The sell-off at the long end of the U.S. bond market is turning into a global bond market rout. Yields are rising in countries around the world; in places where nominal long-end yields were negative, like Germany and Japan, they are flipping positive.
To put it more simply, as U.S. Treasury 10-year notes and 30-year bonds plummet in value, their nominal yields spike higher. That spike is contagious, causing long-end yields everywhere to rise.
There is serious danger here. The risk of market meltdown — or some kind of flash crash, reminiscent of what happened in March 2020 — is now high.
Events are moving so quickly, in fact, that markets could be in full meltdown mode by the time you read this (though hopefully not).
As a side note, we’ve lost track of how many times the “by the time you read this” qualifier has applied to same-day broadcast in the past year — a function of events unraveling at so great a speed that even a four-to-six-hour publishing window can make a huge difference.
The danger stalking markets now is related to something called a “market structure event.”
A market structure event is major price dislocation — like a meltdown or a flash crash — that happens for reasons unrelated to investor sentiment or a long-term fundamental outlook.
Market structure is a reference to the way investors are positioned: Who is long, who is short, how much leverage they have, and so on.
The phenomenon of “portfolio contagion” is notably a market structure issue: If leveraged hedge funds take a large hit to a certain area of their portfolios, for example, they often have to liquidate positions in unrelated areas of the portfolio to reduce exposure across the board.
In this manner, a sell-off can be contagious in that the selling jumps from one area of the portfolio to another; the linking factor is the structure of what the hedge funds are holding.
Portfolio contagion issues can even impact safe-haven assets to the degree that, in the midst of a meltdown, safe haven assets are like cash in a bank account; investors will sometimes sell a portion of their safe-haven positions to raise cash when everything else is going haywire. This is why, say, the price of gold can fall hard with everything else when a real meltdown hits.
Market structure can also refer to complex, behind-the-scenes arrangements comparable to the plumbing structure of a building. Like office tenants in the building, investors don’t see or think about the pipes, which are hidden in the walls; nor do they worry about boiler maintenance, as it is presumably some maintenance person’s job to take care of that.
But the tenants certainly notice when the pipes start to burst or the boiler explodes: If this happens for some technical behind-the-scenes reason, it is a market structure event (which can also be dubbed a plumbing event).
The most dangerous of all market structure events takes place when there is too much leverage in the system and too many market participants with the potential to become forced sellers.
Some version of this happened in the Crash of 1987, for example, as institutional investors tried to hedge their equity portfolio risk by selling S&P 500 futures as a hedge. The more the market dropped, the more that S&P 500 futures contracts were sold; with everyone doing this at once, and buyers stepping back, this caused a drop for the history books.
The danger posed today, in terms of market structure meltdown risk, comes from three areas: The U.S. treasury market; the liquid and beloved FANG names; and highly speculative, highly illiquid tech sector names, many of which are packaged into ETF products.
Here is a quick summary of the market structure risks posed by each area:
- The bond market is exposed to “convexity hedge unwinds” related to the mortgage market. We’ll explain in more detail what that means, but the shorthand version is that, as mortgage rates go up, institutional investors holding trillions of dollars’ worth of mortgage-backed securities have incentive to sell U.S. treasuries, causing yields to spike further.
- The FANG names are the stock market equivalent of high-yielding money market accounts or low-risk zero coupon bonds: Putting a big slug of money in, say, Apple or Google is almost like getting a risk-free return on cash with internal compounding (or that is how investors have acted anyway). The trouble is that, in times of market stress, there is no substitute for actual, honest-to-goodness cash; that in turn means investors could decide to sell FANG shares en masse, not because they have soured on big tech, but because they need liquidity.
- The low-liquidity, insane-valuation, speculative mania areas of the market — think garbage SPAC offerings and companies whose market cap makes no sense — have the potential to face substantial selling in a “zero bid” situation, meaning, investors desperate to get out could be flooding the market with sell orders, but finding no buyers on the other side. The trouble with Buzz Lightyear-type valuations (“To infinity and beyond!”) is that, when euphoria evaporates, the result can be a Wile E. Coyote-style air pocket due to an absence of buyers all the way down.
In some ways, a speculative mania is like throwing a months-long party entirely funded by credit cards.
The whole thing is a blast; the entertainment is free; nearly everyone is having a great time (and seemingly getting rich); and for most of the mania it feels like the party will never end.
The trouble with a mania, though, is that eventually, the bill always comes due.
The bill itself can come in many different forms: Sometimes, as described here, it is the inevitable consequence of an extreme build-up in market distortions, as expressed through the corrective mechanism of a violent market structure event.
Then, too, what’s going on in the bond market right now has the potential to dislocate far more than just stock prices.
If the U.S. treasury market goes further into a violent sell-off for non-economic reasons — meaning, bond market sellers dumping not because of a fundamental view, but for other reasons entirely, like the unwinding of convexity hedges — that means long-end yields (the nominal interest rate on the 10-year and 30-year) could spike to levels that really and truly start breaking things.
It’s really quite the domino chain, because if a market structure event on the bond side makes yields spike in a violent enough fashion, the Federal Reserve might be forced into another nuclear-level intervention on the scale of March 2020 — a kind of do-or-die yield curve control (YCC) at the long end (we’ll have to dive further into the mechanics of YCC some other day).
Just know that if a forced Fed intervention happens — where the Fed steps up buying at the long end of the bond market — it could mean an additional hundreds of billions of dollars, or even trillions of dollars, instantly added to the Federal Reserve balance sheet, with long-run consequences of who knows what.
Paradoxically, in the near term, this whole danger scenario is wildly bullish for the U.S. dollar, in part because other currencies will look so much worse in the event global financial markets go haywire, and in part because panicked U.S. investors will exit their international holdings and repatriate dollars en masse.
In TradeSmith Decoder we recently took an aggressively bullish U.S. dollar position spread across multiple currencies — even though we are long-term U.S. dollar bearish — and today we are feeling better than ever about that long USD exposure.
Our rationale for getting bullish on the USD was rooted in the Quantum Deficit Effect, as explained in these pages on Feb. 10 — but the dollar looks even better now with risk assets going into convulsions. When things get bad — as in really, really bad — everybody wants cash, which still means dollars (though someday it might not).
Circling back around to the unwinding of convexity hedges: The convexity hedge aspect of the bond market helps explain why U.S. treasuries could sell off much harder from here (which in turn could translate to yields spiking higher, which could then wreak further havoc).
The U.S. bond market has gone into convexity-related meltdown mode before (causing long-end yields to spike) with consequences rippling out across bond markets globally. The most famous example of this is probably the infamous “bond market massacre” of 1994, a rising interest rate period in which a slew of overleveraged hedge funds and shadow bank lenders were carried out on a stretcher.
Like earthquakes that vary in size, serious bond market dislocations furthermore seem to occur like clockwork every few years; we saw one in March 2019, and another one dubbed the “taper tantrum” in May 2013. (Then, too, the 1994 bond market massacre got going in March, which means these things tend to happen in a March-April-May timeframe — uh-oh.)
Convexity hedging — we’ll try to keep this simple — is a popular strategy for investors in the mortgage-backed securities market, which is gigantic (the U.S. has nearly $17 trillion in mortgage debt).
For investors in mortgage-backed securities (MBS), there is something called “prepayment risk,” which happens when homeowners decide to refinance and pay off their old mortgage loan, swapping it out for a new mortgage loan at a lower rate.
Institutional investors aren’t happy to see their MBS holdings close out early, because it typically means they will have to reinvest the previously deployed MBS funds at a lower yield. For these investors, prepayment risk means losing an attractive source of income or swapping it for a less attractive one.
What’s more, MBS prepayment risk is greatest when U.S. treasury prices are rising (and yields falling). That is because falling interest rates create a greater appetite for refinancing on the part of homeowners.
As a result of the relationship between rising treasury bond prices, falling interest rates, and increased payment risk, institutional investors like to “convexity hedge” their MBS holdings by purchasing large quantities of U.S. treasury bonds. The idea is that, if prepayment risk is rising, as an MBS investor, you will at least make money on U.S. treasuries going up.
Now, here is where it all goes haywire: When U.S. treasury prices threaten to decline sharply — the way they are now — interest rates start to rise, and prepayment risk goes away. Homeowners lose a taste for refinancing with rates moving higher; at the same time, institutional investors in mortgage-backed securities no longer have a need to hold large quantities of U.S. treasuries as a convexity hedge.
The upshot is that, as U.S. treasury prices start to fall, and yields start to spike, there is the added risk of a massive “convexity hedge unwind” as large institutional investors in the mortgage-backed securities market dump large quantities of U.S. treasuries they no longer need to hold.
Again, the “convexity hedge unwind” aspect of bonds is a strange deal — it gets into the plumbing of the nearly $17 trillion mortgage market — but the most important thing to be aware of is that the dynamic we are describing here has violently rocked bond markets before, on multiple occasions.
The worst-case scenario over the coming days and weeks — which, unfortunately, is all too possible — is a scenario in which all three market structure scenarios come to pass and start feeding on each other.
Imagine a scenario where long-end yields continue to spike higher — thanks to convexity hedge unwinding en masse via the mortgage market — even as panicky investors start liquidating their big tech holdings to raise cash, even as highly speculative low-liquidity tech names go “zero bid.”
Not only could this scenario actually unfold, it really wouldn’t be a surprise if it did; that is because, for one, we have seen the movie (or various forms of it) many times before; and for too, the serious, embedded dangers inherent in markets now are a function of reckoning for speculative overreach.
Sometimes the price paid for overreach isn’t paid by a formal bill, but rather the reaping of a whirlwind brought on by a violent market structure event (which never would have happened if not for a reckless degree of speculative excess build-up in the first place).
