The RPI includes housing costs but that is not the reason why it tends to be 0.7% per annum greater than CPI. It is the method of calculation that drives the difference. For those with Maths A Levels, the RPI is arithmetic whereas the CPI is a geometric calculation. Statisticians say that geometric is the one to go for.
Index linked gilts (“linkers”) are indexed to the RPI, but over the last few months the debate has arisen as to whether the government should change to CPI or indeed whether the RPI should be calculated geometrically. A geometric RPI would result in it being lower by 0.7% per annum.
If that change was made overnight and in a way that was legal and binding, then a simple calculation would show that it would be devastating, particularly to the long dated end of the index- linked market. The 1/8ths linkers of 2068, due to mature in fifty years time, currently cost you £266 based on RPI averaging 3.3% over the life of the gilt. If that value driver suddenly dropped to 2.6% (3.3% - 0.7%) per annum then you would have 50 years x 0.7% less return each year which gives you a fall in value of 35%. Compounded, the effect would be closer to 42%.
In January of this year, Mark Carney told the House of Lords that he wanted it changed and that he would not want to see RPI driving the linker market in ten years’ time. And in July the debate re-surfaced in the House of Lords when the head of the Office for National Statistics, John Pullinger, said he now favoured a change that took effect “more quickly than ten years”. At the same time Liz Truss, Chief Secretary to the Treasury invited the UK Statistical Authority to review the method.
There have been other examples of reform. Ofwat is moving to CPI rather than RPI but then compensating the water companies by giving them greater yields elsewhere. Chris Grayling recently suggested that the rail fares get indexed to CPI to keep down prices; which would in turn mean lower wages for railway workers.
If that change was made overnight and in a way that was legal and binding, then a simple calculation would show that it would be devastating.
My own view is that the government would lose the trust and confidence of investors if it suddenly expropriated large lumps of value from investors to itself by modifying or replacing the current RPI. In the short term there would probably be a buyers’ strike prejudicing the government’s finances. In the long run, investors would demand a premium to compensate themselves for taking the risk of more unilateral and adverse changes to the way that our government borrowing works; and the government’s borrowing costs could potentially rise in the long term to a point where the effect was a greater cost than the original saving on the RPI calculation change.
In my opinion the way to solve the problem could be to issue new gilts linked to CPI rather than RPI and to possibly buy in the longer dated RPI issues over time.