Michael Green published another one of his structural-markets essays this week. While his central argument is something I will leave you to ponder, there are some parts that are worth highlighting.
“Economics often teaches us to think of markets as if they exist first and rules are added later. That is backwards. Every market is shaped by rules. Accounting rules matter. Tax rules matter. Capital requirements matter. Fiduciary standards matter. Index methodologies matter. Even deciding what qualifies as a “safe” asset changes who buys what, when they buy it, and why.
That does not make markets fake. It means market prices reflect more than private opinions about value. This distinction becomes especially important when regulation or institutional design strongly favors one form of investing over another.“
I’d go further, actually: understanding this should be a prerequisite for investing at all. Learn it, and then put your money in.
Take European government bonds through the negative-rate years. The infamous framing — that nominal yield tells you whether a bond is worth owning — is close to useless; what matters is the real rate, and Europe’s negative nominal yields coincided with weak growth and near-zero inflation, so there was a genuine economic story underneath them. But I’ve always wondered how much of that pricing was a preference the market actually revealed, versus an artifact of banks and insurers being required, by capital rules, to hold those bonds at whatever yield was on offer. That’s the general pattern: banking and insurance regulation is one of the most powerful forces determining what sits on a balance sheet, period.
Another great example is the relationship between the regulator and rating agencies. Someone has to analyse credit risk for the whole system, and the agencies do a job nobody else wants, with a business model that’s structurally compromised since they’re paid by the issuers they rate (see 2008). And yet the function is indispensable; you can’t outsource issuer analysis to every buyer individually. The catch is that regulators don’t always defer to the agencies’ output. When it suited the political project of European integration, EU institutions considered Greek debt more or less irrespective of what the rating agencies said, because the single-market concept can’t tolerate discriminating against one of its own.
Sovereign risk-free status works the same way. When S&P downgraded U.S. debt in 2011, the market simply overruled the agency and kept treating Treasuries as the risk-free asset. And regulators, in effect, ratified that judgment. No major institution in the developed world is barred from holding Treasuries on its balance sheet. Then there are the commercial workarounds: JPMorgan runs a money-market fund in China rated “AAA*,” asterisk included, and regulators on both sides of that border are fine with it.
“Traditionally, a bond manager started with the needs of the buyer. How much income does the client need? How much interest-rate risk can the client tolerate? How much credit risk is appropriate? The portfolio was built around those objectives.
A market-weighted bond index reverses that logic. The borrower decides how much debt to issue, what maturity to choose, and what kind of security to sell. The index then gives that debt a weight largely according to how much is outstanding. Investors who are required—or strongly encouraged—to track the index respond to those financing decisions. In effect, the seller of credit increasingly determines the portfolio of the buyer of credit.
That is a very different objective function.“
Borrowers issue according to their own needs, not the market’s, and nowhere is that clearer than Switzerland. Swiss institutions have enormous appetite for Swiss franc bonds; the Swiss government has very little interest in supplying them, because its objective is fiscal and financial stability, not satisfying institutional demand. That asymmetry is also why something like XEON depends entirely on the existence of the EU, the single market, and the euro. No individual European country (outside maybe Germany) could, on its own currency, sustain a debt market deep and liquid enough to backstop a money-market fund at scale. Finland has the rating but not the debt; Italy has the debt but not the rating. Only the euro area in aggregate has both.
An index, as Green frames it, is simply a measure of something. The whole trick is knowing exactly what that something is.
“Again, this is not necessarily wrong. Index investing has many advantages: low cost, transparency, diversification, and consistency. Nor does every benchmarked investor behave mechanically. But once benchmark-following capital becomes large enough, we cannot automatically treat every market price as the result of independent investors expressing views about growth, inflation, risk, or value.
Some demand may simply be the consequence of a rule.“
The tension isn’t “passive investors free-riding on active ones”, that’s the wrong frame. The tension exists because the market is full of participants operating under fundamentally different mandates. A passive investor believes they’re buying the aggregated wisdom of the crowd; what they’re actually buying, in Green’s framing, is a fork from someone whose mandate was to design a spoon.
“A market-value-weighted bond index answers a simple question: how much eligible debt is outstanding, at market value?
That is a perfectly coherent thing to measure.
It does not answer a different question: how much duration should an investor buy?
Nor does it separately choose curve exposure, convexity, mortgage optionality, or credit risk. Those characteristics emerge from what issuers sell, how borrowers refinance, what securities qualify, and where prices move.“
Once issuers understand there’s a structural buyer for anything that qualifies, they’re free to issue whatever suits their own balance sheet, not whatever the buyer actually needs. Untangling who’s really in the driver’s seat is harder here than in equities, because a bond carries more dimensions than a share does.
“When bond fund managers build portfolios, they are typically considering the objective of the CUSTOMER and building a portfolio to suit. When they were the dominant marginal source of demand, it could be argued the “market” reflected the intersection of supply AND demand.“
If active managers are in the driver seat, the index is fine because there is still the price component attached to each security. The lower the price, the lower the weight in the index.
“But the minute the dominant marginal source of demand is determined by benchmark FOLLOWERS, the odd phenomenon emerges that “supply determines demand.” The distinction matters because enormous pools of capital manage relative to these benchmarks.“
Mike is obsessed with passive investors, but I guess the problem here is exacerbated by what he was saying earlier about rules. Let’s not forget it.
“Consider the Bloomberg U.S. Aggregate. In the licensed history I hold, its month-end option-adjusted duration has ranged from 3.63 years in March 2009 to 6.69 years in December 2021. It was 5.80 years on September 18, 2026.“
“Nobody managing the index sat down in 2009 and decided that 3.63 years was the appropriate amount of interest-rate risk. Nobody held a meeting in 2021 and decided that 6.69 years was better. The debt stock changed. Mortgage behavior changed. Issuance changed. Prices changed. The benchmark reported the result. The index describes the market.
Once portfolios are managed relative to the index, description becomes instruction. And once that happens, the mechanical portion of demand is no longer an independent opinion about value. It is demand generated by the rule itself.“
That should worry anyone holding an aggregate bond ETF on the assumption that “I checked, it’s a seven-year duration fund”. That number is a snapshot, not a promise, and the provider isn’t obligated to call when it drifts. There’s good evidence for how badly this can go. Research on issuer behavior, Appleton’s, among others, shows companies systematically extending maturities when rates are low and shortening them when rates are high. Which is exactly why the Bloomberg U.S. Corporate index’s duration climbed from roughly seven years at the end of 2018 to nearly 8.9 by 2020, right as companies locked in cheap long-term financing and before the 10-year yield rose some 450 basis points. Jeffrey Gundlach called this in real time, arguing in a 2020 webcast that the Agg’s yield-to-duration ratio made it arguably the worst moment ever to hold a passive fixed-income fund. The bill came due in 2022, when the Agg posted its worst calendar year since its 1976 inception, down roughly 13%.
Every bond has an issuer on the other side optimizing for their own interest…unless they’re compelled to issue, in which case you may not want to be the buyer anyway. Investment-grade indices have quietly gotten riskier through exactly this mechanism: BBB, the lowest rung still inside investment grade, went from under a fifth of investment-grade corporate bond indices in 2003 to more than half by 2019. No retail investor in an “IG bond fund” voted for that; it happened purely because that’s what got issued and kept its rating. I remember this shift arriving in real time: every wholesaler walking through my office door was pitching it (and nobody wanted to be the one to suggest capping BBB exposure, because that meant giving up carry relative to the benchmark).
The sharper problem hits when a downgrade occurs. This is quite a known issue in the industry (ask Michael Milken about Fallen Angels), which is why mutual funds and separately managed accounts typically preserve manager discretion on timing: a manager can actually talk to a known investor base and agree on an approach. A passive fund holding only investment-grade paper has no such option; it’s structurally forced to dump any bond that loses its rating, at whatever price is on offer, exactly when that credit is under the most pressure.
And then there is the last consideration, the less relevant of all because you should not invest in this type of bonds regsrdless 😉 During the shale boom, energy became one of the largest sectors in the U.S. high-yield index purely because frackers issued so much debt to fund drilling. Energy was close to 15% of the market and the largest issuing sector by 2014. Anyone holding a “diversified” high-yield ETF was, without making a decision, making a large bet on oil. When crude collapsed, the high-yield index fell about 6% while the energy component alone fell 22%: concentration risk manufactured entirely by issuance patterns, not by any active call. Equity indices drift by sector too, but the mechanism differs: sector weight there shifts because investors bid up whatever they find most promising, a demand-side story (though Green would probably push back on how clean that distinction really is).
A bond fund built around explicit duration and credit-quality targets doesn’t solve all of this, but it gets you closer to owning what you meant to own. If everyone converges on the same target — AAA paper at 8-year duration, say — that segment gets bid up and everything else becomes relatively cheap by comparison…but this is a problem for another day.
At minimum, you stop owning 3-year paper only because it’s cheap for the Italian Treasury to issue, or Greek debt only because Greece shares a currency with Germany.
“The broader lesson is that an index rule does not remain a mere description once capital is tied to it. The people managing that capital act on the rule. Methodology becomes market structure.“
What I am reading now:
Follow me on Bluesky @nprotasoni.bsky.social
0 Comments