Designing robust incentives

Incentives are motivational only if their outcomes are within the control of participants. This TTT considers the approaches available to help ensure incentives are not unduly sensitive to uncontrollable and unexpected events.

What is a robust incentive?

Incentives which pay out at one extreme to another from cycle-to-cycle tend to be perceived by participants as a lottery. Across the FTSE250 for example, recent history indicates that vesting of LTIPs is generally significantly more volatile than annual bonuses, with LTIPs being in either the nil or full pay-out zones around 36% of the time, compared to only around 15% for annual bonuses. This all-or-nothing nature can be the result of several different features in the plan design e.g. narrow performance ranges, uncontrollable factors.

Is it better to use multiple measures?

Possibly… more measures can help improve the probability of some pay-out, but only if the measures are not highly correlated. The most effective incentives are typically those which are ‘balanced’, i.e. they capture multiple aspects of performance (internal and external, financial and non-financial, absolute and relative, short- and long-term). Whilst annual bonuses tend to capture several measures, most LTIPs capture only two metrics, increasing the probability of an all-or-nothing outcome.

Can Relative TSR improve robustness?

Any relative performance assessment can help neutralise external factors, but only if such factors impact all companies in the benchmark consistently. Consequently, the use of relative TSR can be helpful in neutralising external factors only if the benchmark comprises companies with similar exposures to industry sector and geographies, and are at a similar stage of growth.

One downside of Relative TSR (at least for listed companies who are under the scrutiny of the major investors) is the very narrow performance range, with threshold required to be set at median and full-vesting most commonly set at upper quartile, i.e. reducing the performance range to only 25% of all possible outcomes. It is notable that US-based companies generally adopt a wider TSR range, with threshold often set at lower quartile and full vesting at 85th percentile.

Can we use relative financial performance?

Comparing financial outcomes against comparable companies can be an extremely effective method to dial-out the impact of external factors on incentive outcomes. As with relative TSR, it relies on the use of a relevant peer group but the number of companies can be lower than is required for relative TSR (which is inherently sensitive to market sentiment). A downside can be the comparability of the financial measure between companies, which may be subject to specific adjustments or assumptions at individual companies.

Could the measure be normalised?

Yes. ‘Normalising’ is simply using a financial ratio, which is inherently neutralised to top line movements. Examples include cash conversion, margin and ROCE.

Can we neutralise the measure to an external factor?

The most common ‘neutralisation’ is to measure financial performance in bonuses on a budgeted FX rate, to avoid the bonus outcome being driven by changes in uncontrollable global macro-economics. Lesser-used approaches include the measurement of performance vs indicators of industry volume e.g. global industrial production (for manufacturers), passenger kilometres (for an airline), etc.

Should the performance range be widened?

Incentive theory suggests the range set for an incentive measure should encompass all possible outcomes, to ensure there is always an incentive to improve performance from the status quo. In practice, extremely wide performance ranges can be problematic as the lower end may be too achievable whilst the top end perceived to be unachievable. However, more volatile companies should aim to set wider performance ranges than the norm to avoid all-or-nothing outcomes.

Should the performance period be shortened?

A short performance period can help ensure the performance range set for the period is realistic as shortening the period over which expected performance is projected helps to reduce the significance of external factors.  However, the reverse is true for a measure like relative TSR for which lengthening the performance period can improve the quality of the performance comparison as it reduces the impact of short-term fluctuations in market sentiment.

Should discretion be used?

In theory, discretion is a perfect solution for ensuring incentive outcomes reflect overall performance and are not unduly impacted by external events. However, in practice, discretion can be problematic if it sets unhelpful precedents or may be mistrusted by participants and/or shareholders if the basis for the discretion is not transparent.

Discretion is most effectively applied when referenced against a set of principles and checked against historical overrides to ensure consistency of decisions.

Is there an alternative incentive design which avoids the need to set targets?

If setting targets is unduly problematic, then a company could adopt a value sharing approach, which requires only to agree a sharing % and a threshold above which the employee share is funded. Several FTSE companies use this approach for the annual bonus, funding a bonus pool by, say, 5% of profit. Other companies have implemented long-term value sharing plans funded by, say, 10% of the incremental shareholder value created above a threshold return.

Restricted stock plans can also help address a perceived lack of robustness in LTIP design, ensuring employees are awarded equity-based incentives but without the complexity of setting performance targets.