A few ideas for fixing GDP

Written by Colby Smith via Alphaville on 8/1/2018

It should come as no surprise to readers of this blog that GDP is a faulty measure of a country’s economic health, and that treating it as the sole barometer for well-being can lead to spurious assumptions.

GDP never claimed to pay attention to this, and the way it’s calculated simply doesn’t allow for holistic measurement. In recent years, some academics have suggested broader formulations of GDP that more precisely measure intangible goods, the distribution of growth and environmental factors. A couple have gained traction.

The first proposal, from a team of economists led by Carol Corrado and Jonathan Haskel, tries to better account for intangibles. They propose “not to demolish GDP”, because as an economic measure, it does two things quite well: it avoids double counting (by subtracting out all the inputs that go into a given output) and it weights the value of products through prices. They do want “to repoint and to extend it”, however, by taking care not to confuse product improvements with price inflation, extending measurement to “non-market labour” like housework, and finally, adjusting for intangible capital and intellectual property.

These changes are meant to more comprehensively cover what the Bureau of Economic Analysis (BEA), the agency that measures US GDP, calls the “digital economy”: (1) infrastructure such as hardware and software, (2) e-commerce and online transactions and (3) digital content and media. In March the BEA released (for the first time) preliminary statistics on the scope and scale of this sector and announced plans to expand its coverage.

That is a good start, but there are structural problems with GDP itself that simply cannot account for changes happening in today's economy, say Corrado and Haskel. Uber drivers provide one example of the challenges in accurately measuring economic output. Unlike traditional taxi drivers, who are on their employers' payrolls, they are considered independent contractors. So if the BEA uses employer surveys to generate industry-level estimates of economic activity, there's bound to be some mismeasurement, points out Michael Redmond of the Richmond Fed. What's more, GDP does not measure home production, and the growing phenomenon of flexible employment is further blurring the line between that and market production.

Corrado and Haskel also believe that what makes an iPad and iPad has less to do with the silica and semi-conductors that go in it, and more to do with intangible assets like design, branding and entrepreneurship. These have typically been treated like “air conditioning or electricity”, they argue, commodities that are used in the course of production as opposed to inputs with long-term value. It is for these reasons that they believe expenditures such as the design of new goods and services, labour training and the curation of big data should be included in national account calculations, just like software, scientific R&D and the creation of artistic originals.

Big Tech's barter economy presents additional problems. Accurate GDP measurements require prices, so what can be done to account for digital services that are “ free”? Corrado and Haskel acknowledge that any old survey won't do, and instead use the approach taken by economist Erik Brynjolfsson. In a recent study, he asked a large sample of online users just how much they would need to be paid to forgo access for a year to specific digital goods. Internet maps fetched $3,600; email, $8,400. Search engines were the most valuable, with survey respondents needing $17,500 to do without. This kind of study is possible on a larger scale, Corrado and Haskel argue, and the resulting values can be added to GDP.

They also see scope for GDP to account for well-being by adding indicators like leisure and financial security to its consumption component. They cite a pair of economists who gathered cross-country data on GDP, hours of work (the inverse of leisure) and aspects of security (mortality, inequality etc) and came up with a final measure of well-being that correlated strongly with GDP per capita. Corrado and Haskel explain how it works:

GDP per capita in the UK is 75% of the US level. But working hours are 34% lower and life expectancy 2% longer, so it turns out that, on this measure, welfare is 97% that of the US level.

These alternatives sound great on paper, but the data points they require are hard to come by. Often, such data are not shared widely between government agencies—let alone across borders. Worse, there’s no standardised method for valuing many of today’s intangible assets, and it is difficult to gauge the accuracy (or gameability) of new indices that measure slippery things like well-being. Corrado and Haskel acknowledge these shortcomings, but implore readers to carry on: “As noted by Read (1898),” they write, “'it is better to be vaguely right than exactly wrong.'”

A second school of thought wants to move beyond a one-number-fits-all approach. These economists say statistical offices should track a dashboard of quality-of-life indicators: inequality, education, social connections, political voice and the environment. Nobel Laureates Joseph Stiglitz and Amartya Sen as well as economist Paul Fitoussi pioneered this thinking in a 2009 report commissioned by then-French president Nicolas Sarkozy.

Economists Diane Coyle and Benjamin Mitra-Khan also support this idea, and suggest a dashboard tracking natural capital (such as clean air), human capital, intangible assets, physical capital, net financial assets and social capital (like trust). Not only do they want to measure each of these resources, they also want to gauge people’s ease of access to them. Access is important, they argue, because “distribution is everything when thinking about aggregate economic welfare”, a fact that “has largely been forgotten”. Researchers at t he Washington Center for Equitable Growth agree. As we covered previously, they say GDP growth should be measured at different distributions (the bottom 10 per cent and the top 1 per cent, for example).

This is all well and good, but Coyle and Mitra-Khan concede that it is prohibitively difficult to obtain this type of data at the moment. No country has an adequate assessment of the quality of its infrastructure and maintenance needs, they say. And the idea of social capital is so abstract that there are not yet “systematic attempts” to measure it. So far, academics have mostly relied on surveys:

While they are encouraged by recent efforts to redress this, one hurdle remains: how to value each resource. Take natural capital:

If the price of oil falls, this translates into reduced natural capital, but intuitively a lower price causing more oil not to be drilled conserves the asset.

Coyle raises two additional points: a single number is very compelling, and making the switch from GDP is hard unless everyone follows suit. “It's like game theory,” she says. “If politicians tried, their critics would say you're only doing that because GDP is not growing enough.”