TTU154: Why Trend Following Isn't Crash Insurance ft. Philippe Jordan

Philippe Jordan · Capital Fund Management (CFM)

Trend following diversifies portfolios but won't pay for a market crash; that isn't its role.

Jordan argues trend following is a valuable portfolio component that diversifies, but it should not be expected to offset losses in a crash. He cites a long-run Sharpe of about 0.5, which he says beats the stock market over the very long term (not over the last decade or two). He frames trend following's role as diversification, not crisis insurance, and says the strategy has been through periods of severe pain, such as the 1994 Mexican fixed income crisis, when it lost heavily.

“It does diversify, but it's not gonna pay for your crash. That's not what it does.”at 0:22

Philippe Jordan · Capital Fund Management (CFM)

Crowded deleveraging amplifies losses: risk systems pile in on drops, then feed negative feedback loops.

Drawing on the August 2007 quant quake, Jordan describes how, during the mortgage deleveraging of bank balance sheets, trading desks needed balance sheet back and sought it across departments. Statistical businesses first treated each drop as a buying opportunity until their risk systems saturated, which produced a negative feedback loop that was amplified each day. He says his firm chose to get out of the way, though getting back in quickly enough afterwards was not perfect.

“people that run businesses statistically will look at any drop as an opportunity to add risk until their own risk systems saturate and you get a negative feedback that gets amplified with every day that goes through.”at 12:01

Philippe Jordan · Capital Fund Management (CFM)

Equity market gains are largely driven by fund flows, and flows reliably move prices.

Jordan says the US equity market's strong performance is in large measure due to flows, and that flows matter and cause reversal flows. He calls the idea that flows move prices unrevolutionary. He links this to the question of lasting edges: high-frequency strategies have a high impact on flows and so are more exposed to them, while slower strategies with turnover have less impact, and strategies with no turnover have a lot of impact.

“The US equity market's magnificent success is in large measure due to, yes, flows... flows do matter. Reversal flows do occur and flows do have impact.”at 36:51

Philippe Jordan · Capital Fund Management (CFM)

AI will become as ubiquitous as electricity or the internet, and everyone will manage AI agents.

Jordan argues that AI is becoming a general-purpose technology like electricity or the internet. He says it is not peculiar to CFM, since everybody will be developing and managing agents. Tools will let people type what they want, and firms will then search the internet on their behalf. Managing money is only one of many areas that will be deeply affected, and he expects the impact on society to be immense.

“I don't think that is peculiar to CFM. I think everybody is going to be doing that. I think everybody is going to be a manager of agents.”at 25:13

Philippe Jordan · Capital Fund Management (CFM)

Within about seven years, not using extensive technology to manage portfolios will seem exotic.

Jordan argues that systematic and technology-driven investing, which was once a hard sell, has become normal as AI has entered people's lives and imaginations. He says institutions and large family offices have already embraced systematic investing, and he expects AI to complete the shift. He predicts that within about seven years the exotic position will be managing portfolios without extensive technology, though venture capitalists in Menlo Park remain reluctant.

“I think that within probably seven years, it's the contrary that's going to be thought of as being exotic. People that don't use extensive technology to manage portfolios.”at 33:14

Philippe Jordan · Capital Fund Management (CFM)

Rising computing power per dollar is the biggest leverage for systematic investing, and it keeps surprising.

Jordan says the scientific framework is leveraged by increasing compute power per dollar spent, and calls this the big leverage. Over 35 years, the trend has kept rising and has repeatedly surprised people who thought Moore's Law would flatten. The latest jump came from graphics cards built for games. He does not know what the next source of compute will be, but he expects the trend to continue.

“So it's the scientific framework over increasing amounts of computing power per dollar spent. That's the big leverage.”at 18:52

Philippe Jordan · Capital Fund Management (CFM)

AI models can be run on short data histories if they are small, diversified risk allocations.

Jordan says AI agents now trawl papers and quickly test them, which saves researcher time. He expects agents to start trading, and says this will spread beyond CFM. His argument on standards: a diversified book of hundreds or thousands of alpha engines can tolerate weaker robustness per model, while a core trend system needs a lot of statistical robustness. An AI model with only about two years of data but good results and under 1% of overall risk is a form of diversification, so the firm is playing the technology through diversification.

“if you have an AI model that doesn't have any robustness because you're not looking at a hundred years worth of data, you might be looking at two, but you have an intuition and you're getting good results... within less than 1% of your overall risk, then you're playing diversification.”at 29:16

Philippe Jordan · Capital Fund Management (CFM)

Western demographic decline and social-contract strain are underpriced long-term risks that will reshape markets.

Philippe names demographics in the West as one of the risks people underestimate today. Falling birth rates make it hard to reverse the trend, and Western societies were built for a middle class with children, so the social contract must change. He is 'rapidly running into' retirement age with the baby boom cohort, and he argues this drives generational conflict and perceived inequality. He cites France as an example: its Gini coefficient peaked in 1970 and has fallen since, and it has one of the lowest inequality levels in the G7, yet anxieties persist. He concludes these are long-term risks that show up in the short term with consequences, are severely underpriced, and will shape markets structurally over the next ten to fifteen years.

“I think these are long term risks, but long term risks have a way of showing up somewhere short term and being quite consequential. So that's something I think that is severely underpriced and will shape markets in a structural fashion for the next ten, fifteen years.”at 41:09

Philippe Jordan · Capital Fund Management (CFM)

Chinese futures contracts are the only truly diversifying 'alternative' markets; other alternatives fail hurdles.

Philippe tested what people call alternative markets against his criteria. These are persistent low correlation to the existing mix, and operational soundness, meaning they are not too illiquid, complicated or hard to process. Chinese contracts were the only ones that passed. They were truly diversifying with decorrelation characteristics to CFM's existing mix, deeply liquid, and operationally sound to access, although complicated because nothing is easy in China. Other alternatives such as electricity did not meet the thresholds that merit investment.

“the only true alternative markets were the Chinese markets. The Chinese contracts were truly diversifying ... they were deeply liquid ... but they they were operationally sound to access.”at 43:51