IEX Group Brad Katsuyama: Masters of Business (Audio)

Episode Summary

Executive Summary: The episode centers on Brad Katsuyama’s path from RBC trader to co-founder of IEX, after he discovered that modern market structure and high-frequency trading let some participants see and react to orders faster than others. He explains how experiments, latency differences, and IEX’s “shoebox” delay were used to create a fairer exchange, while arguing that transparency, accountability, and market design—not outright bans—are the right fixes.

Main Topics: Brad Katsuyama’s background and move to Wall Street (Priority: 4/5): Katsuyama describes growing up in Ontario, moving from RBC Toronto to New York in 2002, and learning Wall Street culture through mentorship and increasingly senior trading roles. Discovery that market structure was disadvantaging traders (Priority: 5/5): He recounts noticing that orders increasingly failed to fill in full after 2006, suggesting that markets were reacting before his trades could execute and that the displayed market was not reliably accessible. Experiments revealing exchange latency and routing advantages (Priority: 5/5): RBC approved controlled testing to understand exchange behavior. Katsuyama and his team used repeated buy/sell experiments to measure how order timing and geography affected fills across venues. The role of technology, geography, and high-frequency trading (Priority: 5/5): The conversation explains how fiber routes, exchange locations, co-location, and microsecond differences created advantages for faster traders, especially when orders hit the nearest venue first. Creation of IEX and the shoebox delay mechanism (Priority: 5/5): Katsuyama explains how IEX was designed with a 350-microsecond delay to neutralize speed advantages and ensure trades occur using up-to-date market information. Fairness, dark pools, and criticism of HFT (Priority: 4/5): He argues that some dark pools and HFT behaviors exploit market structure, but that technology itself is not the enemy; the issue is fairness, transparency, and accountability. Impact of Flash Boys and media attention (Priority: 4/5): Katsuyama reflects on Michael Lewis’s reporting, the 60 Minutes segment, the market controversy it sparked, and how public attention validated but also intensified the mission.

Key Arguments: The market became less accessible to ordinary traders because faster participants could react to visible orders before others could execute them. Latency differences were not abstract; physical geography and network routing directly affected who got fills and who got picked off. Controlled experimentation showed that routing a single order to one venue could produce a full fill, but routing to multiple venues reduced fill rates, exposing hidden speed advantages. IEX’s design is a market-based fix: slow orders just enough to eliminate unfair speed advantages without banning technology or computerized trading. Dark pools can serve a legitimate purpose for large orders, but many have sacrificed quality and transparency in pursuit of volume. High-frequency trading benefits often come from technology and lower trading costs generally, but some strategies extract value from slower investors and should be constrained by market design. The best long-term solution is not more rhetoric but a fair, transparent system with consistent rules and accountability. The exchange and regulatory ecosystem itself helped enable the problem by monetizing speed, data, co-location, and payment structures. IEX is intended to compete as a mission-driven marketplace rather than a maximally extractive platform. Public exposure through Michael Lewis and 60 Minutes mattered because it forced the issue into broader awareness and accountability.

Data Points: RBC experimentation loss limit: up to $10,000 per day - Katsuyama says RBC authorized him to spend this amount to test theories about execution and exchange behavior. Intel offer fill in 2006: 100,000 shares could be bought if desired - He cites 2006 as a time when displayed liquidity was more obtainable. Intel offer fill in 2007: 80,000 shares - He says his ability to buy displayed shares began declining in 2007. Intel offer fill in 2008: 70,000 shares - Continuation of the decline in accessible liquidity. Intel offer fill in 2009: 52,000 shares - He uses this to illustrate worsening execution quality over time. Latency between exchanges: 476 microseconds - Ronan Ryan estimated the inter-exchange latency from BATS to other venues at the time. IEX shoebox delay: 350 microseconds - The coiled-fiber delay built into IEX to neutralize speed advantages. Original target latency for Thor: under 400 microseconds - The earlier RBC router aimed to make orders arrive at multiple exchanges simultaneously. Thor achieved latency: about 290 microseconds - Katsuyama says they got the system down to this level and fill rates improved. First day IEX volume: 560,000 shares - He compares this with low initial expectations and concerns about whether anyone would trade there. Survival volume target: 50 million shares/day - This was described as the level needed for IEX to survive. Recent IEX record volume: 182 million shares - Katsuyama states this was their record a few weeks before the interview. Current IEX market share: around 1% of total market - He describes IEX’s scale at the time of the interview. Volume unavailable to IEX: roughly 50%-60% displayed quotes, 10% open/close auctions and pre/post-market, about 20% retail internalization - He explains why becoming an exchange would unlock more of the market. Number of dark pools in the U.S.: over 40 - He notes the proliferation of dark pools and expects some to disappear with more transparency.

Pivotal Quotes: "Brad, we love what you've done, we love your team, we love RBC. You've only solved 3% of my problem." — Brad Katsuyama: A client’s comment that convinced him no single broker could solve the market structure problem for most investors. "The market's become, at times, more volatile intraday. Volumes are inflated." — Brad Katsuyama: He argues that technology-driven trading can distort market behavior and inflate apparent liquidity. "The problem is diffuse harm and concentrated benefit." — David Einhorn: Katsuyama cites this line to explain why most investors lose a little while a small group gains a lot.

Implications: The episode frames market fairness as a design problem, not just a regulatory one. For investors, it suggests structure can be changed to reduce hidden speed advantages. For the industry, IEX represents a test case for transparent, mission-driven exchange competition.