Author: Jakub Growiec

Professor in Social Sciences (Economic and Finance), SGH Warsaw School of Economics, Economic Advisor (NBP)

Beyond GDP: how to measure economic growth in the digital era?

In economists’ discussions we can sometimes hear that GDP is a poor measure of economic development and that we should look for a better one. Nevertheless, GDP is still widely applied in macroeconomic analysis, and alternative measures are used significantly less often, and with various reservations.
Beyond GDP: how to measure economic growth in the digital era?

(©Envato)

However, the range of the alternative measures of the level and pace of economic development is increasingly broad, and these measures will be probably rising in significance in the digital era.

In line with its definition, GDP describes the aggregated value of final goods and services generated by national and foreign factors of production within the territory of a given country in a specific unit of time (typically, a year). Given this definition, GDP is a good yardstick of the size of the economy; when divided by the population of the country (GDP per capita), it is also an indicator of average wealth, productivity, well-being and the level of the country’s economic development. At this point, however, many questions are raised (probably the most comprehensive and wide-ranging report containing criticism and extensions of GDP has been drawn up by Joseph Stiglitz, Amartya Sen and Jean-Paul Fitoussi).

Firstly, is GDP an adequate measure of the well-being of a country’s population? After all, it cannot capture many aspects of well-being. For instance, it does not include unregistered production, goods and services available free of charge, the value of leisure; neither does it reflect the scale of social inequality and the extent of the risks faced by the public every day, such as illnesses, accidents, environmental pollution and crime. Meanwhile, it seems indisputable that apart from wanting to be as wealthy as possible, we would also like to have access to a good and cheap health service and education, and enjoy free time and good interpersonal relations – we would like to live in a country where we feel safe, free and happy.

Since each of these things can be estimated, synthetic indicators have appeared in economic literature, encompassing an extended range of sources of well-being of a country’s population. This, however, has stirred debates about which measures should or shouldn’t be included, and what weights should be ascribed to them.

For example, the popular Human Development Index (HDI) is a geometric mean of appropriately prepared indices of GDP, longevity and education, which are taken into account at equal weights (each 1/3). Shouldn’t we also take into account other measures, though?

Charles Jones and Peter Klenow propose a measure of well-being derived from economic microfoundations, and covering the following: life expectancy, consumption in relation to income, the amount of leisure, the scale of inequality of consumption and the scale of inequality of leisure. This measure does not include GDP at all, in line with the logic that income is not the end, but a means to an end, this end being a life abundant in consumer goods and pleasantly spent leisure time.

A Solomon solution to the inevitable arbitrariness of the choice of weights has been put forward by the OECD: their Better Life Index  contains 11 components (housing, work, education, civil society, life satisfaction, work-life balance, income, social relations, environment, health, security), whose weights can be determined by each user by adjusting the respective sliders. Finally, we can also decide that the only thing that matters is our feeling of being happy and everything else only helps to achieve that. In which case the measurement becomes easy again (albeit not free from discussions about the cultural determinants): it is enough to ask the respondent how happy they feel.

 Economic activity generates a number of negative externalities not covered by GDP.

Secondly, even if we resist the temptation to treat GDP per capita as a measure of well-being, regarding it merely as an indicator of the broadly understood level of economic development, it is not certain whether it measures all the goods produced in the economy accurately. In particular, the literature raises reservations about the omission of the negative impact of the economy on the environment. Economic activity generates a number of negative externalities not covered by GDP, such as greenhouse gas emissions, build-up of waste,  biodiversity loss, landscape degradation as well as air, water and soil pollution. Therefore in discussions on natural environment protection there are calls for the presently universal desire to maximise GDP to be replaced by a more “sustainable” approach, including also a potential objective of limiting the impact on the environment, by, among others, reducing greenhouse gas emissions or recycling materials. At the same time it is suggested that GDP itself be made “greener” by using the so-called “green GDP”, which includes a correction for biodiversity loss and for the cost of climate change. A similar indicator, the Index of Sustainable Economic Welfare, ISEW, adjusts GDP for the cost of degradation of the environment and the depreciation of natural capital.

Thirdly, and this point has so far been raised somewhat less frequently, but its significance will probably rise over time, GDP also omits changes in the economy occurring on the back of the rising digital economy. Since the 1980s, exponential growth has been observed in the total computing power of computers and the total volume of stored and transferred data. These are growing at a pace one order of magnitude higher than global GDP, doubling not every 20-30 years, like GDP, but every 2-3 years. There is no doubt that this would not be happening if it did not make economic sense. However, the impact of this IT explosion on GDP is not visible. In 1987 the Nobel Prize laureate Robert Solow said that, “you can see the computer age everywhere but in the productivity statistics.” He hardly expected that his words would be repeated to this day.

Although, on the one hand, many digital goods – applications, games, videos, texts etc. – are available free of charge on the Web, which is why they are not covered by GDP, these goods undoubtedly have a concealed value which is directly unmeasurable, but which we can try to estimate indirectly, by, for example, measuring the time and attention devoted to them  (the so-called attention economy). In addition, we also pay for them in a non-financial way, namely by becoming involuntary recipients of advertisements and providers of data. Instant access to information also generates difficult-to-measure indirect benefits related to improved operational efficiency of firms and increased effective scale of production. Yet on the other hand, the Internet may also diminish the productivity of workers who are constantly distracted by leisure-enhancing technologies available online. This is discussed in a very apt and interesting way by Łukasz Rachel. Instant access to information is conducive to entertainment, but also globalisation, international fragmentation of production and the creation of global value chains.

To summarise, the total benefit from digital goods is probably much bigger than what follows from GDP accounts, but difficult to capture due to the zero market price of these goods and absence of reliable valuation of data.

Interesting insights into the strengths and weaknesses of GDP as a measure of economic development are offered by the history of the indicator. Although today we try to retrospectively estimate GDP hundreds, and even thousands of years ago, the measure itself was defined first in 1934 by Simon Kuznets. Its construction is, on the one hand, universal – indeed, GDP comprises the aggregate value of final goods and services produced in the entire economy – yet on the other, it is deeply anchored in the realities of the industrial era economy. The elements most underestimated in GDP, e.g. environmental externalities or digital goods, are exactly the areas that we have only started to understand recently, and where market valuation operates the least efficiently to this day, and market institutions are the least developed.

It is only since the industrial revolution that we have observed a final separation of GDP growth from population growth and the gradual spread of prosperity.

When today we try to estimate what per capita GDP was in Europe in ancient or medieval times, we approximate it with data on agricultural production. This is because in those times it was a key sector of the economy – almost everybody worked in agriculture. Besides, Malthusian population dynamics was in force, mercilessly reducing per capita GDP to the level of subsistence; any potential surpluses ended up in royal or church estates. Starting from the Renaissance, the significance of trade and crafts started to increase, wealthier merchants and burghers emerged, and, in effect, per capita GDP rose somewhat. Yet it is only since the industrial revolution that we have observed a final separation of GDP growth from population growth and the gradual spread of prosperity. And it was only when the industrial economy had developed to a sufficient degree, its institutions had been created and solidified and specialised economists had appeared in the world, did we finally learn to measure its product accurately.

Today in turn we are seeing a decoupling of the pace of growth of computer processing power, and of the volume of collected and transmitted data, from the pace of GDP growth. The pace of digital growth exceeds that of GDP growth by a whole order of magnitude. Also, the institutions of the digital era are only now being created; only now are we becoming familiar with the operation of copyright to works which can technically be disseminated at no cost, or with the protection of personal data and other sensitive data. Also, we still don’t know how to measure transactions in digital goods. Data transfer alone, measured in bits, is not enough, because data is not yet information, and besides, data is often sent and re-sent back and forth.

What would be an equivalent of value added in data processing? How can we measure the final value of digital goods for the final recipient, who gets them free of charge?

There is also one more interesting analogy. Just as, at the dawn of the industrial era, vast fortunes of industrialists and oil company owners were made, today massive fortunes are being born in the IT industry, especially in software. If today “data is the new oil”, the owners of GAFAM companies (Google, Amazon, Facebook, Apple, Microsoft) are the equivalent of the 19th-century industrial and oil magnates (and 15th-century merchants). Their economic activity is also new and sufficiently different from those previously known that we still don’t know how to evaluate or monitor it – or for that matter, how to tax it effectively!

Therefore, looking to the future, I suspect that while GDP will remain a very important measure of a country’s level of development and the level of development of our civilisation as a whole, alternative measures will be increasingly taken into account. Not only in terms of taking into account environmental externalities not covered by GDP – which has been attempted for several decades – but also in terms of products of digital economy.

For example, one of the key conundrums of the modern economy is why GDP per capita in the developed countries is rising more slowly today than in the period1950-1970, despite the unparalleled technological progress in the meantime, the increase in the level of people’s education, and the extension of infrastructure, including R&D infrastructure, plus an unimaginable leap in information processing. I will leave you with an intriguing thought: perhaps the growth of human civilisation has not slowed down at all, but GDP per capita has stopped keeping up with it?

 

The author expresses his own opinions, not the official position of NBP.

(©Envato)

Otwarta licencja


Tags