After an inquiry into its 57-year-old Retail Price Index upended the 290 billion-pound ($440 billion) index-linked bond market last year, the ONS won’t refurbish the gauge, and instead will publish a new measure called RPIJ from tomorrow. Old RPI will remain the benchmark for inflation-tied prices including how much tax gamblers pay and what telephone calls cost, even after a panel said it has “unsuitable” properties, and the ONS said it “does not meet international best standards.”
“They’ve created a great deal of confusion and anxiety and muddle and they’ve made it worse because it’s going to be a long time before they come back to look at it sensibly,” said Kate Barker, a former Bank of England policy maker and an adviser at Credit Suisse Group AG in London. “I felt really blindsided that the national statistician could just reject that advice.”
U.K. National Statistician Jil Matheson headed a review by the Consumer Prices Advisory Panel, which she chairs. It focused on clothing prices, which account for 4.5 percent of the RPI, after changes to collection methods skewed the data. One of the conclusions was that the current Carli mathematical method, which uses arithmetic averages and is used in 27 percent of the inflation measure, is inferior to the Jevons calculation based on geometric averages.
“The RPI is misleading,” said Philip Turnbull, a member of CPAC and the Royal Statistical Society. Continuing its publication “is not consistent with the duty of the national statistician to provide reliable and competent and meaningful indicators following best professional standards and international best practice.”
Matheson wasn’t available for comment. Derek Bird, head of prices at the ONS, said while he acknowledges RPI’s flaws, maintaining the index gives users the continuity they want.
“There’s the requirement that looks at statistical methodology, and there’s the user perspective,” he said. “She’s required to look at both of those elements and to try to find a solution that balances both.”
After the inquiry into the RPI started in mid-2011, investors in the inflation-linked bond market shied away from the securities. The debt had its worst performance in a decade last year, returning 0.8 percent according to Bank of America Merrill Lynch indexes. So far this year they’ve returned 4.2 percent, helped in part by a jump after Matheson’s announcement sent the yield on the 10-year debt to a then record of minus 0.99 percent. Last week the yield dipped to minus 1.4 percent.
“There is broad agreement that the current RPI methodology is not fit for purpose,” said Paul Johnson, a director of the Institute for Fiscal Studies, who expected the review to sanction a change. “Where they’ve ended up is that it’s better to be consistent than right, which is kind of surprising.”
All 10 economists in a Bloomberg News survey had predicted the ONS would change the RPI following its review.
Rejigging benchmarks is a sensitive issue in the wake of revelations that traders around the world manipulated the suite of interest rates known as the London Interbank Offered Rates, which are tied to more than $300 trillion of securities globally. Royal Bank of Scotland Group Plc, Britain’s biggest publicly owned lender, said Feb. 28 it will recoup 302 million pounds from its bonus pool and compensation clawbacks after it was fined for rigging Libor.
Bird at the ONS said it’s impossible to compile a comprehensive list of all the uses of RPI, which include rail fares and water charges, wage deals negotiated by unions, levies on alcohol, tobacco, gambling and fuel, and some student loan and corporation tax rates.
Paul McMahon, director of economic regulation at trade group Water U.K. said the organization was “pleased” with the decision to leave RPI alone, since its members’ income and debts are in part tied to the index.
“Water is hard-wired to RPI,” he said. “The RPI has been in there forever, but it’s been understood by investors.”
Nevertheless, the U.K. Statistics Authority has decided to remove RPI’s “National Statistic,” designation, which it assigns to economic indicators it deems to meet a quality threshold. RPI, the only statistic the ONS has a statutory duty to produce because of its role in the inflation-linked bond market, is the first non-experimental ONS statistic not to have the stamp.
Speaking before the authority’s decision, Turnbull said the loss of the quality mark “could be embarrassing for the ONS.”
The proposals the ONS set out in the consultation for improving RPI, which was 3.3 percent in January, would have delivered a lower inflation rate by between 0.4 and 0.7 percentage point. Estimates published last week of historical data show the new RPIJ measure would have been 2.5 percent in December, against an RPI reading of 3.1 percent. CPI has held at 2.7 percent since October. Tomorrow’s data for all three indexes will be for February.
Samuel Tombs, an economist at Capital Economics Ltd. in London, estimates a change may save the government about 7 billion pounds a year by 2016-2017 at the expense of investors.
“There were some convincing arguments for updating the RPI,” Tombs said. “Full-scale recalculation is unlikely now.”
The U.K. has an alternative benchmark inflation indicator, after introducing the Harmonized Index of Consumer Prices, known now as CPI, in 1996 to meet European Union legislation. In 2003 Gordon Brown, then Chancellor of the Exchequer, changed the Bank of England’s targeted measure of price stability to CPI from RPI. Inflation-linked bonds were left tracking RPI, though in 2011 the government switched to CPI for benefits, tax credits and public-sector pensions.
“There is no index that gives the right answer and all of them are subject to one problem or another,” said Johnson at the IFS.
From the mid-1990s, “the introduction of major methodological improvements in the RPI stalled,” said David Baran, who worked on price indexes at the ONS for 15 years and is now a consultant. “At the same time use of the CPI has increased to the point where it is the government’s and many other users’ preferred measure of inflation, though CPAC continues to provide advice on its development.”
CPI and RPI divergence due to their calculation mathematics -- CPI doesn’t use the Carli formula and RPI doesn’t use the Jevons formula, and other methods used have different weights -- was about half a point until early 2010. It ballooned to a full point after the change in clothing-price collection methods was introduced that year, an episode that taught the ONS some lessons about altering indexes. The ONS didn’t publish the difference between the CPI and RPIJ last week.
“We wouldn’t introduce a change in that way now,” Bird said. “We’ve learned that were we to consider something similar, we’d want to run it in parallel to assess its impact.”
To be sure, the ONS has introduced other inflation changes that haven’t produced such market disruptions. Its annual update of the basket of goods, announced last week, added e-books and white rum, and removed champagne. It also introduced a measure called the CPIH, which adds owner-occupiers’ housing costs to the existing CPI.
The eventual decision to leave RPI unaltered was a “sensible option” for David Dyer, senior portfolio manager at AXA Investment Managers in London, which oversees $1.4 billion. “You could see it being a legal minefield that would have been a real nightmare.” Still, the prospect that the RPI may lose its designation as a National Statistic raises “reputational issues for the U.K.,” he said.
“International investors in a market that has bonds maturing after 50 years are going to want reassurance on how that’s calculated,” Dyer said.
To contact the reporter on this story: Jennifer Ryan in London at firstname.lastname@example.org