The short answer
To understand what a management team is likely to optimise, first look at how it gets paid.
Executive compensation does not mechanically determine every decision. Strategy, competition, regulation, the balance sheet and personal judgment still matter. But a scorecard changes which outcomes receive the most attention, especially when a large part of annual or long-term pay depends on a small number of measurable targets.
The useful framework is:
Incentive pressure = Metric × Weight × Payout curve × Time horizon × Adjustment rules
The metric says what counts. The weight says how much it matters. The payout curve shows how sharply the reward changes around the target. The time horizon affects whether management is rewarded for this year or for durable performance. Adjustment rules determine what can be excluded from the final result.
This is why reading only the headline pay figure is not enough. The scorecard underneath it is often more informative.
Read the scorecard before the strategy slogan
Two companies can both say they prioritise profitable growth while paying executives for very different results.
One may place most of the annual bonus on revenue. Another may emphasise operating profit, free cash flow or return on invested capital. A third may link long-term awards to earnings per share and relative total shareholder return.
Each design creates a different pressure:
- Revenue can reward scale, but weak pricing or expensive growth may still pass through the scorecard.
- Operating profit adds cost discipline, but can encourage management to delay spending that would help future growth.
- Free cash flow rewards cash conversion, but accelerating collections, delaying payments or postponing investment can lift it in the short run.
- Earnings per share combines profit with the number of shares over which that profit is spread, so buybacks can raise the metric even when total profit does not change.
- Return on invested capital asks whether the company earned enough on the capital it used, but delaying new investment or excluding some capital items can make the result look stronger.
- Relative total shareholder return compares the share price and dividends with a peer group, but can still pay well when the whole sector performs poorly.
No metric is automatically good or bad. The issue is what behaviour becomes easier to reward and what important result may be left outside the scorecard.
The metric alone is not the contract
A compensation table usually includes a threshold, target and maximum. That shape matters.
Imagine that a bonus pays nothing below 90% of target, 100% of target pay at 100%, and 200% of target pay at 110%. Moving from 99% to 100% can then be more valuable to an executive than moving from 80% to 90%, even if the operating improvement is similar.
The same metric can also produce different behaviour over different periods. A one-year EPS target creates a shorter decision window than a three-year cumulative EPS target combined with return on invested capital.
Investors should therefore ask five questions:
- What is measured?
- How much weight does each metric receive?
- Where are the threshold, target and maximum?
- Is the period one year or several years?
- Which costs, gains and losses can the compensation committee remove?
What the research finds
Academic evidence supports taking these incentives seriously without turning them into an automatic accusation.
A Journal of Financial Economics study by Bennett, Bettis, Gopalan and Milbourn found that firms just above earnings-based compensation goals showed patterns consistent with managing the measured result, including higher abnormal accruals and lower research and development spending. The shape of the goal mattered, not only the metric’s name.
A study in The Accounting Review found a strong positive association between share repurchases and compensation contracts with EPS performance conditions. It also found that repurchases could benefit shareholders in some settings and did not find broad evidence of investment myopia in its sample.
The balanced reading is more useful: a target changes the pressure around a decision, but the economic quality of the final decision must still be tested separately.
EPS has two moving parts
Earnings per share can rise because the company earns more, because the average share count falls, or because both happen together.
EPS = Net income attributable to common shareholders / Weighted-average diluted shares
Take a simplified company earning $100 million with 100 million shares:
- Before a buyback: $100 million / 100 million shares = $1.00 EPS.
- After repurchasing 10 million shares: $100 million / 90 million shares = $1.11 EPS.
The business did not earn an extra dollar in this example. The same profit was divided among fewer shares, producing an 11.1% mechanical lift in EPS.
That does not make the buyback artificial or automatically harmful. Continuing shareholders own a larger share of the same company. But the source of the improvement matters. Investors should separate profit growth from denominator reduction and then ask what the company paid for the shares, how the repurchase was financed and which investments were not made.
Share-based employee awards and new issuance can move the denominator in the opposite direction. A company may therefore repurchase shares while barely reducing the final share count. The gross cash spent on buybacks is less informative than the change in shares that actually remain.
McKesson: the link is visible in the company’s own filing
McKesson offers an unusually clear case because its fiscal 2026 proxy statement connects the compensation targets with expected capital deployment.
Its fiscal 2026 compensation system had two layers:
- Annual management incentive: adjusted EPS 50%, adjusted operating profit 25% and free cash flow 25%.
- Completed fiscal 2024-2026 performance-share awards: three-year cumulative adjusted EPS 50%, average return on invested capital 25% and relative total shareholder return 25%.
Most importantly, McKesson said the adjusted EPS targets assumed capital deployment through share repurchases. That disclosure does not prove that the compensation plan caused every buyback. It does show that repurchases were incorporated into the performance architecture rather than treated as an unrelated event.
What the McKesson numbers show
McKesson’s fiscal 2026 10-K reported:
- $4.8 billion of common-stock repurchases.
- Net income attributable to McKesson of $4.762 billion, up from $3.295 billion.
- Weighted-average diluted shares of 124.1 million, down from 128.1 million.
- Diluted EPS of $38.38, up from $25.72.
Net income rose about 44.5%, while EPS rose about 49.2%. The lower average share count therefore added to already strong profit growth.
A simple bridge makes the denominator effect visible. If fiscal 2026 net income were divided by the prior year’s 128.1 million average shares, EPS would be about $37.17 rather than the reported $38.38. The roughly $1.21 difference is the estimated contribution from the lower denominator, subject to rounding and the limits of this simplified comparison.
This is a GAAP demonstration using reported net income and reported average shares. McKesson’s compensation calculation uses adjusted EPS, so it should not be confused with the company’s final incentive payout calculation.
Why the other metrics matter
If a company rewarded executives only for EPS, reducing the share count could become unusually attractive. McKesson’s scorecard adds operating profit, free cash flow, return on invested capital and relative shareholder return.
Those measures act as partial guardrails:
- Operating profit requires performance from the underlying business.
- Free cash flow tests whether accounting profit turns into cash.
- Return on invested capital asks whether investment and acquisitions earn enough.
- Relative shareholder return connects the result with what shareholders actually experienced compared with peers.
The guardrails are not perfect. Free cash flow can rise when investment is delayed, and relative shareholder return can be positive even when absolute returns are weak. But a balanced scorecard is harder to optimise through a single financial lever.
Counterexample: O’Reilly reaches buybacks through a different scorecard
O’Reilly Automotive shows why investors should not reverse the logic and assume that every large buyback must come from an EPS-linked annual bonus.
The weights in its 2025 annual cash incentive were:
- Comparable-store sales: 40%.
- Operating income: 40%.
- Return on invested capital: 20%.
EPS was not one of those annual cash-bonus metrics. Yet O’Reilly still repurchased $2.10 billion of shares during the year. It also spent $1.17 billion on capital expenditure and opened 207 net new stores.
The company describes buybacks as the use of excess capital after profitable reinvestment opportunities have been funded. Whether that claim remains true must be tested against store economics, future growth and the price paid for shares. Still, the case shows that identical corporate actions can emerge from different incentive systems.
Compensation changes the probability and attractiveness of a decision. It does not remove strategy, economics or management judgment.
A practical metric-to-behaviour map
When you find a compensation metric, identify what it rewards, how the result can be managed and which numbers can verify it:
| Metric | What it rewards | How the result can be managed | What to check |
|---|---|---|---|
| Revenue | Growth and market share | Discounts or expensive customer acquisition can make growth look stronger than its economics | Pricing, margins and customer acquisition costs |
| Operating profit | Cost discipline | Research, hiring or maintenance spending can be delayed | Research spending, hiring and maintenance investment |
| EPS | Profit growth and per-share results | Buybacks can reduce the denominator even when total profit is unchanged | Net income, average shares and the buyback price |
| Free cash flow | Cash conversion | Collections can be accelerated, payments delayed or investment postponed | Working capital, capital expenditure and one-off cash movements |
| Return on invested capital | Capital discipline | New investment can be delayed or some capital items excluded | The company’s definition, acquisitions and restructuring items |
| Relative shareholder return | Outperformance against peers | Falling less than a weak sector can still produce a high payout | Peer group, measurement dates and absolute return |
The strongest analysis links the pay metric to a real management choice and then tests that choice in the financial statements.
How to investigate a company in 15 minutes
1. Find the proxy statement
For a US-listed company, search the SEC’s EDGAR database for the latest DEF 14A. This proxy statement covers matters put to a shareholder vote, including the board and executive compensation. Investor.gov also provides a short guide to finding the filing. Search within it for “annual incentive,” “performance metrics,” “performance stock unit (PSU),” “threshold,” “target,” “maximum” and “adjustments.”
2. Rebuild the scorecard
Write down each metric, its weight, measurement period and payout range. Separate annual cash bonuses from long-term equity awards. A three-year award should not be analysed as if it were a one-year bonus.
3. Read the definitions
“Adjusted EPS” is not necessarily the same as reported EPS. Identify which acquisition costs, restructuring charges, currency movements or other items the committee can exclude.
4. Test the observed behaviour
Open the company’s 10-K annual report and cash-flow statement. For an EPS-linked plan, compare net income growth, EPS growth, repurchase spending, share-based compensation and the change in weighted-average shares.
5. Check what did not happen
Compare buybacks with capital expenditure, research spending, acquisitions, debt reduction and balance-sheet strength. The opportunity cost is part of the decision.
Seven questions before drawing a conclusion
- Is the metric tied to the annual bonus, the long-term award or both?
- Is the metric absolute, relative to peers, or both?
- Does the company disclose how repurchases affect the target?
- Did EPS grow faster than total net income?
- Did the share count fall enough to justify the cash spent?
- Were attractive investments still funded?
- Did management buy shares below a reasonable estimate of value, or merely support a headline target?
These questions do not produce an automatic verdict. They turn a compensation disclosure into a testable investment hypothesis.
The conclusion
Executive compensation is best read as a map of attention. It shows which outcomes the board has made financially important, which thresholds may create pressure and which time horizon management is being asked to optimise.
The right conclusion is not “EPS targets cause bad buybacks.” It is more precise: an EPS-heavy scorecard can make repurchases more attractive at the margin, especially when payout rises sharply near a target. Operating profit, free cash flow, return on invested capital and long-term shareholder returns can balance that pressure.
You can read a company’s story in its presentation, its behaviour in the cash-flow statement and its priorities in the compensation plan.
Sources and further reading
- US Securities and Exchange Commission, Investor.gov, Proxy Statements: How to Find: investor.gov
- McKesson, 2026 Proxy Statement: sec.gov
- McKesson, fiscal 2026 Form 10-K: sec.gov
- O’Reilly Automotive, 2026 Proxy Statement: sec.gov
- O’Reilly Automotive, 2025 Form 10-K: sec.gov
- Bennett, Bettis, Gopalan and Milbourn, Compensation Goals and Firm Performance, Journal of Financial Economics: doi.org
- Young and Yang, Stock Repurchases and Executive Compensation Contract Design, The Accounting Review: doi.org
Not investment advice; for research and educational purposes.