Incentive performance ranges

The credibility of an incentive plan is fundamentally reliant on the credibility of the performance ranges set for its performance measures. This Time to Think considers the theory and market practice around how performance ranges should be set, and highlights some potential pitfalls.

What do you mean by performance range?

The performance levels which govern the payout of an incentive. For example, the payout of a Bonus plan would typically be based on three performance levels:

• Threshold (the minimum acceptable performance to warrant any payout);
• Target (the expected outcome); and
• Stretch (the level of performance required for maximum payout).

How wide should the range be?

In theory, the performance range should capture all possible performance outcomes, to naturally reinforce outperformance of the status quo, as there is always an incentive to improve performance. For example, a range around Relative TSR should start at 1st percentile (i.e. bottom), and max out at 100th percentile (i.e. top). However, once the range is narrowed, and thresholds and caps are introduced, the risk that participants ‘game’ the incentive increases.

In practice, performance ranges are narrower than this ideal, to account for many additional factors, including investor expectations, and last-year’s outturn.

What inputs should be considered?

Most commonly, performance ranges are set around a company’s Budget for the period e.g. 95% to 110% of Budget PBT.

However, using only Budget is one dimensional, and other perspectives on performance can be helpful in understanding the possible range of acceptable performance outcomes. Other perspectives include expectations for sector peers and long-run extrapolations of historical performance.

Where should Budget sit in the range?

It depends on the toughness of the Budget. Most commonly, performing in line with Budget delivers 30-50% of max payout. However, if Budget is considered tough, it would be more appropriate for the performance range to be skewed to rewarding underperformance of the Budget; conversely, if Budget is considered more achievable, then the performance range should be skewed to reinforce significant outperformance of Budget.

Should the range differ by measure?

Yes. The performance range should reflect the potential volatility in the measure, which is fundamentally dependent on the ‘information’ in the measure. Simple, one dimensional measures such as ‘volume’ would typically have a narrow range, whilst the ranges for ‘fully-loaded’ measures such as cashflow (which is sensitive to volume, pricing, costs, investments) should be wider.

Should Threshold always be higher than last year’s outturn?

Not necessarily. In Ellason’s experience, most companies would default to setting Threshold at a level which exceeds the prior year outturn, to reinforce growth. However, there may be factors which caused an exceptional, and unsustainable, level of performance in the prior year which are not expected to be repeated; and which could justify setting Threshold lower than the prior year outturn.

Is a narrow range, such as a zero/one hurdle, ever appropriate?

Yes. A hurdle (i.e. which leads to a binary – all-or-nothing – outcome) is suitable when the objective being targeted is of a critical nature, e.g. regulatory requirements around the capital base for a bank, maintenance of a stated dividend policy, etc.

Hurdles also make sense when an incentive is based on a broad measure scorecard (common in some sectors such as extractive industries or financial services). In these cases, using a hurdle helps simplify the plan, with sensitivity to all-or-nothing outcomes for each measure mitigated overall through using a higher number of measures.

Should we set the same range width for Group and business units?

This depends on relative size, which will impact the volatility of performance. In theory, Group performance should always be less volatile than an individual business unit’s, on the basis that Group performance is the collective aggregation of multiple BUs. Consequently, the performance range set for Group should be narrower than that set for the BU (assuming, of course, the same measure, e.g. PBT, is used).

There are exceptions: if a Group is subject to additional exposures to certain factors (e.g. regulation, taxes, etc.) not evident at the BU level, then the impact of these on potential volatility at the Group level should be considered when setting the range.

How stretching are ranges in practice?

Evidence across the FTSE350 suggests over the last 5 years that bonuses typically pay out at some level in every 9 cycles out of 10, and would result in nil payout in only 1 cycle. LTIP ranges tend to be tougher, with some vesting in every 8 cycles out of 10, and nil vesting in 2 cycles (see chart below). The typical of a bonus is c.70%, whilst an the typical LTI pay-out is c.60%.

This distinction between bonuses and LTIPs is not surprising. Bonus payouts are more frequent due to: (i) more measures, (ii) wider ranges (as only based on financial targets and not TSR), and (iii) greater prevalence of non-financial measures, which provide for greater flexibility to inherently widen the range to reflect