Is it time to change how we measure inflation?
Measures of inflation are hugely influential on policy but struggle to account for the pace of technological innovation, says Paul Ormerod
During this decade, the rate of inflation has become a key factor in determining living standards. Many benefits are linked to it. It sets a marker for wage demands. And when it goes up, cries of “cost of living crisis” grow ever louder.
But what exactly is inflation? The Office for National Statistics has a variety of estimates, such as the consumer price index and the retail price index. Over time, these all broadly move together.
At any point in time, however, they can and do differ. Debates take place as to which one is the more appropriate for updating whichever particular benefit or subsidy is being discussed.
The answer is in fact much more complicated than simply deciding which of the ONS measures to choose.
Over the past two decades or so, leading economists and econometricians, mainly based in America, have produced arguments which claim that the rate of inflation has been considerably overstated.
A key reason is the massive waves of technological innovation which have taken place. The quality of the products which have become available has grown dramatically. Smart phones, for example, have more computing power than many mainframes of the 1970s. But how do we measure this increase in quality?
This is a very difficult problem for the ONS in the UK and their counterparts elsewhere, such as the Bureau of Economic Analysis (BEA) in America.
When the national statisticians gather information on sales of a product such as Marmite, which has basically remained unchanged for generations, the problem does not really arise.
Comparing like with like
In markets where innovation is rapid, we are simply not comparing like with like. In the early 2000s, top US econometrician Jerry Hausman at MIT thought about how to take into account the impact of a completely new product.
He suggested that a way of estimating the change in its price from the previous period, when it did not exist, is to assume that the price was then so colossally high that no-one bought it at all! So the implicit drop in price this period is huge.
This alone led to estimates that inflation in the US, even before the massive wave of new tech products got underway seriously, was systematically overestimated by at least one per cent a year.
The latest issue of the American Economic Review contains a long and highly technical paper by no fewer than eight leading US economists. The first phrase in the title is “Quality Adjustment at Scale”.
The authors examine two issues with the official measures of inflation. First, changes in quality in general, not just the introduction of new products. Second, what are known as substitution effects. When the relative price of a product increases, consumers make economies by substituting for them in their overall expenditure.
They construct an impressive theoretical framework for how to take these into account.
But a major innovation is to use highly detailed information on consumer spending captured by digital technology. They examine 2.6m products at the level at which retailers themselves use to restock. For example, bottles of the same brand of sparking water can be bought either individually or in packs.
The conclusion of the paper is stark: “Accounting simultaneously for substitution and quality change lowers measured inflation rates substantially across a broad range of goods”.
The economists argue that the use of item-level data, available through advances in technology, can “reengineer” many aspects of the national accounts, not just inflation but productivity and output growth.
Official bodies such as the ONS and the BEA are aware of these issues. But they themselves need to move faster and innovate to incorporate these major developments into their estimates.
Paul Ormerod is an Honorary Professor at the Alliance Business School at the University of Manchester. You can follow him on Instagram @profpaulormerod
