SI426: The Hard Part Isn't Buying Trend Following ft. Yoav Git

Original Media · 1h 11m · English · Podcast · Top Traders Unplugged

TLDR

In this episode of Top Traders Unplugged, host Alan Dunn and guest Yoav discuss how to construct CTA portfolios for different objectives — standalone alpha, equity convexity (crisis alpha), or bond convexity. They define two types of convexity: return convexity (holding assets negatively correlated to equities, like bonds in low-inflation regimes) and exposure convexity (mechanically reducing equity beta as trends reverse). CTAs are capital-efficient for bond exposure but less so for equities due to risk-balancing across asset classes. Commodity trend adds the most diversification to a liquid CTA base (low correlation, boosts Sharpe from ~0.6 to ~0.8), while fast trend only helps against very sharp, short-lived equity shocks. Inflation regime is the key driver: bond trend offers positive convexity across regimes, commodity trend shines when inflation rises, and equity trend struggles due to mean-reversion. The optimal CTA mix depends on whether the investor wants absolute return, equity downside protection, or bond downside protection.

For you

PRO members can sign in and request more information — or their own version of this digest, tuned to what they care about.

Sign in to unlock

Surprising Insights

•

Mainstream CTAs provide far less equity exposure convexity than most investors assume — a 10% allocation yields only ~3.5% equity beta reduction due to risk-balancing that delevers equities.

•

Commodity trend adds zero equity convexity but is the best diversifier for standalone Sharpe improvement, while alternative trend adds neither equity convexity nor incremental Sharpe once commodity trend is included.

•

Fast trend's convexity advantage disappears at quarterly horizons because its aggressive risk management keeps average equity exposure lower than slow trend's.

•

Equity trend as a standalone strategy has historically had near-zero or negative Sharpe (0.1-0.2) due to persistent negative autocorrelation from value buyers, while bond trend has been significantly higher — contrary to the intuition that equity trends should be the core of trend following.

•

Bond trend maintains positive convexity to equity drawdowns even during high-inflation regimes when long bonds fail, because CTAs mechanically flip to short bonds as yields rise.

Key takeaways

•

CTAs provide two distinct types of convexity: return convexity (holding assets that move opposite equities, immediate but not guaranteed) and exposure convexity (mechanically shorting equities as they fall, guaranteed but delayed).

•

A standard 15% vol CTA delevers equities and levers bonds to equalize risk, making it ~3x more capital-efficient for bond exposure than equity exposure — a 10% CTA allocation gives ~10% bond beta but only ~3.5% equity beta.

•

Commodity trend is the single best diversifier to a liquid CTA base: historically ~20-30% correlation to mainstream trend, boosting portfolio Sharpe from ~0.6 to ~0.8, while alternative trend adds almost nothing once commodity trend is included.

•

Fast trend acts as a first responder — it flips positions quickly and provides double the convexity of slow trend in one-month equity crises — but its lower average commitment means it matches slow trend only at quarterly horizons and underperforms in protracted drawdowns.

•

Bond trend delivers positive convexity to equity drawdowns across all inflation regimes because CTAs flip bond positions when correlations shift; equity trend historically has low/negative Sharpe due to mean-reversion (value buyers step in), while commodity trend is the primary inflation hedge.

•

Inflation regime drives everything: rising inflation hurts long bonds but helps commodity trend; low inflation helps bonds but hurts commodity trend. A bond + commodity trend combination creates an 'inflation plus 4%' all-weather defensive block, while equities are a growth block (inflation plus variable growth premium).

•

For equity convexity, investors need large allocations to dedicated equity trend (near 1:1 to desired beta reduction) or accept that mainstream CTAs provide limited equity exposure convexity; for bond convexity, a dedicated bond trend allocation is vastly more capital-efficient than any alternative.

•

The number and mix of CTA managers should match the objective: a liquid CTA index replicator as cheap beta, plus commodity trend for alpha, plus optionally bond trend for bond convexity, equity trend for equity convexity (if sized large), and fast trend only if insuring against sharp, short shocks.

Notable quotes

“CTAs are difficult to live with. You go down the stairs, come up the elevators, you might have three years with no performance. Your investment committee will look at you like an idiot.”

▸ Notable quotes · 0:01

“Convexity is something nicer. So convexity, well, positive convexity is something nicer. It's a little bit like buying an option. It makes money both on the upside and on the downside.”

▸ Notable quotes · 12:05

“If you look at the total portfolio approach, everybody wants, you know, you actually want a lot of CTAs in your portfolio. You put it into your Claude, and Claude will tell you you want 20 CTAs.”

▸ Notable quotes · 9:49

“Most CTAs, let's call it, run at 15% vol. What they will do is they will actually delever their position in equity and lever up the exposure in bonds so that both asset class internally within the CTA are running at 15%.”

▸ Notable quotes · 21:04

“If you add commodity trend, it adds no value whatsoever to add alternative trend. Like the additional sharp from alternative trend after you add commodity trend is like literally zero.”

▸ Notable quotes · 61:00

“Equities are a growth game. So what you see is when inflation is going down or inflation is low, equity will more than give you will have a very will give you inflation plus 10 or 12%. But when you think about, but when you think at a high inflation environment, equities are kind of barely are still able to give you inflation, but they're not going to give you the inflation plus.”

▸ Notable quotes · 42:32

Chapters

•

0:00 — Intro and CTA Performance Update: Alan Dunn and Yoav open with casual conversation about travel and internships, then review November CTA performance: SocGen CTA Index down 3.2% MTD but up 10.2% YTD, driven by bond selloff and commodity reversals post-midterms.

•

12:06 — Defining Convexity: Return vs Exposure: Yoav defines positive convexity as option-like payoff on both tails, then distinguishes return convexity (holding negatively correlated assets, immediate but regime-dependent) from exposure convexity (mechanically reducing beta as trends reverse, guaranteed but delayed).

•

21:00 — Capital Efficiency Across Asset Classes: Because CTAs risk-balance at 15% vol, they delever equities and lever bonds — making a 10% CTA allocation equivalent to ~10% bond beta but only ~3.5% equity beta. Speed of trend further modulates average exposure: slow trend ~60% committed, fast trend ~30%.

•

27:00 — Inflation Regimes and Asset Class Convexity: Bond-equity correlation flips positive above ~4-5% inflation. Bond trend stays positively convex across regimes by flipping positions. Commodity trend mirrors long bonds: shines when inflation rises, struggles when low. Equity trend offers almost no bond convexity historically.

•

57:00 — Building the CTA Portfolio: Alpha vs Convexity Objectives: For standalone alpha: liquid CTA base + commodity trend (biggest Sharpe boost) + spread trend. For equity convexity: need large dedicated equity trend allocation; fast trend only helps for sharp 1-month shocks. For bond convexity: dedicated bond trend is unmatched in capital efficiency.

•

44:00 — Stagflation, Industrial Production, and AI Capex: Yoav discusses stagflation concerns, the Man Group paper on inflation/industrial production regimes (skeptical of its granular conclusions), and his bullish view on AI infrastructure capex driving industrial commodities — commodity trend wins in both overheating and stagflation if inflation persists.

•

67:00 — Manager Selection and Behavioral Risks: Investors chase CTA performance like trend followers themselves (60% slow trend exposure to trend following), which is destructive. Yoav hints at a future discussion on optimal manager count and portfolio stability across managers.

Speakers

•

Alan Dunn — Host

•

Yoav — Guest

⋮