ETFs are eating the bond market

At the start of the century, Barclays Global Investors launched a weird little fund in Canada, a backwater of global finance. Today, its myriad offspring are rewiring swaths of the $130tn bond market.

BGI’s creation was an exchange-traded fund, which occupied a pretty exotic corner of the investment industry. ETFs had only been around for a decade and there were less than 90 in existence, with total assets of $70bn. But this was a niche product even for a niche industry: rather than some big stock market index, this ETF would track the Canadian bond market.

The iShares Core Canadian Universe Bond Index ETF — launched on November 20, 2000 — was the first of its kind to invest in fixed income rather than equities. It represented a trial balloon for BGI, the asset management arm of the British bank.

Another two years had passed before BGI launched a smattering of bond ETFs in the US, to feeble interest. Some ETF believers were sceptical that the vehicles would ever prove successful in fixed income. They were wrong.

This is no longer a niche. In fact, fixed income ETFs — now a $2tn asset class — are shaking up the old order in a shadowy but important pillar of finance that has long been ruled by big banks and investment groups.

Even the assets under management chart above understates how powerful this trend is. After all, the past few years have not been kind to the bond market, depressing the value of most fixed income ETFs and obscuring huge inflows. Even in 2022 — one of the worst years in history for the asset class — bond ETFs attracted $245bn of investor money. They have taken in another $195bn so far this year.

As a result, new players are coming to the fore. Just a decade ago, only one of the 20 biggest bond funds in the world was an ETF (and it was just a share class offshoot of a larger Vanguard fund) In the top 50 there were just three in total. Today, five of the 20 largest bond market vehicles are ETFs, and there are 18 in the top 50:

This worries quite a few people, with the European Central Bank, the IMF, several academics and a horde of investors concerned that fixed income ETFs could somehow prove hazardous, perhaps even systemically so. The investor Carl Icahn once called them “extremely dangerous”.

At the same time, the ETFs’ unique mechanism of tracking indices has reinforced and accelerated several other major bond market trends. The outcome is some of the biggest changes to affect the global fixed income markets in a generation.

FT Alphaville has written about various aspects of these issues before, but a deeper, more comprehensive look is warranted. After all, bond ETFs are now increasingly shaping the underlying markets, rather than just trying to mimic them.

The effects are most apparent in corporate bonds (which this post is primarily but not exclusively focused upon) and in the US, but they are beginning to become noticeable in most corners of the global bond market, according to Jeff Johnson, head of fixed income products at Vanguard. He said:

The first bond index funds and ETFs were focused on higher-grade more liquid parts of the markets, like Treasuries and investment-grade corporate debt, but today there’s just as much money in fixed income ETFs that follow more satellite portions of the market.

Perhaps most intriguingly, some even think they are easing a challenge that has dogged the bond market since the global financial crisis of 2008 — the trading difficulties introduced after banks were slapped with more onerous regulations, which curtailed their ability to intermediate between bond buyers and sellers.

As Gregory Peters, co-chief investment officer of PGIM Fixed Income, told FTAV:

The technology around ETFs — even more than the ETF itself — has transformed fixed income trading . . . As a consequence the liquidity of the bond market has improved pretty significantly.

It’s even more efficient than what we had before the global financial crisis, when we simply relied on dealer balance sheets. This utilises technology which will only get better, quicker, faster, and stronger. And I think that’s the transformational part of the ETF chassis.

How investors (mostly) learned to stop worrying and love the bond ETF

The night before her interview with BlackRock in 2015, Samara Cohen had to Google how ETFs worked.

Cohen had actually started her career with the asset manager in 1993, but at the time it was merely a small, traditional bond investment house, and the first US ETF had been around for only one inglorious year (the first ever ETF was actually launched in Canada in 1990). The company that she was now thinking of rejoining — after a long stint at Goldman Sachs — was a very different beast, having acquired passive investing giant Barclays Global Investors.

Over breakfast, at Casa Lever on New York’s 53rd Street, she confessed her ignorance to Mark Wiedman, who led BlackRock’s iShares business at the time. He was unconcerned, arguing that she could probably pick up ETF expertise pretty quickly. He was right: Cohen is now chief investment officer of BlackRock’s $7tn of ETFs and index funds, including over $900bn of bond ETFs.

Cohen reckons there were two major events that have helped stoke the growth of fixed income ETFs — the first of which happened just months after she boomeranged back to BlackRock.

In December 2015, a corporate bond fund run by Third Avenue Management collapsed and had to freeze investor withdrawals. That sent a shiver of fear through credit markets. Suddenly, investor interest in bond ETFs exploded, according to Cohen:

Up until that point a lot of the institutional money in bond ETFs were tourists — people who were mostly used to equities and just liked that they could access bonds in a stock market product. 

That changed very meaningfully at the end of 2015, because there were so many bond funds that underperformed the broader high yield market. We started getting more incoming calls rather than outgoing calls, from people that wanted to know how bond ETFs worked.

The second big moment came in March 2020, as Covid-19 lockdowns sent financial markets into a tailspin.

The bond market clogged up dramatically: everyone started to sell and few wanted to buy, with even US Treasuries becoming scarily illiquid for a period. The arbitrage mechanism that happens under the hood of ETFs also gummed up as a result, but — and this is crucial — the actual shares of bond ETFs kept trading despite insanely high volumes (much more on this later).

This was more than could be said for large sections of the underlying fixed income markets. As JPMorgan’s analysts noted at the time, this was “the largest ever stress test for ETF markets . . . [and] other than a few glitches, ETF markets generally functioned as intended and held up well”.

For BlackRock’s Cohen, this was the point when fixed income ETFs truly arrived, proving their mettle and gaining broad acceptance among both fund managers, as a valuable trading tool; and investors, as a fund structure to embrace:

It was a huge moment for fixed income ETFs. Traders were packing up their Bloomberg to work from home, and couldn’t trade bonds. But you could trade ETFs. They provided really important price discovery during that volatile period, and helped create more resilience for the bond market. It was a huge change. 

Some sceptics argue that it was only the Federal Reserve including bond ETFs in its massive crisis-fighting stimulus package that prevented ✌️something✌️ in the industry from breaking. But most independent analysts — like the Bank of Canada — have concluded that bond ETFs were not merely resilient: their continuous trading was actually valuable.

The Fed only bought less than $9bn of bond ETFs, which was quickly unwound. The simple fact that the US central bank was willing to buy them was a big endorsement for a lot of institutional investors.

However, fixed income ETFs haven’t just become a big deal for the investment industry. The most intriguing development is how the underlying ETF technology is starting to affect the bond market itself. To understand why, we have to look under the hood.

The bond market’s two-lane speedway

James Mauro’s day starts horrifically early. He arrives at BlackRock’s San Francisco office by 5.30am, where he starts poring over how his carefully-constructed portfolio performed overnight.

But Mauro doesn’t scour the market for interesting tactical opportunities thrown up by a central bank meeting, or contemplate strategic shifts necessitated by rising corporate defaults. Any macroeconomic reading is mostly recreational.

For more than a decade, Mauro has managed the iShares Core US Aggregate Bond ETF, a $104bn behemoth that goes by its stock market ticker AGG. It’s BlackRock’s single biggest fixed income fund, and the fifth biggest in the world.

Although AGG is entirely passive — striving to reflect the performance of the Kuhn Loeb Lehman Barclays Bloomberg US Aggregate Index, the dominant American bond index — there’s still a lot of hard, finicky work involved.

The first thing on Mauro’s agenda is to analyse what the portfolio looks like relative to its index, and slight tweaks that might be needed. Then the team turns to inflows or outflows that must be immediately dealt with, discusses looming bond sales that AGG may have to participate in, and finally publishes a list of acceptable raw material for constructing new AGG shares.

This is one of the most crucial decisions of any given day. Inclusion on these lists can have a huge impact on individual bonds, as well as the fund itself. As one senior asset management executive told FTAV:

The underlying bond market has become almost like a two-way speedway. Bonds that are on those investible lists are increasingly liquid, because you can create them into an ETF or break the ETF apart and take them back. But bonds that aren’t on these investible lists have become more illiquid.

Let’s unpick this a little, as there are both some ETF peculiarities and bond market idiosyncrasies at play here. If you’re already familiar with both you can skim the next few paragraphs.

Think of an ETF like a warehouse, with shares listed on a stock exchange. You can put virtually anything inside this warehouse — large American stocks, emerging market bonds, oil futures or physical gold bars. But there are two distinct ways that these shares can change hands.

“Secondary market transactions” are the ordinary buying and selling of the ETF shares throughout the day. These typically represent the majority of all ETF transactions. For example, monthly secondary market trading in Mauro’s AGG has averaged nearly $16bn a month this year, making the bond ETF more actively traded than the likes of Morgan Stanley and ConocoPhillips, and not far off PepsiCo and McDonald’s.

Then there are “primary market transactions”, where an elegant arbitrage mechanism known as creation/redemption ensures that the value of the shares in the warehouse consistently reflects the value of its inventory.

If rising demand means that the warehouse’s shares become more valuable than the securities it contains, then market-makers known as “authorised participants” — big banks like Goldman Sachs, or trading firms such as Jane Street — can buy the individual securities that match its holdings, go to the warehouse operator and exchange this representative bundle for freshly-minted shares in the warehouse.

Conversely, if selling pressures mean the value of the warehouse shares drift below that of its holdings, an authorised participant can go to the warehouse and exchange shares for a representative slice of the underlying securities inside the warehouse — and sell them off piecemeal to capture the price difference.

That way, any discrepancy between the ETF and its holdings should be continuously and hopefully quickly arbitraged away. The whole set-up looks a bit like this:

You are seeing a snapshot of an interactive graphic. This is most likely due to being offline or JavaScript being disabled in your browser.

For big mainstream stock market ETFs — like the $500bn SPDR S&P 500 ETF Trust — this is super easy.

Market makers can simply snap up shares in the 500 members of the index and exchange them for a freshly baked slice of SPY (as the ETF is known), or do the opposite, and buy shares in SPY and redeem them for a proportional amount of the 500 underlying stocks, which they can instantly sell for a profit. This arbitrage opportunity keeps an ETF’s trading in line with the index it tracks throughout the day.

However, bond ETFs are in a radically different situation, because fixed income markets are much bigger and more complex.

Finicky income markets

Take the corporate bond market, often just referred to as “credit” in the finance industry. A company usually only has one class of shares trading, but could have dozens of bonds, each with different maturities, interest rates and legal protections. For comparison, in the US there are more than 500,000 registered corporate bonds, but only about 8,000 stocks.

But this is the proverbial tip of the iceberg. Governments, states, development banks, cities, metro systems, international organisations, counties, and even universities and churches issue bonds. You can also issue bonds backed by mortgages, student loans, aircraft leases, art, solar power, catastrophe insurance and music rights, and then slice them up into smaller idiosyncratic bond “tranches”.

All told, there are many millions of individual bond securities around the world — each as unique as a fingerprint.

As a result, the bond market is also far less liquid than stocks. Setting aside Treasuries and a few other very active government bond markets, probably only about 1-2 per cent of bonds trade at least once day. A decent chunk go for months without any trading activity whatsoever.

Here’s a Citi chart on US corporate bond trading volumes from 2018 that illustrates the problem (note, the number of corporate bonds it found doesn’t match the number above, which is from CUSIP Global Services and captures a lot of smaller, even less liquid bonds):

It’s just too hard to try to buy every single bond in whatever index it tracks (often it is actually impossible). Instead, bond ETFs typically do something called “sampling”: fund managers assemble a representative slice of more actively-traded bonds that mimics the index as closely as possible in financial and fundamental terms — for example duration, yield, credit rating, industry, convexity, geography yada yada.

In some cases, this subset can be reasonably close to the real deal. For example, the Bloomberg US Aggregate index includes about 13,400 securities, while BlackRock’s AGG has roughly 11,700 holdings. But in many markets the sample is by necessity much narrower.

Sometimes the ETF may even hold more securities than its underlying index to find a blend that will perform similarly. The pioneering Canadian bond ETF holds 1,624 bonds, compared with the 1,529 securities in its benchmark, while Vanguard’s $18bn mortgage-backed securities ETF holds 1,442 bonds, compared with its index’s 972 members. Most of the time, this gets an ETF pretty close to the underlying index, albeit often not perfectly so in some hairier markets like junk bonds.

Here’s a compare and contrast between JNK — State Street’s $8bn high-yield bond ETF — and its index, the Bloomberg High Yield Very Liquid Index (itself a subset of the broader high-yield bond universe):

On any given day, ETF managers publish creation and redemption lists with specific bonds or broader parameters for types of bonds they’ll accept in return for making new shares, or what securities they will give in return for shares handed to them. Trading shops like Jane Street seek to oblige.

As Steve Laipply, co-head of fixed income ETFs at BlackRock, told FTAV:

There are small underweights and small overweights at a given point in time and ideally tweaking the creation basket will constantly help fill in small holes or level things out. You’re constantly trying to smooth out the fund’s profile to match the index as closely as possible.

Harder, better, faster, stronger

This machinery has come a long way since the early days of bond ETFs, but Mauro says the improvements have been particularly stark since 2017, when many in the industry began investing heavily in technology underpinning the process.

Creation baskets — the menu of bonds that ETF managers like Vanguard and BlackRock will accept in return for shares — were much smaller, and might only list 50 bonds. Each creation could take up to two hours. Across BlackRock’s entire ETF complex, the asset manager was probably only issuing 10 such baskets on a typical day day.

BlackRock can now do baskets with 350-400 securities in just five minutes, according to Mauro, and does about 70 of them every day thanks to widespread automation. “We have a lot of simple algorithms that together help a very complex investment process,” he said.

However, as the two-lane speedway comment above indicates, the growing size of bond ETFs mean that they are having a mounting impact on the underlying fixed income market — especially in areas like corporate debt. If a bond can be delivered as part of a creation basket, it becomes more tradable. If you want, you can even deliver a massive portfolio of bonds to exchange for ETF shares.

How great is that impact? In 2021, Barclays’ global head of research Jeffrey Meli estimated that transaction costs for bonds included in LQD fell by 3.5 per cent, even when adjusting for the bias of the ETF to choose more liquid bonds.

In other areas, the impact has been even stronger. Transaction costs for junk bonds included in HYG — BlackRock’s main US high-yield bond ETF — fell by more than 14 per cent. In other words, in addition to being easily-tradable instruments themselves, ETFs actually help make the underlying bonds more liquid (the full Barclays report can be read here).

BlackRock reckons that the impact is even more powerful. Research shared with FT Alphaville indicates that average daily volumes of ETF-included bonds is more than 60 per cent greater, and the spread between the prices at which people are willing to buy or sell a bond are much narrower.

But the result is to make bonds that are a bad fit for ETFs relatively less liquid. “There is an increasing contrast between bonds in ETFs and bonds left behind,” admitted Brett Pybus, the other co-head of fixed income ETFs at BlackRock. “But ETFs haven’t created the problem, they’ve merely accentuated it,” he stresses.

Bond villains?

Some analysts think the impact is more pernicious. As we mentioned earlier, fund managers have themselves become increasingly enthusiastic users of bond ETFs. Barclays’ Meli reckons this means ETFs might, in practice, suck some activity away from the underlying bond market.

FT Alphaville’s emphasis below:

The potential substitution of ETF trades in place of trades in the bond market raises the possibility of a dark side to the liquidity story. Institutional investors, faced with declining liquidity in individual corporate bonds, trade ETFs when managing liquidity needs instead. These ETF trades inevitably come at the expense of corporate bond trades that would otherwise have happened, thus further reducing bond liquidity. It is possible that aggregate liquidity deteriorates due to increased presence of ETFs, even if the individual bonds that are included in ETFs still benefit relative to those that are not included.

More worryingly, some researchers suggest that bond ETFs might actually worsen the liquidity of the bonds they include at times of crisis.

Four academics — Naz Koont, Yiming Ma, Lubos Pastor, and Yao Zeng — studied the performance of US corporate bond ETFs, their creation baskets and the impact on their underlying securities over 2017-2020, and found that the generally positive impact flipped in March 2020.

As their 2023 paper Steering a Ship in Illiquid Waters: Active Management of Passive Funds put it:

In normal times, a bond’s inclusion in an ETF basket makes the bond more liquid because shocks to investors’ demand for ETF shares are largely idiosyncratic. A random mix of creations and redemptions across ETFs increases the trading activity in basket bonds, improving their liquidity. That is not the case, however, in periods when investors’ liquidity shocks are systematic, resulting in imbalances between creations and redemptions.

For example, large redemptions move many redemption basket bonds to APs’ balance sheets. The APs, who also tend to act as market makers in these bonds, may then become reluctant to purchase more of the same bonds, reducing their liquidity. The effect of basket inclusion on bond liquidity is thus state-dependent: positive in normal times but negative when there is large imbalance between creations and redemptions.

. . . The COVID sub-period provides preliminary evidence that inclusion in [creation/redemption] baskets can hurt a bond’s liquidity when redemptions are systematic and persistent. In this sub-period, many investors experienced liquidity shocks that led them to sell ETF shares. This selling pressure was met by APs who purchased many ETF shares from investors, redeemed them, and then tried to sell the bonds acquired through RD baskets. Bonds heavily represented in RD baskets thus became heavily represented in APs’ inventory.

Given their balance sheet constraints, APs became reluctant to purchase even more of the same bonds in their role as market makers. Bonds present in RD baskets thus lost their most natural buyers. When its own market makers do not want to buy it, a security can become quite illiquid.

Not everyone is convinced that this is accurate. The increase in illiquid bonds in redemption baskets and market-maker balance sheets may have simply reflected that all relatively less liquid securities can become extremely difficult to trade in times of tumult. So, naturally, ETF providers will offer more of them through redemption baskets to avoid tracking errors, and more will tend to clog up the balance sheet of dealers. But an ETF provider cannot compel an authorised participant to accept anything.

And even if it’s partly true it’s probably less bad than it seems in practice, and mostly a reflection of just how chaotic March 2020 was — when even US Treasures suffered an unnerving bout of illiquidity. And that had nothing to do with ETFs.

Crucially, as mentioned earlier, bond ETF shares themselves kept trading smoothly even at the worst of the market tumult. They in practice became quasi closed-end funds, with creations and redemptions slowed down by acute underlying bond market illiquidity.

Yes, their prices sometimes therefore tumbled far below the theoretical but stale net asset value of their holdings — unnerving some observers — as the creation/redemption mechanism gummed up because of the illiquidity of the underlying bonds . . .

. . . but investors could still sell bond ETFs at a time when they struggled to sell the underlying holdings. To take a concrete example, on March 12, 2020 — one of the worst days that tumultuous month — LQD shares traded about 90,000 times, but its five biggest holdings traded only 37 times each on average. 

As Jane Street’s fixed income head Matt Berger pointed out:

If corporate bond outflows are greater than the ability of the market to absorb the selling then prices are going to move. And if there’s a severe mismatch of buyers and sellers then both ETFs and the underlying bond market will struggle. But I’m confident that the ETF mechanism isn’t going to break down.

If this was all there is to the bond ETF story, they would still represent a major evolution for fixed income — a shift with many positive sides and, yes, some potentially negative ones.

But it is how bond ETFs have helped accelerate other trends that really makes this a revolutionary development.

Electronic trading

This is strictly speaking a GIF showing a Citicorp FX trader back in the 1980s, but it’s a good way of illustrating how bonds have historically traded as well:

The phone remains a major way things get done, coupled with Bloomberg chats. But that is now changing. The majority of trading in major government bonds now happens electronically, and, slowly but surely (though faster than most people in the bond market would have expected just five years ago), other corners of fixed income are following.

Here is a 2023 chart from Flow Traders, a European market-maker that specialises in ETFs, showing their estimates for electronic trading shares for various asset classes (you can read the full report here):

Things are progressing particularly fast in the US, where these numbers are already of date.

Coalition Greenwich, a data analytics company, estimates that 49 per cent of US investment-grade corporate bond trading is now electronic and 35 per cent of US high-yield bonds now trade by algorithm. That’s nearly double the pre-pandemic levels. Even US municipal bonds — probably the most archaic part of the entire fixed income world — have seen a shift, with 16 per cent of trading now electronic.

Some e-trading may still involve humans, but Coalition Greenwich estimates that 28 per cent of all US corporate bond transactions are now fully automated, having almost doubled over the past five years. They have to automate because:

. . . top-tier U.S. corporate bond desks see roughly 30,000 inquiries per day on average, making it now virtually impossible for humans to manually respond to all (or in some cases, even some) of those client-generated requests-for-quote (RFQs).

If you want to buy or sell a large position — say, $40mn of an IBM bond — you’ll still probably pick up the phone, call a big dealer and ask for a quote. But for smaller chunks, electronic trading has become far more viable. You can therefore also slice up a big trade into a series of smaller electronic ones.

As a result, average trade sizes for US corporate bonds have fallen by more than a third over the past decade, to under $400,000 this year — even as overall trading volumes have more than doubled — according to Coalition Greenwich.

This is a secular trend that probably would have happened even in a world without bond ETFs. But the growing heft of ETFs has been a huge accelerant, according to traders, fund managers and industry executives.

“The evolution of fixed income ETFs and electronic trading in bonds are inextricably linked and have been for over a decade,” said Leland Clemons, co-founder of BondBloxx, a bond ETF specialist. This is both because of the liquidity that ETFs bring to the party, and because of the new entrants they have brought into the ecosystem.

To facilitate bond ETF arbitrage, specialists like Jane Street and Flow Traders have stepped into the underlying bond market, creating a “virtuous cycle”, according to Kevin McPartland, head of research for Coalition Greenwich:

To do that trade, they had to get better and smarter at trading not only the ETFs, but also the underlying corporate bonds. To some extent, the rest is history.

That in turn forced banks to make sizeable investments in their own technology to compete with this new breed of market-maker, while bigger asset managers bulked up on their own trading processes to take advantage of the emerging ecosystem.

Ramon Baljé, head of fixed income at Flow Traders, said that the banks were now starting to catch-up with their non-bank competitors in terms of electronic bond trading, contributing to increased liquidity across the fixed income ETF ecosystem. “This becomes very automated, and bid offers are very tight,” he said.

So-called “high touch” human trading will probably always prevail when really big blocks change hands, and for more recondite slices of fixed income. Habits die hard, after all. Amazingly, some players still use fax machines as part of completing a fixed income trade, according to Alex Morris, chief executive of F/m Investments, which runs a series of ETFs that invest in US Treasury bonds.

But the shift towards faster, smaller and more electronic trading is still in the early stages, argued Chris Concannon, chief executive of MarketAxess:

The introduction of fixed income ETFs has really accelerated the electronification of the bond market. And we’re still in the early minutes of a soccer game. There will probably always be some human engagement in the bigger trades, but all the execution will eventually be electronic.

Portfolio trading

Imagine you’re the chief investment officer of a big pension plan. You think interest rates are going even higher, and this will crush the economy. So you plan to ditch a big chunk of your corporate bond exposure — and fast.

But you have hundreds of bonds, each unlike any other. Selling them off piecemeal will take ages. Before, you might have asked a bank for a quote on the whole lot and girded yourself for the inevitable gouging, but banks don’t have the balance sheets to warehouse enormous bond bundles any more.

Today, this process can take a matter of minutes — even if you are shifting thousands of bonds, collectively worth billions of dollars — thanks to another phenomenon unlocked by the advent of fixed income ETFs: portfolio trading.

As Barclays analysts put it in a recent report (with our emphasis):

Portfolio trading is the latest link in the evolution of the corporate bond market. In a portfolio trade, investors bundle a set of corporate bonds into one basket and execute the entire basket as a single piece of risk, with one market-maker. Portfolio trading has grown from 2% of the market trading volume six years ago to more than 9% today.

Corporate bond liquidity has been historically constrained by the large number of instruments, each of which has a different maturity, coupon, seniority or optionality. The large number of bonds makes it difficult to match buyers and sellers. Portfolio trading has turned this constraint on bond liquidity into a strength, combining traditional credit trading and equity principles into one product.

Investors typically own thousands of bonds in their portfolios, and there are many aspects of corporate bond portfolio management that require trading a large number of bonds, such as adjusting curve, sector or rating exposures. This explains the popularity of portfolio trading.

That’s pretty remarkable: a type of trade that almost didn’t exist a decade ago now accounts for nearly a 10th of all US corporate bond trading volumes. Thanks to the creation-redemption mechanism of ETFs, it’s now arguably easier and cheaper to trade large, diverse portfolios of bonds than it is to trade individual ones.

That’s a game-changer for the investment industry. As Jane Street’s Berger noted:

ETFs, portfolio trading and the growth of electronic trading have incentivised people to hold much more diversified portfolios. If you had 500 bonds you needed to move it used to be pretty difficult. Now you can shift risk quickly, easily and cheaply, even in times of stress.

Portfolio trading has grown particularly strongly in the US, but — as M&G Investment’s “Bond Vigilantes” team noted earlier this year — it has fast become a phenomenon in Europe as well.

In fact, portfolio trading is now transcending the original ETF technology that enabled it. Thanks to the wider acceptance and use of pricing tools like Bloomberg’s BVAL and CBBT, more investors feel comfortable buying entire portfolios even without going through the ETF share creation process. As Bondbloxx’s Clemons said:

If portfolio trading of bonds was necessary to trading ETFs, now it’s become its own thing. If we were talking about that 20 years ago, people would have laughed us out of the room for being ridiculous. I think sometimes we take it for granted, the access to markets that ETFs have allowed.

This might seem humdrum to non-credit nerds — people have been able to do this for decades in equities — but the ability to shift huge, multi-faceted chunks of fixed income risks quickly, efficiently and at reasonable prices is a game-changer for the corporate bond market.

Moreover, the remarkable growth of portfolio trading is another crucial interlocking factor that is helped by and in turn helps the growth of electronic trading and bond ETFs. “It’s like a golden triangle, or three legs of a stool, with the increase in algorithmic trading, the rise of ETFs and portfolio trading all working together”, said BlackRock’s Pybus.

And now a fourth leg has emerged to add even more support.

Systematic investing

Historically, the bond market has been dominated by former jocks with MBAs at big banks and investment companies, but the ETF-accelerated growth of more algorithmic trading in smaller chunks of debt is attracting new players on to the field.

Quantitative investors with PhDs in mathematics, physics and computer science — long ascendant in the stock market — are now eyeing opportunities in fixed income.

Quants have been active in major government bonds, which are more transparent and tradable, for decades. Surfing trends in the “rates” market is a particularly old strategy. But their model-driven, systematic approach has often struggled in the balkanised, illiquid pockets of fixed income, like corporate debt.

That is now changing thanks to improving liquidity and better data, which in turn is begetting more liquidity, as Barclays noted in a report in May. The bank’s own emphasis in bold below.

The number of funds and the total deployed assets associated with systematic strategies have ballooned recently. Up until the past few years, systematic strategies occupied a relatively small corner of the credit market. This has been changing rapidly, and the effects on liquidity and price action are already apparent.

We view this as a net positive for credit investors. The models employed by these strategies screen over a much larger universe of bonds, which should inject additional liquidity into the market. Their ability to identify outliers at scale should also make markets more efficient, as they provide more opportunities for price discovery.

Barclays analysts estimate that somewhere between $90bn and $140bn is now in systematic credit strategies. This is obviously small given the size of the market, but the bank’s analysts note that it will “likely only grow from here”.

Moreover, the higher frequency of their trading means that they are playing an outsized role in the corporate bond market. Here’s Barclays’ estimate for how much of cash credit trading they accounted for in 2023:

In a sign of the times, both Blackstone and Ares have acquired systematic credit specialists in recent years. Many other big, established quant funds are also becoming more active, as well as established credit-focused hedge funds setting up dedicated systematic strategies.

But the crucial bit is how all these trends are now reinforcing each other, with the bond ETF as the crucial pillar. As Joshua Barrickman, head of US fixed income index funds at Vanguard told FTAV:

In capital markets you often need something big to shake things up, to move away from the old way of doing things. That’s what ETFs have done.

They’ve really enriched the fixed income market structure. There’s a whole ecosystem that now revolves around bond ETFs, like portfolio trading, algorithmic trading and systematic investing. Everyone has ETFs in their toolbox today.

CONGRATULATIONS, you have (nearly) reached the end!

Of course, the conjoined growth of bond ETFs, electronic trading and systematic investing might not be seen as positive by everyone.

The overall result is that more parts of the fixed income market are beginning to resemble the stock market — what some insiders have dubbed its “equitisation”. The implication is that the bond market will also become increasingly prone to whiplash moves, and — in extremis — stomach-churningly fast crashes.

As Jane Street’s Berger said:

There’s always been a debate about whether electronic trading is a good thing. I think most people have concluded that it is. But it does mean that crises play out far more quickly than in the past . . . People have to get used to the fact that price discovery happens faster than it used to.

The problem is that bond markets matter far more to the real economy than the stock market, even if they attract far less mainstream attention. Flash crashes in equities are mostly scary; flash crashes in fixed income are dangerous.

But, on the whole, there seems to be a growing consensus (even among some former sceptics) that the emergence of the fixed income ETF ecosystem has been a boon to the broader bond market. The most intriguing suggestion made by some people FTAV spoke to was that bond ETFs are de facto playing the role that big investment bank balance sheets did before the financial crisis. As one senior industry executive told us:

Before companies used to have these big data centres, and now everything is in the cloud. Before banks used to keep all these bonds in their inventory. Today, they’re stored in ETFs. The cloud is the ETF. Bonds are just stored in ETF that can be easily accessed

Or, in diagram form from a Bank of Canada report on the subject:

Fear of bond market illiquidity has been such a staple of the post-2008 era that Bloomberg’s Matt Levine for a period even had a semi-regular “people are worried about bond market liquidity” slot in his newsletter. If bond ETFs actually end up solving — or at least ameliorating — this problem, then it would be quite the twist, given that many people have expected them to make it worse.

Finally though, the ETF-enabled equitisation of fixed income might also have some awkward consequences for the people that work in it.

PGIM’s Peters recalled how he used to take the daily ferry from Manhattan to New Jersey with a horde of New York Stock Exchange floor traders, and kept wondering why on earth all of them had jobs when a computer could do what they do far better, cheaper and faster. And then, almost all of a sudden, they all disappeared.

The parallels to his own industry were uncanny, Peters noted:

Oftentimes these things take far longer than we think it will, but when it goes it goes quickly, and I think that’s where we are on the fixed income side. . . . . Fixed income markets have always been somewhat of a late adopter, but I think we’re on the precipice of radical modernisation.

Anyway, now that you’ve made it to the end, an apology for perhaps spending more time on this subject than most might think it warrants. Thanks for staying with us!