Certain people ~cough~ at FTAV Towers are above-average interested in indexing, and how indices have evolved from snapshots of markets into a phenomenon that enjoys enormous influence over them.
Stock market indices get most of the attention. That’s understandable, given how most of the multitrillion dollar passive investing industry is in equity index funds. But as we’ve written (in possibly nauseating detail), we’re now seeing a similar trend play out in the bond market.
We therefore thought it would be fun to chat with Steve Berkley. He retired as CEO of Bloomberg’s index business in 2023, and for a few years worked as COO of BlackRock’s Aladdin platform in the noughties. But Berkley is mostly known for helping launch and then supervise the Aggregate family of bond market benchmarks for nearly four decades.
They now live inside Bloomberg, after the data giant bought the bond index business from Barclays, which in turn had acquired it from the US shell of Lehman in 2008. These are now pretty much shorthand for the bond market, in the same way that the S&P 500 is synonymous with the stock market. And in fixed income indexing, no one has seen more than “Mr Agg”.
So Steve, how did you end up on Wall Street?
I was probably one of the first folks who actually graduated with a computer science degree, back in 1979. There was great demand for computer scientists at the time, as most people in the field were actually mathematicians who dabbled in it.
My first job was for Grumman Aerospace, which later became Northrop Grumman. I then joined Sperry, which became Unisys. One day I answered my buddy Marty’s phone, as he was out to lunch. The gentleman on the other side wanted to know if I knew anything about computer systems. I said yes, but I’d get Marty to give him a call back. When Marty wasn’t interested I called myself, and the guy said: “You wanna work on Wall Street?” I replied that I didn’t even know where Wall Street is, but he replied that I shouldn’t worry about that.
I was young, I had no responsibilities, so I figured I’d take my shot. And so I got to Wall Street, building trading systems for Citibank. And after about five years I got another call, this time from Lehman Brothers, which was looking for technologists to work on a new concept of “indices”. I said I didn’t know anything about indices, but I knew how to build systems. They said great, and that’s how it started.
What was it like at what was then called Shearson Lehman?
It was at a time of craziness on Wall Street. You know, it was a time when everything was fine — smoking on the trading floor, drinking, carousing. It was just like how Michael Lewis described it in Liar’s Poker.
Why was the Agg started?
The US Agg was basically just the government and corporate indices, plus mortgage bonds. This was important because investors had begun to use mortgage securities to outperform. By using out-of-index bonds they could enhance their relative returns. It’s a trick fund managers have always used.
My job was to standardise the indices and make them the benchmarks for the investment community. We had a rules-based methodology, which I always thought was really important. You can’t have any biases in your products, especially if you’re trading it yourself.
The Agg was launched in 1986 but it was backdated till 1976, so we had 10 years of data, for historical purposes. It gave people a sense of how the market performed during various time periods.
How was the Agg actually produced back then?
It was a really difficult process. It kept a lot of people out of the business.
We had an entire team of people that read prospectuses and pulled out the information as best they could, and that would form the data part. Then, we had to get prices for the bonds — once a month at the start. And I can honestly say we were the one team within Lehman Brothers that the traders feared. Because they hated publicly pricing their bonds, even if it was just once a month.
The rules at the time were that anything with an outstanding of $1mn or more was included in the index, but a lot of the times the traders didn’t know what these bonds were. Many were small and illiquid. Over time things became more refined and we took liquidity into account, but in the beginning we just included every bond that existed into the index.
The traders would fill their best guesses for spreads on green bar paper, and this would sent up to the Edit Room. There we had about 10 guys with Monroe calculators who would put in the maturity date, the coupon, the yield and come up with a price and calculate the index. For outliers we’d go back to the trading desks and check. The process took five days — and lots of midnights.
How did you ensure that the Agg got traction?
The idea was that when people start using indices they’re going to need information about what’s in them, how is it changing, and so on. There were no fees — we didn’t charge for the data — but the information was reserved for our best customers. Trade with us and we’ll give it to you, was the pitch.
But we did give it away to the pension consultants, which our salesmen naturally hated. I had to explain how the consultants were a very important part of the decision-making process as to what benchmarks investors would choose for their funds. So by making it easy for the consultants to do their job, they were going to recommend people use Lehman’s Agg.
So that was a very important strategic step in our ability to leapfrog the main index at the time, which was produced by Salomon Brothers, and called the BIG index — it stood for broad investment grade.
Salomon was an amazing organisation, with brilliant people working there. But we were able gradually chip away at it by working with the consultant community, and ultimately providing them with the tools that helped establish the Lehman indices as the bond market standard.
How much input do investors have in the process?
I once made a big mistake. The US government started issuing inflation-linked bonds, one bond at first. My view was that it was a Treasury, so we just throw it in there, like it was grandma’s soup.
It was like I had committed mass murder. Even if it was just one security, people said we couldn’t just put something in there like that, that it was bad process. It was philosophical. People just thought it wasn’t a traditional Treasury security, and treated including it as almost a capital offence.
They were right, though. We needed a process. So we set up an index advisory council, which brought together key clients, key investors and people whose thoughts we really valued. We’d meet once a year and discuss what markets should be included, and how.
They were our indices so we made the final decisions, but it was very helpful for us. It created a sense of shared ownership.
Was this what came into play in March 2020?
Yes. The bond market had pretty much frozen. Our clients were pointing out there were no bids or offers. So people argued we couldn’t rebalance our indices at the end of the month, when there was no liquidity.
For example, most of our indices exclude bonds that have less than a year to maturity. But in many cases that would have meant forced selling when there were hardly any buyers around, and that would have crushed them.
We didn’t really have the regulators involved at the time, but there were a lot of government people in the process who didn’t want to see any more chaos in markets either.
And so we did a partial rebalance. We retained for another month the bonds that would normally fall out because of having less than a year to maturity, while still excluding those that has been downgraded. We ran that by our clients and gave everyone ample notice on what would happen. It was the right thing to do.
This is interesting because it goes to the heart of how indices have gone from something that were just supposed to be a measure of the market, to something that can actually affect the market itself. When did that start becoming apparent?
It kind of grew over time. Things like the launch of ETFs was a watershed moment. I remember very clearly ringing the bell of the American Stock Exchange for the launch of the first fixed income ETF. Things then picked up, there was momentum.
I’m not sure there were any specific discrete events where it became clear that this was happening. But the growth of passive investing gave a lot of prominence to the benchmarks and how they work.
Did you ever think that passive investing would enjoy the kind of growth in bonds that we’ve seen in equities? They’re obviously very different markets.
I bet my career on it. So the answer is yes, right? I just knew — and I don’t mean to be obnoxious about it — that technology would change how people would access the market.
What are some of the events that stick in your mind from over the years?
We often had to think on our feet. Situations that should never happen, do happen. Like yields going negative. Much of the software was built years ago, when it was inconceivable that bond yields would ever go negative. So we have to deal with all these crazy events that come up.
There were so many different crises over the years. Like 9-11. That was a very difficult time. It really impacted us [Lehman] because we were blown out of our building. The markets were closed for a few days, but we couldn’t get the product out. We didn’t have any servers. We didn’t have anything. But it was also an amazing time, because clients and even our competitors were offering us office space and anything else we needed.
Though I do remember one client, who called us on September 12 and said, “Hey, we didn’t get our files last night.” And I said: “I’m not sure if you heard, but New York had a bit of a problem yesterday.” And he’s like, “Oh, yes, yes, yes, I understand. When are we going to get them today?” I’ll never forget that conversation. I just said we’d get them out as soon as we can.
One of the most common criticisms of bond indices is that by being cap-weighted, they are in effect favouring the biggest borrowers, whether they are countries or companies. And that makes them worse indices than stock market ones, where cap-weighting makes more sense?
It is what it is. If you wanted an index that had a different weighting scheme that can be created. We were like Baskin Robbins. You can come in and pick whatever flavour you want. But in general, if there’s an overweight in certain securities, then, yeah, that’s because that’s what exists in the marketplace.
It’s a valid point. I’m not saying it’s right or wrong. I’m just agnostic. If you want a cap-weighted index, fine. You want an index that has static sector weights, OK. There’s not just one way to do this. This is where modern tools and technology really help.
But the US Agg is now the default for a lot of people, and almost half of it is US Treasuries. And if you add in mortgage-backed securities — much of them guaranteed by various government agencies — then it comes to nearly 70 per cent of the entire index.
These weightings shift over time though. I remember that when Bill Clinton was president we discussed what we would do when the US government had no debt! That was a legitimate question at the time.
There are always these things that come up. So I guess over the years you just get immune to these issues. A lot of process and weighting questions have come up over the years. But we try to reflect the whole market as it is, and then pull an investor council together to deal with the trickier questions.
Doesn’t a Treasury-dominated bond index make it easier for fund managers to outperform over time by simply weighting a bit more towards higher-yielding corporate bonds?
I just don’t think it’s that easy. There are a lot of factors at play. You have sector weights, you have duration bets, you have credit questions. There’s just no simple formula to beat the index.
At the end of the day, most active bond investors also underperform the index, and that tells you something.
So what does the future of bond indices look like?
I think the next steps are going to be realtime indices — the ability to measure how the Agg is performing intraday — and then the creation of a futures market, just like we have in the equity market.
I also think AI can have a tremendous impact in fixed-income space, whether it’s in production or reordering the indices that are being used by folks. It’s one of the most exciting things going on right now.
You know, everybody throws “AI” around without really understanding what it can do, like it’s a magic potion of some kind. But there are tangible things that it can do that really help in the index space, whether it’s mining data, improving quality or suggesting investment ideas with historical index data as inputs.
The likes of BlackRock will occasionally suggest that their big bond ETFs now are the new de facto realtime fixed income indices.
It’s better than nothing. They’re a better measure than waiting to the end of the day. It’s a good step forward, just like daily pricing for the Agg was a step forward from monthly pricing. People use ETFs because they’re the best thing there is today, for a liquid, intraday, realtime measure of what’s going on.
But I’m not sure I’d agree that they now are the benchmark. It’s a bit like the tail wagging the dog. ETFs are close to it, but they’re not really benchmarks per se. In my mind, the future is realtime indices and a futures product.
I know you’re trying to visit every country in the world. Now that you’re retired, how many are you up to?
It’s been a little static because of some family health issues, so I’m still stuck at 167. So I need another 26 more, the last I checked. Maybe you want to join me in Somalia or North Korea!