A small detour from analysing single companies, and book reviews have always been one of my favourite type of writings: I have a penchant for the inside stories and views of famous investors.
Mary Childs’ “The Bond King” chronicles the birth and rise to dominance in fixed income markets of Pacific Investment Management Company, but mostly of its charismatic founder William (Bill) Hunt Gross, the chief investment brain, public voice and institutional face of Pimco until his exasperated partners kicked him out in September 2014 (although officially he resigned to join a much smaller competitor, Janus Capital Group, before retiring four years later).
Some of the stories are already well known to people in finance: for example, how Gross learnt risk management by reading Ed Thorp’s book “Beat the Dealer: A Winning Strategy for the Game of Twenty-One” and playing blackjack. And the fierce competitiveness of Pimco, both inside and outside the firm, is also legendary.
Other anecdotes are nice to read but more suited for the gossip pages and much less interesting from the financial markets point of view: the details of the acrimonious battle with Mohamed El-Erian (co-CEO and co-CIO until he left in January 2014); the “hunt” for the moles that were leaking these facts to journalists; the internal fights and Gross’ aggressive and erratic behaviour, including the suits with his neighbours and the pettiness of the divorce from his wife.
While probably not in my top 5 finance books, it is still an entertaining reading and definitely recommended to anyone interested in investment management.
Here are the best parts in my opinion.
How it all started
What we take as “normal” today, it wasn’t few decades ago: bonds have always been an important part of finance, but only in the past half-century they became a colourful one.
Out of business school, Gross started as a private placement loan officer at a staid, old insurance company, Pacific Mutual Life Insurance Company: his job was to evaluate to whom to lend money, and among the proposals he assessed there were a young man named Sam Walton in Arkansas and Warren Buffett’s Berkshire Hathaway.
At the time, the business of life insurers was easy: estimate how long customers will live and how much you have to pay out when they die. They could then take the premiums and invest the money in bonds that throw off interest payments until they mature and return the money approximately when the insurer expects to need it back.
Gross saw before everyone the chance to do something with bonds besides collecting interest payments: with interest rates being more volatile, a knowing investor could swap one security for another, exploiting pricing anomalies to pick up extra income and capital gains.
It was not stimulating work. Gross found himself with uninteresting jobs like clipping bond coupons in Pacific Mutual’s vault, snipping the little tags off the bottom of corporate certificates and mailing them in for interest payments. That was all a person would do with bonds at the time. For the first few years, Gross couldn’t wait to get transferred into the stocks division. But then, at the start of the 1970s, he convinced his boss to let him try a radical concept: trading bonds.
At the time, Pacific Mutual was a small regional player (it wasn’t even in Los Angeles, but two hours away in Newport Beach), with a credit portfolio of well under $1 billion, but the management consultants at McKinsey had already recommended a strategy to grow: by investing. So they created a shell company (Pacific Equity Management Company) that quickly became Pacific Investment Management Company.
Pimco started really small, with just $5 million from its parent’s portfolio.
So they took the $5 million and the empty corporate shell and ran with it, a stealth project inside a sleepy insurer. The three fell into different roles naturally: (Jim) Muzzy didn’t love investing, managing portfolios. But he was warm and good with people, and he had some ideas to jazz up marketing (i.e., actually explain to the clients what they did with the money). Gross hated traveling, talking to people - he could do it, and clients found him charming, but it took him away from his desk and the more pressing business of trading. (Bill) Podlich was so obviously brilliant at business strategy that adding any other function would have been a glaring misallocation of his time.
Thus, they easily became a “three-legged stool” - Gross in bond trading, Muzzy in client services, and Podlich in business strategy. They had balance.
A very competitive workplace
Gross soon became the face of a revolution: actively trading those pieces of paper in pursuit of price appreciation. And he was good at it. He reigned over the bond market for decades: from the outside, he seemed folksy, if a bit eccentric, kind of the bond market’s Warren Buffett. The truth was more complicated.
For his own admission, Gross has never been an easy person to interact with, and from the beginning his motivation was to be famous, much more than having power or money. His chant in the 1990s was: “Grow or die, grow or die!”
Anyone who knew Gross in the flesh knew the façade didn’t match up. In real life, Gross was exacting, uncompromising, unwilling to back down. And he had built Pimco in his image. Normal trade floors, at banks or other money managers, were jumbles of loud televisions and yelling and slamming phones and crude jokes and debates - but not at Pimco. Everyone emailed.
The thick carpets muted the sound of any foot traffic; the clicking of keyboards was the only exception to the library silence. If a trader had to speak, to call a bank and get an order of bonds, he did so quietly, breathing words into the telephone. […] Emailing was always better, even if the email recipient was sitting three feet away. Especially if the recipient was sitting three feet away.
[…] The early crew had seemed to enjoy tormenting each other on the trade floor. They rotated who bore the brunt - though, more often than not, it landed on the same people.
The sense of superiority was still evident years later, when Pimco was a powerhouse with 2.000 employees and over $2 trillion in assets under management.
In December 2009 it hired Neel Kashkari as Head of New Investment Initiatives, basically Pimco’s first push to enter the equity markets. Kashkari (who is now the president of the Federal Reserve Bank of Minneapolis) was a Goldman Sachs technology banker that in 2006 cold-called Hank Paulson (former Goldman’s CEO and at the time George W. Bush‘s Treasury Secretary) because he wanted to “learn how government worked”. He became Paulson’s aide and was instrumental during the 2008-2009 crisis with drafting the Troubled Asset Relief Program (TARP) legislation.
It could be hard to get the rhythm of Pimco’s culture. Especially for someone like Neel Kashkari. He came from investment banking, which is driven by personal relationships. He was formal, respectful, well dressed, good-looking. He was unflappable, as evidenced in a heated cross-examination before Congress when he led TARP.
[…] Most problematic was Kashkari’s warmth, his diplomacy. He was polite to the point of kindness. He made eye contact with Bill Gross, daily. He spoke to the assistants, even. For one former Pimco executive, a single moment captured it: when walking up to the building, a more junior person always held the door for more senior people, to allow them to walk frictionless and first into the capacious lobby. The senior people took this as expected, a function of the natural order; usually they’d pass without a nod or other acknowledgment.
Not Kashkari. One day, this lower-ranking executive recalls seeing Kashkari approach the building behind him, and the man of lower rank held the door as expected. Kashkari sped up and, upon walking in, turned his head and made eye contact.
He said, “Thanks.”
The executive froze in shock. This guy will never make it here, he thought.
Kashkari did indeed leave in 2013, with Pimco still not getting any traction for its three new equity funds.
How to build a bond manager: “structural alpha”
One of Gross’ signatures in building his long-term track record has always been buying “riskier” higher-yielding bonds that were not included in benchmark indices that he was asked to beat, like mortgages when these were still little understood or international bonds when most asset managers were not yet looking outside of the US.
That little extra fairy dust of risk meant a little extra fairy dust of return, which most of the time meant beating the index’s returns.
Many of these “tricks” were mechanical, small trades that added up, sort of way to replicate his card-counting stint in Las Vegas and allowing him to keep a slim but consistent statistical advantage over fellow gamblers.
One of the simplest tricks was exploiting the difference between cash and cash equivalents, holding higher-yielding short-dated corporate bonds where competitors held actual cash. Another was selling volatility, in which a trader sells derivative contracts to bet that prices will keep trading within a certain range.
The Lambda Cash strategy was developed in the 1980s. Fund mandates limited how much leverage Pimco could use, so it had to figure out a workaround: its enthusiastic embrace of derivatives became a habitual, persistent and long-established tools.
Because derivatives don’t involve buying the underlying thing now, they require much less money up front.
[…] Gross loved a bargain, so he took to derivatives wholly and vigorously.
But leverage is the cocaine of the investing world: leverage adds to returns, but the price for using it might be higher than you can pay. Just using the word leverage can make conservative managers of mutual and pension funds nervous.
Sitting on the trade floor in the early 1990s, Brynjo explained this accounting system to the members of the Investment Committee. “… So, we’ll just call that leverage,” he concluded.
Every face went white(r). “That’s not going to work,” one of them said.
“You want me to change the formula?” Brynjo asked.
“No. No, the formula is fine. It’s the name.”
“Well, what else are we going to name it?”
As finance nerds love the Greek alphabet, they settled on Lambda Cash.
Just Lambda Cash could add 0.25 percent, 0.4 percent a year - which, in fixed income, is everything. And Pimco could do it forever. In any asset class, that’s how you win: if you just don’t lose all your money on some big, dumb trade gone wrong; if, instead, you steady-eddy along, eventually you’ll be number one in the long-term rankings.
“Strategic mediocrity,” Pimco’s self-deprecating junk bond manager Ben Trosky used to call it, his own plan never to be number one in a given year, but also never to blow up. Simply staying in the game long enough meant you won. And that’s what clients wanted: a good track record, over time; a manager who beat those arbitrary benchmarks, even if sometimes that meant gaming them a little - how would the clients know better? Or, rather, what did it matter if it was working?
Taken together, all these reliable trades comprised what Gross called structural alpha: whatever can be squeezed out of a market beyond the market’s own rally, what every fund manager seeks; and structural meaning replicable, persistent.
Structural alpha trades were supposed to generate 0.5 percent, 1 percent in performance a year.
Such trades helped especially in those periods when a manager randomly lost his touch forecasting the direction of interest rates or picking the right credit over the wrong one. Which happened to everyone: it was unavoidable.
So, when Gross was off, these structural trades helped provide a cushion to fall back on.
The GNMA heist
Pimco always prided itself on finding a way by burrowing into the particulars and coming out with profits no one else thought existed.
It really put itself on the map in the 1980s with a complicated but elegant feat, one trade that established its reputation on Wall Street as an intimidating trading partner.
In 1983, the little band of traders orchestrated a perfectly legal stunt in the mortgage futures market. The business was beginning to catch traction, but the brilliance of this one trade established a reputation for Pimco as shrewd, meticulous, and with a stomach for very high risk.
In 1975, the Chicago Board of Trade (CBOT) introduced the Government National Mortgage Association (GNMA, known affectionately in markets as “Ginnie Mae”) Collateralized Depositary Receipt (CDR) futures contract. The GNMA CDR is a contract on mortgage-backed securities and, as such, was the first interest rate futures contract.
As with all futures, each contract expires on a set date, at which time the buyer can settle in cash or can “roll” into a new contract with an updated expiration. But the market for trading futures was still developing, there was also another option, one people hadn’t paid attention to: the buyer could take delivery of the underlying security, in this case a bundle of home mortgages backed by Ginnie Mae.
Usually, the higher the coupon on a bond, the more it was worth. This becomes truer as interest rates fall - high-coupon bonds become even more valuable, because if new bonds today get you only 8 percent, you’ll pay more for an old bond that yields 16 percent. But mortgage bonds are different, because homeowners can prepay their mortgages, and often do. When interest rates fall, homeowners with expensive 16 percent mortgages will refinance, taking out new loans at lower rates and repaying their old ones, killing the payment stream on the old bond. Investors receive the principal, but lose out on all the future years of valuable high interest payments. So, investors who own mortgage bonds with high interest rates don’t get the full benefit when interest rates fall.
When a CDR buyer demanded delivery of the underlying security, the seller could choose which Ginnie Mae mortgage bond to deliver. The market was aware of this: futures traders generally assume that they’ll get the “cheapest-to-deliver” security. But the formula didn’t account for that unique problem of mortgage prepayment: high-coupon mortgage bonds are less valuable than standard twenty-year bonds that don’t have the pre-pay risk. Because of this flaw, in practice, it would always be much cheaper for a seller to deliver high-coupon Ginnie Maes: the futures formula treated them as more valuable than they were and, so, required sellers to deliver fewer of them.
The problem was that there were only so many of the highest-coupon Ginnie Maes in the world. When rates where sky-high there had been ample supply of them; but rates had begun to fall in 1982, so the sun was setting on future supply of those high-coupon bundles, the cheapest to deliver. They were going to be scarcer.
“The high-coupon bonds were small fractions [of the universe of securities], but the traders kept using that algorithm for their pricing. We stood back, and our thought was, You shouldn’t be using that, because not enough exists. If you want to play the game that way, we’ll jump in.”
The market didn’t yet realize its mistake, that it was pricing the futures as though there was an unlimited supply of high-coupon, cheapest-to-deliver bonds.
In finance, there is not really a mechanism to say, “Excuse me, but you have made a mistake in your model,” other than to exploit the mistake for all it’s worth. That’s basically the theory that markets are efficient: some asshole in New York or Chicago or Tokyo or Orange County will correct your mistake, albeit at great cost to you and your clients.
There was also another strange wrinkle to the contract: it gave the choice of converting the CDR into a perpetual security, locking in a set coupon payment of 8% for the rest of its life, which became particularly powerful if rates declined. At Pimco, traders realised that this was effectively an option-loaded contract: if interest rates went down, and the mortgages prepaid in a very short period of time, the holder could choose to collect a perpetual 8% coupon paid by the seller. If interest rates went up, the holder could keep the collateral, which would likely outperform twenty- to thirty-year Treasuries.
The next step was obviously to start buying up Ginnie Mae futures, as many as they could: and the market was more than happy to feed its insatiable appetite, not realizing it was digging itself into a hole. Pimco was also aware that it might be perceived as so disruptive that CBOT might put a stop to what they were doing: they could probably pull this trick only once…
By the time the buying spree ended, Pimco had amassed some $2 billion in notional exposure. That was about the same as its entire assets under management at the time, an incredibly high-risk gamble - or, it would have been if it weren’t such a slam dunk of a trade.
With interest rates rapidly falling, Pimco told its counterparties that it wanted to exercise its right for delivery: and suddenly everyone was forced to take a closer look at the contracts.
“We got lots of lower-coupon Ginnie Maes delivered to us, which were just like gold,” Meiling recalls.
“They might have been worth twenty percent more than a high-coupon Ginnie Mae.”
[…] In the end, the trade generated an estimated $70 million for clients. While that wouldn’t move the needle for Pimco today, at the time, that figure was huge. It remains probably the biggest trade by relative value in Pimco’s history. For the brave clients who participated, the trade added some two hundred basis points to that year’s performance.
The Total Return ETF
Pimco had always avoided going into passive strategies, but in the end it caved in and in March 2012 it introduced the Total Return ETF, managed by Gross with more or less the same strategy employed in the world’s biggest mutual fund which had reliably delivered returns for institutional clients since the 1970s.
The new ETF was not materially cheaper than the mutual fund: total expenses came to 0.55%, compared with 0.85% percent for Total Return fund at the time, depending on the share class. And competitors ETFs, tracking the index that Total Return aimed to beat, charged just 0.10%. But actively managed ETFs, with a human being at the helm, were only a small sliver of the market: and none of the others had Bill Gross.
For these reasons, the Total Return ETF had to have a good performance since the very beginning, to make it palatable to retail investors. There was, however, a big difference with the Total Return fund: the smaller amount of derivatives (futures, swaps, CDS, …) that the ETF could hold.
When Pimco’s brand-new must-succeed mom-and-pop Bond ETF was rolling out in 2012, and Bill Gross was going to have his derivative-trading hand tied behind his back, he was in the market for good ideas. If their usual magic tricks were muted, they would find other loopholes, other ways - securities in the market with embedded leverage, or language in the documents that didn’t explicitly rule out something.
The solution was found in “odd-lot positions”. Big institutional investors buy bond in large lots, at least $1 million and round multiples. But mortgages get paid down over time, and mortgage-backed securities have “odd-lots” that float around the market, unloved and largely forgotten. As they trade infrequently, there are often no available market prices: asset managers rely on outside pricing services to estimate the value of each bond each day, using a mixture of things like old trading prices, hypothetical prices at which banks say they’d transact, and comparisons to similar bonds.
But pricing services generally account only for round lots: clever bond experts could buy a bunch of “odd lots” for cheap, slot them into the pricing system, and watch as they were rounded up to the full-size lot price. Instant profit. And it’s not illegal: rule 17a-7 of the Investment Company Act of 1940 allows for cross-trading among a family of funds, provided they move the security at the market price.
On March 2, Gross sent a handwritten note to the trading desks with instructions in his characteristically stilted cadence. “Today – ASAP within the next 2 hours - find 1-2 million bonds in your area that are 2 points or more cheap to how they would be marked by pricing services at close tonight.” The desks did as they were told.
On March 9, Pimco bought an odd lot at $64.9999 and plopped it into the system, where it was valued at $82.7459. A tidy instant gain of 27 percent, for zero work. The ETF’s “net asset value” (the cumulative value of everything it held) jumped by almost $0.02 per share in one day, thanks to that trade alone. And that was one odd lot, among many.
The internal pricing group was immediately alerted, as they are responsible for checking that there wasn’t anything fishy that might cause problems with trading partners or, much wore, regulators. The pricing difference was flagged, but the Pimco traders confirmed that those same bonds were trading in the “low 80s” and so it was ok to mark them up in the Total Return ETF.
The fluke was working: one month after launch, the Total Return ETF gained 1.6%, while its twin Total Return mutual fund eked out just 0.04% (the benchmark Barclays Capital Aggregate Bond Index lost 0.6%).
They kept up the strategy: in the ETF’s first four months, Pimco bought $37 million worth of odd lots, more than 150 bundles in total. This would be a drop in the bucket for the Total Return Fund, or for Pimco as a whole, but in the still-small ETF, it made all the difference.
By the end of June, the divergence was gaping: the Bond ETF had generated 6.3 percent since its birth, while the Total Return mutual fund it was supposed to track had managed to generate just 2.8 percent.