Locked in?

It is well known that the triple lock on the State Pension – which states that it rises each year by the highest of earnings growth, inflation and 2.5 per cent – makes its value dependent not just on the general pace of growth in prices and wages, but also on their volatility. The next two years look set to provide a case study in why this means the triple lock is not a sensible mechanism for pensions uprating. This is because the temporary impacts of the lockdown are set to drive nominal earnings down this year (we estimate that the year-on-year change in the triple-lock-relevant May-July period will be a fall of 3.3 per cent), before rising sharply next year (by around 5 per cent). The triple lock will turn that into State Pension growth of 7.6 per cent over the next two years, compared to 2.5 per cent price growth and 1.5 per cent earnings growth.

This outsized growth in the State Penson would take place alongside much slower growth in working-age benefits. With the £20 per week uplift to the main rate of Universal Credit and tax credits due to expire in April 2021, we expect most working-age benefits to be only 5 per cent higher in April 2022 than they were in April 2019, compared to an 11.8 per cent increase in the State Pension over that period.

The policy answers to these challenges are to replace the triple lock, and pay more attention to the contrast with working-age benefits. A minimal and temporary fix would be to operate the triple lock over the coming two years as a whole, so that the State Pension would be likely to rise by 5 per cent (twice 2.5 per cent) over two years (given this should exceed two-year growth in earnings or prices). A better approach would be to replace the triple lock with a ‘smoothed earnings link’, which maintains the peg to earnings over the medium term, but allows short-term deviations to protect the State Pension’s value during periods of weak wage growth, or fast price growth. If the Government wants the State Pension to continue to rise faster than earnings, it should in addition set a clear objective for the level of the State Pension relative to pay – as it does for the National Living Wage – and supplement the ‘smoothed earnings link’ with a fixed additional annual rise until that target value is reached. Finally, policy makers should not go ahead with the big cut to working-age benefits planned for next year.

The triple lock is a messy tool for achieving a higher State Pension

The health impacts of coronavirus are particularly acute for pensioners but, to date, the economic effects of this crisis have not fallen heavily on this group. But they will have a major impact on pensions in the years ahead, thanks to their interaction with the triple lock on the State Pension. This note sets out what that impact might be, and the choices facing the Government.

The triple lock was introduced in 2010, and states that the basic State Pension and new State Pension will be uprated each April by the highest of:

• Earnings growth, based on ‘total pay’ (including bonuses), in the Office for National Statistics’ (ONS’s) Average Weekly Earnings (AWE) data, measured for May-July in the previous year;
• Inflation, based on annual growth in the CPI series in the previous September; and,
• 2.5 per cent.

This was part of a long-overdue move to restore the link between the level of the State Pension and average earnings. In normal times we’d expect nominal earnings growth to outpace both inflation and 2.5 per cent. But the last decade has not been normal when it comes to pay, with unprecedentedly weak earnings growth meaning that the inflation or 2.5 per cent components of the lock have applied in seven separate years. The result is that the triple lock was estimated to have cost around £6 billion per year by 2015-16, relative to earnings indexation.

There is an important debate to be had about the right level for the State Pension. Challenges around the adequacy of future pensions and the level of risk being borne by individual pension savers are good reasons to conclude that the State Pension in 2010 was too low as a fraction of earnings. However the juxtaposition of the triple lock and major cuts to working-age benefits has been problematic, especially given the much stronger growth in pensioner than working-age incomes over the past two decades. And the mechanism itself is a messy way of achieving the objective of a higher State Pension, providing an upwards ratchet that is dependent on not just the general pace of growth in prices and wages, but their volatility too.

The problems with the triple lock will be on full display as nominal earnings fall this year and then rise sharply next year

The seismic effects of the current crisis on our jobs market mean the messy nature of the triple lock is about to manifest itself once again, and in a very significant way. The reason for this is that earnings are set to fall this year (throughout this spotlight, unless otherwise stated, we deal with nominal changes in pay, the State Pension and benefits) and, because of that fall, rise sharply next year.

The Coronavirus Job Retention Scheme and pandemic-related hours reductions are the main drivers of this unusual path for earnings. Many workers currently furloughed have experienced pay cuts, reflecting the fact that government support covers 80 per cent of previous earnings, up to a cap of £2,500 a month. Although they are not working for it, employees’ furlough pay still counts in the Average Weekly Earnings series that underpins the earnings component of the triple lock. Temporary lockdown-related hours reductions will also push down on pay levels this year.

When combined with broader weakness in pay in the face of a major recession, the result is that the Bank of England’s illustrative coronavirus scenario in its latest Monetary Policy Report has total pay, as measured by the AWE, falling by 2 per cent in 2020.[i]

The consequence of this 2020 trough is likely to be stronger nominal earnings growth in 2021, as the furlough scheme is phased out and temporary hours reductions are unwound post-lockdown. The Bank of England, for example, expects 4 per cent nominal earnings growth in 2021.

To understand what these patterns mean for the State Pension, we need to consider how they translate into annual growth in earnings in the May-July period that matters for the triple lock. In particular, May-July 2020 covers the months in which the Job Retention Scheme is likely to be near its peak, and activity and travel still relatively constrained. This means we expect to see a more exaggerated fall-then-rise in earnings for triple lock purposes than implied by the Bank’s estimates for annual changes over the whole (calendar) year. Given we already have official earnings data up to April 2020, and an early estimate for May based on HM Revenue and Customs’ real-time information, we can use the Bank of England’s scenario to come to a reasonable estimate of these growth rates.

Our estimate, shown in Figure 1, suggests that earnings will fall by around 3.3 per cent annually in May-July 2020, before rising by around 5 per cent in May-July 2021. That would be an unprecedented fall in average weekly earnings, followed by the highest growth rate recorded (in those months) in 20 years. It is worth noting that a contributing factor to this volatility may be the precise choice of the earnings index that feeds into the triple lock – total pay including bonuses – as this sometimes registers larger swings than other series, and appears to be doing so at the moment.[ii]

Figure 1: In mid-2021, annual earnings growth is expected to be its highest in two decades

Inflation is currently very low – the CPI series used in the triple lock grew by 0.5 per cent in May.[iii] So, with price growth expected to stay low through to September,[iv] and a likely fall in average earnings in the data covering May-July, it seems almost certain that, under current policy, the State Pension will rise by 2.5 per cent in April 2021. In April 2022, it would then rise by the growth in earnings recorded in May-July 2021, or around 5 per cent. Compounded, that’s 7.6 per cent growth in the State Pension over two years, and 11.8 per cent over three years accounting for the 3.9 per cent (earnings-driven) increase in April 2020.

This representants a significant increase in State Pension levels, despite the underlying picture of the next few years being one of slow growth in earnings, and, indeed, prices. Figure 2 shows that this 7.6 per cent figure is more than three times the growth in prices (2.5 per cent) or earnings (1.5 per cent) over the two years.

While these may seem like small margins, they have non-negligible effects on public spending. The 7.6 per cent growth in the State Pension between April 2020 and April 2022 will mean spending £3 billion more in 2022 (and every following year) than if the State Pension had kept pace with earnings over this period, and £2.1 billion more than if it had kept pace with inflation (which is what would have happened under a ‘smoothed earnings link’ uprating mechanism – see below).[v]

Figure 2: The State Pension is set to rise three times as fast as underlying price and earnings series over the next two years

These changes to the State Pension stand in stark contrast to the expected path of working-age benefits

The outlook for working-age benefits is very different. The basic element of Universal Credit (UC) and its equivalent in tax credits have been increased by a very significant £20 per week this year. But that increase is temporary, with benefit levels set to return to their inflation-driven path in April 2021 (most working-age benefits are uprated each April by CPI growth in the previous September, like the inflation component of the triple lock). By 2022, our expectation, based on recent inflation forecasts, is that most working-age benefits will be only 5 per cent higher than three years previously, compared to the 11.8 per cent rise in the State Pension. Between April 2020 and April 2022, the currently planned unwinding of the coronavirus uplift to UC and other benefits would mean that the 7.6 per cent rise in the State Pension is matched by an 18.9 per cent fall in the value of the main rate of unemployment benefit for a single adult.

As Figure 3 shows, such an outcome would continue a pattern of diverging paths for the value of the State Pension and working-age benefits. While the State Pension has grown rapidly relative to prices, years of uprating only in line with price growth meant that the value of working-age benefits was largely unchanged over two decades before the 2016-2019 benefits freeze cut their real value. In 2014, the basic State Pension was 56 per cent higher than the main rate of unemployment benefit; in 2019 it was 77 per cent higher; and by 2024 it is expected to be 91 per cent higher.[vi]

Figure 3: The basic State Pension is expected to rise to almost double the value of the main rate of unemployment benefit in the mid-2020s

Figure 4 shows this same pattern relative to earnings. In 2022, once the temporary uplift to UC has been removed, the main rate of unemployment benefit will be at its lowest value ever relative to earnings. The basic State Pension, however, will be at its highest level relative to earnings since 1987. Whatever your view on a rising State Pension relative to earnings, the contrast with the continued erosion of the value of working-age benefits is indefensible.

Figure 4: The State Pension has risen relative to average earnings in recent years

Policy makers should focus on fixing the triple lock’s problems in the years ahead

The policy answers to these developments are to replace the triple lock, while paying more attention to the contrast with working-age benefits.

A short-term fix would be to enforce the rules of the triple lock not in each year but over the coming two years as a whole, to avoid the artificial volatility in measured average wages. For example, the lock could operate as planned in the coming year, with a likely 2.5 per cent increase in April 2021. The 2022 value of the State Pension would then be determined by the highest of two-year growth in earnings, two-year growth in the CPI growth, with a backstop of a 5 per cent increase on the 2020 value – with this backstop being very likely to be the element that takes effect, given the weak forecasts for earnings and price growth shown in Figure 2.[vii] Such an approach would narrow – but not remove – the expected gap between increases in the State Pension, earnings, prices and working-age benefits over the next two years.

A far better approach would be to move permanently to a more sustainable ‘smoothed earnings link’ uprating mechanism for the State Pension, as previously proposed by the Work and Pensions Select Committee and the Intergenerational Commission. This maintains a peg to earnings over the medium term, but allows short-term deviations to protect the value of the State Pension during periods of weak growth in wages, or fast growth in prices.

Importantly, this approach would be preferable to the triple lock, whatever your view on whether the State Pension has or has not yet reached a sufficiently high level relative to earnings. If policy makers wanted to continue to increase the value of the State Pension relative to earnings – as happened over the past decade – then this could be achieved by adding an additional annual uplift. Such a ‘smoothed-earnings-link-plus’ approach could be followed until the target value is reached (as also proposed by the Intergenerational Commission). So, while some defend the triple lock on the basis that its ratchet mainly benefits future pensioners, their objectives can be better realised in a far less messy way.

Finally, reflecting our second critique of the triple lock, it should be clear that we cannot continue with such big discrepancies in how we uprate the State Pension and working-age benefits. This is especially true during a recession whose financial impacts are very heavily focused on the working-age population. The crisis in our jobs market will be far from over by April 2021, so we should be looking to maintain the uplift to the basic element in Universal Credit and tax credits. With child poverty having risen, there is a strong case for restoring some of the value of child-related benefits, too.

Good policy making requires separating our objectives from the tools we use to achieve them. Whatever your objective for the State Pension, the triple lock is a poor tool for achieving it, as the next two years looks set to make very clear.

[i] We base our earnings estimates on the Bank of England’s illustrative scenarios, not the Office for Budget Responsibility’s (OBR’s) coronavirus reference scenario. This is because the latter is based on a National Accounts measure of wages and salaries, and, more importantly, assumed that Job Retention Scheme payments would be recorded as transfer payments from government to households and therefore not show up in earnings series (the ONS has since clarified that they will be treated as a subsidy to companies and therefore enter earnings data). This means that the OBR’s reference scenario presents a much steeper earnings fall in 2020, and much stronger rebound in 2021, than is likely to be recorded in the outturn data used to calculate the triple lock.

[ii] The AWE series is based on the ONS’s Monthly Wages and Salaries Survey, which samples businesses registered for VAT and/or PAYE (so misses the very smallest firms), asking them about their pay bills and employee headcounts (with the former divided by the latter to create average earnings data). The triple lock is based on total pay (including bonuses) in this data, which can be quite volatile given the irregular nature of bonus payments. Indeed, it appears that ‘regular pay’ (excluding bonus payments) has dipped by less than total pay recently – the single-month nominal annual change in AWE regular pay was 0 per cent in April this year, compared to a 0.9 per cent fall in the total pay series. Annual Survey of Hours and Earnings data covering a week in April each year and published in the autumn provides a measure of median earnings, which can similarly be less vulnerable to volatility than the mean earnings measures that the AWE data is, by design, limited to. The implication is that the choice of earnings series used in the triple lock, as well as the time period it relates to, may well be amplifying the fall-then-rise through this year and next.

[iii] It’s worth noting there is a parallel debate to that on earnings data, above, on the correct inflation measure to use in benefits uprating, which we do not discuss here.

[iv] In their coronavirus scenarios, the OBR and the Bank of England differed in their inflation expectations compared to those before coronavirus (the Bank expected much lower inflation in 2020 and 2021, whereas the OBR scenario entailed higher inflation in 2021 than in its March 2020 Economic and Fiscal Outlook). More recent forecasts have shown inflation expectations trending downwards from those of both the OBR and other forecasters in April. In the costings and benefit projections presented here, we use the Q4 CPI forecasts for 2020 and 2021 (we assume these also apply to September in each year) in HM Treasury’s June 2020 average of independent forecasts: 0.8 per cent CPI growth in 2020 and 1.7 per cent in 2021. It is worth noting that no forecasts have inflation rising above 2.5 per cent over the next couple of years, so this choice is not material to our expectations for the triple lock.

[v] Estimates are based on RF analysis using the IPPR tax-benefit model, which accounts for the offsetting effects of higher State Pension spending on means-tested benefit spending and Income Tax revenues.

[vi] These figures refer to the old basic State Pension. The new State Pension is more generous, at £168.60 per week in 2019 and £175.20 this year. In 2019, the new State Pension was 131 per cent higher than the main rate of unemployment benefit.

[vii] Similar temporary fixes have been suggested by others.