Waiting is not performance: rethinking the deferred bonus

An executive has a maximum annual bonus of £1 million. Performance during 2026 produces an award of £600,000. £300,000 is paid in cash and £300,000 is converted into shares and deferred for three years.

That introduces a longer-term economic element. The executive remains exposed to what happens to the share price between 2026 and 2029. If shareholders prosper, the deferred shares may increase in value. If value is destroyed, they may fall.

But the performance assessment is still essentially backward-looking. The £600,000 bonus has been determined by reference to 2026 performance. The number of deferred shares is not ordinarily retested against performance during 2027, 2028 or 2029.

Being blunt, deferral extends economic exposure. It does not, by itself, extend performance measurement. And that distinction matters because deferral is often expected to do several different jobs.

There are at least three good reasons to defer bonus.

It can extend an executive's exposure to shareholder returns and help build meaningful share ownership. It can create retention value where the deferred award remains conditional on service. And it can keep remuneration available for adjustment if something later emerges that engages the malus provisions.

Those objectives overlap and there is nothing wrong with one instrument doing several jobs.

But there is no obvious reason why each job requires the same amount of remuneration to remain outstanding for the same period.

That is where the conventional architecture becomes more interesting.

Consider alignment. For a newly appointed CEO with little share exposure, converting part of the annual bonus into shares may materially change their economic exposure. For an executive already holding several times salary in shares, the marginal alignment produced by deferring another £1 may be much smaller.

Market practice is beginning to recognise that distinction. The Investment Association's current guidance recognises that a lower proportion of bonus may be deferred where an executive has built a significant long-term shareholding. Rolls-Royce takes a graduated approach, reducing deferral as share ownership increases. Unilever has gone further: under its new policy, bonus deferral ceases once the relevant shareholding requirement has been met.

That makes intuitive sense if deferral is principally an alignment mechanism.

But alignment is not its only function.

Deferral does not simply postpone payment. It can keep remuneration within reach.

If something later emerges that engages the malus provisions, an outstanding award may be reduced before it is delivered. Once value has been delivered, the company may instead have to recover it through clawback.

Those are not equivalent positions. Recovering cash or shares already received can raise practical questions around contractual power, evidence, tax, set-off and enforcement, particularly after an executive has left or moved jurisdiction. Reducing an award that remains outstanding may be considerably easier.

That makes Unilever's approach particularly interesting. Its move away from bonus deferral once the shareholding requirement is met was accompanied by a review and strengthening of its malus and clawback arrangements. For more on malus and clawback click here.

There is therefore a broader design point here. Reducing the role of one part of the remuneration architecture can increase reliance on another.

Retention is different again.

Looking at a single deferred bonus award can understate its effect. An executive may have several overlapping annual cohorts outstanding at any one time. Collectively, they can create a substantial and continuously replenished amount of value that may be forfeited on departure.

The right question is therefore not whether deferral “supports retention”, it is whether the whole outstanding stack is large enough and sufficiently conditional, to change the economics of leaving.

Leaver treatment matters. So does the willingness of a prospective employer to compensate the executive for forfeited awards.

There is also an important distinction between performance having been determined and the executive having an unconditional entitlement to the resulting value. Where deferred remuneration remains conditional on continued service, that conditionality is itself part of the retention mechanism.

Simply making an executive wait does not necessarily retain them.

For many companies, the most compelling justification for deferral may be simpler: some things take time to emerge.

A bonus is normally determined shortly after the end of a financial year, relevant information may surface later. Accounts may be restated. Misconduct may be discovered. A risk-management failure may become apparent. Data underlying the original award may turn out to have been wrong.

None of those events automatically reduces an award. The rules still need an appropriate trigger and an effective power. But if the conditions for adjustment are satisfied, keeping remuneration outstanding can make intervention materially easier.

The recent banking reforms are instructive on this point, which you can read more about here and here.

In financial services, certain senior material risk takers were previously subject to minimum deferral periods of seven years; the PRA and FCA have now reduced the relevant minimum to four. The PRA's internal analysis found that around 70% of adverse outcomes from risk-taking were discovered within four years, while fewer than half emerged within three.

That does not mean listed companies should adopt four-year bonus deferral. Banking risks are different. If anything, the evidence could be used to argue that three years is too short.

The regulator asked how long the consequences against which remuneration was intended to remain exposed actually took to become visible, and calibrated the period accordingly. That is surely a better starting point than market convention.

If three years is intended principally as a risk-adjustment period, the interesting question for a remco is therefore: what evidence is there that three years corresponds to this company's risk horizon?

Longer is not automatically better. The regulators also recognised that deferral has costs. Very long periods can reduce the perceived value of variable remuneration, complicate recruitment and retention and increase pressure on fixed pay.

The answer cannot simply be more deferral.

A remuneration committee reviewing bonus deferral can usefully separate four questions.

Performance: over what period are we actually measuring performance?

Exposure: how much continuing exposure to shareholder returns do we want after the bonus has been determined?

Retention: how much genuinely forfeitable value needs to remain outstanding to affect the economics of leaving?

Risk adjustment: for how long does remuneration need to remain within reach because material adverse information may emerge?

There is no particular reason why the answer to all four questions should be three years and nor is there any reason why the optimal amount should always be 50%.

For alignment, the answer should reflect the executive's existing equity exposure. For retention, it should take account of the entire stack of outstanding awards and the leaver provisions. For risk adjustment, it should reflect how long the relevant risks take to crystallise and the strength of malus and clawback.

And if the objective is genuinely to assess performance over a longer period, deferring a bonus determined by reference to one year's performance does not achieve it. That requires a different design choice.

None of this means that 50% for three years is wrong.

Market conventions have value. Investors understand familiar structures. Standardisation reduces complexity and conventional answers are sometimes conventional because they work reasonably well.

But prevalence cannot establish the right answer for a particular company. The amount and duration of deferral should follow from what the company is trying to achieve: the executive's existing economic exposure, the value genuinely required for retention, the nature and timing of relevant risks, the strength of the adjustment architecture and the economic cost of delaying remuneration.

That analysis may still lead a remuneration committee to 50% for three years.

The important thing is that it gets there at the end of the analysis, not the beginning. And where the real objective is long-term performance measurement, that is a separate design choice.

A three-year wait can be useful, but surely its value lies in what happens during the wait, not in the waiting itself.