Long-term growth effects of advancing digitization

About the Author: Dr. Thieß Petersen is a Senior Advisor of the Bertelsmann Foundation in the “Global Economic Dynamics” project and lecturer at the European University Viadrina in Frankfurt an der Oder.

Digitization has now reached many areas of the economy. In the future – assuming society decides to do so – it will penetrate the economy even more. The author finds both growth-promoting and growth-dampening effects. In the long term, the development may be accompanied by a significant drop in employment, which would have a negative impact on consumption. In the short term, he sees positive growth effects from the necessary investments as part of the digital transformation. However, it is quite possible that although there is a decline in the measured gross domestic product, social prosperity increases due to the time gained for self-determined activities.

What is digitization about? The term digitization is understood comprehensively here. It essentially describes the global spread of information and communication technologies, i.e., the increased use of these technologies in all areas of human existence. This use affects not only the economic production processes but also consumption (online shopping), education (e-learning), political participation (e-governance), transport (ticket machines and e-tickets), health care (telemedical procedures), and much more up to the leisure and communication behavior of the people (social media).

Only a few developments that are already indicated today or can even be found in practice illustrate how much digitization can change economic processes in the future: ¹

Modern industrial robots are already producing products, preliminary work, machines, and, in turn, industrial robots with less and less human support. Driverless forklifts and warehouse databases have largely replaced human labor in fully automated warehouses.
Self-driving vehicles currently come in semi-automated and highly automated forms. The last remaining step is fully automatic driving, in which a vehicle automatically masters all conceivable traffic situations. In the future, this will replace the truck, bus, and taxi drivers. The same can be expected for rail transport and shipping.
Legal pattern recognition software quickly finds precedents, making many lawyers and legal professionals redundant. Medical diagnostic software programs search databases and make medical diagnoses in seconds. In addition, surgical robots take over the activities of surgeons.
Even more powerful translation software takes over the work of interpreters and translators. For example, at the end of May 2014, Microsoft presented the first test version of a simultaneous translation program for video calls, “Skype Translator”.
Writing robots are increasingly becoming a competitor for journalists. In the field of business journalism, for example, software is used that evaluates companies’ quarterly reports, analyzes other company data from the Internet, takes into account the figures of competing companies, and then writes an article about it. The record for completing all of these tasks is 15 seconds, i.e., 15 seconds after the official publication of the company’s data, a print-ready text is available.²
In the financial services sector, online banking, online insurance, online securities trading, and online lending are replacing online providers such as Auxmoney for bankers, insurance brokers, and stockbrokers. This trend in the field of securities trading is reinforced by computer programs, which already in 2012 executed two-thirds of all stock orders in the USA. ³
Online reservations and travel and hotel booking portals are replacing travel agencies and their employees in the tourism sector.
Care robots are used in elderly care and childcare.
Finally, 3D printers will increasingly ensure that consumers can produce products themselves. In perspective, this can lead to the end of numerous production and trading companies.

Even if some of these developments appear utopian from today’s perspective, it can be assumed that they will become a reality sooner or later. However, it should be pointed out that the developments outlined are not based on the laws of nature since they are shaped and influenced by socio-political decisions. Since there is no certainty about the concrete effects of digitization, the following statements are rough guidelines for possible developments.

Medium and long-term growth effects of digitization

The increased use of digital technologies influences the economic growth of an economy through numerous channels. Among the most important are: the need for investment to build up the necessary digital infrastructure, the technological progress brought about by ongoing digitalization, the trend towards the sharing economy, the employment effects of ongoing digitalization, and the question of how citizens use the time savings caused by technology.

Investment needs, technological progress, and growth

The need for investment to set up the digital infrastructure increases the overall economic demand for goods in the short term. Companies adjust their production levels to the increased demand, resulting in stronger economic growth. Necessary public investments also have a positive effect on growth and employment.

The technological advances associated with digitization also have a growth-enhancing effect: Technological advances mean that production costs fall and companies can offer their product on the market at a lower price. As a result, the market price for the product in question falls. Normally, consumers respond to a falling price by asking for more units of the product in question. If companies adapt to the increasing demand, production will increase. For the economy as a whole, this means an increase in the goods and services produced, i.e., an increase in the real gross domestic product (GDP).

However, it must be considered as a limitation that the value of the goods and services produced can fall: the value results from multiplying the number of goods by the corresponding market prices. Suppose the fall in price as a result of technological progress is relatively large and the associated increase in the number of goods produced and demanded is relatively small. In that case, the value of the goods produced and demanded will fall. In terms of GDP, this means that nominal GDP can fall as a result of the technological progress associated with ongoing digitization. If the fall in prices is not properly taken into account when converting nominal GDP into real GDP, the actual increase in the number of goods can be underestimated, and thus also the real increase in GDP can be underestimated. To make matters worse, technological progress changes not only the prices of goods but also their quality. The correct measurement of real gross domestic product requires these quality differences to be taken into account. In order to meet this requirement, the Federal Statistical Office uses procedures within the framework of hedonic price measurement that enables a corresponding quality adjustment.⁴ However, if the monetary value of quality progress is underestimated, the GDP growth reported by the national accounts will be lower than it actually is.

Sharing economy and growth

Another growth-influencing development concerns the trend toward the sharing economy. This means that consumers no longer buy certain products themselves but rent them for a certain period of time. A prominent example of this is car sharing. With this form of use, people become members of a car-sharing network that purchases a certain number of automobiles. The members of the network can then use these automobiles by paying a membership fee and/or a usage-based fee.

There are already numerous goods that are shared. In addition to car-sharing networks, there are these facilities for bicycles, apartments (e.g. Airbnb), tools, household appliances, toys (e.g. Rent That Toy! and Spark Box Toys), designer ties (e.g. Tie Society), designer clothes, handbags or jewelry, as well as in in the areas of media and entertainment (e.g. Spotify and Netflix). ⁵ With a view to the level of economic growth – defined as an increase in GDP over time – there are three main consequences of this form of use:

The trend towards the sharing economy lowers economic growth in terms of the national accounts because the demand for consumer goods falls: If four people share a car, only one car is in demand and produced, but no longer four.
The trend towards the sharing economy further reduces economic growth if consumers bypass the traditional channels of the market when using the sharing networks: ⁶ When tourists access private apartments in the city they are visiting, they pay a far lower fee than they would for a hotel room. In terms of quantity, the quantity of goods demanded remains constant. However, since GDP values the goods and services consumed at their market prices, the substitution of private housing for hotel rooms leads to a decline in nominal GDP. The drop in GDP is even more serious if the private owners of the home (or other goods with corresponding networks) do not declare their income on their tax returns. In this case, the use of a private dwelling is not recorded in GDP at all.
If these two effects result in economic growth slowing down or GDP even shrinking, this impacts investment activities: When demand for consumer goods falls, it makes no sense to increase production capacity in companies. As a result, investments are falling. A falling investment demand means that companies in the capital goods industry can sell fewer machines and other means of production. Consequently, the capital goods industry reduces production, further reducing economic growth.

All three growth-dampening effects then also have a repercussion on the labor market: companies adapt to the lower demand and reduce their production. This also reduces the level of employment. If the workers released as a result do not find new jobs, their disposable income will fall. As a result, the purchasing power of these people decreases, so that they have to limit their consumption. The associated reduction in private consumer demand reduces production in the consumer goods industry, which further weakens economic growth.

Employment effects of advancing the digitization

A purely quantitative analysis shows that technological progress in many areas of activity has meant that machines have largely replaced human labor: ticket machines and ATMs are replacing counter clerks, fully automated production systems produce goods, the logistics industry works with highly automated facilities, “in which Software, computers, and robots play a much larger role in their interaction than humans” ⁷ – the list could be continued at will. The consequence is a fall in labor input measured in hours worked, even as GDP grows.

Economic growth without a simultaneous increase in employment or a decrease in unemployment was already observed in the USA at the beginning of the 1990s. ⁸ It was repeated during the recovery after the recession that began in 2007: Between Q4 2007 (the start of the recession) and Q3 2012, real GDP in the US grew by around 2.2%. Due to the high productivity advances, however, the number of people in employment fell by 2.6% in the same period. In this context, Mark J. Perry speaks of a “jobless recovery,” which can be attributed to the high productivity and efficiency gains. ⁹ The phenomenon of “jobless growth” can even be found in less developed economies such as China and India. ¹⁰

The question of whether there will also be a downward trend in need for labor in the future because capital and technology are being used more and more and are replacing human labor is answered inconsistently in the literature:

In 2013, Carl B. Frey and Michael A. Osborne published a widely acclaimed study in which they calculated the probability that certain jobs in the USA would be computerized by 2035. Based on 702 jobs, they estimate that by 2035 around 47% of American workers could be replaced by computers. ¹¹ Applying the method used by Frey and Osborne to Germany, “it turns out that 59% or over 18 million jobs are at risk” ¹². A paper by the Center for European Economic Research (ZEW), which also applies the study to Germany, comes to the conclusion that the probability of automation affects 42% of employees, especially low-skilled and low-income people. At the same time, the authors point out that this potential for automation does not necessarily have to lead to an actual drop in employment because there are “social, legal, and ethical hurdles to the introduction of new technologies” that must be taken into account. ¹³
Even greater are the technology-driven job losses that Jeremy Rifkin predicts. As early as the mid-1990s, he assumed that we would run out of gainful employment as a result of technological progress: “By the middle of the next century, there will be no more workers; they will all have fallen victim to the Third Industrial Revolution.” ¹⁴ Roughly Twenty years later, he reiterates this thesis: “Disregarding the possibility of an unforeseen setback, as we move into the mid-21st century, we will increasingly rely on intelligent technologies operating under the under the supervision of small groups of highly skilled human and technical workers.”¹⁵

In addition, there are also studies that do not see any displacement of workers through digital technologies, at least in the medium term, but even expect employment growth:

A study published in July 2016 on behalf of the Federal Ministry of Labor and Social Affairs produced a forecast of the effects of advancing digitization on the German labor market up to the year 2030. The authors also assume that there will be a high level of layoffs in fields of activity that can be digitized, as you can see, but at the same time, a higher need for “coordinating, researching, communicative, creative and decision-making activities” ¹⁶. Between 2014 and 2030, the number of jobs that will be made redundant will be almost 1.11 million people. The areas of transport and logistics, raw material extraction, production and manufacturing, agriculture and forestry, commercial services and sales, as well as the hotel and restaurant industry, are particularly affected. These job losses are offset by gains of around 1.35 million jobs, mainly in the areas of health, social affairs, teaching and education, business organization, natural sciences, IT, media, art, and culture as well as the language, humanities, and social and economics. On balance, this results in an increase in the number of employed by 240,000. ¹⁷
In its calculations from July 2016, the Boston Consulting Group assumes that around 600,000 jobs will be lost in Germany between 2015 and 2025 as part of the transition to Industry 4.0. At the same time, however, around 1 million new jobs will be created; an increase in jobs of around 400,000 can be expected by 2025. ¹⁸
A simulation calculation by the Institute for Labor Market and Occupational Research (IAB) assumes in a basic calculation that around 490,000 jobs will be lost in Germany from 2015 to 2025 (mainly in the manufacturing sector), but at the same time, 430,000 new jobs will be created ( especially in the service sector). The loss of jobs is, therefore, around 60,000 jobs. ¹⁹

How are these diverging theses to be assessed? Advancing digitization will tend to result in physical capital and digital technologies replacing human labor in the production processes of highly developed industrial nations such as Germany. In the next ten to 15 years, the associated job losses will still be moderate. In the long term, however, it can be assumed that there will be significant job losses – both in manufacturing and in the service sector. This primarily affects activities with low qualification requirements but increasingly also demanding professions. Despite the considerable potential for savings in human work that can be expected as a result of ongoing digitization, the work is coming to an end, which implies fully automatic production in all areas of human existence, but which cannot be seen in the long term either. However, the savings in human labor will be so massive that in the long term (i.e., from 2040/2050) in developed economies, they could reach around 50%, calculated by Frey and Osborne.

This release of workers in turn has repercussions on the demand for goods: the increased use of capital and technology means that workers tend to be replaced by capital and the level of employment thus falls. This also increases productivity. This makes it possible to produce an increasing amount of goods and services with less and less labor input. Consequently, in developed industrial societies, the need for human labor (measured in hours) is declining. This reduces wages as the price of labor is a production factor. At the same time, the disposable income of those who cannot find paid employment is falling.

If both the amount of labor employed and wages decrease, the macroeconomic distribution of income shifts in favor of capital: Those earning capital income receive an increasing share of aggregate income, while the share of income earned by wage earners decreases. An increase in capital income at the expense of labor income does not automatically lead to greater income inequality within society. If wealth, which forms the basis of capital income, were distributed more or less equally in a society, every citizen could compensate for the loss of income he suffers as an employee with higher capital income. In fact, however, wealth in developed economies like Germany is very unequally distributed. As digitization progresses, increasing proportions of overall economic income flow to relatively few high-income households. These are characterized by an above-average savings rate. This leads to a decline in aggregate consumer demand.

In addition, the growing income gap between full-time and part-time workers and the unemployed should be considered: if technology reduces the need for human labor, the number of people who are either unemployed or only able to work part-time increases. This reduces the annual income of these people.

As a result, in the course of advancing digitization, there is a weakening of mass purchasing power and an overall economic loss of demand; companies are adapting to this with falling investments and lower production. Seen in this way, there is a demand-related brake on growth.

Time wealth and growth

If the thesis of increasing productivity is correct, it means in the long run that, advanced economies can produce a constant amount of goods and services with less and less labor input. Whether the GDP of society then increases depends essentially on how people use the time gained. Four basic developments are conceivable here:

People use the time they have gained for private leisure activities and take advantage of commercial offers. Examples of such offers are cruises, city trips, and vacations; visits to leisure and amusement parks or wellness hotels; visits to sporting events and major sporting events such as the Olympic Games, Formula 1 races, and football world and European championships; visits to Concerts, operas, museums, and art exhibitions, commercial training opportunities and much more. ²⁰ This trend towards a commercial event and experience economy promotes growth because market prices are paid for such offers, and these offers are therefore included in the national accounts.
Citizens use the time gained for private leisure activities without resorting to commercial offers. They play music, play theatre, work in their own garden or play sports. Apart from the products necessary for these activities, such as musical instruments and sportswear, no other products are required. There is, therefore, no significant increase in GDP.
The time saved can be used for volunteer work. As there is no market price for these activities, the national accounts do not record these activities. GDP is, therefore, not increasing. If voluntary work crowds out services traded on the market, GDP actually falls.
Finally, it is also conceivable that people use the additional time for gainful employment. If the additional products produced are then sold on markets, there is an increase in GDP. The prerequisite for this, however, is that there is also an affluent demand for these additional products. However, this is not guaranteed: If there is a significant reduction in gainful employment due to the increased use of capital and the distribution of income thus becomes more unequal, there is a lack of corresponding mass purchasing power.

It is impossible to predict which of these four development trends will actually become a reality. The values of the citizens are decisive for this. In any case, it cannot be ruled out that, ceteris paribus, the time savings associated with advancing digitization in highly developed economies will lead to a tendency towards stagnation or even a decline in GDP.

Even if the GDP falls, it does not necessarily mean that people’s living conditions will deteriorate. On the one hand, there is an increase in the real quantity of goods, which is not fully covered by the national accounts, but is nevertheless present and improves people’s material living conditions. On the other hand, citizens can use the time gained for self-determined activities, which can increase their life satisfaction. People’s prosperity – defined as life satisfaction or happiness – then increases even though GDP falls.

Conclusion

The advancing digitization has both growth-promoting and growth-dampening effects. The necessary investments as part of the digital transformation and technological advances have a growth-promoting effect. The trend towards the sharing economy and the shedding of workers reduce the overall economic demand for goods and consequently have a dampening effect on growth. In the short term, i.e., up to around 2025/2030, it can be assumed that the growth-promoting effects will predominate in developed economies such as Germany. In the medium and long term, i.e., from 2040 to 2050, however, the growth-dampening effects are likely to gain the upper hand, and GDP could tend to fall in terms of the national accounts.

Two limiting aspects should be pointed out. First, the developments outlined only describe the isolated effects of advancing digitization. It is quite possible that other growth-relevant aspects, such as export surpluses and a population that is increasing due to immigration, can overcompensate for this growth-dampening effect in the long term. Secondly, it should be remembered that socio-political decisions significantly influence the concrete effects of digital developments. The social acceptance of the technologically possible developments also plays an important role here: If society does not allow self-driving vehicles for safety reasons, for example, truck, bus, and taxi drivers will not be made redundant either.

Source:

1- Cf. N. Boeing: New jobs for robots, Zeit online from 2014-02-18, (2017-01-24); C. Kurz, F. Rieger: Arbeitsfrei – A voyage of discovery to the machines that replace us, Chapter 4, Munich 2013; D. Marin: The brilliant robots are coming! A. Rolf, A. Sagawe: Google’s core and other spider webs – The architecture of the digital society, Constance, Munich 2015.
2- See L. Jensen: The typewriters, in Brand eins, No. 7, 2015, (24.1.2017).
3- See P. Lassen: High-frequency trading – the turbo algorithm in the stock exchange network, FAZ online of November 28, 2012, (January 24, 2017).
4- See S. Linz, T. Behrmann, U. Becker: Hedonic price measurement for IT capital goods, in: Economics and Statistics, H. 6, 2004, pp. 682-689; and Federal Statistical Office: Information on methods – Effects of digitization on price statistics, Wiesbaden, November 22, 2016.
5- See J. Rifkin: The Zero Marginal Cost Society, Frankfurt, New York 2014, pp. 331-345; W. Eichhorst, A. Spermamann: Sharing Economy – Opportunities, Risks and Design Options for the Labor Market, IZA Research Report, No. 69, Bonn 2015, p. 4; T. Theurl: Sharing Economy: beneficiaries or victims of institutional inconsistencies?, in: Wirtschaftsdienst, vol. 96 (2016), no. 8, p. 605.
6- See J. Rifkin, loc.cit., p. 339.
7- C. Kurz, F. Rieger, loc.cit., p. 142.
8- See T Khemraj, J Madrick, W Semmler: Okun’s Law and Jobless Growth, Schwartz Center for Economic Policy Analysis Policy Note, New York 2006, p 3.
9- See MJ Perry: The US economy is now producing 2.2% more output than before the recession, but with 3.84 million fewer workers, AEIdeas of November 6, 2012 (January 24, 2017).
10- See S Mehrotra et al., Joblessness and Informalization: Challenges to Inclusive Growth in India, IAMR Occasional Paper, No 9/2012, Institute of Applied Manpower Research Planning Commission, Government of India, New Delhi 2012; J. Rifkin, loc.cit., p. 183.
11- See CB Frey, MA Osborne: The Future of Employment: How Susceptible are Jobs to Computerisation?, Oxford 2013, p. 44.
12- C. Brzeski, I. Burk: The robots are coming, ING DiBa – Economic Research, Frankfurt aM 2015, p. 2.
13- See H. Bonin, T. Gregory, U. Zierahn: Transfer of the study by Frey/Osborne (2013) to Germany, ZEW brief expertise for the Federal Ministry of Labor and Social Affairs, No. 57, Mannheim 2015, p. 23.
14- J. Rifkin: The end of work and its future, 4th ed., Frankfurt am Main, New York 1996, p. 107.
15- J. Rifkin: The Zero Marginal Cost Society…, op.cit., p. 195.
16- K. Vogler-Ludwig, N. Düll, B. Kriechel: Labor market 2030: Economy and labor market in the digital age – Forecast 2016, Munich 2016, p. 12.
17- Ibid., p. 16.
18- See The Boston Consulting Group: Inside Ops – Are your Operations ready for a digital revolution? Boston 2016, pp. 6-7.
19- See I. Wolter et al.: Industry 4.0 and the consequences for the labor market and the economy, IAB Research Report, No. 8/2015, Nuremberg 2015, p. 63.
20- See A. Rolf, A. Sagawe, loc.cit., pp. 205-208.