Verdict:
MIXEDHow your thesis holds up against the knowledge base
After the historic drawdown in long-duration government bonds in the early 2020s, investors will treat long bonds as a broken safe haven and shift “anywhere but bonds” for safety going forward.
You brought a drawdown chart to a regime fight. Bonds can be a terrible investment and still be the safe haven—sometimes because growth collapses, sometimes because the state forces the issue. Your claim needs a forward mechanism, not PTSD.
“BofA's Chief Investment Strategist Michael Hartnett, has declared that the first half of the 2020s delivered what he terms “bond-market humiliation,” with long-duration government debt suffering unprecedented damage. The numbers are staggering. The iShares 20+ Year Treasury Bond ETF, a proxy for long bonds, lost a massive 31% in 2022 alone; one of its worst years on record. The maximum drawdown from its 2020 peak through late 2025 was a catastrophic -47.8%. As capital increasingly looks for a safe haven it will almost surely be "anything but bonds."”
Source: Hartnett (BofA)
Where the knowledge base challenges your reasoning
You’re extrapolating a regime outcome from a drawdown statistic (pain ≠ permanent death).
seriousYour mechanism is basically: “bonds got wrecked, therefore they can’t be a safe haven.” That’s not a mechanism; it’s trauma. The bond market can absolutely reassert itself as a hedge if the tightening finally bites and the market starts pricing weaker growth/inflation and an eventual pivot. In that setup, the long end can stay anchored or fall even while officials talk tough—i.e., duration works again precisely when the economy rolls over. The thesis needs a forward-looking driver (growth/inflation/policy path), not a rear-view mirror drawdown chart.
📚 2000 tightening into the dot-com bust: curve flattened and long yields didn’t validate ‘new economy’ optimism; recession followed.
📚 2006–2007 pre-GFC flattening/inversion: long yields stayed low despite hikes, foreshadowing housing/credit stress.
📚 Japan 1990s–2000s: repeated policy efforts met structurally low long yields as growth/inflation expectations stayed depressed.
📚 2018 flattening before the 2019 pivot: long yields rolled over as markets priced the end of the hiking cycle.
“Anything but bonds” ignores the ugly reality: in some regimes the state forces bonds back into the ‘safe’ bucket—by design.
seriousIf deficits/issuance and geopolitical/industrial priorities dominate, the adjustment channel may not be austerity; it can be yield suppression, regulatory nudges, and variants of financial repression. In that world, investors may hate bonds in real terms, but they can still become the system’s ‘safe’ asset in nominal terms because policy makes them so (explicitly or implicitly). Your thesis asserts capital ‘almost surely’ flees bonds; the counter is that policy can corral domestic balance sheets into holding them anyway, especially if national-security/industrial spending is treated as non-negotiable.
📚 US 1942–1951: yield caps and repression produced deeply negative real returns for bondholders during inflation spikes
📚 UK 1940s–1970s: long periods of negative real gilt returns amid inflation and policy constraints
📚 Japan 2016–2023: formal YCC suppressed yields and distorted term structure (different inflation outcome, same mechanism of capped yields)
📚 US 2020–2021: heavy QE and suppressed term premia alongside rising inflation—real yields deeply negative
Safe-haven demand can still show up as a bond bid—just not necessarily in the part of the curve you’re talking about.
minorEven if long duration has ugly supply/term-premium dynamics, stress can still drive demand for sovereign paper—often concentrated in bills/short coupons if the buyer base prefers liquidity and low mark-to-market risk. That doesn’t validate “anything but bonds”; it’s more like “bills, not 30s.” If you don’t specify *which* bonds and *what* stress regime, you’re making a tradeable claim untradeable.
📚 1994 bond massacre: leveraged positions and convexity hedging amplified rate volatility
📚 1998 LTCM: levered relative-value unwind transmitted stress into sovereign curves
📚 March 2020 Treasury selloff: levered basis trades and dealer balance-sheet limits worsened dysfunction
📚 Eurozone 2011–2012: periphery debt became pro-cyclical as the marginal buyer vanished until ECB backstops arrived
What you should verify before putting money on this
📊You haven’t established whether the economy is actually in the setup where the long end should stay low/fall during tightening (the classic ‘bond market is indicting growth’ signal).
→ What to check: Is the curve flattening/inverting with the long end staying anchored *despite* restrictive policy and upbeat official rhetoric, and is that being confirmed by weakening activity/credit spreads?
📊You haven’t shown that the fiscal/issuance + term premium regime is still the dominant driver (the condition that makes duration structurally poor risk/reward).
→ What to check: Are deficits and long-end issuance still persistent/large *and* is term premium rising/remaining elevated alongside a weaker stock-bond hedge relationship?
📊You haven’t verified the buyer-base deterioration claim that would make long Treasuries fail as a haven in stress.
→ What to check: Are official holders concentrating in bills while long-end auctions lean more on mark-to-market/levered intermediaries, and do risk-off episodes coincide with Treasury selling rather than buying?
What your thesis needs to be true to work
- •This thesis requires long-duration sovereign bonds to remain an unreliable hedge in risk-off (i.e., stock-bond correlation stays unstable rather than reverting to the old negative-correlation regime).
- •This thesis requires fiscal/issuance pressure and inflation uncertainty to persist enough that term premium stays meaningfully positive (so duration continues to have poor risk/reward versus cash-like instruments).
- •This thesis requires the marginal buyer of sovereign duration to remain more mark-to-market/levered and less price-insensitive (so ‘safe-haven’ flows don’t automatically translate into a clean long-end rally when stress hits).
Conditions that would upgrade or downgrade this verdict
✅ Upgrade if:
- In the next equity drawdown, long-duration Treasuries fail to rally (or sell off) while measures consistent with term premium/supply pressure remain elevated.
- Auction outcomes and flow data show the marginal bid for duration is increasingly mark-to-market/levered while official/price-insensitive demand concentrates in bills.
- Deficits/issuance remain historically large with no credible consolidation signal, and stock-bond correlation stays persistently unreliable.
❌ Downgrade if:
- A tightening-to-slowdown sequence emerges: activity weakens, credit spreads widen, and the long end rallies meaningfully as markets price a pivot/cuts (duration ‘wins’ again).
- Inflation uncertainty credibly re-anchors and term premium compresses materially, restoring the classic risk-off bid for duration.
- Policy shifts toward explicit/implicit yield suppression (repression/YCC-style tools), making bonds ‘safe’ in nominal terms even if real returns are ugly—undercutting the idea that capital can simply choose ‘anything but bonds.’