
Convexity Resilience and Building Better Brakes with David Dredge
What this covers
This is a single-guest conversation on the State Street "Street Signals" podcast with David Dredge, CIO of Convex Strategies. He builds a case against the way fiduciaries construct portfolios. The starting point is the Sharpe ratio and mean optimization. He notes that the S&P 500's 7.9% average annual return since 1929 was realized only three times in 96 years, and that geometric compounding is driven by the large losses and large gains rather than by the average. He also cites Benoit Mandelbrot on why the underlying assumptions are wrong.
The middle of the conversation looks at where systemic risk hides. Dredge singles out correlation as the most underappreciated one, because VaR models, liability-driven investing and risk parity all assume it is stable. He says correlation breakdowns came before LTCM), the August 2007 quant blow-up and the 2008 crisis. He also revisits the 1987 crash, when portfolio insurance assumed continuous liquidity. He argues that regulation assigning zero risk weight to AAA super-senior CDO tranches let banks pile up leverage in assets that looked riskless. In his view, accounting and capital rules also push regulated entities to sell volatility, which leaves convexity underpriced.
The prescriptive part is more contested. He wants investors to add convexity and to free capital from bonds, citing gold's long-run edge over Treasuries and doubting that FX reserve managers' bond holdings make sense. That capital would go into growth equities, tech and Bitcoin. He criticizes the 60/40 portfolio and current "total portfolio approach" advocates, and he faults tail-risk managers who try to offset negative carry. He says allocators resist dropping bonds because admitting past error is politically hard. For retail investors, he suggests buying call options and holding cash. Throughout he uses a racing metaphor: stop predicting the track and build better brakes so the car can go faster. He also argues that accountability and skin in the game are the only reliable foundation for risk management.
David Dredge argues that portfolio resilience requires explicit convexity through long volatility exposure, not traditional mean-optimization approaches, enabling investors to simultaneously protect downside and accelerate upside participation.
- Sharp ratio and modern portfolio theory treat volatility symmetrically and optimize to unmeasurable 'expected returns,' destroying long-term compounding paths
- Low-volatility assets create leverage and fat-tail risks that regulation masks with zero risk-weighting, making them dangerous not safe
- Buying cheap convexity where systemic leverage has built up (not where you predict crises) provides asymmetric payoffs while freeing capital for market participation
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
Annual returns of the S&P 500 since 1929 have averaged 7.9% but realized that average return only 3 times in 96 years, while the standard deviation of 18.8% captures the actual distribution; variance, not mean, drives compounding, yet fiduciaries continue optimizing portfolios to the mean rather than to managing variance.
“The annual returns of S&P going back to 1929. And the average return as people talk about all the time 7.9%. And the frequency of that average is three times in 90 however long that is 96 years three times. The standard deviation is 18.8%. Which what's driving the returns? The mean or the variance? obviously the variants massively”
In geometric compounding, the negative numbers and the big numbers have much larger impact on terminal outcomes than numbers near the mean because you need a 100% gain to recover from a 50% loss, making volatility drag on compounding path the critical performance driver that calendar-year-focused fiduciaries ignore.
“geometric compounding path that is absolutely killing you where the big numbers have a much bigger impact than the numbers near the mean and the negative numbers have a much bigger impact than the positive numbers. If you lose 50, you've got to make a 100 to get back to break even. And so you're constantly rolling around this concave return dynamic and you're decaying away from the desired compounding path.”
Correlation breakdowns predictably preceded both the LTCM collapse (several months before), the 2008 financial crisis (well before bank failures), and the 2007 quant fund blow-up (August 2007), indicating that correlation instability is a leading indicator of systemic stress.
“I can go back through all of history and remind those that are old enough that correlation got wacky several months before we unwound LTCM. And correlation got wacky way before the banks went out of business in '08. The systematic funds all blew up in August 07.”
Benoit Mandelbrot proved 50 years ago that the Sharpe ratio's foundational assumptions—normal Gaussian distributions, linear regressions, arithmetic vs geometric means, and time-averaging instead of ensemble-averaging—are knowably wrong, yet these flawed assumptions remain the basis of fiduciary asset allocation.
“The fiduciary industry, rightly or wrongly, has adopted what I call sharp world financial economic methodology. These assumptions that are, you know, knowably wrong. Benoia Mandelro proved them all wrong 50 years ago. normal Gaussian distributions and and linear regressions and arithmetic as opposed to geometric means and erotic as opposed to non-erotic path you know ignoring time looking ensemble averages instead of time averages”
The most underappreciated systemic risk is correlation: everybody has built assumptions about correlation into their risk models (VaR at banks, liability-driven investment at pension funds, risk parity structures), but when correlation becomes unstable, these models fail and force synchronized derisking across the system.
“the risk that they're not measuring is the biggest risk and that's correlation. And so everybody has built into their risk methodology. So value your risk at a bank and and the the market risk equivalent close-up of that into risk weighted assets and how pension funds are measuring their risk whether they're running a risk parity structure or a liability driven investment type structure. Everybody's making assumptions about correlation.”
Managers who try to mitigate negative carry (theta bleed) in tail-risk strategies are misaligned with their clients and focused on managing their own compensation structure rather than their clients' capital; the client wants a goalkeeper (protection), not a goalkeeper who tries to score goals (opportunistic yield harvesting).
“if somebody's doing strategies to mitigate the bleed of providing the breaks, they're trying to manage their compensation, their performance, not trying to manage their clients. The client wants breaks. He wants a goalkeeper. It's on him to take advantage of that and drive faster or go goal score more goals. He doesn't want his goalkeeper running up and trying to score goals because he think it's opportunistic.”
The Sharpe ratio defines risk as volatility and treats upside and downside volatility symmetrically, which is fundamentally wrong because it punishes strategies for capturing positive upside moves as much as it penalizes downside losses.
“the sharp ratio is defining risk as volatility, which is obviously nonsense because it's treating as symmetrical upside volatility and downside volatility.”
Portfolio construction optimizing to the mean via Sharpe ratio methodology creates concave return structures where investors foregoing upside (bonds have no upside beyond coupon) while only probabilistically reducing downside, leading to volatility drag that decays geometric compounding paths over multi-decade horizons.
“the whole system is trapped in Sharpe World...the simple 60/40 balanced portfolio is to simply reduce the amount of risk you're taking. So you take 40% out of the engine and bet that driving at 60% speed you probably won't crash based upon a lap and a half look back of historical curves...And so you're always driving too slow in the good parts and barely hanging on on every curve”
The October 1987 crash (Black Monday in the US, Blacker Tuesday in Asia) revealed that portfolio insurance models had made false assumptions about continuous liquidity, showing that theoretical hedges fail when most needed.
“I got here the first Monday of October 1987, just in time to catch the what is commonly known as Black Monday in most of the world, the October 87 crash, which we call Blacker Tuesday out here... I was a guy studied financial mathematics at Berkeley under Mark Rubenstein of the famed Cox Rubenstein binomial model and port at the peak of his portfolio insurance business. So, I got a good firsthand look of the implications of making assumptions about continuous liquidity that were not true.”
The regulatory framework assigns zero risk weighting to super-senior CDOS tranches because they are rated AAA and can be structured to be mathematically riskless at origination; banks then leverage these 'riskless' securities infinitely under leverage limits, creating massive fat-tail leverage that materialized in 2008.
“So you can say your super senior tranches of subprime CDOS's because they're AAA weighted and you've tunched them just like this are a zero risk weighted asset. So there no risk and then you can lever them infinitely. Turns out what what is the risk? It's the fat tail created by that leverage.”
The solution to the flaws of mean-variance optimization is to add highly positively convex, negatively correlating assets, free up capital trapped in traditional bonds and diversification strategies, and allocate more to participatory opportunities like growth stocks, tech, and Bitcoin that participate in bull markets.
“the way you build convexity is you add highly positively convex negatively correlating stuff. You free up all the capital that's been burned and tied up into traditional absolute return sharp ratio bonds strategies. and you put more of your capital to work in the things that participate in good markets. You own some NASDAQ instead of just S&P. You own tech stocks, you own growth, you own Bitcoin, and you go find those things that are thin tailed.”
Gold has outperformed US Treasury bonds over the last 30-40 years on a total return basis, making it a more appropriate replacement for bonds in portfolios seeking inflation-resistant diversification than traditional Treasuries.
“the obvious one to replace that, you know, obviously I've been an enormous advocate of forever and I'm sure everyone's trying to figure out one of the obvious ones to replace bonds is gold. You've been way better off owning gold for the last 30, 40 years than owning US Treasury bonds.”
If the objective shifted from optimizing arithmetic mean returns to optimizing actual geometric compounding, all fiduciary behavior would flip; currently the incentive structure prevents this shift despite knowing it's optimal.
“If you could turn it around so that geometric paths non-orgotic geometric paths were what people were driving towards everybody's behavior would flip.”
Dredge's core advice is to stop trying to predict the unknown shape of the racetrack (future markets) and instead focus entirely on managing the resiliency of the car through good brakes, while exploring how to drive faster through better aerodynamics, engine power, steering, and transmission—all of which are controls within the investor's domain.
“Stop trying to predict the future unknown shape of the raceourse and manage what you could control the resiliency of your car. Put good brakes on your car and go out and explore how to drive it faster. Can you get more aerodynamic? Can you get a bigger engine? Can you get better steering? Can you get better transmission? But all of those things as you're exploring that, first thing you want to make sure you have in place is good brakes on the car.”
Allocators repeatedly ask how to fund tail risk hedges; the correct answer is simple—eliminate inefficient diversifiers like bonds—but allocators refuse, saying it's politically impossible within organizations to admit past strategy was wrong.
“one of the things I get asked all the time by that crowd is, well, you know, how should we fund the the tail risk? Well, easy. Get rid of all this. Oh, we can't do that. That's going to be because the whole point is if you're if you're dragging the car down by just impeding the capacity of the engine and you're not going to fix that, what's the point of the better brakes?”
Banks use value-at-risk and option models to hedge Bermudan swaption exposures, but these models don't capture tail exposures, causing basis risk when fat-tail events occur.
“the banks when they're bringing in those Bermudant swap options going out and hedging them their risk is being based upon value at risk and option models and stress scenarios that also doesn't necessarily account for tail exposures that they'll end up with.”
The traditional 60/40 portfolio approach reduces risk by driving at 60% speed rather than managing downside explicitly, leaving portfolios perpetually slow in good markets and barely hanging on in bad ones, trapped between insufficient upside capture and probabilistic downside reduction that fails at tail events.
“the fundamental risk methodology in the financial industry, so think the simple 60/40 balance portfolio is to simply reduce the amount of risk you're taking. So you take 40% out of the engine and bet that driving at 60% speed, you probably won't crash based upon a, you know, a lap and a half look back of historical curves within 95 percentile confidence levels. And then you'll kind of map to the average lap speed given the sort of normal lap shape. And so you're always driving too slow in the good parts and barely hanging on on every curve hoping that it's not, you know, a two percentile curve relative to what you've seen in a far too short historical path.”
FX reserve managers globally (China, Russia, India, Singapore, etc.) own massive amounts of US government bonds accumulated through currency intervention, but as countries debase their currencies in ongoing 'currency wars' to keep exports cheap, holding other nations' bonds becomes a losing proposition—yet they continue accumulating due to structural FX reserve mandates.
“who owns relatively speaking, who's the most overweight government bonds in the world? Well, FX reserve managers, right? They've been doing QE but because of the size of their markets their QE is building FX reserves is buying somebody else's bonds. China and Russia and India and Singapore and so these guys own so many bonds and you know you don't need to be sort of a a scholar of history to figure out that the guy selling you the bonds is debasing them and you've got too many and you're just going to get more cuz you're constantly in the currency war and the Asian central banks are the bastard of keeping their currencies relatively weak and they just keep accumulating them I mean how many do you need”
Even current vocal advocates of 'total portfolio approach' aren't actually cleaning up the capital tied up in traditional concave strategies; they're just adding negatively correlating loss mitigation on top of the existing drag, failing to address the core problem of capital inefficiency in bonds and traditional diversification.
“even the current vocal advocates of total portfolio approach are really just saying, 'Well, we're going to add on some negatively correlating explicit loss mitigation. We're not actually going to go and clean up all the junk has been a drag on our performance for the last 30 years.' And there's a huge reluctance to clean out the capital.”
Buying put options or explicit tail-risk hedging requires active management because of theta decay and changing delta; you must rebalance, adjust position sizes as prices fall, and manage where you sit within your delta allocation, unlike a simple stock position which can be perpetually held.
“buying puts or you're buying explicit protection is becomes very actively managed because you're back into this time problem, theta problem. You your your delta exposures changing whether you're long calls or your long stocks and long puts. So it requires once you get into the business of hedging it becomes more active because you've got to do stocks you could just sit on they're perpetual but options are all fixed so you need to manage maintain exposure roll think about where you are within the delta appetite that you have do you want to release some cash and add some more calls because prices have fallen”
The regulatory and accounting structure incentivizes volatility selling by allowing regulated entities (insurance companies holding callable notes, banks hedging with value-at-risk models) to recognize option premium and basis risk as calendar-year revenue while not capitalizing proportionate tail risk, creating systematic oversupply of convexity at depressed prices.
“the insurance company that's buying the callable note structure is being allowed to account for it to the hold date as though that's the risk and not account for the short optionality and the terminal duration date should interest rates go up and the banks when they're bringing in those Bermudant swap options going out and hedging them their risk is being based upon value at risk and option models and stress scenarios that also doesn't necessarily account for tail exposures that they'll end up with. And so if you have enough participants in the food chain that get to recognize the repackaged option premium or the carry or the basis risk involved at edging them as calendar year compensation year revenue but don't have to capitalize significant risk in the tales almost without fail you'll get supply and demand imbalances”
Individual retail investors without access to OTC derivatives and ISDA accounts can gain convexity exposure by buying call options on equities and index options, using the calls as their participation vehicle and holding the rest in cash and money market funds, immediately creating a convex payoff function.
“any investor can think about structuring the payout function of how they manage their investments. Most simply put you can buy call options. Again, without getting into too much detail of how we see the world, but certainly from sort of middle of last year, topside volatility in equity markets was incredibly cheap. But we always say, you know, we're only buying what we rank as the cheapest in the world. So for a significant chunk of time, that was the cheapest thing in the world from about middle of last year. And so anybody who has access to exchangeraded products can repackage their desired risk taping taking in call option form.”
If allocators optimize portfolios to the terminal (40-lap race backwards from the end) rather than to the beginning (one lap at a time), they would buy straddles (long volatility) rather than sell them, because the only outcomes that matter to long-term wealth are the fastest and most dangerous parts of the return distribution.
“I say all the time you imagine that sort of 40lap race. If you're compounding from the beginning, if you're optimizing from the beginning, one lap at a time, modern portfolio theory, it's going to tell you to sell the straddle around the expected return, which is exactly what modern portfolio does. It's optimizing to expected return, a a madeup precise number. But if you were optimizing from the terminal 40 laps backwards, they tell you to buy the straddle because it'll say the only thing that matters is the fastest part and the most dangerous part. all the other stuff around the middle doesn't make any difference.”
Accountability and skin-in-the-game are the only reliable mechanisms for effective risk management; models, volatility estimates, and correlation assumptions cannot substitute for direct personal responsibility for losses, which is why traditional banks operating under Glass-Steagall with partnership structures outperformed the post-deregulation regulatory-accounting model.
“I always end with the same slide if I'm giving a risk-management presentation. The key to risk management is accountability because you can't model risk. You can't know the only thing that will make people behave appropriately is skin in the game.”
Convex Strategies' process is entirely bottom-up and agnostic to underlying market direction; they buy volatility and convexity when the price is right based on supply-demand dynamics, not based on where they think risks will materialize, allowing price discovery of risk locations rather than predictive positioning.
“we buy options because the price is right. And over time, that probably tells us some idea of where the risks are because we can see where the leverage has built up in the system, where the left tails exist, where the uncized risk that makes the system vulnerable is. So again, for us, it's very bottom up.”
Allocators shouldn't pay hedge fund fees for positively correlated risk (like equity market exposure); they can get this exposure cost-efficiently through ETFs or direct equity ownership; they should only hire hedge funds to provide uncorrelated defensive services (tail risk protection), because that's what's hard to do.
“I'm always adamant that allocators shouldn't be paying hedge fund fees for positively correlated risk. They can get all the positively correlated risk they want for free in the form of ETFs or equities or gold or futures or whatever. Decide how much risk you want and take it cost efficiently and then hire people to to play goalkeeper to be the brakes because that's hard. It takes a lot of you know day in and day out work to manage the the cost the data the lost sensitivity through time to ignore calendar years and manage you know nobody in the real world nobody wants their insurance to decay to zero 12 11 10 9 8 7 6 5 4 32 1 month and then start over again.”
Even poorly implemented convex portfolio construction will improve outcomes significantly because adding any amount of asymmetric risk reduction enables better long-term compounding, making the strategy robust to execution errors.
“the beauty is about adding convexity to your investment portfolio is even doing it poorly will have a fantastically positive impact even you know moderately adding something ineffectively is going to improve your outcome because again the more you can become resilient to divergences the better you're going to do.”
Institutions relying on computers and optimization algorithms will resist moving away from Sharpe-world methodology because doing so requires admitting the PhDs hired to build these systems were working on solved false problems.
“they want to whip out their optimizer calculator and say, well, where do I put in the the number? We don't know the number. But what we do know is that we'll have an improved compounding path. They're like, but we have to have a number because if we don't have a number, what are we going to do with all these computers we bought that optimize to that number? and we're gonna have to tell somebody what we've been doing all these years and we got PhDs in was was wrong. And there's a huge reluctance to do that.”
If major FX reserve managers (China, Russia, India, etc.) shift from accumulating Treasury bonds to accumulating gold through passive reserve flows, it could create an inelastic market dynamic where gold rises continuously as trillions of reserves gradually reallocate, similar to Mike Green's thesis on passive equity inflows driving stock valuations.
“I wrote this month about some of the talk about the inelastic market hypothesis and my friends Mike Green's commentary about passive flows into equity markets. Well, imagine the world's biggest central bank reserve managers and a sort of passive flow dynamic into gold. You kind of imagine getting what we we're getting and they would just kind of keep going.”
Convex Strategies doesn't actively manage negative carry (theta decay) on long volatility positions; instead, they construct the fund as a perpetual private-equity-like vehicle where annual investment budgets accumulate positions with 20-100x potential payoffs, so the cost of insurance is funded through eventual multi-year payoffs rather than quarterly P&L optimization.
“we're tasked by our investors to build and optimize and grow and manage through infinite time asymmetry in we're not tasked with taking risk. So we only risk mitigate. Now the best way to mitigate the cost is to buy things efficiently in efficient price and efficient structure efficient time horizon. But you know our our our in a sense our job is to invest. So our job is to go out and invest. We think of sort of the fund as kind of like a a private equity vehicle where every year we're drawing you know our spending budget, our investment budget. Some people think of it as a bleed, but given the you know multi-year duration of our book, it's not really. And we're just stacking year after year of investments and you know 20 30 40 50 100x potential payoffs that all inevitably will be tied together through correlation if and when there's a systemic market type event.”
Dredge's personal blind spot is the tendency to apply his own views about specific central banks or markets to investment decisions, when his discipline should be to buy anything cheap regardless of whether he likes or dislikes the market, because all the information needed is in the price of convexity, not his opinions about policy makers.
“the mistake I make is think that I know something and the only thing that I need to know is if it's cheap. If it's cheap, we should buy it. And I I shouldn't apply, well, I I like that central bank. I think what they're doing is right. I I like that market or I don't like that market. I shouldn't buy something that's not cheap because I don't like the market. And I shouldn't reject something that is cheap because I do like the market. All the information I need is in the price of convexity.”
Insurance should be persistent through time, not decaying to zero every month and resetting; the market structure incentives that pay traders/managers monthly compensation create unnatural pressure to allow insurance to decay rather than maintain constant protection.
“nobody in the real world nobody wants their insurance to decay to zero 12 11 10 9 8 7 6 5 4 32 1 month and then start over again. You'd like your insurance to be constant through time if not growing.”
Convex Strategies takes 8-18+ months from initial identification of a market structure (like Korean callable note demand) to executing the first trade, requiring fundamental analysis of flows, structuring mechanisms, hedging practices, vega duration impacts, and how leverage is building in the tail before deployment.
“there's a particular example that we're very active in at the moment that we started really working on, you know, pretty constantly in January 2024 and printed our first ticket in August 2025. So, it takes that long. It's that, you know, think of us as fundamental investors. We're working on the flows, the structuring, the hedging, how it's impacting the market, what the bank's willing to do from the basis risk they're taking. How is it building up exposures? How is that tail getting more and more fat to the left side?”
Convex Strategies internally maintains a whiteboard tracking central bank policy changes and major political events to identify regime shifts; they previously named the regime 'steepening cycle' (mid-2023) and recently renamed it 'correlation is going to be a [problem]' (January 2024 onwards).
“we have a great big whiteboard here in the office that we just know sort of central bank policy changes or political, you know, big events through the year and as we sort of see what regime we're in, cuz we we don't forecast anything, we we sort of name the board. the the previous board name I think that went up in sort of middle of summer 23 was the steepening cycle and then in January this year we changed the board name to excuse my language you know correlation is going to be a so right now correlation is driving things”
The vega and vol-of-vol (third and fourth moment) exposures are preferable to delta and gamma for long volatility strategies because they have longer duration, are more efficiently priced, and provide asymmetric returns over extended time horizons.
“we don't like short-term options because we really want to own the Volvall and the Vega, which is sort of a longdated leverage factor, not not delta, gamma. Delta and gamma are not asymmetric enough for what we're interested in. They take too much work, too.”
Emerging markets in the late 1980s had significantly fatter tails than US markets due to pegged currencies, limited participant access, and information asymmetries; applying US Treasury and US equity risk models to emerging markets was inappropriate because those models assumed efficient, liquid markets with thin tails.
“The currency was pegged. So the market didn't get to decide the price or there was limited access for participants or limited access to information. Now eventually I grew up and realized how naive I was and found out that all markets are that way particularly once they get regulated. You know, I learned as I went, you could see, particularly in the emerging market activities of Bank of America and others, you they're very much caught in this trap of, you know, making a little bit of money over a few years and losing a lot when something happened. they had the massive the turkey at Thanksgiving problem”
Nassim Taleb worked at Banker's Trust at the start of his career, where he learned a principle central to risk management: 'not trying to forecast the market, but managing your payout function to X,' which is the core distinction between prediction-based and outcome-based risk management.
“this is you know going back to another Nasim TBism. So Nasim also got at the start of his career banker's trust you know that that it's not trying to estimate X. It's not trying to forecast the market. It's manage managing your payout function to X. And that's really I guess the number one thing I learned in that sort of transition from how do we account for risk in a very regulated structure at Bank of America to how do we actually manage risk and manage our payout function to risk in a place like Bankers Trust”
Dredge's career has focused on understanding how assumptions about continuous liquidity, correlation stability, and normal distributions fail; this is the through-line from 1987 Black Monday through current volatility imbalances.
“I got a good firsthand look of the implications of making assumptions about continuous liquidity that were not true...I guess my biggest sort of shock into wait there's a different way to look at it was when I met and eventually went to work for the guys at Banker's Trust”
Portfolios with convexity had significantly better outcomes across 2019-2025: they participated in 2019 gains, cut drawdowns in Q1 2020, stayed invested for the 2020-2021 rally, mitigated 2022 losses, skipped the recession fears of 2023 and fully participated in 2023's 54% NASDAQ rally, the 2024 25% S&P rally, and the 2025 gold and equity rallies, demonstrating multi-year compounding advantage.
“if you had more participating risk in 2019, you had a really good year. And if you had something really efficient to cut off the draw down in Q1 2020, you didn't lose that compounding path. And then you stayed in the market and participated in the next three quarters of 2020 and then participated in 2021 and then had something that cut off the negative compounding in 2022. So you come into 2023 still fully risked up. You're not sitting there listening to the, you know, the bearish siren song of, oh, there's going to be a recession and they hiked rates. You come in and you get that 54% up year in NASDAQ in 2023 and you're still actively taking risk in 2024 and bada bing S&P goes up 25% two years in a row and the whole world's crying recession or the whole world's moaning and groaning about geopolitics and blah blah blah and then boom 2025 guess what happens you know it's another upyear gold goes up 60% bang you're in it”
Volatility selling through single-name autocallable ETFs represents massive leverage being added to the system: 40 new single-name autocallable ETFs were recently launched, with the market shifting from index-based autocallables to single-name structures beginning in 2024, creating exposure concentration and amplifying dispersion risk.
“V selling is a form of leverage which is the one we're focused on. So you've seen this massive explosion. I'm sure you saw just last week that you know they came out with 40 new names of auto callable singleame ETFs that are getting launched and that's just a massive form of leverage of selling pouring into the single names. the Asia structured product autocallable machine which dwarfs everything else in the world shifted from indices to single name stuff beginning of last year and that has just been building and building and building”
Dredge uses turbulence indices and Mahalanobis distance-based measures (similar to Mark Kitzman's research) to monitor volatility and correlation unusualness, trying to measure fragility in the system without making forecasts, though price is ultimately more practical than these models for actual trading decisions.
“I I think they can be very useful and and we have some, you know, funky maths that we play around with that that we're looking at. You know, we always in fact my my most recent note is titled initial conditions. So we're always trying to to understand where we are, how fragile is the sand pile. We're not forecasting. We're not making any predictions. We're just trying to measure fragility. So that you can imagine there's a lot of curvature in the math that we're doing to measure stuff.”
Most US and European regulatory frameworks are now allowing actively managed ETFs to run tail risk strategies (with constraints), making it possible for retail investors to access more sophisticated convex exposure than was previously available, though these are still limited compared to hedge fund approaches.
“there are increasingly in the US and somewhat in in in Europe in USITs form actively managed ETFs. They're still not allowing that here in Singapore yet, but we're working on it.”
Banker's Trust and Dredge pioneered structured products and derivatives hedging in emerging markets (interest rate swaps in Korea, callable notes, Bermudant swapions) with the advantage of having positive convexity and long optionality; this gave them a competitive advantage in entering nascent swap markets before competitors.
“it was much easier for them to buy into this concept. Hey, let's create these pools of positive convexity and use those than to be much more aggressive in our willingness to go and enter into new markets and create new, you know, create interest rate swap curves because we done the first callable note structures in Korea and Juan and we're now long optionality on interest rates. So, we're happy to go and start providing a bid offer spread in a swap market that had never existed before.”
Building a core risk business at Banker's Trust was based on the premise that in complex markets with actively managed pegged currencies and offshore/onshore dynamics, assumptions about historical volatility, correlation, and continuous liquidity are 'pretty stupid'; the business was structured to embed positive convexity while exploring new markets.
“what we constructed as what we called a core risk business with this sort of premise because of the complexity of risk around these actively managed pegged currencies, onshore, offshore markets that any assumptions about historical volatility and correlation and continuous liquidity were pretty stupid. And and so we sort of built the premise on the business on this premise of embedding positive convexity while we went out and figured out how to make markets and structure derivatives and underwrite risks and stuff.”
Volatility markets operate continuously (246 trading days per year at minimum, likely moving toward 24/7 with tokenization), requiring constant evaluation of supply/demand dynamics rather than periodic rebalancing; this continuous nature requires dedicated expertise.
“We're going to trade not we're going to tokenize everything and trade 24/7”
The allocation industry is slowly transitioning from silo-based asset allocation (evaluating each asset class separately) to 'total portfolio approach' (evaluating contributions of each holding to overall portfolio performance), but this transition is proceeding slowly because it requires admitting that 30 years of Sharpe-ratio-optimized holdings were mistakes.
“Fortunately, that's breaking up and I'd like to take some credit for it because I've been pounding these guys for 35 years and more and more people talking about, you know, the new catchphrase as though somehow it's this revolutionary innovation total portfolio approach and that we should actually evaluate everything on its contribution to the overall portfolio and that our objective should be the performance of the portfolio.”
Dredge worked for Baker's Trust, ABN Amro, and then retired to the buy-side to found Convex Strategies; the strategy he runs now is operating in the same 'food chain' of structured product volatility selling and convexity recycling that Banker's Trust pioneered in the early 1990s, which was originally Asia-dominated but has become increasingly global.
“Then did that and then did it again for AB and Amro later on and and then retired to the the buy side. And now we sit with our little business here, Convex Strategies, where we just really sit in the same food chain that my bankers trust guys and I were innovators of back in the early '9s of this structured product volatility selling behemoth that now is really global, but for the, you know, 35 years was an Asia dominated thing.”
Dredge spent his early career making poor forecasts (predicting he would avoid Black Monday by staying close to home, then being sent to Singapore; arriving in October 1987 just as the crash began), which taught him not to forecast future paths but instead to build resilient portfolios.
“I came out originally with Bank of America in 1987, just in time. I got here the first Monday of October 1987, just in time to catch the what is commonly known as Black Monday in most of the world, the October 87 crash, which we call Blacker Tuesday out here. I had, you know, graduated from uh business school of Berkeley and had taken a job with Bank of America in San Francisco because I didn't want to get further away from home, which is in Salt Lake City. I turned down New York jobs because I wanted to stay close to home and 3 months later they sent me to Singapore. So my my ability to forecast the future was laid bare fairly early on”
Janet Yellen's economic models and beliefs that Dredge studied under at UC Berkeley proved to be 'mythological' and wrong in practice, particularly when applied to emerging markets; this informed his skepticism toward models and his emphasis on empirical observation of real-world risk.
“I studied economics at Berkeley under Janet Yellen in my grad school days and uh learned over the years how mythological most of her beliefs were and her economic models and beliefs of the world. And that sort of just started me down this path of trying to figure out what risk was”