Is inflation overstated?



Are hundreds of millions of dollars of pension payments being made unnecessarily? The truth may be politically unpalatable, but it could lie with inflation and its likely overstatement over several decades.

Consumer price inflation is one of the most widely used economic statistics and has a direct impact on the payment of billions of dollars of benefits, pensions and salaries. However, its measurement is open to interpretation and often controversy. Increasingly, researchers believe that the failure of inflation indices to capture changes in what we buy and how we buy it has resulted in reported inflation being higher than it really is.

The most influential recent study was the Boskin Report, commissioned by the US Senate in 1996. This concluded that the widely-used US consumer price index (CPI) overstated inflation by 1.1 percentage points a year. The report found several biases, the main one being an inability to account for the changing quality of goods and services and new goods coming on to the market. The index also failed to adequately account for consumers switching between different expenditure categories (upper-level substitution bias) and between different goods within expenditure categories (lower-level substitution bias). The rise of discount stores has further muddied the picture (outlet bias). The chart shows the breakdown of these biases. Despite recent improvements to its calculation, the index is still estimated to be biased upwards by between 0.5 and 1.0 percentage points a year.  

Figure 1: Contributions to the Total Upward Bias in US CPI according to the 1996 Boskin Report

Source: Greenlees, 2006.

Adjusting for improving quality is a difficult task. Advances in internet speeds, digital content, smart phones and computers are some of the many technological improvements that are hard to capture in an index. The iPhone is a classic example of how technology increases quality and convenience by incorporating a whole host of items that were formerly supplied separately, including telephones, cameras, CD players and GPS systems. Few of these improvements are picked up by CPI¹. So whilst the price of an iPhone now is surely higher than a typical mobile phone 10 years ago, the additional benefits should outweigh the measured price increase. In their struggle to capture the enhanced quality of products like the iPhone, indices like the CPI inevitably create an upward bias. The problem may have been compounded if, as some have argued, technological change is taking place at a faster rate today than in the past.

The implications stretch widely. If CPI is overstated, then real GDP will be understated – one estimate puts the difference at 0.25 percentage points a year². And if the CPI bias has increased recently, it may go some way to explaining the productivity slowdown seen in the US lately. The overstatement may also have cost the US government significant amounts of tax revenue, as tax brackets have been over-indexed. With a third of US federal budget outlays estimated to be influenced by CPI³, it means that hundreds of millions of dollars of government debt may have been – and will continue to be – accumulated unnecessarily. This extends to the many private employment contracts and pension payments that are determined by CPI. Any upward bias in the price index is likely to mean that corporate pension fund liabilities – and deficits – are overstated.

The likely upward bias in CPI has profound consequences for our understanding of our own financial health and that of the wider economy. As long as indices such as the CPI fail to accurately capture the effects of changing consumer preferences and technological advances, they are likely to overstate inflation. As a result, economic measures of standards of living, such as GDP or real wages, will continue to portray a worse picture than is actually the case.  

¹ Davies, Gavyn (2015), The greatest unknown - the impact of technology on the economy, Financial Times, 15th June 2015

² Boskin, Michael.J (2004), Causes and Consequences of Bias in the Consumer Price Index as a Measure of the Cost of Living, 58th International Atlantic Economic Society Conference

³ Boskin. Michael & Jorgenson, Dale (1997), Implications of Overstating Inflation for Indexing Government Programs and Understanding Economic Progress,  American Economic Review 87, no. 2 (1997): 89-94.


Read the full report

Is inflation overstated? 5 pages | 666 kb