Incentives, Investor Behavior, and AI with Devin Shanthikumar
Smart people make devastating investing mistakes not because they lack intelligence, but because they don't understand the incentives embedded in financial systems. Devin Shanthikumar's research began with a simple observation during the dot-com bubble: Technically sophisticated individuals at the cutting edge of technology lost substantial money because they didn't understand how markets actually work. That question became her career focus: what mistakes are people making, and why?
The answer reveals how sell-side analysts navigate competing pressures from institutional clients seeking valuable insights, retail investors generating trading commissions, and companies controlling access to management. This creates what Devin calls "speaking in two tongues"—recommendations skew positive to generate trades while earnings forecasts reflect different accuracy incentives. Understanding these dynamics matters more than ever as investment commentary shifts toward social media, though early research offers surprising hope: engagement on online platforms may moderate views rather than amplifying extremes.
Her current work examines how AI is changing analyst research, with early findings showing improved accuracy and faster processing of complex filings. Most surprising: AI enables bolder forecasts rather than forcing consensus, freeing analysts to focus on irreplaceable human judgment.
Listen or Watch
Choose your platform:
Apple Podcasts | Spotify | YouTube | PodBean
What We Cover in This Conversation
How the dot-com boom and bust shaped Devin's research question: what mistakes are smart people making, and why
The multiple constituencies analysts serve simultaneously: institutional clients, retail investors, companies, and trading desks
How access to management creates incentive conflicts that shape analyst research outputs
The "two tongues" framework: why recommendations and earnings forecasts reflect different constraints
Why analyst research provides the ideal laboratory for studying AI: clear outputs, measurable accuracy, quarterly evaluation
Early AI findings: improved forecast accuracy, faster updates after earnings, better processing of complex filings
The counterintuitive discovery that AI enables bolder forecasts where human judgment matters most
How Devin approaches teaching in the AI era: foundational knowledge, critical thinking, using AI as a learning partner
The challenge of reduced entry-level positions and what it means for career development ladders
Key Takeaways
The dot-com bubble revealed how intelligence without understanding incentives can lead to disaster. Smart, technically sophisticated individuals in the Bay Area made painful investing mistakes. This pattern recurs: technical competence doesn't transfer to understanding market mechanics or the business models producing investment commentary.
Analysts "speak in two tongues" because recommendations and forecasts serve different purposes. Recommendations skew heavily toward "Buy" and "Strong Buy" because they can generate retail trading commissions. Earnings forecasts face different incentives—accuracy matters, but slight conservatism that allows companies to beat estimates strengthens analyst-management relationships. These aren't conscious manipulations but responses to institutional pressures analysts navigate daily.
AI improves analyst research quality while enabling greater boldness in human judgment. Devin's research shows AI tools enhance forecast accuracy, accelerate updates after information events, and improve processing of complex filings. The surprising finding: rather than driving consensus as automation might suggest, analysts using AI produce bolder forecasts, particularly for companies where private information and judgment matter most. AI handles routine processing, freeing analysts to focus on the irreplaceable human work of interpretation and insight generation.
Timestamps
00:00 - Introduction: Devin's path from Berkeley engineering to Stanford finance PhD
02:00 - Growing up as child of immigrants, discovering academia through teaching
04:00 - Early 2000s dot-com boom and bust: smart people making painful mistakes
06:00 - Harvard Business School postdoc and learning the case method
08:00 - First research question: are small investors naive to analyst incentives
15:00 - Trust, cynicism, and the risk of moving from conflicted advice to unregulated commentary
17:00 - The disintermediated world of finfluencers and social media investment commentary
21:00 - Why analyst research is the ideal laboratory for studying AI impact
23:00 - Early AI findings: improved accuracy, faster updates, better processing of complex filings
26:00 - Teaching in the AI era: foundational knowledge, critical thinking, smart tool use
32:00 - Building human community as essential in an AI-mediated world
34:00 - Rapid-fire questions
About Devin Shanthikumar
Devin Shanthikumar is faculty at the UC Irvine Paul Merage School of Business and Associate Dean for Undergraduate Studies. She earned her undergraduate degree in engineering and computer science from UC Berkeley and her PhD in finance from Stanford, where she studied with Ulrike Malmendier, current president of the American Finance Association. Her research examines how incentives, information design, and interpretation shape investor behavior and market outcomes, with recent work focusing on AI's impact on financial analyst research and performance.
Stay Connected
Want insights like this delivered to your inbox?
Twice-monthly analysis on markets, behavioral finance, and wealth strategy. Join 1,000+ subscribers who receive sophisticated thinking without the superfluous complexity.
Share This Episode
Found this valuable? Share it with someone who'd benefit.
Disclaimer
The content of "Treussard Talks" is for informational and educational purposes only and should not be considered financial advice. The views expressed are those of the host and guests and do not necessarily reflect the opinions of Treussard Capital Management or its affiliates. Consult your own financial advisor before making any investment decisions. For full disclosures, visit treussard.com.