Overconfidence: Why Tech Employees Hold Employer Stock Too Long

The finance literature on overconfidence has three well-established findings. People overestimate the accuracy of their own forecasts. People believe they know more than they do. People believe their own decisions are better than average. Each finding operates robustly in individual investing, and each compounds when the investor is making decisions about an asset they work on every day: their employer’s stock.

A senior engineer at a public tech company sees product roadmaps, hears internal metrics conversations, observes hiring patterns, and participates in strategy discussions. The information feels like an edge. “I know this company better than any outside analyst,” the employee reasons. “I should hold my concentrated position because I can see what’s coming.”

This is usually wrong. The employee has partial information that the market has priced, often inaccurately on the pessimistic side for strong companies and accurately across the full picture including risks the employee has not focused on. The confidence is real. The edge is not. The result is holding a concentrated position through ordinary market cycles at a significant risk-adjusted cost.

The Empirical Evidence on Employee Stock Prediction

Academic studies consistently find that employees are no better than average at predicting their own company’s stock. Meulbroek’s 2005 study of insider-restricted employee ownership found that employees holding concentrated positions underperformed a diversified alternative by 2-4% annualized, net of any informational advantage.

Benartzi’s 2001 study of 401(k) employer-stock allocations found that employees heavily invested in their employer’s stock had no forecasting advantage. The employees tended to buy the stock after it had already outperformed (recency bias) and hold through underperformance (status quo and loss aversion), producing net returns below diversified alternatives.

A 2023 Vanguard analysis of 401(k) plan participants with employer-stock concentrations over 20% found that over 10-year rolling windows, concentrated holders underperformed by a median of 2.1% annually versus participants without concentration. The concentration premium was slightly positive in a small number of outlier companies and strongly negative on average.

The pattern is not surprising. Market-efficient pricing means that the information an employee has (positive outlook on products, positive view of management) is already in the price. What employees systematically lack visibility into: macroeconomic shifts, competitive displacement by firms they don’t compete against directly, regulatory changes, and management mistakes that internal culture downplays.

What Employees Actually See

The specific information a senior IC has about their employer:

Product roadmap. What is shipping in the next 6-12 months.
Hiring pace. How aggressively the company is hiring.
Internal metrics. DAU, MAU, retention, conversion, revenue growth seen from inside.
Leadership quality. Whether the CEO and exec team are making good decisions.
Team morale. Whether engineers and product managers believe in the company.

This information feels rich. It is also:

Late. The market has priced on forward-looking expectations that reflect similar information from thousands of employees at peer companies. Your read is one data point among many.
Selection-biased. You joined and stayed because you believe in the company. The people who left have different views you don’t hear.
Narrow. You see your team, your products. You don’t see competitors’ products in development, regulatory risks, macro shifts.
Correlated with price you can’t act on. Most employees are subject to §10b5-1 constraints and can only act on information through pre-approved plans, eliminating the edge even when it exists.

The employee’s information edge, net of market pricing efficiency and trading constraints, is close to zero for most equity positions.

Calibration Studies and the “Planet-Level” Confidence

Cognitive psychology research on calibration asks people to answer factual questions and provide confidence intervals they are 90% sure contain the right answer. Well-calibrated people have their intervals contain the answer 90% of the time. Overconfident people have intervals that contain the answer 40-60% of the time.

Applied to stock forecasting: an employee who says “I am 90% sure this stock will be higher in 12 months” and runs this judgment across 10 years of positions typically sees the stock higher in only 55-65% of the 12-month windows. The employee’s subjective confidence exceeds the empirical success rate by 25-35 percentage points.

This miscalibration is robust. Training in statistics reduces it slightly. Experience does not automatically reduce it (experienced investors often show the same pattern). Explicit feedback loops (tracking predictions against outcomes) reduce it, but the feedback loops are rarely maintained.

The “I Work Here So I Should Hold” Fallacy

A separate but related bias: treating employer loyalty as investment logic. The employee reasons, “I work here, I believe in the mission, selling my stock would be disloyal or would signal lack of faith.”

This reasoning conflates employment and investment. Employment is a relationship that provides salary, career development, and (for equity comp) additional ownership stake over time. Investment is a decision about how to allocate existing capital.

A loyal employee who sells a concentrated position to diversify is not less loyal. The diversification decision has no effect on employment commitment. The conflation exists purely in the employee’s head.

Public-market employees are particularly prone to this bias because their vest dates put them through an emotional “I am now an owner” moment every quarter. The ownership feeling builds identification with the stock beyond the financial position. Private-market employees see less of this cadence and often diversify faster at liquidity events.

The Cost Across a Career

A senior IC who joins a high-growth tech company at age 28 with a $300K total comp (roughly $150K salary + $150K equity) and receives refreshers of $100K per year, over 10 years:

Cumulative vests: $2M
Starting basis at vest: $2M
If held as concentrated: depends on stock performance
If sold and diversified at vest: grows at diversified return

For a high-performing single stock (say, Nvidia from 2015-2025), the concentrated strategy dominates massively. For an average single stock, the strategies are roughly equivalent on expected return but differ sharply on risk-adjusted return. For a below-average single stock (most stocks), diversification wins by meaningful amounts.

The problem: the employee does not know ex-ante whether their employer is the next Nvidia or the next Peloton. Overconfidence suggests Nvidia. Base rates suggest average or below.

Expected-value analysis: base-rate-weighted outcomes across all tech IPOs from 2010-2024 show that a strategy of holding all concentrated stock post-vest underperforms a strategy of selling 50% at vest by roughly 1.5-2.5% annualized over 10 years, with dramatically wider outcome distributions. On a $2M cumulative-vest basis, that’s $300K-$500K of expected underperformance over 10 years.

What Calibrated Confidence Looks Like

A calibrated employee distinguishes what they know from what they guess:

“I know our next product launches in Q2.”
“I guess the launch will meet market expectations.”
“I know my team is technically strong.”
“I guess competitors have weaker teams.”

The “know” statements are informational. The “guess” statements are forecasts, and forecasts are subject to base rates and uncertainty that should lower confidence.

A calibrated allocation decision: hold a modest overweight (perhaps 10-15% of net worth) in employer stock to reflect the genuine information you have, but do not concentrate at 40-60% on the belief that your forecasts are reliable. The 10-15% overweight captures whatever edge exists. The rest of the portfolio is diversified against your forecasting errors.

Implementation

Write down your thesis. If you hold concentrated stock, articulate the specific reasons. “The company is great” is not a thesis. “In the next 18 months, product X will capture 20% of market Y, driving revenue growth from 30% to 45%” is a thesis.
Track your predictions. Keep a log of positions held versus positions diversified. Compare outcomes over 2-year windows. Most employees discover their forecasts are no better than 50-50.
Sell into strength. When the stock is “higher than expected,” that’s exactly when selling reduces concentration risk most efficiently. Selling at all-time highs feels wrong because it requires disagreeing with the recent trajectory, but that’s when the expected value of diversification is highest.
Use structural constraints. 10b5-1 plans, scheduled sells, exchange fund contributions. These don’t require you to override your overconfidence; they operate without requiring the decision.
Benchmark to diversified alternative. Track what your concentrated position is worth versus what a diversified version would have been worth. If the concentrated is winning, you can rationalize holding. If it’s losing, you have evidence of your forecasting error.

Frequently Asked

Doesn’t overconfidence sometimes pay off? Yes, tail outcomes exist. A founder who concentrated in their startup from day one and held through a 100x return clearly beat diversification. The expected value calculation accounts for tail outcomes but recognizes that the median outcome is worse than diversification.

If I have no concentrated position, am I under-exposed to my employer’s upside? Vesting is your exposure. Over four years of RSU vests, you gain employer exposure equivalent to 4x your annual equity comp. That’s already substantial. Concentration beyond the vest schedule requires an active holding decision.

What about the founder’s concentration? Founders face different calculations. Their concentration reflects lock-up, control issues, and tax constraints (QSBS holding periods, for example). They also face higher upside tails. But even founders should be diversifying at liquidity events when they can.

Does the Dunning-Kruger effect apply here? Somewhat. Newer employees often show highest confidence. Mid-career senior ICs show confidence moderated by exposure to company struggles. Executives at public companies often show appropriately calibrated confidence (or use their calibration to exit positions through §16 officer planning).

Is the overconfidence different for private-company employees? Private-company employees have less frequent price signals, which reduces the “see-the-price-daily” reinforcement cycle. They often show less active-forecasting bias but more strategic-holding bias (“I’ll diversify after the IPO”). The concentration risk is real in both cases.