Serious Errors on UK Telecoms Data: Prices could have fallen by 90% more than the official price index
Brian Sturgess - April 2018
The Office of National Statistics (ONS), Britain’s official data agency, has admitted that it has made serious errors in its estimation of the output and productivity of the telecoms sector. A paper co-authored by Richard Heys, deputy chief economist of the ONS which compared approaches to deflating the output of the telecoms industry across countries found a “wider disconnect between the technological performance and economic measurement of the industry in the UK.”1 The authors reported that the deflator used in the UK telecoms sector was biased upwards.
The ONS mismeasurement problem arose from the use of inaccurate price indices to deflate the nominal output of the sector in order to gain an estimate of real output. Official data shows no increase in telecoms prices between 2010 and 2015, but given there was a slight fall in industry turnover or nominal output, this means that real gross value added fell by 4 per cent pulling down productivity. The method of measuring prices does not measure quality changes including the growing bundles of text, calls and data offered on mobile and fixed networks. Improving the price deflator to include these would have shown a price fall of 35 percent or a significant rise in the industry’s real output. Heys admitted that transmitted data usage in the form of calls and texts had risen by 900 per cent from 2000 to 2015 so converting these into a digital format, would imply a fall in the price per bit of more than 90 percent.
These problems mirror the fairy tale2 methods to measure output in the financial services sector based on interest rate spreads which gave the paradoxical result of a rise in real output at the time of the financial crisis of 2008. In the telecoms sector it means that given the significance of the sector, (at around 2 per cent of Gross Domestic Product), economic growth could be higher over that period and inflation lower pushing up GDP per capita.
Such mismeasurement mistakes are frequent in developing economies particularly in Africa.3 Correcting errors in measuring the contribution of the telecoms sector in Nigeria were partly responsible for a one off uplift of 89 per cent in national income in 2014 after the country rebased its national income statistics. In that case the problem was a result of a failure to take note of the growth in the mobile telephony sector. Many emerging markets simply cannot afford to take frequent and extensive surveys of economic activity.4 In contrast, the measurement of economic data is supposed to be more accurate in the developed world where far more resources are available to government agencies.
The estimation mistakes of the size of the telecoms sector were announced by the ONS, (with a budget of £207 million) were not trivial. Fixing the problem creates another conundrum. If the productivity and real output of the telecoms sector is shown to have risen and the overall index of producer prices used to deflate nominal GDP is left unchanged, estimated productivity and value added in other sectors would have to fall. From the perspective of real spending by consumers on telecoms, changing the measurement of telecoms in the consumer price index to reflect falling output prices would raise conceptual and practical issues for the Consumer Price Index (CPI) according to Heys. Social security benefits and many other contracts are tied to changes in the CPI never mind the fact that the index is a major input into monetary policy decisions on interest rates made by the Bank of England. There are also legal and political difficulties. The methodology used to calculate the CPI is intended to produce a harmonised set of inflation statistics across the European Union and is laid down by Eurostat.5 The only inflation index that the ONS has discretion over is the Retail Price Index (RPI), but the less said about the recent history and misuse of that index the better.
In brief government data which are used as a foundation for contracts and economic decisions have been proven to be wildly inaccurate. This distorts the allocation of resources and it undermines faith in government. The dissemination of economic data by the government is supposed to be a public good. But what if it is a public bad? This was the view of Sir John Cowperthwaite, Financial Secretary of Hong Kong throughout the 1960s, the period of its economic miracle, who banned the publication of all, but the basic minimum of economic statistics. He thought publically endorsed economic data were dangerous and distorted.
The embarrassing revelation by the ONS about its faulty telecoms data was made ironically at a seminar hosted by the spanking new Economics Statistics Centre of Excellence. One of the purposes of the Centre is to improve the quality of economic statistics produced by the ONS and to make them relevant to the changes in economic activity wrought by the digital economy. The authors of the paper lamely concluded that the bias in the deflator used suggested “the need for continued research in this area.” The experts’ clarion call for job creation and more resources!
Fighting the Gender Wage Gap: The Bahrain Experience
Can Big Data Improve Economic Measurement?
Trade Data: Use with Care
Debt to GDP Ratio: Use with Care
The Flaws in Population Data
The UK Retail Price Index: Broken, Inaccurate and Unfair