Pay raises by formula vs by manager — how the choice shapes the gap

Once a year, in most companies, every employee gets a raise. The mechanism by which that raise is decided varies — and the variation produces measurable differences in how the pay gap evolves over time.

At one end, the raise is set by formula: a percentage based on band position, performance grade, and tenure, calculated centrally, applied uniformly across the population. At the other end, the raise is set by manager: each line manager allocates a discretionary pool to their direct reports based on their judgment of contribution and retention risk.

Both approaches are defensible under the directive when the methodology is documented and applied consistently. Both have predictable effects on the per-category gap. The choice matters less than knowing which effect the chosen approach produces.

Two mechanisms, two effects

What formula-based raises tend to do

A well-designed formula reduces the variance in raise outcomes for employees with similar profiles. Two employees in the same role, same level, same performance grade, same tenure receive the same raise — modulo any explicit factor in the formula like geographic adjustment.

This is mechanically gender-neutral when the inputs are gender-neutral. Performance grade is the key input that needs scrutiny: if the grading process embeds bias, the formula propagates it without dilution. The gap that emerges from a formula-based system is the gap embedded in the grading.

The advantage is auditability. Every raise can be traced to its inputs. An Article 10 assessment can examine the formula directly — and either confirm it's gender-neutral by design or identify exactly where the input bias enters. The defence is structural.

The trade-off is rigidity. The formula cannot react to context — a high-performer who is about to leave doesn't get a retention-adjusted raise unless the formula has a retention input, which most formulas don't. The system optimises for consistency over individual responsiveness.

What manager-discretion raises tend to do

Manager discretion allows responsiveness. The high-performer about to leave gets the retention adjustment. The employee whose role scope just expanded gets recognition outside the formal promotion process. The structural-flex case is handled by structural flex.

The variance, however, lives in many places. Three patterns surface across the directive-sensitive aggregate:

• Negotiation-driven differentiation. Employees who explicitly raise pay expectations during the review tend to receive higher raises within the same pool. The pattern is gender-asymmetric in incidence: men negotiate raise outcomes at higher rates than women in comparable cohorts.
• Visibility-driven differentiation. Employees with high in-the-room visibility — frequent stakeholder interaction, work that is observed by the manager directly — tend to be rated higher on the qualitative dimensions of contribution. Visibility correlates with role type and, in many cohorts, with gender mix.
• Pool-anchoring. Managers allocate within a fixed budget. The mental model is "I have X to distribute" — and the distribution tends to concentrate around the manager's perception of the team's relative contribution, which is itself influenced by the two factors above.

None of these individual decisions is biased. The cumulative result, observed across the per-category aggregate the directive examines, often shows a small but consistent skew. The defence in an Article 10 assessment is harder because the rationale is distributed across many individual decisions, each with its own narrative.

The practical middle

Most companies operate a hybrid. A formula-derived baseline raise (often 60-80% of the pool) plus a manager-discretion component (the remaining 20-40%). The hybrid balances the auditability of the formula with the responsiveness of discretion.

Three design choices distinguish defensible hybrids from un-defensible ones:

1. The formula component is published. Employees can see how the baseline is calculated. Managers cannot override the formula floor.
2. The discretionary component is constrained. Managers operate within a defined range (e.g., 0-3% additional above the formula baseline) and submit rationale for any allocation outside the band.
3. The discretionary component is audited. The distribution of discretionary raises is reviewed annually for patterns — by gender, by role type, by manager. The audit produces a record. Patterns that emerge can be addressed before they propagate to the next cycle.

The cumulative effect

A raise cycle is repeated annually. A 0.3% gender-asymmetric drift per year, compounded over a decade, produces a 3% structural gap on tenure-comparable employees. That 3% becomes visible in the per-category Article 9 metrics; it cannot be explained by tenure (the employees have similar tenure); the rationale has to be reconstructed from a decade of manager-discretion decisions.

The reconstruction is what an Article 10 assessment will demand. The methodology that produced the drift is what makes the reconstruction possible — or impossible.

One offer letter touches one employee. One raise cycle touches all of them.

Where the diagnostic starts

Detecting raise-cycle drift requires the per-category data to be calculable. The category structure, the documented methodology, and the historical compensation record have to exist before the drift can be measured. PayGapCheck™ produces the analytical report on a salary file — including the per-category quartile distribution that surfaces the cumulative effect of raise cycles year on year.