Index-linked gilts and the end of RPI

The full article from this blog post was published on The Economic History Review, and it is now available at this link: https://onlinelibrary.wiley.com/doi/full/10.1111/ehr.12875

By Michael J. Oliver and Janette Rutterford, Open University

On the 25^th November, 2020, the Chancellor of the Exchequer announced that index-linked gilts — government bonds whose payments increase with inflation — will no longer be linked to the Retail Price Index (RPI), but the newer Consumer Price Index including housing costs (CPIH). This switch will take place from 2030 and will save the government an estimated £2 billion a year. Holders of index-linked gilts will receive no compensation.

In our recent article (Oliver and Rutterford 2020), we described how and why index-linked gilts were introduced in the UK in 1981. The UK was the largest country to issue an index-linked financial instrument, despite opposition by the Bank of England. Mrs Thatcher was persuaded by the then Chancellor, Nigel Lawson, to use them at a time when investors were almost on ‘strike’: even though gilts offered nominal interest rates exceeding ten per cent, this was still less than the rate of inflation. Consequently, real yields on gilts were negative.

The first index-linked gilt, 2% IL Treasury 1996, was sold only to pension funds and enabled them to hedge their long-term pension payments with increases linked to earnings. The 2% coupon was chosen to match those increases in real terms. Nonetheless, there was considerable uncertainty about how much these gilts were actually worth. Consequently, the government allowed non-binding bids, with all successful bidders allotted bonds at the lowest accepted bid price. This price was par, £100, despite offers having ranged from £80 to £200. Evidently, the 2% coupon was about right for index-linked gilts.

There was debate on which index to use: a GDP deflator, an earnings index, or a cost of living index (RPI). It was feared that few would understand the GDP deflator. The earnings index, favoured by the Wilson Committee, appealed to pension funds, but it was measured once a year and, between 1972 and 1982, was 10% higher than the RPI, which would have vastly increased the cost of government borrowing. The RPI  was viewed as a good cost of living estimate. Its   other advantages were that it was already being applied to index-linked savings, and it was calculated monthly with only a two-month delay.

The first issue was a success, and it was quickly followed by further issues open to all types of investor. There are now 60 index-linked gilt issues, with a total value of approximately £400 billion, which represents a quarter of total government debt. Currently, all the interest payments, and the final repayment of these gilts, are linked to the RPI. But now there is an alternative index which has two versions: the CPI, which excludes housing costs, and the CPIH, which includes housing costs expressed as the real or implied rental value, plus council tax.

The RPI includes the costs associated with buying a property, including mortgage payments, but it does not reflect the fact that about 40 % of households are mortgage free.  There are also other problems in how the RPI is calculated from its constituent parts and a general agreement that CPI and CPIH are subject to fewer ‘technical’ problems.

The deciding factor for the government’s switch to CPI is that it is consistently lower than the RPI by around one percentage point. Switching to CPI would save the Exchequer £2 billion a year. Obviously, no inflation index will be perfect, and the cost of living varies substantially within the UK. But there are precedents for switching to the CPI. For example, in 2010, the coalition government started using the CPI for benefits, tax credits, and public sector pensions, and National Savings index-linked certificates were also linked to the CPI rather than the RPI, from 1 May 2019. Some have called this approach ‘inflation shopping’ as the government seems to prefer using the CPI when it is applied to costs rather than revenue. But previous switches to the CPI also generated a demand for gilts linked to the CPI: KPMG have estimated that £300 billion of private-sector defined-benefit pension liabilities, and £200 billion of public-sector pension liabilities, are linked to the CPI.

Three out of the 60 issues of index-linked gilts require the government to compensate investors for implementing any changes which will make them worse off. These three issues will have matured by 2030, the date for the switch to the CPI. Although it is rational for the government to reduce the costs of financing debt, it is worth remembering that the consensus long-term forecast for inflation in 1981, when these gilts were designed, was 9% rather than the 1% or 2% of today. What happens to government borrowing costs on one quarter of its total debt if inflation reverts to a higher rate.?

To contact the authors:

Michael J. Oliver, michael.oliver@open.ac.uk

Janette Rutterford, janette.rutterford@open.ac.uk