the market move by all means. And this is what we call the negative gamma. And I often refer to that as the most powerful brick and my book and my YouTube channel. And the reason I call it the most powerful because the gamma hygiene is totally price and sensitive. So the dealers who collected their insurance premium, they're essentially in a profit preservation mode. Why would I, I would even say they're basically fighting, it's an existential threat for them. So they have to be extremely disciplined and extremely aggressive in managing their risk. And they don't really care at what price they do it because their profit is ultimately coming from the option side. And that's why keep emphasizing the importance of that factor. And it's not a free lunch by any means. Absolutely not. Let's now shift to the discussion a little bit to the role of algorithms. In one of your papers, you discuss two conflicting strategies that coexist. Some will use the oil in order to hedge for inflation, but also some would use oil perhaps to hedge for the risk of a recession. How do these two strategies interact to determine the oil price? And you do kind of confess that article you published I believe about six months ago at least. But things keep changing, things evolve. There are many, many other algorithms, the two algorithms that we describe in the article. To be honest, they shifted a little bit to the back burner. What we were discussing back then was a tug of war between inflation hedgers and recession hedgers. So the inflation hedgers were basically with parity funds that own the portfolio of stocks and bonds highly leveraged on bonds. And they need to hold inflation sensitive assets like oil to hedge their bond portfolio. So now the inflation is high, elevated, but it's fairly steady. And the market genuinely expects inflation to stabilize and to interest rates to come down. So they were basically clearly under hedging that risk. And a bit similar on a recession side. So recession back then we described in the form of a crisis, a crisis, alcohol algorithms. But now recession peers also subsided as well. But I think there are other algorithms that kind of evolve in from these two. I wouldn't really call it an algorithm. But one area where I'm paying a lot of attention now, we haven't written anything on that yet, is a potential, I would call it a rebalancing algorithm. A rebalancing of essentially massive profits that the industry accumulated in metals market. The astronomical profits now are seeding in gold, silver, copper, aluminum, you name it. And I know a lot of people showed this graph myself and included that oil to gold ratio is at the hundred year low, except for COVID episode. So some people discount that saying, you can't compare it's meaningless. I don't discount it. I take it very seriously. I think oil as measured in terms of other commodities is extremely cheap. And I think there is a delayed reaction to potentially rebalancing effect or spillover effect. And it also related to a weaker US dollar. So the weaker US dollar hasn't really impacted oil yet. And I think there is a delayed reaction to that. Because at the end, the weaker dollar makes oil cheap outside of the United States, which means higher demand for oil. And it also increases the local production cost outside of US. That means lower supply. I think it's still coming. And oil is an asset like gold. And there is basically a shortage of investable assets in the world. And I think we need to pay attention at what point this money is going to be taking profits in gold, silver, and copper. And basically selling the leaders and buying the lager, which is oil. There is a lot of money in that space. Just to finish up on that topic, again, I was tweeting recently another semi-joke. So when oil hedge funds trained 100,000 contracts in the wake, a lot of people getting a lot of fundamental people panic. But 100,000 contracts is only $6 billion in the national term. And just to put things into perspective, a tiny silver market on one of the recent days traded $40 billion in the national. So we panic when oil trades $6 billion, but silver traded $40 billion. That means there is like seven times more money trading silver. And all that money can easily move into oil and add on top of this all the gold money and copper money and aluminum money. There is a ton of money sitting in various forms in precious and industrial metals. And I think to me, this is the most important algorithm to watch now. At what point they're going to start basically-- essentially global rebalancing, where people are going to start selling crypto, selling equity markets, selling gold, selling silver, basically selling whatever went up and buying the lager, which is oil. I imagine that those volumes actually are quite large compared to what we see flows in the oil market. And as a result, those actually can move prices. Is that correct? Yeah, absolutely. That's my point, because oil for now it's kind of boring. That's why all the money is currently in metals and crypto and equity markets. But at some point, things can reverse. And the catalyst on that could be in the completely different market. It could be in equity. So it could be in crypto. It could be somewhere else. And all of a sudden, money is going to be shifting into safer assets like oil. So that's something we have to watch for sure. In the last few weeks, we have also seen a sharp increase in the open interest for call options. I believe that this is very much related to the risks of US attack on Iran. And some interpret this in a way that those risks are pushing algos to become more bullish. Do you agree with this interpretation that the algos are becoming more bullish? And more generally, how do this mechanism interact to influence the oil price? Yeah, algos are machines, really reactionary. They never act as a catalyst or anything. So there are really several forces in play here. But they all sort of feed off each other. So the initial catalyst come from geopolitics, of course. It could be Iran, it could be Israel's. And then some macro trainers initiate in the market by buying, for example, a large quantity of call option to hedge their macro portfolio. So that's your initial catalyst. And then essentially, the option sellers, those the negative gamma people, they are the ones who are creating the trend. So the macro trainer experts by buying call options, then the option hedger, the gamma hedger, is buying a lot of futures that create an upward trend. And then once the trend is established or started, then it triggers the algos. So there is a three-step process. First, it's a macro trader by the call option, then it's a gamma hedger by end futures, and then CTAs react to that. But again, their role is purely reactionary. The interesting thing here, though, that single option that the macro trader may buy, the macro trader may just do it once, but the option may still be there until expires in that constantly impacts the futures market until it expires, because it's going to be creating the trend up and down. And yes, you absolutely right. The option is used to the upside, recently in the record highs, calls relative to the boots or near record highs. In the past, historically, those spikes in call options, it was a fantastic opportunity to take the other side, because the geopolitical risks historically hasn't been realized. But as I mentioned earlier, we all used to have a buffer from producer hedgers selling call options. Now that buffer is much smaller. So selling this calls you gets a little bit tricky. I still think it's probably an OK trade, but it's a much riskier trade, because you don't have a buffer from producers anymore. It's much, much weaker. And Ilya, do these mechanisms work in the opposite direction? So for instance, if geopolitical risks paint, do these basically work in the opposite direction as well? Yeah, absolutely. I mean, first of all, if geopolitical risk fades, you have to unwind all these users the same way that you bought. It's fairly symmetric. And on top of that, a single put option can generate a lot of future selling as well. I wouldn't go.
I would say somewhat symmetric in the current market, it can go both ways, but the catalysts, again, could easily be triggered by a single large option transaction. Sometimes we don't see those transactions, they do happen over the counter as well. In your latest paper, you identify a momentum strategy in which investors mimic the large funds and therefore they reinforce trends and I think you refer to that as follow the flow. Then you identify a contrarian strategy where investors bet the case extreme positioning, perhaps expecting mean reversion in prices and you refer to that strategy as fade the crowd. You argue that in recent months the latter strategy has become more profitable. Why is that the case? Is it due only to the fact that we have seen some extreme short positioning on some of the oil contracts? Although we have seen recently this extreme positioning now has eased, how are these two different strategies now interacting in the current context? It's generally a good idea to take the other side of the crowd it trade across all markets, but to be honest, the strategy worked really well to take the other side when the CTA is a too long or too short. But to be honest, if anything, that's more like a side effect of the range bound markets because we know the CTA is predominantly trade the momentum strategy and the markets will range bound. So by taking the other side of CTAs, you're effectively taking the other side of a momentum which means you're betting on the range bound market and the market wasn't in the range bound because that's again where your fundamentals come into play because you essentially have a fundamental cap here from oversupply and a little bit from Washington as well and we know that if the market rally a lot, Opik is going to probably increase their production again. And the same thing, we have a somewhat lower flow because of a Chinese buying for a strategic petroleum reserve and which genuinely escalates as the prices get cheaper and cheaper in the addition, the retail buying as well. So you have a floor and you have a ceiling basically now, fundamentally. So the momentum strategy haven't really been working well and then by taking the other side of people who trade the momentum, you basically make money and that's why that strategy made money was the last few years. But I would just warn people, CTAs, they're going to adjust. So they see the patterns, they see the patterns are changing, so they're going to diversify their strategy away from being a pure momentum strategy as well. So it's not necessarily that strategy is going to continue in the future. It's quite interesting, Geliha. You say that CTAs adjust in the last few months. How they have been adjusting to a range bound, oil market where prices have been trading within relatively narrow range? Yeah, I would say a lot of them are trading a bit faster now. So the fast CTAs, I actually make it money. The slow CTAs are not. But the slow CTAs do not adjust as quickly. So because ultimately they need a longer back test and we just haven't accumulated enough of this longer term patterns for them to adjust. But if we stay another couple of years in the range bound market, then this pattern is going to be included in their models. Do you haven't really seen that yet? Geliha, it's very difficult nowadays not to talk about AI. And one of the question to you is how do you think AI will shape the financial oil trading? And do you think it will make the market more or less volatile? Or actually it will not have much impact. What are reviews on that? Extremely but some preliminary view. I think it's a bit too early to say as we all are learning, as we go. So at this point, I'll probably just stay neutral on terms of marketing back and maybe all have another discussion in the year or so. In fact, I mean, the way how I look at AI or machine learning, traders and analysts usually use it in two different ways. One way you use AI for prediction and one for inference. So prediction is hard. Prediction is very difficult because everything changes, or regime changes in oil all the time. I'm not saying it's impossible. It's just difficult. But I see a lot of passion. I teach it NYU and several other places. And I see so much passion among students to dig into that. There are some interesting observations, but nothing there yet in terms of robust and best of all strategy. So it's work and progress. One good thing about this type of machine learning models for prediction, it helps to identify the importance of different factors. We call them feature selection. So it helps you to explain the market behavior because oil is driven by different factors at different times. And some of these machine learning algorithms are pretty good in telling you what's the main factor is today. They're less good and forecast in the price, but they're good in factor identification. And then the other type of models that people use for inference, this is actually related to the work. We are doing this with your team at OIS where we're essentially trying to replicate the hedge fund behavior. As I mentioned earlier, that's what we need to do. Instead of just simply counting barrels, we need to see how different market participants react to different factors. So we're trying to replicate the hedge fund behavior where essentially we're trying to reconstruct the function y is equal to f or x where x are your historical crisis or some combination of them, some momentum signals, your y, your outputs positions, so your inputs are prices, your outputs are positioned and the function itself is your behavior. So you know inputs and you know outputs, you're trying to reconstruct the behavior. So on that in that direction, we actually achieve some interest in the results and hopefully we'll write about them in the near future. But that's not a prediction. So that's more like a reconstruction of something. Earlier you mentioned the rebalancing point, which I think is a very very important point. Don't you think AI actually could make those rebalancing more extreme given the ability to connect across different markets? What are your thoughts on that? I do not know yet because again that's obviously if all the managers that have these portfolios going to adopt this. Clearly there is a potential in huge role AI that can play in portfolio optimization. I don't know to what extent it's being used currently. Yes, I didn't mention that. It's not again for price prediction but there is a role for AI to play in constructing optimal portfolios. So earlier my last question. So for someone who's following oil prices and oil markets in 2026, what should really they pay attention when it comes to all goals, hedge funds and their investment strategies? So let me put the question perhaps differently. What trends will you be monitoring in 2026? As I mentioned, besides the rebalancing one that we have already discussed. So keep an eye on what's going on in other commodity markets for sure for any size of spillover. But then I will continue to emphasize the options market because you asked, you brought it at this point yourself. We keep setting record volumes almost every month. Huge interest. As you know, I recently launched the YouTube channel and I was thinking about doing just one quick lecture on options. But the demand was so high. So I already recorded four lectures on options and people keep asking and those are the people who are relatively new to the oil market. They're people who trade in other asset classes. But they are interested in applying their skill to the oil market because they are learning that the oil options market is quite big. So again, I kept joking recently that a few more months of this option volumes and oil may soon be taking the silver road. So you know that the silver was quite crazy recently. So I guess people should be paying as much attention to anything that's happening in the oil gamma space watching Putin calls as they historically did for counting like pushing barrels. So we go back to the original message, Ilya, look at fundamentals but you cannot basically ignore what's happening with all of these alcohol search funds that investment strategies. As they will continue to have an impact on the oil market. And as you said, you know, the fundamentals don't change on a daily basis but there are so many other moving factors. This is all what we have time for. As ever, Ilya, thank you so much for fascinating discussions and insights. And thank you to our listeners for tuning into this podcast. I encourage all of you to visit the oil
my S website and read the papers we discussed today. And as Ilya promised, there will be more forthcoming paper this year as well. Thank you for listening to this podcast from the Oxford Institute for Energy Studies. You can find other podcasts, as well as our written research on our website at www.oxfordenergy.org. If you would like more details about our energy transition, gas, oil, electricity or China research programmes, then please contact us at
[email protected].
PodcastTranscript AI