Why investors see what they want to see, and often get it wrong
Recently, there was considerable discussion about discount brokers losing lakhs of active investors in early 2025. The narrative was compelling and widely accepted: Sebi’s stricter future and options (F&O) regulations had finally caught up with speculative traders, forcing them to abandon their platforms.
Market commentators, social media influencers, and financial journalists all nodded in agreement. The data seemed to confirm what everyone expected would happen after the regulatory changes.
There was just one problem: the underlying data told a completely different story. The term “active investor” in NSE reporting means someone who has executed at least one trade in the preceding twelve months. This means that an investor is only dropped from the active count one full year after they stop trading. The decline in active investors reported in the first half of 2025, therefore, reflected individuals who had stopped trading in the first half of 2024, long before the recent regulatory changes took effect.
This episode illustrates one of the most pervasive cognitive traps in financial analysis: confirmation bias. We instinctively seek information that supports our existing beliefs, while ignoring or dismissing contradictory evidence. When data appears to validate our preconceptions, we rarely dig deeper to understand what it actually means.
Finding the truth
The financial world is particularly susceptible to this bias because markets are complex, data is often ambiguous, and there’s usually some evidence to support almost any viewpoint. Bull markets produce bullish data, bear markets generate bearish statistics, and both sides can find compelling evidence for their positions. The challenge isn’t finding supporting data — it’s finding the truth.
Our increasing reliance on artificial intelligence for research and analysis has amplified this tendency. Many people now routinely ask AI chatbots to evaluate their investment ideas or market opinions. The responses they receive often seem sophisticated and well-reasoned, lending credibility to their original thinking. However, there’s a fundamental flaw in this approach: AI systems are designed to be helpful and agreeable, making them natural sycophants. They are what we call ‘yes-men’ in corporate life or politics.
When you present an AI with your conclusion and ask whether it’s correct, you’re essentially asking it to confirm your bias. The system will typically find ways to support your position, collect supporting evidence, and present a coherent case for your viewpoint. This creates a dangerous illusion of validation–you feel more confident in your opinion because an apparently intelligent system has endorsed it. I may have coined a phrase there. AI = Apparently Intelligent.
The solution isn’t to avoid AI altogether, but to use it more effectively. Instead of asking “Is my analysis correct?” try asking “What are the flaws in this argument?” or “What evidence contradicts this conclusion?” Better yet, present the AI with neutral inputs and ask it to identify multiple possible interpretations without revealing your preferred outcome.
This approach — actively seeking disconfirming evidence — is fundamental to sound investment research, whether you’re using AI or traditional methods. Professional investors understand that their first instinct is often wrong, and they’ve developed systematic ways to challenge their assumptions. They maintain devil’s advocate processes, seek out opposing viewpoints, and constantly question their conclusions.
Confirmation bias in investing
The confirmation bias problem extends beyond individual analysis to market-wide phenomena. When a particular narrative gains traction, it becomes self-reinforcing. Media outlets report stories that fit the prevailing theme, analysts produce research that supports the consensus view, and investors make decisions based on what “everyone knows” to be true. Alternative explanations are dismissed or ignored until reality eventually intrudes.
The most significant danger of confirmation bias in investing is its influence on portfolio decisions. Investors who believe technology stocks are overvalued will focus on metrics that support that view, while dismissing contradictory evidence. Those convinced that a particular sector is undervalued will find reasons to justify additional purchases, even as prices continue to fall. Both groups will feel intellectually satisfied with their positions right up until the market proves them wrong.
Developing immunity to confirmation bias requires conscious effort and systematic processes. Force yourself to engage with viewpoints that make you uncomfortable. The investor who admits uncertainty and hedges their bets often outperforms the one who is absolutely certain and completely wrong.
Conviction is admirable and feels heroic, but flexibility is more profitable.
Dhirendra Kumar is founder and chief executive officer of Value Research, an independent investment advisory firm. Views are personal.
