How much would an average American, whose annual disposable income is $42,300, need to be paid in order to be persuaded to give up their mobile phone and access to the internet, for a full year? Would it be more, or less, than $8,400 for the year? Ponder that question – its importance will become apparent later.
The question is relevant to a much more familiar issue. Why has productivity growth slowed down so much in all major economies (both advanced and emerging) in the past decade?
To some extent, this has been a continuation of a much longer term trend in the advanced economies (see box below), but some of it is a new phenomenon. The slowdown in labour productivity growth accounts for most of the massive disappointment in output growth in the world since just before the Great Financial Crash in 2008. In the US, for example, productivity growth since 2005 has been 1.8 per cent per annum lower than it was in the prior decade, a deterioration of momentous proportions.
An optimistic explanation for this change is that the arrival of the internet and mobile technology has led to output gains that are not being correctly identified in the national accounts. If that is the case, then both real GDP and labour productivity might be higher than shown in the official GDP statistics, and the “slowdown” in productivity growth might not be genuine after all.
Recently, important new evidence has been published on the mismeasurement hypothesis (MMH). As yet, there is no unanimous verdict, though the bulk of the evidence suggests that it can explain only a fairly small part of the productivity puzzle. The bulk of the slowdown seems genuine.
Interest in the MMH went mainstream about a year ago, when Martin Feldstein argued that the official statisticians were substantially under-recording the contribution of the digital economy to GDP and productivity in the US. This made intuitive sense, since it appeared to explain the “productivity puzzle 2.0″ – i.e. the co-existence of weak productivity data with everyday experience of the enormous impact of digital technology on people’s lives. The economics profession is only now developing a considered response to this issue.
There are many ways [1] in which the digital economy might distort the productivity data, but only two are really important.
The first is the possibility that the ongoing improvement in the quality of digital products is being underestimated, so the official price indices applied to such products are systematically too high. If that is the case, then inflation is lower than the official data show, and real output is higher. The productivity puzzle automatically diminishes if this is the case.
Several new, semi-official studies have recently opined on this issue. They have reached slightly different conclusions.
A comprehensive research paper by David Byrne, John Fernald and Marshall Reinsdorf, just published by the Federal Reserve Bank of San Francisco, argues that inflation in the digital economy is indeed being understated, but not by as much as it was in the decade before 2005.
As a result, they believe that recent productivity growth has been slightly better than officially estimated (by about 0.3 percent per annum), but this makes no difference to the slowdown in productivity in the recent data compared with the prior decade. In fact, it makes the slowdown slightly more serious.
There seems to be an emerging consensus in favour of this judgment in the US. For example, the Economic Report of the President for 2016 shares this assessment.
But a recent inquiry into UK statistics conducted by Charles Bean reached a different, or at least more nuanced, conclusion. In a very interesting chapter about the impact of the digital economy, Bean suggests that annual UK consumer price inflation might have been overstated by over 1 per cent as a result of digital distortions.
Although this does not automatically read across to the productivity data, it would raise the possibility that real GDP growth has been much higher than the official data show, so a sizable chunk of the UK productivity puzzle might be linked to this phenomenon.
This and other factors lead Professor Bean to conclude that the case is unproven: official statistics agencies in the UK and other countries need to do far more work to understand this problem better [2].
A second possible cause of the MMH raises even thornier questions. This is the notion that much of the welfare benefits from consuming digital products is missing from the GDP data.
In the UK, an average citizen spends 20 hours per week using the internet, about double the time spent in 2005. Much of the content consumed is charged at extremely low or zero marginal prices in the market, so it is given very little weight in the GDP data.
Even those economists who are sceptical about the MMH tend to concede that GDP may be severely underestimating consumer welfare, especially for products that did not exist a decade ago (eg Facebook and You Tube).
But they also point out that GDP has always been intended to measure market activity, not consumer welfare, and that the absence of non market activities (like domestic work within families) has frequently “distorted” the data in the past. They are sceptical whether these distortions have increased in the digital era, which is what would be required to explain (or excuse) the productivity slow-down.
Chad Syverson at the Chicago Booth School of Business has done the numbers on this. He calculates that the productivity slowdown in the US is equivalent to about $2.7 trillion of lost output per annum by 2015. Even on the most generous method that he can find to calculate the extent of the underestimated consumer surplus from the digital economy, he reckons that only about one third of the productivity gap can be explained in this way.
That brings us back to our original question. Chad Syverson reckons that the unrecorded value of the digital economy to the average citizen would need to be $8400 per year in order to explain the entire productivity gap. This is one fifth of net disposable income per person. He suggests, on prima facie grounds, that few people would value their access to the digital economy at one fifth of their disposable income.
Maybe, but what is the appropriate figure? Casual observation suggests that most people are now extremely reliant upon, or addicted to, the internet, especially via their smartphones. Faced with the choice, I doubt whether they would be prepared to be transported back to the obsolete technology of a decade ago in exchange for an annual payment of less than, say, a few thousand dollars a year – ie far less than than the value currently accorded to digital activity in GDP.
If that conjecture (and it is no more than a conjecture) is valid, there might be something in the mismeasurement hypothesis after all. But the evidence suggests that it is not the main reason for the productivity debacle .
——————————————————————————————
[1] For example, the so-called sharing economy might depress measured productivity by shifting output towards under-recorded sectors (like rentals offered by households on Airbnb, or car rides on Uber). Current evidence suggests that these effects are still quite small, though they may be growing rapidly. See here and here.
[2] Chad Syverson (2016) shows that the extent of the slowdown in productivity growth among countries since 2004 is not related to the size of the digital sector in each economy. This suggests that the digital factor is probably not the major driving force behind the slowdown. The same is true in cross-sectional studies of US states, recently published by IMF economists Cardarelli and Lusinyan (2015).
Copyright The Financial Times Limited . All rights reserved. Please don't copy articles from FT.com and redistribute by email or post to the web.