Phase, Don't Flip
Big workplace changes need a published glide path, not a short-notice cliff edge
Bell repeatedly defends phasing as a policy virtue, not a sign of weakness. Big changes — contribution rates, retirement ages, mandation — affect how people get paid, how they plan, and how markets price risk. Flipping a switch on these without a glide path destroys trust even when the destination is correct.
The framework has three components: a clearly named destination, a published timeline, and protection of the cohorts who can't adjust. Australia took longer to reach today's contribution levels but started earlier; the UK went faster but later — both used phasing. Where phasing fails is when it's confused with delay, or when it's reversed at short notice, as Bell argues happened with George Osborne accelerating women's state pension age rises.
The operational test: can a 25-year-old today look at the published timetable and see exactly what age they retire at? If yes, the phase is honest. If no, it's procrastination dressed up as caution.
- Workers, employers and markets all need lead time to absorb structural change.
- Short-notice change to long-horizon decisions is a trust violation, not just an inconvenience.
- A published destination is the difference between phasing and stalling.
- Phasing is only honest if the cohorts closest to the cliff are protected hardest.
- Speed of the rollout matters less than the credibility of the timetable.
- Name the destination publiclyState exactly what the end state is — the retirement age, the contribution rate, the asset allocation. No ambiguity about where you're going.
- Publish the glide path with datesLay out the year-by-year sequence so individuals can plan against it. The UK state pension age timetable (66 → 67 by 2026-28, 68 in early 2040s) is the model.Pro tipTie milestones to legislation so future governments can't quietly pull them forward.WarningMultiple sequential dates without legislative anchoring are a wishlist, not a plan.
- Protect cohorts inside the danger zoneAnyone within 5-10 years of the change has the least time to adjust. Either exclude them, grandfather them, or give them disproportionate notice. The closer to the cliff, the more protection.WarningFailing this step is what blew up the 2010s women's state pension age changes.
- Build a mandatory review cadenceSchedule a legally required review at fixed intervals (UK does six-yearly state pension age reviews). This forces honest reassessment rather than ad-hoc tinkering.
- Resist the short-term saving temptationAccelerating phasing usually saves money in the short run and destroys trust forever. Bell explicitly contrasts Osborne's boast about 'biggest saving he ever made' with the lived consequence for affected women.
- Communicate the timetable on every relevant surfaceState pension age, contribution rate, mandation deadlines should appear on payslips, government statements, pension portals. The credibility of the phase comes from its visibility.
Australia started earlier than the UK and took longer to reach today's higher contribution levels. Bell cites this as proof phasing works internationally — bigger structural change, longer phase, durable outcome.
Coalition government pulled forward planned increases at relatively short notice. Bell describes it as Osborne 'boast[ing] about it being the biggest saving he ever made' but at huge cost to affected women's life plans.
The 10% / 5% pension asset target is dated to 2030 with a backstop power that doesn't activate before then. The five-year runway is the phasing — and the backstop sunset is the protection against re-phasing.
Bell's view comes partly from watching the political fallout of Osborne's accelerated state pension age changes for women — saving short-term money but breaking trust with people who had life plans built on the previous timetable. He frames the lesson as: 'this is a load of people's lives that you just changed at very short notice, so we're not doing that.' Phasing is now baked into the revived Pensions Commission's terms of reference.