UK inflation – “bad decisions” from “misleading” figures

More bad news on inflation. Not only is the rate going up but experts warn that confidence in the inflation measure itself could be fading.

The January figures will be released tomorrow and are expected to see the CPI rise to 4.0% up from 3.7% intensifying the policy makers dilemma. That’s a big problem but there could be a larger one looming. The respected IAOS journal (for matters relating to official statistics) has just published a stinging attack on the accuracy and usefulness of the UK’s figures. There is a fear that the real rate of inflation, as felt by consumers, is above the published number and that the conditions are in place for a loss of public confidence in this most important of statistics.

The paper concludes: “The official price indices currently available for the UK are misleading the general public and are of doubtful relevance for policy purposes.” It explains that this is not an ethical issue in the sense that government statisticians or politicians are deliberately misleading the British public about inflation. Rather, it says that the statisticians have given “more weight to questionable economic theory than to the public’s need for a clear and transparent measure of price inflation.” The end result is that “politicians may make bad decisions because they are using bad statistics and the general public loses faith in the statisticians because of the gap that they see between their own daily shopping experience and the official measure of inflation.”

Concern about the inflation numbers has been simmering since the government announced that it was shifting from the RPI to the CPI to uprate benefits. Pensioner groups and others dependent on such upratings have been voicing their concern at what looks like a money-saving move – as the CPI generally shows a lower annual rate, its use saves the government money.

CPI and RPI, selected headline indices, UK from Timetric

The authors describe the detrimental pressures on government statistics as falling under four categories:

1.  Politics and institutions

The course of history is littered with examples of where those in power have interfered directly with all forms of prices. No single issue affects households more, say the authors, than a fall in their real incomes brought about by price rises while wage increases or job prospects are limited. “Not surprisingly”, the authors say, “those in authority have always had a vested interest in defusing this situation by showing slower price increases.” They say that these days, in a democracy: “this is accomplished simply by changing the ‘formula’ rather than more obviously fixing the numbers themselves, defining what the ‘preferred’ official inflation measure is and by trying to conceal the full extent of domestic price escalation.” Changes to the index which are apparently ‘technical’ end up suppressing the rate of inflation – that is what is happening in the UK with the widening gap between RPI and CPI.

2.  Theoretical and conceptual concerns

A major concern of the paper’s authors is “whether the much-vaunted recent harmonisation of index number methodology with economic theory, as it is conventionally recognised, is really appropriate or a cynical act of chauvinism.” Understanding this issue, requires the drawing of a clear distinction between a ‘pure’ price measure and a cost of living index (COLI); the first calculates the change in prices of a given bundle of goods and services, while the second reflects how users, consumers and other purchasers respond to such price changes. The authors say that “the distinction, unfortunately, is not always completely clear nor totally unambiguous.”

Even assuming it is agreed that these quality improvements should indeed be used to reduce any price increase that accompanies the launch of the new, improved model, the paper argues that one of the methods presently used – hedonic regression – is “questionable”. They also say that the implied so-called ‘upper level substitution effects’ are likely to be minimal. The classic textbook case of substituting margarine for butter, when the price of the latter rises, is “purely mythical” in their view.

These considerations suggest that a better measure of inflation – and certainly a more transparent one – would be an index that keeps its weights fixed for several years at a time. The expenditure weights of both the UK’s RPI and CPI are changed annually on the grounds that they need to capture substitution effects as consumers replace more expensive items by cheaper ones. But, the authors say: “Suppose that Granny Smith apples rise sharply in price prompting a switch to a less desirable apple but one whose price is not rising. An index with frequently changing weights may show no apple inflation at all implying that people are as well off as before. In reality of course they are worse off because they are denied access to the better quality apple. An index whose weights were fixed for say four or five years at a time would better reflect the loss of welfare from price inflation of the more desirable apple.” Annual updating of weights for price indices is now standard in most OECD countries, but they say: “This does not necessarily make it right.”

3.  Definitional ambiguities

The UK government’s preferred measure of inflation, the Consumer Price Index (CPI), is “little more than a political convenience“. It serves, they say, as an indicator necessary to facilitate comparability with other EU states and to ensure conformity with EU regulatory macro policy targets. “The CPI purports to measure the general increase in prices when, in fact, it leaves out major costs like housing that are of primary concern to households.” Housing costs form a large share, often up to a third or more of total income, of many people’s spending and, historically, have risen much faster than most other prices.

An important housing cost – mortgage interest – is included in the RPI, but the CPI excludes all housing costs on the “curious grounds that it was not possible to reach agreement among the EU member countries on how to measure them”. Even if it is difficult to find a comparable way of measuring housing costs in a group of countries, “there is no reason why housing costs should not be included in a CPI designed to measure inflation in the UK”. While housing costs are so important, “people can have little faith that an index excluding this item can be relevant to their inflation concerns”. The government’s belief in the relative stability of ‘core’ inflation, excluding fuel prices for example, is also worrying. From a policy perspective, this index says very little about the fundamental forces that determine inflation.

4.  Technical and operational considerations

Features of how a market actually works and how goods are put up for sale at various prices at different types of outlet have an important effect on inflation but whatever would be the best way to measure national price changes has to be qualified by what is practical, feasible and within the bounds of the available official budget. This always means making compromises.

For example, spending patterns and prices paid can be affected by a supplier offering across the board discounts through special credit arrangements or the provision of preferential payment options. Or perhaps a retailer might mark down items that show signs of physical deterioration, notably in the case of perishable products, or are deemed obsolescent – end of range – and thus are put on one side on a special ‘sale’ shelf or rail.

“How relevant are the UK’s official measures of price change?“,
Statistical Journal of the IAOS 27 (2011) 31-37

PS Even if you don’t believe the indexes, the individual components of the inflation basket are available at inflationstats.com.