Overpaid S&P 500 Ceos Underperform Cheap Ones Like Elon Musk by a Landslide

More CEO Pay Does Not Buy More Stock Return Across the S&P 500's Largest Companies

FMP Executive Compensation Data, FMP/OpenBB Price Performance. 82 S&P 500 companies with CEO data, April 2021 to April 2026.

The $16.5 Million Consensus Has No Empirical Basis

Median CEO compensation in the S&P 500 hit $16.5 million in 2024, a 6.6% increase over 2023 and an all-time record.2 The compensation consulting industry treats this figure as a benchmark. Boards set pay packages relative to peer medians, consultants calibrate long-term incentive grants to "market," and proxy advisors evaluate plans against sector norms.

The entire apparatus rests on an assumption that higher pay attracts and retains talent that generates superior shareholder returns. Five years of data across the 82 largest S&P 500 companies with available CEO compensation records suggests the opposite.

The correlation between cumulative five-year CEO compensation and five-year stock returns is -0.109. Negative. Not zero, not slightly positive, not "mixed."

The regression line tilts downward. Every additional $100 million in CEO pay over five years is associated with lower, not higher, stock performance.1

The Consensus and Where It Breaks

The dominant narrative from compensation consultants like Pearl Meyer, Pay Governance, and Semler Brossy holds that "pay-for-performance alignment" is strong because CEOs earn more when stock prices rise and less when they fall. A March 2026 Harvard Law School study reinforced this by showing strong correlation between relative Cumulative Adjusted Pay (CAP) and relative total shareholder return across peer groups.3

The narrative breaks in two places. First, the comparison is always relative, measured against industry peers, not absolute. A media CEO whose stock falls 49% but "only" trails peers by 20 percentage points can still receive a pay increase.

Disney's Bob Iger demonstrates the problem perfectly. His compensation jumped 30% to $41.1 million in fiscal 2024 while the stock declined 21% and the peer group returned 79.6%.4

Second, the Virginia Tech finding that CEO compensation has become 24% more similar across all public firms since 2006 explains why the correlation is negative.5 If every board benchmarks to the 50th-75th percentile of peer medians, and every peer group does the same, the result is a ratchet that pushes compensation upward regardless of performance. Boards are not paying for alpha. They are paying for the privilege of not being below median.

The Value Destroyers Got Paid Like Value Creators

Intel's executive suite collected $298.7 million in cumulative compensation over five years while the stock lost 23% of its value.1 Pat Gelsinger, who was forced out in December 2024, earned at least $46 million during his tenure and left with $10 million in severance, while Intel's share price fell 55% from his appointment date.6 The board replaced one highly compensated CEO with another and the structural issue persists.

Disney paid Robert Iger $200.5 million over five years while the stock fell 49%. Only 80% of shareholders approved Disney's executive compensation package in 2025, far below the S&P 500 median of 92.7%.4 Shareholders expressed displeasure with their votes, but the pay kept rising anyway.

ServiceNow stands as the single most extreme case in the dataset. Bill McDermott collected $356.1 million in five-year cumulative compensation while the stock returned 1%.1

That is $356 million for a rounding error of shareholder value creation.

The Value Destroyers Versus the Efficiency Kings

Source: FMP Executive Compensation, FMP/OpenBB Price Performance. April 2021 to April 2026. Ticker labels include 5-year cumulative CEO compensation in parentheses.

The Names That Prove the Inverse Thesis

Elon Musk collected $100,000 in total reportable compensation over five years at Tesla while the stock returned 53%. The contrast is imperfect because Musk's true economic interest runs through a massive stock option package worth tens of billions. But the proxy-reported number matters for this analysis because it reflects what the board chose to disclose as CEO pay, and Tesla's shareholders have explicitly voted to affirm that structure.1

Alexander Karp at Palantir took home $18.1 million in five-year cumulative pay while the stock returned 536%. Michael Dell collected $16.7 million while Dell Technologies returned 282%.1

Jayshree Ullal at Arista Networks earned $57.6 million against a 556% return. These CEOs generated 10 to 30 times more shareholder value per dollar of compensation than the median S&P 500 CEO.

The pattern is clearest in the semiconductors. Jensen Huang at NVIDIA earned $159.9 million over five years, but the stock returned 1,182%. That produces a pay efficiency ratio of 7.4% return per $1 million of comp.1

Hock Tan at Broadcom earned $286.8 million and delivered 550%, yielding 1.9% per $1 million. Compare that to Intel, where every million dollars of CEO compensation bought shareholders a negative return.

The Cheapest CEO Quartile Outperformed the Most Expensive One by 21 Points on Median and 71 Points on Mean

82 S&P 500 companies sorted into quartiles by five-year cumulative CEO compensation. Q1 is the lowest-paid quartile. Source: FMP, OpenBB.

The Mechanism Behind the Disconnect

Three structural forces explain why higher CEO pay correlates with lower returns. The first is adverse selection in the compensation ratchet. Boards that overpay relative to performance signal weak governance.

Weak governance compounds over time through poor capital allocation, empire building, and misaligned incentives. The pay package is not the cause but the symptom.

The second force is dilution. The majority of mega-cap CEO compensation arrives as stock awards and options. When CEOs receive massive equity grants, existing shareholders absorb dilution.

A $160 million stock grant does not appear on the income statement, but it reduces per-share value. Companies that hand out smaller equity packages preserve more ownership for shareholders.

The third force is the "founder premium." Four of the top ten performers by pay efficiency (Musk, Karp, Dell, Huang) are founders or co-founders who own substantial equity positions independent of their annual compensation. Their incentives are structurally aligned with shareholders because their net worth depends on the stock price, not the proxy statement. Founder-led companies tend to accept below-market pay packages because the CEO's real compensation is capital appreciation on existing holdings.

The Counter-Argument

The pay-performance correlation has legitimate confounders that prevent a causal interpretation. Survivorship bias in the performance data is the primary one.

Companies in the top quartile of CEO pay include firms like Broadcom (+550%), Palo Alto Networks (+188%), and Apple (+105%) that delivered strong returns despite massive compensation packages. The data shows that some boards write enormous checks and get strong results.

The "tournament theory" of executive compensation holds that high CEO pay motivates the entire C-suite to compete harder for the top job, generating organizational returns that exceed what the CEO individually produces. Academic research from Lazear and Rosen provides theoretical support, and empirical work from Kale, Reis, and Vaidyanathan finds that larger pay gaps between CEOs and the next level of executives are associated with higher firm values. This theory would suggest that cutting CEO pay could reduce organizational performance even if the CEO themselves is not the primary driver of returns.

There is also a composition effect. The highest-paid CEOs tend to lead companies in transition. Bob Iger returned to fix Disney's streaming losses, Pat Gelsinger attempted to rebuild Intel's manufacturing capabilities, and Brian Niccol was recruited to revive Starbucks.

Turnaround CEOs command premiums precisely because the job is harder and the probability of failure is higher. Measuring their pay against stock performance during the turnaround period may penalize exactly the kind of bold leadership that boards are trying to incentivize.

Finally, the five-year window creates timing distortions. A CEO hired in 2023 with a large sign-on grant may show high cumulative pay but has only had two to three years for strategy to take effect.

Measuring compensation over a full business cycle, seven to ten years, might show different results. The negative correlation could narrow or disappear with a longer measurement period.

Key Data Table

Catalyst Map

• 2026 Proxy Season (April-June 2026). Say-on-pay votes at Disney, Intel, ServiceNow, and Comcast will test whether shareholders enforce consequences for the pay-performance disconnect. Disney's 80% approval rate in 2025 is trending toward potential failure territory.
• SEC Pay-vs-Performance Disclosure Rule. The rule requiring companies to disclose Compensation Actually Paid (CAP) versus TSR is entering its third full year. Institutional investors have access to standardized data that makes pay-performance gaps visible. Proxy advisors ISS and Glass Lewis are incorporating this data into recommendations.
• CEO transitions at Intel, Disney, Starbucks. All three value-destroyer companies are in active leadership transitions. New CEOs will set the tone for compensation expectations. Boards that reset pay lower will signal governance improvement.
• Passive indexer engagement. BlackRock, Vanguard, and State Street collectively control 20%+ of votes at most S&P 500 companies. Their voting guidelines on executive compensation are tightening. A single large indexer shifting from "approve" to "abstain" on compensation packages at underperforming companies could push multiple say-on-pay votes below 70%.

Sources