Long-term projections require simplifying assumptions. By assuming that GDP grows in line with its historical trend, our central projections in effect imply 47 years of uninterrupted trend economic growth. But history tells us that the actual path of output will not be smooth.
We assume that most receipts and benefits are uprated with earnings, which negates any effects from cyclical swings in productivity. One key exception is the uprating of state pensions, where we apply the triple lock, uprating payments by the highest of CPI inflation, earnings growth or 2.5 per cent. The effect is asymmetric: during recessions, when earnings growth might fall below inflation or 2.5 per cent, the triple lock acts as a floor for the uprating of pensions; but it does not act as a ceiling when wages grow faster than either 2.5 per cent or inflation during a boom. We incorporate this effect into our central projections by assuming that the triple lock will, on average, push the annual pension uprating almost 0.4 percentage points above average earnings growth. Without this effect, our debt projection would be lower, ending the period at 61 per cent of GDP, in contrast to our central projection of 87 per cent of GDP.
Cyclical movements in economic activity would be expected to have broader effects on the public finances. We assume that spending on public services is linked to GDP per capita each year. In practice, spending plans are set out in advance and do not automatically adjust to temporary changes in GDP within each period. But receipts move more closely with the economic cycle, as do some parts of welfare spending. To explore the sensitivity of our projections, we illustrate alternative paths for debt as a share of GDP under two stylised economic cycles:
- a symmetric cycle, with the economy alternating above and below its trend level every three years, with peak differences of around 2 per cent of GDP. The average length of this business cycle, and the size of the shock, have been loosely informed by movements of actual GDP over the past relative to a statistically filtered trend; and
- an asymmetric cycle, whereby recessions are deeper and sharper than booms. The recessions reduce GDP by around 3 per cent relative to its potential, and the booms add around 1 per cent, but last almost twice as long. We also assume that the final cycle is bigger than the rest.
In mapping the fiscal implications, we apply our usual cyclical adjustment ready-reckoner, which assumes that a 1 per cent change in GDP will result in a 0.7 per cent of GDP change in borrowing after two years. The cost of the triple lock has not been modelled separately. The actual change would depend on many factors, including the specific nature of the economic shock and the composition of receipts and spending at the time.
This produces the illustrative projections in Chart B. The main message is that a symmetric cycle would lead to debt bouncing around our central projection (below it in the chart, as we assume that the first economic cycle begins with a boom), but that negatively asymmetric shocks would lead to permanently higher debt, as cyclical changes in the primary balance would have permanent effects raising the amount of debt interest.
Underpinning these projections is the assumption that spending continues to grow in line with potential output. But economic shocks may also have structural consequences that are not automatically fed through to changes in spending (and/or have longer-run effects on receipts), which could amplify the swings in borrowing and affect how debt accumulates over time.