The CEO Pay Machine
Expose how peer benchmarking mechanically escalates executive pay
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 6
- Confidence
- 98%
The CEO Pay Machine explains executive-pay inflation as the cumulative result of peer selection, percentile benchmarking, performance-linked awards, and weak board resistance. Consultants begin with a comparison group that often contains highly paid executives. A company then places its CEO near the upper quartile because directors consider their organization above average, even when no corresponding performance test exists. That elevated target becomes another observation in future compensation studies, pushing the entire comparison set upward. Stock options and annual incentives add further upside while insulating executives from equivalent downside. The framework therefore examines the complete mechanism rather than judging each component separately: rules that appear reasonable in isolation can collectively guarantee escalating compensation regardless of results.
Origin
Stephen Clifford developed the model after serving on compensation committees, questioning their standard consultant-designed systems, and researching why the resulting incentives appeared irrational. Extracted from Masters in Business.
Core principles
- 01Internal value matters more than comparisons with unrelated CEOs
- 02Selected peer groups are not neutral market evidence
- 03Pay levels should follow demonstrated performance, not corporate pride
- 04Individually plausible rules can combine into a destructive system
- 05Compensation design shapes executive attention and company investment
How to run it
- 1
Map the compensation system
List salary, bonuses, options, restricted stock, benefits, performance measures, and consultant recommendations. Record who proposed and approved each component.
Pro tip Review several years of proxy statements to reveal recurring increases and changing rationales.
Watch out Do not evaluate one award in isolation from the rest of the package.
- 2
Audit the peer group
Examine how comparison companies were chosen and whether they match the company in size, industry, complexity, and executive labor needs. Flag peers included mainly because they pay more.
Pro tip Compare the selected group with an independently constructed group.
Watch out A consultant-generated data set is not automatically objective.
- 3
Test percentile claims
Match the CEO's targeted pay percentile against company performance on relevant measures. Require evidence before treating the company or executive as top quartile.
Pro tip Use several operational and financial measures rather than stock price alone.
Watch out Corporate pride is not a performance measure.
- 4
Separate skill from luck
Estimate how much of the result came from the overall market, sector conditions, or inherited assets. Attribute rewards only to outcomes the executive could materially influence.
Pro tip Compare performance with close competitors operating under the same conditions.
Watch out Absolute stock gains can reward an underperforming CEO during a rising market.
- 5
Trace incentive effects
Ask what behavior each measure will encourage when a large payment depends on hitting it. Look for pressure to manipulate timing, cut investment, repurchase shares, or neglect unmeasured priorities.
Pro tip Model the easiest way to hit each target without improving the underlying business.
Watch out Powerful incentives can narrow attention more than they improve effort.
- 6
Break the escalation loop
Set compensation from responsibility, internal equity, long-term value creation, and verified performance rather than automatic peer escalation. Document why every increase is necessary.
Pro tip Require directors to explain what additional value shareholders receive for additional pay.
Watch out Continuing to benchmark above the median recreates the same upward spiral.
In the wild
A board chooses highly paid comparison companies and targets its CEO at the seventy-fifth percentile because it considers the business superior. The committee never tests whether revenue growth, returns, or market-share gains also rank in that percentile. The higher package then enters the data used by other boards, helping raise the next round of benchmarks.
→ Pay rises through the benchmarking mechanism without evidence of superior executive performance.
An executive receives options when the stock is at twenty dollars. Seven years later, a broad market and sector rally lifts it to fifty, allowing a large gain even though the company trails its peers. The reported grant-date expense captures only a fraction of the eventual shareholder cost.
→ The executive receives a windfall largely attributable to market conditions rather than company-specific value creation.
Common mistakes
Treating consultant data as objective
Peer groups reflect selection choices and can be deliberately tilted toward highly paid executives. Independently test their comparability before using them.
Paying for rank without requiring rank
Boards often target upper-quartile compensation without requiring upper-quartile operating performance. Link any premium to demonstrated relative results.
Reviewing each rule separately
A component may possess surface logic while still contributing to a system that guarantees escalation. Evaluate the package as an interacting mechanism.
Is it for you?
Best for
It is best for directors, investors, governance teams, and policymakers reviewing executive-pay proposals.
Not ideal for
It is not ideal for setting ordinary employee salaries in active labor markets with readily transferable skills.
From the transcript
“Then the company quote benchmarks their CEO.”
“They guarantee with mathematical certainty that the CEOs pay will escalate and I regardless of performance, regardless of performance.”
“One, the individual part of the machine had to have a certain surface logic, and to it had to increase the CEOs pay.”
From the episode
Steven Clifford Says You Don't Need a Compensation Consultant
Steven Clifford