A deep look at how financial markets are designed

Financial markets are fast-moving, complex, and opaque. Even the U.S. stock market is fragmented into an array of competing exchanges and a set of proprietary "dark pools" run by financial firms. Meanwhile, high-frequency traders zoom around buying and selling stocks at speeds other investors cannot match.

Yet stocks represent a relatively transparent investment compared to many types of bonds, derivatives, and commodities. So when the financial sector melted down in 2007-08, it led to a wave of reforms as regulators sought to rationalize markets.

But every financial market, reformed or not, has its quirks, making them all ripe for scholars to scrutinize. That's what Haoxiang Zhu does. The Gordon Y. Billard Professor of Management and Finance at the MIT Sloan School of Management is an expert on how market design and structure influence asset prices and investors. Over the last decade, his detailed theoretical and empirical studies have illuminated market behavior and gained an audience - scholars, traders, and policymakers - interested in how markets can be structured.

"When we need to reform markets, what should we do?" asks Zhu. "To the extent that something is not done perfectly, how can we refine it? These are very concrete problems and I want my research to shed light directly on them."

One award-winning paper Zhu co-wrote in 2017 shows how transparent, reliable benchmark prices help investors efficiently identify acceptable costs and dealers in many large markets. For instance, in 2012, LIBOR, the interest-rate benchmark applied to hundreds of trillions of dollars in derivatives, was shown to have had price-manipulation problems. Zhu's work emphasizes the value of having robust benchmarks (as post-2012 reforms have attempted to address) rather than scrapping them altogether.

Another recent Zhu paper, published this past September, looks at the way the Dodd-Frank banking legislation of 2010 has changed the trading of some credit default swaps in the U.S. - by using centralized mechanisms to connect investors and dealers, instead of the one-on-one "over-the-counter" market. The new design has been working well, the paper finds, but still has room to improve; investors still have no easy ways to trade among themselves without dealer intermediation. Additional market-design changes could address these issues.

Many of Zhu's results are nuanced: One 2014 paper he wrote about the stock market suggests that privately-run dark pools may unexpectedly help price discovery by siphoning off lower-information traders, while better-informed traders help determine prices on the bigger exchanges. And a 2017 study he co-authored about the optimal trading frequency of stocks finds that when it comes to setting new prices, smaller-cap companies should likely be traded less frequently than bigger firms. Such findings suggest subtle ways to think about structuring stock-markets - and indeed Zhu maintains ongoing dialogues with policy experts.

"I think this sort of analysis does inform policymaking," Zhu says. "It's not easy to do evidence-based rulemaking. It's costly to discover evidence, it takes time."

Solving one problem at a time

Zhu did not fully develop his interest in finance and markets until after his college days. As an undergraduate at Oxford University, he studied mathematics and computer science, graduating in 2006. Then Zhu got a job for a year at Lehman Brothers, the once-flourishing investment bank. He departed in 2007, a year before Lehman imploded; it had become overleveraged, borrowing massively to fund an array of bad bets.

"Fortunately, I left early," says Zhu. Still, his short time working in finance revealed a couple of important things to him. Zhu found the daily routine of finance to be "very repetitive." But he also became convinced there were compelling problems to be addressed in the area of market structures.

"I think part of my interest in the details of market design has to do with my industry experience," Zhu says. "I came into finance and economics viewing it somewhat from the outside. I looked at it more as an engineer would. That's why I think MIT's a perfect fit, because of the engineering way of looking at things. We solve one problem at a time."

Which is also to say that Zhu's research is not necessarily intended to produce overarching conclusions about the nature of all markets; he investigates the mechanics of separate markets first and foremost.

"It's hard to get very deep if you start too broad," says Zhu, who earned tenure at MIT last year. "I would argue we should start with depth. Once you get to the bottom of something, you see there are connections between many different issues."

Zhu received his PhD in 2012 from Stanford University's Graduate School of Business, and joined the MIT faculty that same year. Along with his appointment in Sloan, Zhu is a faculty affiliate in the MIT Laboratory for Financial Engineering and the MIT Golub Center for Finance and Policy.

Among the honors Zhu has received, his research papers have won several awards. The paper on benchmarks, for one, was granted the Amundi Smith Breeden First Prize by the Journal of Finance; the paper on optimal trading frequency won the Kepos Capital Award for Best Paper on Investments, from the Western Finance Association; and Zhu's dark pools paper won the Morgan Stanley Prize for Excellence in Financial Markets.

Like a start-up

Much of Zhu's time and energy is also devoted to teaching, and he is quick to praise the students he works with at MIT Sloan.

"They are smart, they are hard-working," Zhu says. Of his PhD students, he adds, "It is always a challenge to go from being a good student getting good grades to producing research. Producing research is almost like starting up a company. It's not easy. We do our best to help them, and I enjoy interacting with them."

And while continuing to study financial market design, Zhu is expanding his research portfolio. Among other projects, he is currently looking at the impact of new payment systems on the traditional banking industry.

"I think that's really a fantastic area for research." Zhu says. "Once you have a [new] payment system, people's payments get diverted away from the banks. … So we basically look at how financial technology, in this case payment providers, siphons off customers and information away from banks, and how banks will cope."

At the same time, Zhu's work on market structures continues to have an audience in the finance industry and among its regulators, both of which he welcomes. Indeed, Zhu has written several comment letters to regulators about proposed rules that could have material impact on the market. For example, he has argued against certain proposals that would reduce the transparency of the corporate bond market, the swaps market, and investment managers' portfolio holdings. But he is in favor of the U.S. Treasury's innovation in issuing debt linked to the new U.S. benchmark interest rate that is set to replace LIBOR.

"In market design the message is often nuanced: There are advantages, there are disadvantages," Zhu says. "But figuring out the tradeoff is what I find very rewarding, in doing this kind of work."