What Is a Bond Index?
Bond index measures the performance of a defined set of bonds, and most are weighted by the amount of debt outstanding rather than by anything about quality. Tracking one therefore means lending the most money to whoever has borrowed the most, which is a deliberate choice that is rarely examined.
An equity index and a bond index share a word and not a method. The difference decides where your money goes, and it points somewhere most people would not choose on purpose.
How it works
Weight is proportional to debt outstanding. An issuer with 50 billion of bonds in the index carries ten times the weight of one with 5 billion.
So borrowing more increases your weight in the index. An issuer that doubles its debt doubles its share of everyone’s tracker fund, and nothing about the issuer needed to improve for that to happen.
An equity index weights by market value, which is the market’s assessment of what a company is worth. Bigger weight follows from the market thinking more of you.
In bonds the causation runs the other way. Bigger weight follows from owing more, which is closer to the opposite signal.
The duration drifts on its own
Issuers choose when and how long to borrow. In a low-rate period, issuers lock in cheap money for longer — so the index fills with long-dated bonds and its duration rises.
Which means your interest-rate exposure moved without you deciding anything. The tracker followed the index, the index followed issuance, and issuance followed what suited borrowers.
The effect is perverse in exactly the wrong direction. Duration lengthens when rates are low — precisely when a rate rise would do the most damage, which bond duration works through.
A worked example
A government-heavy aggregate index in a low-rate decade. Governments issue thirty-year debt cheaply, that debt enters the index at its full size, and the index’s average duration climbs from 5 to 8.
A holder who bought a tracker for stability now owns a duration-8 position they never chose. Rates rise a point and the fund falls about 8%, when the exposure they thought they bought would have fallen about 5%.
Nothing malfunctioned. The index did what it says it does, the fund tracked it faithfully, and the risk carried changed by more than half without any decision being made.
Why it is still used
Because the alternatives are worse in different ways. Equal-weighting thousands of bonds is impractical; weighting by credit quality requires judgements the index would then be making on your behalf.
And because it is investable. Amount-outstanding weighting has the property that the index can actually be held at scale — there is enough of each bond to buy in proportion. Most cleverer schemes cannot be implemented.
So it is a defensible compromise rather than a good idea, and knowing which of those it is changes how much faith to put in it.
The original data
This site’s thirty-year fee measurement: 5 basis points costs 1.5% of the final balance, 20 costs 5.8%, 75 costs 20.2%, 150 costs 36.5%.
Index bond funds sit at the cheap end of that table, which is their strongest argument — the fee saving against an active bond manager is large and certain, and it partly compensates for the weighting problem above.
And the index rebalances constantly as bonds mature and new ones are issued — each change costing a spread inside the fund, on top of the published charge.
The alternatives people have tried
GDP weighting allocates to countries by economic size rather than by debt outstanding, which stops the most indebted governments dominating. It requires a judgement about which measure of economic size to use, and it is hard to implement at scale.
Fundamental weighting uses issuer characteristics — cash flows, assets, ability to service debt — instead of amount borrowed. It is closer to what most people assume an index does, and it involves the provider making credit judgements on your behalf.
Equal weighting within credit bands is simpler and produces a portfolio with far more small issues, each of which is expensive to trade.
None has displaced the standard approach, mostly for the practical reason above: amount-outstanding weighting is the only scheme where the index is reliably buyable in proportion. The default persists because it is implementable rather than because it is right, and that is worth knowing when somebody describes it as neutral.
What the name does not tell you
“Aggregate” is not a defined term. Two funds both tracking an aggregate bond index can hold materially different things — one dominated by government debt, another carrying a large share of mortgage-backed securities with prepayment behaviour that is nothing like a plain bond.
Nor is “global”. A global bond index may be currency-hedged or unhedged, and that single choice usually moves the fund more than the bonds inside it do. An unhedged global bond fund is substantially a currency position wearing a fixed-income label.
And inclusion rules vary. Minimum issue size, minimum remaining maturity, and which credit ratings qualify all differ between providers, and each rule quietly changes what you own.
The factsheet answers all of this in a paragraph. Reading it takes a minute and is the only way to know whether two similarly named funds are comparable at all.
When it fails
The characteristic failure is assuming a bond index works like an equity index. The reasoning that makes equity indexing compelling — you own the market in proportion to what it is worth, nobody is making a judgement on your behalf — does not transfer. A bond index makes a judgement, and the judgement is to lend most to whoever owes most. That is a real position, taken by default, by people who believed they were avoiding taking positions.
A second failure is not checking the duration periodically, since it moves without notice.
A third is treating an aggregate index as diversified across risk types. Many are dominated by government debt, which is one exposure rather than many.
A fourth is comparing two bond indices without checking what they include — investment grade, high yield, global, hedged or unhedged all wear similar names.
And a fifth is assuming the index is neutral. Every weighting scheme is a choice, and this one favours the most indebted borrowers by construction.
Related
Bond fund covers the vehicle that tracks one. Bond duration covers the exposure that drifts. And bond market covers where the constituents actually trade.
Everybody understands what an equity index does. Almost nobody notices that a bond index does something structurally different — it allocates by how much somebody owes, not by how much they are worth. Once you see that, index bond investing becomes a choice rather than a default.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.