Digital World Economy. An economic perspective on industry 4.0

Table of Contents

List of Tables

Table of Figures

Formula Symbols and Abbreviations

1 Introduction
1.1 Hypothesis testing

2 Current State of Research
2.1 Literature Overview
2.2 Taking Stock

3 Theory: Solow Growth Model
3.1 Applied Models: Cobb-Douglas-Production-Function

4 Empirical Investigation
4.1 Methodology correspondent to empirical investigation
4.2 Definitions
4.3 Case Studies of Industry 4.0
4.4 Data
4.5 Results

5 Policy Conclusion
5.1 Outline
5.2 Open Fields and Regulation

List of References

Appendix

Abstract

“The recognition of the importance of the contribution of science and technology to economic growth was exemplified in the frantic search for explanations of the ‘residual’, that part of growth which could not be explained by labour and capital accumulation.” (Freeman & Soete, 1997).

The publicized cost reduction and deregulation of the internet access (Welfens, Internet-economics.net: Macroeconomics, Deregulation, and Innovation, 2002) developed to a considerable extent. Even the poorest economies are catching up. In addition to that, the globalization (e.g. FDI) enabled so-called “spillover-effects”, so that the pillars for a fourth industrial revolution are standing on a fundamental basis. Recent technologies such as the Cyber-Physical Systems arise the question in which dimension the impact of industry 4.0 is observable nowadays and how the future working environment could look like. The internet age 2.0 offers new opportunities for growth, international cooperation, and democracy, but also raises major challenges.

Researcher teams are still discussing the impact of information and communication technology (ICT) on the econimic growth, while the industry 4.0 is resounded throughout the land. The aim of the present paper is to provide an economic perspective of the current development of industry 4.0. A theoretical basement is discussed and adjusted for an empirical investigation. The insights should help to calibrate the actual debate about industry 4.0 into a solid quantified direction and to set up a fundamental basis for feasible forecasts on the “industry 4.0 impact”.

Keywords Industry 4.0 Development ◌ Economic Future ◌ Industrial Revolution ◌ Economies of Scale 4.0 ◌ Economic Pespective on Industry 4.0

List of Tables

Table 1: Estimates of trends in per capita GNP (1750 – 1977)

Table 2: Definition of required Cobb-Douglas model variables

Table 3: Results of the regressions for Germany; France
Source: Own calculations on EU KLEMS database

Table 4: Results of the regressions for the UK; USA
Source: Own calculations on EU KLEMS database

Table 5: Literature Overview of much quoted research papers, 01/2018

Table 6: Growth accounting approaches of the EU KLEMS release

Table 7: EU KLEMS industries

Table 8: EU KLEMS aggregates

Table 9: Results of the regressions for Germany; France

Table of Figures

Figure 1: Cluster analysis with total number and fraction of relevant articles

Figure 2: The development of industry 4.0 from its roots

Figure 3: Mondi, Machine Learning, First Case Study

Figure 4 : Horizon Wind Energy, Forecast & Risk Analysis Tool, Second Case Study

Figure 5: Gross output development for five selected countries

Figure 7: Non-ICT-capital and industry 4.0 capital contribution to value added growth of Italy

Figure 8: Non-ICT-capital and industry 4.0 capital contribution to value added growth of the USA

Figure 9: 2000 to 2014 Profitability & Asset Turnover for selected Countries.

Figure 10: Three options for development

Figure 11: Contributions to VA growth, LP1 growth, LP2 growth, and GO growth by country and year

Figure 12: Positive prospects for talent and employment

Formula Symbols and Abbreviations

Formula symbols:

Abbildung in dieser Leseprobe nicht enthalten

Abbreviations:

Abbildung in dieser Leseprobe nicht enthalten

1 Introduction

The elementary technology the “Internet” enabled the fourth industrial revolution. It changed the way people work. The collaboration in server-clouds and the so-called “Internet of Things” allows even machines to interact. Figure 3 provides the historical development of the fourth industrial revolution initiating in the 1970’s. Many questions are still unsolved, e.g. the key technologies or their impact on the economic growth of economics.

Nowadays the transformation process is ongoing while the impact of this technological change is still unclear. The McKinsey consulting group estimates that just the internet accounts for 21 percent of the GDP growth in major economies over the past five years (Manyika & Roxburgh, 2011).^[1] Furthermore, researcher mention that the share of ICT capital is often underestimated. In real terms (evaluated in prices in 1995), the share of ICT was roughly twice as high, because of the continuously falling prices for ICT technologies and services (Schreyer & Colecchia, 2002) (Welfens, et al., 2012).

Fifteen years prior, the landscape internet-driven companies seemed to stop the growth path after the “Dotcom-Crash”.^[2] Before, the expectations of immense productivity gains were higher than the possibilities of the technology. Also, the average employee was not ready to handle the challenges of digitalization, robotic and the pace of technological change. In 2017 digitalization and automatization literature provides a wide range of IT-specific but also managerial-specific guidance. Available economic papers refering to industry 4.0 are comparatively few. And if available, then are more focused on the impact of ICT on economic growth.

This paper tries to fill the gap to the point of a political-economic perspective on industry 4.0. The object of interest is particularly productivity changes caused through digitalization and automatization which can be summed in the main driving forces of industry 4.0. The initial point of the investigation is the theoretical framework of the Cobb-Douglas-Production-Function. It will be converted in an applied model form, which could be tested with the EU KLEMS database. To justify a claim of this kind, growth accounting estimates are used. The traditional neoclassical methodology determines a contribution to growth from an exogenous technological change in the Solow residual estimate of total factor productivity (TFP) growth. The Solow model applies the standard supply side Cobb-Douglas production function . When the Solow residual was first proposed the startling, a claim was over eight decades a very high proportion, perhaps 90 percent of labour productivity contribution of technological change or a statistical measurement mistake caused by omitted variables. For instance, education or more specific “ICT-related” education.

Afterwards, the reflection of technological change (A) was utterly underappreciated. The best-known example is the so-called Solow Productivity Paradox coined in 1987. In this time computers started to diffuse in most business areas, while the productivity statistics did not reflect a productivity increase (Solow 1987). Some questions still need answers.

1.1 Hypothesis testing

Since the early days of growth accounting, there have been several important developments. Nicholas Crafts pointed out three major developments to modern days economic perspective of industry 4.0 (Crafts, Quantifying the Contribution of Technological Change to Economic Growth in Different Eras: A Review of the Evidence, 2003). First, there is now a wide array of historical national accounts and a great many growth accounting exercises have been carried out so that a much broader historical perspective is now possible. Second, better measurement of the key aggregates is now feasible in terms of considering changes in the quality of inputs and many kind of cross-checks have been developed that can be used to evaluate the plausibility of estimates of TFP growth. Third, the theory of growth accounting and its relationship to growth economics is better developed so that the distinction between TFP growth and underlying technological change is now much clearer.

Therefore, the opportunity to review what quantitative economic history nowadays could deliver about the contribution of the technological change to the growth of economy should be explored more in detail. The time series evidence can provide a key ingredient for the assessment of several claims in the endogenous growth literature. Accordingly, in the context of this work two hypotheses will be addressed:

(1) Based on traditional growth accounting, by how much percentage of the contribution to productivity growth is impacted because of TFP growth? (Impact of industry 4.0 should be included somewhere in TFP)

(2) Accepting traditional neoclassical methods, would the resulting TFP growth estimate a reliable guide to the contribution of technological change to productivity advance?

2 Current State of Research

This chapter provides an overview of relevant studies, the actual theoretical discussions and their empirical applications in the field of industry 4.0 research. In 2.1 a literature overview sums up the most important findings plus issues in this research field to classify the current state of industry 4.0 research in chapter 2.2.

2.1 Literature Overview

Besides the examples from empirical investigations, several studies engage in the topic-complex. In general, for the research field industry 4.0 holds that there exist just a few specific and comprehensive literature refering the economic perspective of industry 4.0. In most cases, the literature covers technological topics or custom-designed approaches. However, studies and research papers engage in this broad field. Notably, in the last years the quality of ICT-related publications improved just as well as the econometric application. One reason is the actuality of the topic but also the rising database for ICT and industry 4.0 sectors. Hence, nowadays the precision and quality of economic perspectives on industry 4.0 should be higher than one decade before. Also, the companies attend to the digital transformation. Their research and investments enlarge the horizon as well.

On the other hand, there are still difficulties located in contradicting study results. Therefore a literature overview including the most important components of the research field should help to give a comprehensive view of the theoretical background of the topic. Table 5 raises this claim. It presents the often-quoted references about industry 4.0 or related topics such as ICT. The assessment of this overview takes place in chapter 2.2, “Taking Stock” to provide the actual status of industry 4.0 research state and the connecting factor to the present paper.

In 2014 some researchers already did this work before^[3] in a more technical related context of industry 4.0 topics (Brettel, Friederichsen, Keller, & Rosenberg, 2014). Their results are furnished in the following Figure 2. It seems that the most important topic is the “individualized production” in this the research field with approx. 42% of the articles. On second place is the “horizontal integration in collaborative networks” (35.6%). On the last place with 23.4% is the “End-to-End digital integration” topic.^[4] However, the fraction of relevant articles is quite low. This approves the broad uncertainty about the object of research. Some assumptions and forecasts possess a lack of research work.

Moreover, it must be stressed that measuring the macroeconomic impact of the internet, ICT and/or industry 4.0 transformation requires a large, consistent, underlying dataset to produce econometrically solid and robust results. Given that the internet of Things (IoT) is a relatively new, and rapidly growing phenomenon this data requirement cannot be fully met. Hence the results presented above cannot be interpreted as a robust confirmation of the causal effect that the IoT has on economic growth, but as a preliminary indication of such an effect.

Extended studies and data could close the gap in the next years. The EU commission provides the EU KLEMS database concerning this issues. This data source forms one of the pillars of the present study, because it is good conditioned and relatively actual since the update from September 2017. The data analysis happens in chapter 4 “Empirical Investigation”. The following subchapter 0 “

Taking Stock” is going to illustrate a more current and more economical literature overview to converge to the object of interest of this paper and to take stock the concerned research gap.

2.2 Taking Stock

Table 5 constitutes the basis of this sub-chapter because it contains the actual papers and opinions about industry 4.0 or related topics. Not surprisingly there are several opinions about industry 4.0 caused by the complexity of this topic. Basic research must be done further for several aspects of industry 4.0 topics. Chapter 5.2 “Open Fields and Regulation” finishes the present paper and illustrates the open fields of research that are recommended.

Some papers have attempted to measure organizational complements directly, and to determine whether they are correlated with information technology investment, or whether firms that combine complementary factors have better economic performance. Finding correlations between information technology and organizational change, or between these factors and measures of economic performance, it is not sufficient to prove that these practices are complements unless a full structural model specifies the production relationships and demand drivers for each factor.

Other studies have attempted macro-economic theories like growth accounting approach, which seems to generate the most solid results. In Table 5 is also a summary of the main findings and the proceeding included. The most studies such as ( O'Mahony, Matteucci, Robinson, & Zwick, 2005) used the growth accounting approach, which is applied in this paper.

The Divergence of Firm Level and Aggregate Studies on Information Technology, Industry 4.0 and Productivity

While the evidence indicates that information technology has created substantial value for firms that have invested in it, it has sometimes been a challenge to link these benefits to macroeconomic performance. A major reason for the gap in interpretation is that traditional growth accounting techniques focus on the (relatively) observable aspects of output, like price and quantity, while neglecting the intangible benefits of improved quality, new products, customer service, and speed. Thus, some papers just illustrating the growth rates of industry 4.0 (or ICT) related sectors.

Similarly, traditional techniques focus on the relatively observable aspects of investment, such as the price and quantity of computer hardware in the economy, and neglect the much larger intangible investments in developing complementary new products, services, markets, business processes, and worker skills (Antonelli, 2003). Paradoxically, while computers have vastly improved the ability to collect and analyze data on almost any aspect of the economy, the current computer-enabled economy has become increasingly difficult to measure using conventional methods. Nonetheless, standard growth accounting techniques^[5] provide a useful starting point for any assessment or for the contribution of information technology to economic growth. Several studies of the contribution of information technology concluded that technical progress in computers contributed roughly 0.3 percentage points per year to real output growth when data from the 1970s and 1980s were used (Jorgenson and Stiroh, 1995; Oliner and Sichel, 1994; Brynjolfsson, 1996). Much of the estimated growth contribution roots directly in the large quality-adjusted price declines in the computer producing industries^[6] (Schreyer & Colecchia, 2002). The nominal value of purchases of information technology hardware in the United States in 1997 was about 1.4 percent of GDP. Since the quality-adjusted prices of computers decline by about 25 percent per year, simply spending the same nominal share of GDP as in previous years represents an annual productivity increase for the real GDP of 0.3 percentage points (that is 1.4 multiplied by 0.25 = 0.35).

Another point that must be considered is the difficult separability of industry 4.0 and ICT topics, because of two reasons. First, the internet is one of the pillars of both “topics” which cannot be allocated exactly (in terms of fractions) to more ICT or more industry 4.0. Second, the lack of data for industry 4.0 calculations requires some pragmatic approaches, e.g. the use of a digital index as a proxy for technology in the growth accounting approach (Moroz, 2017).

Table 5 presents some other issues like the inconsistent meanings of several topics in the field of industry 4.0 research like composition of variables, theory approach, weighting of industry impact (e.g. some state that ICT should be higher weighted as a contributor of growth, because of the ICT price decline (Welfens & Perret, Information & Communication Technology and True Real GDP. Economic Analysis and Findings for Selected Countries, 2014)). After considering the actual research discussions, chapters 3 and 4 are based on the Solow Growth Model including an applied Cobb-Douglas-Production-Function.

3 Theory: Solow Growth Model

The basic version of the Solow Growth Model is a suitable anchor point for an economic perspective on industry 4.0. It provides the theoretical framework within which policy issues are considered. In general, the model is written as:

Where is real output (measured as real value added (VA)), and are capital and labour inputs, respectively. , representing technology, is the crucial input factor where the focus of analysis would be. It includes ICT, which also covers large parts of industry 4.0. The neoclassical assumptions of the model are well known; and therefore, the application of the model for the empirical investigation is conducted accordingly. For comparison between different economies, the model inputs are described as per head factors:

(2)

Here the small letters stand for the per-headfactor. Robert Solow and Moses Abramovitz (Romer, 2002) (Solow, 1957) applied the Growth Accounting form of the model to basically analyze to what extent the economic growth would be explained by changes on capital, labour and, other input factors.

(3)

The variable is the elasticity of production in relation to capital. Then, the economic growth can be divided in per-head-growth, because of the per-head-capital accumulation and an additional term , the so-called “Solow-Residuum”. It can be intepreted as the contribution of technological progress to growth. In fact, it is a collective term for all factors, which lead to economic growth. Growth accounting literature (Solow, 1957) provides a straightforward way of decomposing output growth into numerous factors in the aggregate production function:

Abbildung in dieser Leseprobe nicht enthalten

Where α is the fraction of output that is contributed by the capital input. (1-α) is the fraction that is contributed by labor input. The output growth is segregated into three factors: (K) is capital input, (L) labor input and (z) is the total factor productivity. Total factor productivity (TFP) is the so-called “Solow residual” since it is measured as a residual in the Cobb-Douglas Production Function: (5)

As a residual, TFP includes all the factors except capital and labor input, such as technical change. Hence, digitalization and automatization impacts, which are the most important components of industry 4.0, should be included here. The key insight of the Solow Growth Model is that, if the growth in TFP continues, capital per unit of labor will increase continuously. So does output per unit of labor. This is because, given the quantity of capital and labor input, an increase in TFP will increase the marginal product of labor. Consequently, a measurement by using an applied Cobb-Douglas-Production-Function is the pragmatic approach to measure the impact of industry 4.0.

3.1 Applied Models: Cobb-Douglas-Production-Function

The growth accounting approach to TFP estimation has been extensively employed to estimate the impact of ICT capital deepening on output and productivity growth (see e.g. (Timmer, O'Mahony, & Ark, 2007); (Oliner & Sichel, 2000); (Crafts, Quantifying the Contribution of Technological Change to Economic Growth in Different Eras: A Review of the Evidence, 2003)). It is useful because it allows for the decomposition of output and labour productivity growth into contributions from factor inputs and underlying productivity growth or TFP. Assuming the production function of an economy (j) can be written as:

(1)

Where Q is real output (here measured as real value added (VA)), K and L are capital and labour inputs, respectively, and A is an index of technical progress or TFP. Under assumptions of perfectly functioning markets and constant returns to scale, differentiating (1) with respect to time, yields an index of TFP growth known as the Divisia index (Star & Hall, 1976). This index is a valid measure of TFP growth in continuous time, regardless of the functional form of the production function. In practice, as changes are not observed continuously, an approximation is required. Assuming a translog production function, the Törnqvist index is the appropriate approximation of the Divisia index (Jorgenson et al., 1987), and output growth may then be decomposed into its various components in the following way:

(2)

Term expresses the share of labour in VA averaged over two-time periods. Incorporating quality adjustments to inputs is originally developed by Jorgenson and Griliches (Jorgenson & Griliches, 1967). In studies relating to the impact of ICT on productivity, this has involved quality adjustment of capital, accounting for substitution between modern technology and traditional capital. This approach is also adopted here but in addition, this analysis includes a labour quality adjustment, which is a refinement to many of the earlier studies on the impact of ICT. To incorporate quality adjustments to inputs, the growth in aggregate labour and aggregate capital can be estimated as Törnqvist indexes of their components.

Suppose there are l types of labour and k types of capital. Then these indexes are given by:

Abbildung in dieser Leseprobe nicht enthalten

Where the share of type l labour in the total is wage bill and is the share of type k capital in the value of capital. Thus, if the employment of (highly paid) skilled labour is growing faster than unskilled labour, weighting by wage bill shares leads to faster growth in labour input than a simple count of hours worked.

The study uses a standard supply side Cobb–Douglas framework. In its general form equation (2) can be log-linearized, in a form amenable to direct estimation, where is considered as the technological impact of industry 4.0 on industry growth. Although most research papers apply Cobb-Douglas Production function for estimating the impact of ICT on productivity growth, on this work a practical approach of reconstructing the ICT sector by applying some adjustments to transform same to an industry 4.0 approach would be applied.

First, the industries which are not considered as industry 4.0 related, e.g. the telecommunication industry, are excluded. Second, additional industries like robotic are higher weighted. The exact definition of industry 4.0 is stated in the following chapter 4.2. Accordingly, the regression model is adjusted by including the error term to model (4). This model constitutes the first model for the empirical investigation and it is Hicks neutral^[7]:

(4)

The second model uses already assigned Cobb-Douglas-Function (derived from (Welfens, Innovations in Macroeconomics, 2008) where the input factors define the productivity level. This version of production function is Harrod neutral and in dependence on (Aghion & Howitt, 2007). An increase in one of the factors leads to a Gross Output (Q) increase. The focus is on the use of industry 4.0 technology not only a simple (biased) focus on production of these technologies. Rewriting this function induces the following equation, which is the starting point of the derivation of model two:

(1)

Capital is splited into non-industry 4.0 related capital ( and industry 4.0 related capital . The elasticities α and β are required to measure the impact of industry 4.0 on the productivity. Whereas term (1 – α – β) measures the part of productivity which is the residual of the remaining input factors knowledge (A) and labour (L).

Three proofs can be derived from equation (1) to allow the calculation of growth rates for each input factor:

Abbildung in dieser Leseprobe nicht enthalten

(2)

Equation (2) provides information about the growth rate of industry 4.0 related capital ( and equation (3) estimates the next growth rate for non-industry 4.0 related capital (. Finally, equation (4) provides the remaining part of the growth rates ( for input factors (A*L).

Abbildung in dieser Leseprobe nicht enthalten

(3)

Abbildung in dieser Leseprobe nicht enthalten

(4)

Thus, equation (5) is considered as the second model for the empirical investigation. The share of industry 4.0 would be indicated by the elasticity β and share of other capital would be represented by elasticity α.

(5)

The sum of all growth rates adds up to the growth rate of the productivity (5a):

(5a)

Where the term (A+L) is measured in wages. In addition to a wage increase, this reflects the productivity increase that is caused mainly by technological improvement (A).

Considering proof (1), this equals the growth rate of non-Industry 4.0 capital leads likewise to the growth rate of Q:

(5b)

Exact definitions are included in the next chapter 4.2, while detailed information about the composition of variables and the data are incorporated in chapter 4.4.

[...]

^[1] For the period from 2005 to 2010.

^[2] (Geier, 2015)

^[3] Using a cluster analysis for 5911 articles

^[4] Average calculations

^[5] E.g. the Solow Growth Accounting Approach

^[6] Or more general: ICT hardware

^[7] Cp.with (Welfens, Innovations in Macroeconomics, 2008)