Amar Kuchinad has built an electronic bond trading platform that may very well uproot the most outdated, entrenched part of the financial services industry.
It sure looks like the Golden Age for bonds. The $7.7 trillion U.S. corporate fixed income market is the largest source of liquidity on the planet for companies, and individual investors, pension funds, and endowments are flocking to bonds as never before.
So it’s hard to believe that anything this important could be so trapped in the past. At America’s biggest, most-tech savvy asset managers, traders speed-dial their favorite Wall Street salesman to place their biggest orders over their trademark headphones, just as in the Liar’s Poker era. The electronic platforms that transformed the equity markets decades ago mainly never arrived for the bond market. Relative to stocks, big-ticket fixed-income trading is stuck in the Stone Age.
Who’s paying the untold billions of dollars in extra costs imposed by this antiquated system of matching buyers and sellers? It’s us, the folks who own bonds in their 401ks, mutual funds, and pension accounts.
Naturally, the beneficiaries are the investment banks who charge fat markups and, frequently, their hedge fund clients, who feast off of the constant leaks on who’s buying and selling big chunks of bonds, information that Wall Street firms use to cement their most lucrative relationships.
Now, a Wall Street renegade is poised to revolutionize the market for those mega-trades and, in the process, potentially lower trading costs by 80%. Amar Kuchinad, 40, worked as a fixed-income trader at Goldman Sachs from 2005 to 2011, then spent more than a year as a senior advisor at the SEC, studying regulatory measures for streamlining bond transactions. Kuchinad is a brainiac applied math major from Harvard and a student of game theory. He has a passion for designing systems that even the shrewdest traders cannot “game,” environments that keep such traders from figuring out who’s buying and selling, then using that knowledge to move prices against them, a practice that swells trading costs.
“Wall Street is interested in maintaining the status quo and their margins in the voice market,” says Kuchinad. “Those with the expertise to set up new systems were well compensated not to set them up.”
Now that tight regulations have curbed Wall Street’s role in fixed income trading, rebels such as Kuchinad are using the view from inside the machine, the experience gleaned on trading floors, to uproot the most outdated, entrenched bastion in financial services.
In March 2014, Kuchinad founded Electronifie, a venture backed by $12 million raised from family and friends. Based in lower Manhattan, Electronifie has designed a potentially groundbreaking electronic bond trading platform that’s ready for action. By the end of April, it will be fully operational, directly matching buy and sell orders from the world’s biggest owners of bonds, asset managers, foundations, sovereign wealth and hedge funds. It’s impossible to know whether a system this disruptive will succeed. But the timing is right. And if Electronifie does flourish, it’s likely that big-ticket fixed income trading will, finally and belatedly, join equities in the electronic age.
Of course, electronic bond trading is nothing new. The leading pioneer is MarketAxess, a highly successful exchange that now executes around 15% of all investment-grade transactions. MarketAxess specializes in orders up to around $5 million. It primarily matches bids and asks placed by “dealers,” meaning banks and brokers, and has enormously lowered costs for executing these smaller and medium-sized trades.
Electronifie is aiming at a different market: extra-large transactions of $5 million to $50 million, or even more. Those trades account for around 50% of all investment -grade transactions, measured in dollars. “We’re solving a different problem,” says Kuchinad. “Portfolio managers change their strategies, and have to buy or sell tens of millions of dollars in a single bond. Or they experience big outflows from investors, and have to sell large blocks of bonds.”
Electronifie wants to fundamentally change the way those big blocks of bonds change hands. Today, the vast bulk of transactions go through intermediaries, banks or brokers that take an order from the party that actually owns the bonds, such as a BlackRock or Fidelity, and matches the order with another party with a big fixed income portfolio.
The “end-users” seldom trade directly with one another. Electronifie connects the asset managers, pension funds, and the like so that they buy and sell from each other without an investment bank serving as a go-between. Electronifie is focusing on the market where “voice” is most prevalent, those giant trades. “We expect close to 100% of our trades to be between the principals,” says Kuchinad. “And we eventually expect the average size to be around $20 million.” If Kuchinad succeeds, it will be a phenomenon fixed income trading has never witnessed. Electronifie is also likely to draw competing platforms that, in turn, will accelerate the shift in which, say, CalPERS —California’s massive public employee pension system—can sell blocks of bonds directly to Fidelity, with neither side having the slightest clue of who’s taking their order.
In the past few years, Wall Street’s role in bond trading has changed substantially, providing a wide opening for Electronifie. As a former Goldman Sachs trader, Kuchinad watched the transformation first hand. In the pre-2007 world, the banks, or “dealers,” held giant portfolios of bonds. The asset managers could sell and buy multi-million dollar blocks directly to a Goldman or Bank of America. “The banks would warehouse risk and give the asset managers immediate liquidity,” says Kuchinad. “It was expensive, but the asset managers were getting big quantities of bonds off their books quickly, in a single transaction. They were willing to pay a lot for that service.”
With Kuchinad as guide, let’s examine the typical workflow prior to the financial crisis. A trader at asset manager XYZ seeking to sell $25 million in GM bonds would typically call salespeople at three investment banks and ask each one for their best price. According to Kuchinad, the highest quote was mostly likely to come not from dealers who already owned a lot of GM bonds but instead from the broker who held little or none. “If I own a lot and take more, and if I’m wrong, I’m really wrong,” says Kuchinad. Say a dealer already holds $100 million in GM bonds. He’ll eventually have to sell his position in GM, so if he raises it to $125 million, he’ll have more to sell. As a result, he’ll probably get poor pricing on the last $25 million tranche. So he’ll offer a lower price to asset manager XYZ than a dealer with slim holdings of GM.
Let’s say dealer No. 1—who is light on GM securities—offers the highest quote. The trader would then inform dealers 2 and 3 that they’d lost. “Every time I search for something, I show my poker hand,” says Kuchinad. “These losing dealers are playing in the same game.” They now know that the winning dealer just bought a big GM block and will soon sell it. And they can profit from this information in three ways. First, they could “front run” by dumping their own GM bonds before the winning dealer sells—remember, they may very well be large holders of GM bonds. Second, if they don’t own GM bonds, they could use derivatives to short the securities and pocket a profit when the winning dealer sells and prices fall. Third, they could curry favor with other asset manager-clients who own GM by leaking that a big block of GM will soon go on sale.
A typical, highly valuable tip, says Kuchinad, would be informing a prized client that “a large west coast asset manager not in Pasadena just sold $20 million in GM bonds.” In broker-speak, that identifies the seller as PIMCO. Now the client not only knows that a dealer will soon be selling GM bonds purchased from PIMCO. He also knows that, given PIMCO’s size, PIMCO will probably be selling more GM bonds. Hence, he could benefit by rushing to sell his GM holdings first.
Wall Street traders call such inside tips “color.”
Let’s return to the winner. Dealer No. 1 knows that the losers will work against him by pushing down the price of GM bonds. So he bakes the anticipated drop into his price. Two factors determine his pricing. First, he looks at the bid-offer spread. Say it’s 25 cents on a $100 bond. Today, he could sell the bond at 25 cents more than the purchase price. The dealer will want to pocket that spread as his minimum profit. Second, he also must charge a premium to compensate for the anticipated fall in the price. That anticipated drop is what’s called “market impact,” and it’s caused by the information gleaned by the two losing dealers.
So the winning dealer bids 1/8 of a point, or 12.5 cents, below the offer as a cushion for market impact. In effect, he’s charging in advance for the predicted fall in the price of GM bonds. This is where the fun begins. In the old system, the dealers were in no hurry to sell. “They would never show desperation,” says Kuchinad. “They’d wait out the storm of selling.” In most cases, the price would then rebound to its former level. “Before 2007, it was a matter of getting past the temporary trading pressure until the bond regained its fundamental value,” says Kuchinad. “In most cases, the trader pocketed not only the bid-ask spread, but recouped the premium for market impact as well.” It’s what skilled trading was all about. And for most Wall Street firms, the formula made the pre-2007 era incredibly lucrative.
Once again, the asset managers didn’t mind paying those premiums for market impact, since they were handing the risk to Wall Street. The system was far from ideal because of the leakage. But the spread of information was somewhat contained since the asset manager tipped his hand only to the two losing dealers.
The financial crisis uprooted the old system. The Volcker Rule imposed tight limits on proprietary trading, and the Basel III regulations forced banks to radically lower their leverage. As a result, the banks shed most of their bond trading portfolios. Suddenly, big fund managers could no longer count on Wall Street to buy and sell big blocks of bonds. The “immediate liquidity” they prized became much harder to find.
Banks still buy bonds for their own account, but most of those purchases are on small trades. “Of the greater than $5 million trades, the banks will take maybe 25%, where they’d take most of them before 2007,” says Kuchinad. “It’s highly unlikely an asset manager will get a dealer to bid on a $20 million block.”
Instead of acting as a principal, banks are serving mostly as the equivalent of brokers. In the old system, shrewd traders limited the market impact. In the new regime, the degree of information leakage, and the damage it causes, is far worse. Here’s how today’s trading works: Once again, the asset manager calls three dealers, as in our previous example, to sell $25 million in GM bonds. But the dealers don’t offer to buy the securities. Instead, they say they’ll “work the order” by calling customers they think might be interested. They also propose a price where the bonds are likely to sell. The asset manager may agree to sell at that price or simply tell the dealer to come back with a “call price,” meaning the dealer will call around for the best bids, then give the customer the option of accepting or declining.
Let’s assume dealer No. 1 wins, meaning he’s chosen to shop the $25 million order. The two losers know a big GM block will soon be sold and can do just what they did before, either sell their own bonds if they own any, short them, or leak the news to their favorite clients. But now, it’s not just the two losers who learn about the pending sale. The winner will immediately start calling one fund manager after another to make the sale. Now, it’s the asset manager, not the dealer, who suffers when the leaked information pushes down the price. “The dealers become quick, aggressive sellers on behalf of their clients,” says Kuchinad. “Sure, they want to please the client, but they have a lot less incentive not to appear too desperate than when they owned the bonds themselves.” According to Kuchinad, the dealers may call as many as 20 clients and find only five takers. The 15 who don’t bite now know a big GM sale is coming. “It’s like viral marketing,” says Kuchinad. “Instead of three people working against you, suddenly you have dozens.”
By Kuchinad’s reckoning, the market impact—the change in price caused by information leakage—is now 25% to 50% greater than it was during the pre-2007 period.
That’s the problem Electronifie aims to solve. “The idea was, let’s create a broker-dealer with no conflicts that eliminates market impact,” says Kuchinad. “We don’t own or short bonds. We don’t benefit from information.” In June 2014, he started mock trading sessions on the new platform, asking traders at the big asset managers to try to game their system. “Nothing thrills a trader like figuring out a loophole,” says Kuchinad. Over the following 10 months, he used their feedback to eliminate any flaws that could allow buyers or sellers to derive valuable information. Although Electronifie is designed so that end customers can trade with one another directly, dealers do play a role by providing bid and offer pricing for a share of the revenues. “That fulfills the compliance requirements for the asset managers, showing they are buying and selling at prevailing market prices,” says Kuchinad.
The Electronifie platform is designed so that once an asset manager enters a buy or sell order, they can discern nothing about the identity of the other buyers or sellers, or even if they are buyers or sellers, or the quantities they’re aiming to divest or accumulate. Identity, side (buyer or seller), and size are all hidden.
Let’s take an example of a trade under the post-2007, dealer-as-broker regime that sprays information, versus a trade on Electronifie. XYZ asset manager wants to net $99 selling $25 million in GM bonds. The current market price is $99 3/8. The dealer needs to call dozens of buyers to place the bonds. As the news spreads, the price keeps dropping. He only succeeds at garnering an average price of $98.75. The dealer keeps A 25 cent markup, so XYZ nets $98.50. Hence, the total market impact costs a total of 62.5 cents.
Here’s how Electronifie lowers costs, both by minimizing spreads and market impact. Once again, the current market price is $99 3/8; Electronifie takes a spread of just 4 cents, instead of the 25 cents, in our example, extracted by the dealer. XYZ puts in an order to sell $25 million at $99 3/8 (Electronifie takes its 4 cent cut once the deal is done.) XYZ won’t sell less than $10 million, but will sell more if enough buyers appear. All of the participants on the Electronifie platform are notified that a customer has expressed “interest” in this GM bond. They don’t know if the customer is a buyer or a seller, or how much they have to buy or sell. All bids and offers are “executable,” meaning once you enter a quantity and price, anyone who matches it automatically buys your shares, or sells you theirs.
Let’s say asset manager ABC wants to buy $12 million of the GM bonds. But he also wants to “fish” first, by seeing if this interest is coming from a big player with lots to sell. In that case, ABC would lower its price and grab a bargain. Even now, ABC has little incentive to game, because it doesn’t know if the interest comes from a buyer or seller. Even so, it enters an offer to buy $3 million. A third fund, QRS, enters an order for $10 million. Since the total exceeds $10 million, XYZ sells $13 million of its $25 million block at $99 3/8.
Then, Electronifie sends a blast message to all participants: GM bonds just sold at $99 3/8. A 10-minute “matching session” opens where all buyers and sellers are invited to participate. Say that the “gamer,” ABC, tarries in entering his order for the remaining $9 million. Lots of buyers jump in ahead of ABC, grabbing the entire $12 million that XYZ has left to sell, all at $99 3/8. At the end of the matching session, ABC has bought only $3 million of the $12 million it aimed purchase. In the process, it’s learned nothing. It doesn’t know if the seller wanted to shed $3 million bonds, or $50 million, or anything in between. Fishing gets you nowhere on Electronifie. As ABC discovered, the best practice is putting in your full order at once, as did QRS.
Nor is gaming possible in the matching session. The participants don’t learn how much of their orders are filled, if any, until the session is over.
Because the $25 million sale didn’t move prices, XYZ collected $99 3/8, minus the 4 cents charged by Electronifie, for net proceeds of $99.335. That’s a huge improvement over the $98.50 XYZ would have received using a dealer. Our non-Electronifie example is at the extreme end, assuming lots of market impact. On average, Kuchinad reckons that Electronifie reduces costs from around 35 cents per $100 in spread and market impact, to just its 4 cent markup.
Electronifie is going live at a propitious moment. Big asset managers are worried that as interest rates inevitably rise, retail investors will start dumping bonds. Without the Wall Street banks there to buy them, they fear that buy-sell spreads will balloon because of a severe imbalance between desperate sellers, and a small pool of buyers. Electronifie and other principal-to-principal markets, could contain the damage, in two ways. First, by linking asset managers to other asset managers, they create a far more liquid market. Second, the market impact is far worse in periods of heavy selling. If one dealer or manager learns a big name is dumping a position, they’ll dump too, often at depressed prices, driving the market ever lower. If the sellers remain anonymous, waves of panicky copycat selling become less likely.
The question remains, why did it take so long for the bond market, which is manned by so many brilliant people, to reach this level of modernization? “The balance has shifted,” explains Kuchinad. “In the past, it was hard to recruit people from Wall Street who understood bond trading. Now, the opportunity costs of leaving Wall Street have shrunk.” The once-sumptuous compensation in fixed income and the available jobs have dropped. But surely, Kuchinad could have kept making millions on Wall Street. “It’s a job satisfaction thing,” he says. “Wall Street’s not nearly as respected as it used to.” As Victor Hugo once wrote, nothing is stronger than an idea whose time has come. We’ll soon see if this one makes history.
Watch more business news from Fortune: