Investment Planning; Government Debt.

Why Rishi Sunak Has Sleepless Nights - Government Debt

(AF4, FA7, LP2, RO2)

Government debt is not as long term or fixed rate as you might think.

It is often said that the £2tn+ debt that the Government has built up is not really a problem because Government borrowing is mostly long-term and at low fixed rates. As the graph above shows, the net interest cost of Government borrowing for 2020/21 is in cash terms slightly lower (£23.9bn v £24.4bn) than in 2004/05 when total debt was just a fifth of current levels (£2,198bn v £436bn). As a proportion of GDP, the net interest cost is even smaller – 1.1% now against 1.9% in 2004/05.

If something looks too good to be true, it probably is and in the case of Government debt interest, there are reasons to be circumspect:

1. Net interest and quantitative easing (QE) The graph shows net interest, which is a slippery concept in the world of QE. The way in which QE works is that when the Bank of England buys gilts, its pays for these by newly (electronically) created reserves which pass to the gilts’ former owners. The Bank pays base rate on these reserves. The interest on the gilts acquired by the Bank is passed to the Treasury, which then refunds the base rate cost of the new reserves to the Bank. While the interest rate on the gilts purchased is higher than base rate, the Treasury wins. In 2020/21 the Treasury’s net profit from QE will be £16bn, so without QE the gross debt interest would have been about £40bn. The Office for Budget Responsibility (OBR) gives a useful illustration of this in the March 2021 Economic and Fiscal Outlook (EFO), based on the latest QE programme having been completed (hence the £875bn total). It uses the Bank’s label for QE, Asset Purchase Facility (APF):

The buying spree since QE/APF resumed in March 2020 has largely matched the fresh gilt issuance from the Government, prompting some commentators to suggest that the whole process was little more than money printing. The total amount of QE was £750bn at the end of 2020 and by the time the latest round ends in 2021/22 the Bank will have bought £875bn of gilts. Viewed another way, that £875bn is linked to base rate.

How the gilt market will react when its biggest buyer disappears next year will be interesting (sic) to watch.

1. Index-linked stock Around a quarter of Government debt is index-linked bonds, hardly any of which have been purchased under QE. Service costs for these bonds are driven by the rate of RPI inflation (currently 1.4%), although from 2030 the link will effectively become CPIH. A 1% rise in inflation would add a little over £5bn a year to debt servicing costs. In its Budget forecast the OBR incorporates an assumption that the RPI rises to 2.5% in 2021, falls back to 2.0% in 2022 and then increases over the next three years to 3.0%. If those figures prove an underestimate, then debt servicing costs will be higher than projected.

Inflation and interest rates are closely interlinked, but they do not always move in tandem. The Bank will hold interest rates when inflation is rising, e.g. as in 2011, if it thinks hiking rates will make no difference.

2. Bond duration The UK Government has been successful in selling long-term bonds – the longest conventional bond matures in 2071 while the longest index-linked bond runs to 2068. The result is that the average term to maturity of Government debt is often quoted as around 15 years – impressively long by global standards. However, as the OBR highlighted in the March 2020 EFO, that average is a misleadingly attractive figure, distorted by “a tail of smaller very long-dated gilts”.

If the median is taken as the measure – the point at which half of the debt will have been repaid (and refinanced) – the figure drops to 11 years. The median drops to just four years if it is adjusted further to allow for QE. The reason is, as explained in 1 above, that QE effectively turns a long-term gilt liability for the Treasury into one linked to the short-term Bank of England base rate.

The OBR calculates that if non-gilt sources of Government finance, such as National Savings & Investment products, are also taken into account then the median maturity will be less than one year by March 2022.

This may all seem hypothetical, but an element of reality has already hit. The OBR prepared its EFO with a cut-off date for assumptions of 5 February. However, it noted in the EFO that if it had used market-implied interest rates for just three weeks later, on 26 February, the projected debt interest spending would have been £6.3bn higher by 2025/26.

Source: OBR EFO 3/3/21