MEASURING THE COST OF LIVING: FAILURE TO TAKE PROPER ACCOUNT OF TECHNOLOGICAL PROGRESS LEADS TO OVERESTIMATES OF INFLATION
Measures of the cost of living, like the retail price index (RPI), are inadequate, failing to reflect fully the impact of technological advances on our standard of living. This leads to a substantial upward bias in our estimates of inflation, perhaps as much as 1.6% a year. That is the contention of Professor William Nordhaus of Yale University in the latest issue of the Economic Journal. If he is right, then we may have to rewrite history:

Increases in the price of lighting services since 1830 may have been overestimated by as much as a thousandfold!
US real wage growth between 1959-95, currently measured at a very modest 10%, should be revised to a healthier 70%.
And estimated average annual rates of US productivity growth of 0.6% between 1973-95 should nearly be tripled.
Nordhaus notes that consumer price indices like the RPI are some of the most important measurements generated by economists and statisticians. Ideally, they are designed to measure the cost of attaining a given level of economic well-being. In practice, statisticians take a ‘basket of goods’, which represents the consumption patterns of the ‘average consumer’, and measure how the cost of this fixed basket changes over time.

This statistic is used to define ‘inflation’, and hence determines changes in a wide range of inflation-indexed state payments and benefits, as well as setting the background for pay settlements. It is also crucial for measuring the real growth of the economy, a key statistic in assessing the economic and political performance of the economy and government policies.

Nordhaus argues that the current methods for measuring the cost of living are inadequate and fail to reflect fully the impact of technological advances on the standard of living. One of the basic problems with the ‘basket of goods’ approach is that consumption patterns change: while this is partly due to changes in taste and fashion, it is also due to the impact of technological change on the products we consume.

At present, since there is little change from year to year, statisticians simply adjust the reference basket of goods little by little, hoping that this gradual change will track what is really happening. But Nordhaus argues that there is a consistent bias, which means that the current consumer price statistics underestimate the impact of new products.

There are two main reasons for this. First, the basket of goods only takes up new products with a lag: in periods of rapid change and innovation, this can lead to big errors. Second, there has been an in-built tendency to underestimate the services provided by new products and developments of existing products. New cars provide a range and level of services far superior to the cars of 10 or 20 years ago. Nordhaus argues that if you look over a longer historical timescale, then the comparisons can be completely misleading. Increases in the price of lighting services since 1830 may have been overestimated by as much as a thousandfold! While lighting may be an extreme example, the same principle applies to many other consumer goods.

If Nordhaus is right, then we may have to rewrite history. In the US, this issue was examined by the Boskin Committee, which found that the US inflation figures overestimated the true annual rate of inflation by 1.3%; this upward bias will continue using current methods of calculation. (The Book Review section of this journal issue discusses the Committee’s Final Report.)

The overestimation of inflation implies a corresponding underestimation of real growth in wages and output. Over the period 1959-95, the increase in real wages is currently measured at a very modest 10%: it should be revised to a healthier 70%. Estimates of productivity growth over the period 1973-95 indicate an average annual rate of 0.6%: this should nearly be tripled.

‘The fact that we may be getting such an important statistic as the RPI wrong by so much indicates that we really need to look again at the way it is calculated in the UK’, claims Professor Huw Dixon of the University of York and CEPR. ‘Since so much depends on the inflation rate measure, we need to make sure we are getting it right’.

Note: ‘‘Traditional Productivity Estimates are Asleep at the Technological Switch’ by William B. Nordhaus is published in the Controversy section of the Autumn 1997 issue of the Economic Journal. Nordhaus is Professor of Economics at Yale University.

For Further Information: contact Huw Dixon on 01904-433782/788 (home: 01904 423118; email: hdd1@york.ac.uk); William Nordhaus on 001-203-432-3587; email: william.nordhaus@yale.edu) or RES/ESRC Media Consultant for Economics Romesh Vaitilingam on 0171-878-2919, 0117-983-9770 or mobile telephone 0468-661095 (email: rvaitilingam@cepr.org).