h1

Fingering the index. A proposed technical change that is hugely important

September 8th, 2019

There was some important news this week. You probably missed it in all the nonsense about Brexit. It is quite boring but that doesn’t excuse you from the responsibility of knowing about it. This news will almost certainly affect you directly financially.  

For a decade, statisticians have been undergoing an extended collective nervous breakdown about inflation and how to calculate it. Britain had been very early to look at this problem and had come up with a solution which it badged under the name of the Retail Prices Index (RPI to most of us). It has its origins before the First World War.  

Candidly, it is now showing its age. There are substantial problems with the way in which it is compiled, but perhaps its most fundamental flaw is mathematical. You and I think of producing an average by totting up a total and then dividing it by the number of items you added up. This is called the arithmetic mean and this is how RPI is produced. Unfortunately, when you are calculating an average rate of inflation, this method will, all other things being equal, overstate the underlying rate of increase, because people will tend to buy cheaper goods instead of paying for the more expensive item.

In these circumstances, the generally accepted better method is to use what is known as the geometric mean. You can’t do this in your head. You multiply all the different numbers together and then take the nth root of the product, where n is the number of items.

More modern methods of calculating inflation use the geometric mean. This underpins the Consumer Prices Index, which is the EU’s standard way of calculating inflation. As a result, CPI is thought to be a more accurate measure of inflation. More accurate, however, is not perfect. There are problems with the compilation of CPI too (less serious ones, admittedly). The difference is on average something of the order of 1% a year.

This would have been a matter of technical interest only, were it not for a grenade thrown by the coalition in 2010. It elected to move the increase of all state benefits and pensions to CPI, citing its superior accuracy. The massive long term saving in costs for a government that was committed to austerity no doubt had nothing to do with it. 

Curiously, the government continued to increase rail fares annually by RPI and student loans continue to bear interest by reference to RPI. You will note that the government benefits from paying its obligations by reference to CPI but collecting its due by reference to RPI – a neat if unscrupulous form of arbitrage. The government has continued to issue gilts by reference to RPI.

And there the government left matters, airily claiming that it was not for the government to disturb private arrangements. This was of no help at all to private pension schemes, many of which found themselves lumbered with a measure of inflation that was officially spurned but which their rules, usually drafted generations earlier, required them to use (we have seen a string of cases in the High Court about how to interpret pension scheme rules on this point – as a pension lawyer, I can only applaud government actions that result in increased need for pension law advice, but my clients no doubt feel less enthused about this development). Many long term private contracts have pricing structures that are denominated by reference to RPI. The government had opened a can of worms and had not provided a terrarium.

Ever since then, the geeky great and good have been trying to fill the gap the government created. New and different measures of inflation have been conjured up and later discarded. One consumer committee advised officialdom that RPI should be redefined to reflect the modern era. Not only was this advice spurned, the committee was promptly disbanded following its unhelpful suggestion.

To be fair, the cautious statisticians had a point. If RPI is redefined in a way that reduces its rate, the returns on gilts will be correspondingly reduced.  Buyers of gilts might well argue that such jiggery-pokery constituted a partial default on Britain’s debts – social death for a G7 country and something that might affect credit ratings and future borrowing costs.

So the story of the last decade has been one of slow decline for RPI, as official bodies progressively downgraded it and yet no one had the heart to euthanase it.

This week, the Statistics Authority pulled off a coup de théâtre. It disclosed that it had advised in March that RPI should be discontinued, and in response the Chancellor has asked it to consult with a view to redefining RPI so that it is, from a date between 2025 and 2030, to be calculated on the same basis as CPIH (a modified form of CPI that unlike CPI allows for housing costs). 

Six years after the consumer committee had been disbanded for the temerity of suggesting that RPI should be redefined, the Statistics Authority have basically taken up their suggestion. The difference between RPI and CPIH is roughly 0.7% a year on average.

The intention is that by phasing in the change over such a long time period, the accusation that Britain is defaulting on its debts can be avoided. It will probably work (and with gilt yields so low right now, the government can probably afford the risk of a bit of a fuss even if that’s wrong). If the consultation goes as suggested, that will be good news in the long term for students with debts and regular rail passengers.

What of pension schemes? Well, those pension schemes whose benefits are still calculated in part or in full by reference to RPI will see a saving in their costs in the long term. On the other hand, they will almost certainly have substantial gilts holdings that will produce in the long term lower returns for the same reason. They may lose on the swings what they gain on the roundabouts, or perhaps even more.  

And anyone who has a pension that is index-linked by reference to RPI is facing a hidden loss of value.  The change is right in principle, but a lot of pensioners face losing out in the long term. Because it’s boring and technical, most of them almost certainly won’t notice at all.  But the sums could be very substantial indeed: 0.7% a year adds up to a lot over time.

Anyway, I’m sure you’d rather talk about Brexit. Just check your wallet before you do.  

Alastair Meeks