ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846
Reader Mode
-A A +A

Robot Apocalypse

How Will Automation Affect India’s Manufacturing Industry?

Sunil Mani (mani@cds.edu) is with the Centre for Development Studies, Thiruvananthapuram.

Anxiety about the prospect of technology displacing jobs on a large scale is currently dominating academic and public debate. A number of different occupations are likely to see an increased rate of automation in the near future. However, while studies have shown that this is likely to have an adverse effect on employment, they have all used the occupation-based approach to arrive at their conclusions. A task-based approach is used to arrive at a more accurate estimate of the effect of automation on manufacturing employment in India. Employing a comprehensive data set from the International Federation of Robotics, the nature and extent of diffusion of industrial robots into the manufacturing industry in India is also analysed.

Earlier versions of this paper were presented at the Centre for Development Studies, Thiruvananthapuram and at the National Graduate Institute for Policy Studies, Tokyo, Japan. Comments received from the anonymous referee have also been helpful in improving the arguments and presentation.

The initiation of the “Make in India” programme is an indicator of the current Indian government’s desire to increase employment in the country through the manufacturing route. Under this programme, the manufacturing sector is expected to contribute to at least a quarter of India’s gross domestic product (GDP) by 2020. However, due to the capital-intensive nature of manufacturing, employment generated by the sector so far has been minimal. The pessimism surrounding this issue has been accentuatedby the increasing amount of automation in manufacturing processes elsewhere in the world. Industrial automation is thought to have a deleterious effect on the creation of employment in different sectors of the economy, manufacturing included. This has given rise to an important debate, primarily in the context of developed countries where industrial automation has diffused manifold over a long period of time. This debate, which began in the popular press, has now been brought to the formal academic table by the publication ofan influential and highly-cited piece of research by Frey and Osborne (2013). Subsequently, the Journal of Economic Perspectives organised a symposium on “automation and labour markets” in its summer 2015 issue.1 In the wake of the symposium, a series of studies by academic economists and multilateral institutions, such as the Organisation for Economic Cooperation and Development, have also been published (Acemoglu and Restrepo 2017; Autor 2015; Brynjolfsson and McAfee 2014; Chang et al 2016; Hallward-Driemeier and Nayyar 2018).

Given this context, this paper seeks to understand the extent of diffusion of automation technologies in Indian manufacturing and analyse its effects on employment in the sector. The paper is structured as follows. The first section discusses the concept of automation and identifies the specific automation technology that we consider in the present study. This is followed by a discussion of the motivation for the present study and the significance of the issues being dealt with. The next section delineates the major research questions raised, the methodology adopted to answer them, and the data sources employed. The section that follows engages with existing literature on the diffusion of automation technologies in manufacturing. The next section reports the main findings of our analysis with respect to Indian manufacturing. Finally, the implications of future developments in automation technologies on the conclusions reached in the previous section are discussed.

Understanding the Concept of Automation

Industrial automation involves a range of hardware and software technologies. The employment implications of these automation technologies vary considerably. The specific automation technology that has the most direct impact on employment is the multipurpose industrial robot. The International Federation of Robotics (IFR 2014) defines an industrial robot as “an automatically controlled, reprogrammable, and multipurpose [machine].” In other words, industrial robots are fully autonomous machines that can be programmed to perform manual tasks such as welding, painting, assembling, handling materials, or packaging. Unlike other automation technologies, such as machine tools, programmable controllers, or computer-aided design equipment, industrial robots do not require human operators. They can perform reliably and consistently in harsh, and constrained environments in which a human worker cannot function satisfactorily. Therefore, robots represent the most advanced and flexible form of industrial automation that can be envisioned. So, in the present study, we focus on industrial robots. In addition to industrial robots there are service robots as well. There are two conceptions of industrial robots: delivered (flow) and operational stock (stock). Since we are interested in employment implications, our primary focus is on operational stock of multipurpose industrial robots in the manufacturing sector in India, as this should give us a more accurate picture on employment implications.

Significance of the Study

In recent years, there has been a revival of concerns that automation and digitisation might result in a future with fewer and fewer jobs. These fears have been fuelled by studies from the United States (US) and Europe which have argued that a substantial share of jobs is at “risk of computerisation.” Adopting the occupation-based approach proposed by Frey and Osborne (2013), these studies assume that whole occupations rather than single job-tasks are automated by technology. This paper argues that this approach might lead to an overestimation of job automatability, as occupations categorised as high-risk often involve a substantial amount of tasks that are hard to automate.

In the Indian context, understanding the relationship between automation and employment is essential for the following reasons:

(i) Globally, fears about the effect of automation on employment are increasing. An extension of the earlier Frey and Osborne (2013, 2017) study showed that a whopping 69% of all jobs in India are considered automatable.

(ii) Industries such as computers and electronic products, electrical equipment, appliances and components, and transportation equipment and machinery are the most prone to automation. In many countries, including India, these four industries—in particular the transportation equipment industry—have been emphasised heavily in industrialisation strategy.

(iii) Automation potential is concentrated in countries with large populations or high wages. India, therefore, is a good candidate.

(iv) Of late, India’s economic policy has focused on raising employment by promoting the growth of the manufacturing industry, but there has been a steady decline in the labour intensity of manufacturing employment (Sen and Das 2014).

(v) Within India, there is now a debate as to whether employment has really increased in recent times, especially after economic liberalisation. The term jobless growth has been used to refer to a situation where the employment elasticity of output has been shown to be declining (World Bank 2018).

The rate of diffusion of automation technologies is likely to increase in the manufacturing sector in the near future. The following factors highlight the significance of the study in this context:

(i) A country such as India, which has chosen to focus on manufacturing quite late in its development, can skip stages and start with the latest manufacturing technologies.

(ii) Due to globalisation, there is increasing pressure on manufacturing companies to be more productive and internationally competitive. Therefore, the pressure to adopt productivity-
enhancing technologies is much greater now than ever before. According to estimates by Boston Consulting Group (2015), use of robots can decrease labour costs by as much as 16%.

(iii) With recent developments in artificial intelligence (AI) and machine learning, the variety of tasks that machines can do has seen a quantum jump. For instance, industrial robots are now much more intelligent and can perform a number of operations which they could not do earlier.

(iv) The declining cost of automation and its increasing availability is another factor that can hasten the rate of diffusion of robots. According to Boston Consulting Group (2015), the average price of industrial robotic systems has declined from $1,82,000 in 2005 to $1,33,000 in 2014 (Sirkin et al 2015).

These issues motivate us to understand how automation is being used in Indian manufacturing, and its actual and potential effects on employment in the sector.

Research Methodology and Data Sources

The study attempts to answer the following research questions. At what rate have automation technologies diffused into Indian manufacturing over the period since the liberalisation of India’s economy? What effect has this diffusion had on manufacturing employment? What is the relationship between the rate of diffusion of automation and the intensity of manufacturing employment? What are the likely trends in this relationship in the years to come, when the size and composition of manufacturing is bound to increase and become more sophisticated?

In order to understand the diffusion of industrial robots into manufacturing, we adopted a task-based approach because an occupation may contain several tasks that are not prone to easy automation. Therefore, the task-based approach may provide us a more accurate picture of the diffusion of automation. We measure diffusion through the operational stock of industrial robots per unit of employment. This framework is based on the work of Arntz et al (2017).

The primary data source is World Robotics, an annual survey of robotics that covers 32 countries, including India. It presents industry-wise and country-wise annual and time series data on the number of delivered robots and the operational stock. It also offers data on the task-wise distribution of robots within occupations. When calculating operational stock, it is assumed that the average service life of a robot is 12 years and that there is an immediate withdrawal of the robots after the specified period. Where countries actually do surveys of robot stock or have their own routines for the calculation of operational stock, such as in Japan, the resulting data is used. The survey has been published since 2006, with the latest edition covering providing information up to 2016. It is based on consolidated data provided by industrial robot suppliers worldwide, which is collected by national robot associations and robot suppliers, and then processed by IFR’s statistical department. The data is internationally comparable and thus allows analysis of the distribution of robots worldwide and in individual regions and countries. Further, it provides data on the density of robots per unit of employment in the industrial sector and tasks within industries where these installations are available. For India, data on operational stock and shipments is available since 1999. Industry-specific data is available since 2006, while task-specific data is only available since 2011. The source of the data on employment that is used for computing the density of robots is not always mentioned. Data on manufacturing employment has been taken from the Annual Survey of Industries, which offers industry-wise data up to and including the fiscal year 2014–15.

Literature on Automation and Employment

The literature on the effect of automation on employment in manufacturing is a subset of the larger discussion about the effect of technological development on employment creation. This literature has developed in two phases. The first phase began in the late 1980s, when the first of the international studies on the effect of industrial automation on employment was completed (Flamm 1988). The second phase began in 2013 with the publication of the Frey and Osborne (2013) study. The publication of this study unleashed a wave of concern about the deleterious effect of faster diffusion of automation technologies on employment in manufacturing. In turn, this has spawned a number of studies analysing the effects of automation on employment. These studies can be divided into three sets: the first set of studies analyses the diffusion of industrial robots in a range of countries, the second set shows an inverse relationship between the extent of diffusion of automation and employment in manufacturing, and the third set shows that increased automation has not really resulted in hefty job losses. In the following section, we review these studies at length.

Diffusion into developed countries: One of the earliest studies on the changing patterns of industrial robot use is the one by Flamm (1988). He analysed the rate of diffusion of robot use in Belgium, France, Germany, Italy, Japan, Sweden, the United Kingdom (UK), and the US between 1970 and 1984. His survey focused on two issues: how and where industrial robots were being used in manufacturing and how robot use in the US compared with manufacturing practices abroad. Robot use is uneven across industries, with their use being confined to or concentrated in certain specific tasks and industries. Historically, they were first used—in relatively small numbers—in hazardous and unpleasant operations associated with metal processing. After 1975, Japanese auto manufacturers began to use them in large numbers in spot welding operations on their assembly lines. Later in that decade, they expanded the field of application to arc welding. Their foreign competitors followed suit. In fact, it was welding activities that hastened the diffusion of industrial robots across the developed world. After 1980, more sophisticated industrial robots began to be used in the electrical and electronics industries, with Japanese manufacturers at the forefront again. According to Flamm (1988), the majority of industrial robots are found in electronics assembly and automotive welding. The use of robots has not diffused further because there are only a handful of major use cases in which they are currently a cost-effective solution to manufacturers. In fact, industrial robot use has not increased consistently, but in fits and starts. In this context, Flamm is of the opinion that

one would be well advised to be sceptical of technological optimists who, on the basis of broad statistical job classifications for industrial workers, project veritable tidal waves of robots inundating manufacturing in the medium term.

Another interesting finding is that the diffusion of robot use has lagged in the US manufacturing compared to Japan, Sweden, and Germany. Global variations in the relative prices of capital, labour, and other factors of production do not seem to explain the differential rate of diffusion. The shift to larger product varieties that require more flexible manufacturing plants may be a plausible explanation.

Inverse relationship between automation and employment: The World Economic Forum (WEF 2016) conducted a large-scale survey of major global employers across 15 major developed and developing nations—which included the 100 largest global employers in each of the WEF’s main industry sectors—to estimate the expected level of changes in job families between 2015 and 2020, and extrapolate the number of jobs gained or lost. It found that automation and technological advancements could lead to the loss of more than 5.1 million jobs due to disruptive labour market changes between 2015 and 2020 (WEF 2016). A total of 7.1 million jobs are expected to be lost—two-thirds of which are concentrated in the office and administrative job family—while a total of 2 million jobs will be added.

The McKinsey Global Institute (2017) conducted a similar survey covering 46 countries which account for about 80% of the global labour force. The study showed that almost half of all work activities can potentially be automated using current technology. Technically speaking, automatable activities touch 1.2 billion workers and $1,404 trillion in wages. China, India, and the US account for over half of the automatable jobs. The study also notes that automation could boost global productivity by 0.8% to 1.4% annually.

A more systematic study of the diffusion of robots in the US manufacturing was conducted by Acemoglu and Restrepo (2017). It analysed the impact of robot use on the US labour market between 1990 and 2007. Using a model in which robots compete against human labour to perform certain tasks, they showed that the local labour market effects of robots can be estimated by regressing the change in employment and wages on the exposure to robots in each local labour market based on the national penetration of robots into each industry and the local distribution of employment across industries. Using this approach, they estimated large and robust negative effects on employment and wages across commuting zones resulting from robot use. They supplemented this evidence by showing that the commuting zones most exposed to robots in the post-1990 era did not exhibit any differential trends before 1990. The impact of robots is distinct from the impact of imports from China and Mexico, the decline of routine jobs, offshoring, other types of IT capital, and total capital stock (in fact, exposure to robots is only weakly correlated with these other variables). According to their estimates, one more robot per thousand workers reduces the employment to population ratio by 0.18–0.34 percentage points and wages by 0.25%–0.5%.

Occupation-based vs task-based approach: The main problem with these studies is that they consider very broad occupations and not tasks within occupations. In short, they follow the occupation-based approach of Frey and Osborne (2013). Very often, the assumption that whole occupations can be automated by technology is invalid. Rather, it is typically only specific job-tasks that are prone to automation. This approach may lead to an overestimation of job automatability as occupations labelled as high-risk often still contain a substantial share of tasks that are hard to automate.

An important international study that used a task-based approach was conducted by Arntz et al (2017). It concluded that automation and digitisation are unlikely to destroy large numbers of jobs. However, low-skilled workers are likely to bear the brunt of the adjustment costs as the automatability of their jobs is higher compared to highly skilled workers. Therefore, the likely challenge for the future lies in coping with rising inequality and ensuring sufficient retraining, especially for low-skilled workers.

Graetz and Michaels (2017) conducted another study that used IFR data on robot adoption from 1993 to 2007 to analyse industrial robot use across 17 countries. Employing panel data on robot adoption within industries in those countries, and new instrumental variables that relied on the robots’ comparative advantages in specific tasks, the study found that increased diffusion of robotic technology contributed approximately 0.37 percentage points to annual labour productivity growth. Simultaneously, it raised total factor productivity and wages while lowering output prices. Further, the estimates suggest that robots did not significantly reduce employment, although they did reduce low-skilled workers’ employment share.

The following inferences can be drawn from the studies that we have reviewed here:

(i) Industrial robots are used in specific industries such as automobile, electrical and electronics, and metal working. Even within these industries, they are used only for certain tasks—like spot and arc welding—which are harsh on human beings and are highly repetitive tasks. In fact, their usage does not seem to have diffused into other manufacturing industries over the last four decades.

(ii) Studies which have analysed the relationship between diffusion of automation and employment have got results which are diametrically opposite. Some studies have got an inverse relationship between the two variables, while others have not detected any such relationship. Careful analysis of the former studies shows that they have used an occupation-based approach while dealing with employment as opposed to a task-based approach. The occupation-based approach tends to exaggerate the impact of automation on employment.

(iii) The proxy that is used for identifying automation has varied across studies. Some studies define automation in terms of computerisation, while others identify it in terms of use of industrial robots.

(iv) All the studies—without exception—deal with developed market economies. None of them refer to any developing countries.

Thus, our engagement with the literature shows that there is a real need for a study analysing the relationship between automation and manufacturing employment in the context of a late-industrialising country such as India. Further, in our study, we define automation in terms of its highest form, namely the use of industrial robots. And we use a task-based approach to measure the effect of automation on employment, which should provide more meaningful results.

Main Findings for India

The operational stock of industrial robots in manufacturing has been increasing in India as well as globally. According to the IFR (2017), five major markets—China, Korea, the US, and Germany—accounted for 74% of the total sales volume in 2016. China has become the largest market, representing almost a third of the total market in 2016. At a total of 87,000, the number of industrial robots sold in the country is almost equivalent to the total number of industrial robots sold in Europe and the US together. Apart from China, the other important markets for industrial robots are Japan, the US, the Republic of Korea, and Germany (Figure 2). In India, operational stock increased from just 70 in 2000 to 16,026 in 2016—an annual growth rate of 44% (Figure 1, p 43). Industrial robot sales in India increased by 27% in 2016.

Despite having the largest operational stock of robots, China was still below the world average in terms of robot density in 2016. India appears to have the lowest density, although this is the result of an underestimation which will be made clear in our own estimates of the density of industrial robots. The Republic of Korea has had the highest robot density in the world since 2010 (Figure 3). In 2017, the country had 710 industrial robots in operation per 10,000 employees—up from 311 units in 2010. The increase was due to the continued installation of large volumes of robots, particularly in the electrical and electronics industry, and in the automotive industry. Singapore ranked second with 658 robots per 10,000 employees in 2017. The country’s high robot density is down to the low number of employees in its manufacturing industry—approximately 2,40,000 according to the International Labour Organization—and the large number of installed robots, which has increased significantly in recent years. About 90% of the robots in use in Singapore are installed in the electronics industry. It is interesting to note that China has a density of 97, well above the world average of 85 per 10,000 employed.

It will be instructive to analyse whether the industries where robots are used—and the robot-dependent tasks within these industries—have undergone any changes since Flamm (1988) made his observations in the late 1980s. Three industries account for about 80% of the entire operational stock of industrial robots: metal, electrical and electronics, and automotive (Table 1, p 44). Among these industries, the automotive industry accounts for about 43% of robot usage. In fact, it is the only industry which has increased its share over the years. Interestingly, the same industry accounted for the largest share in the 1980s as well (Flamm 1988). We shall now analyse the tasks or application areas within these industries where industrial robots are used (Table 2, p 44).

Industrial robots are primarily used in two main tasks: handling operations/machine tending, and welding and soldering. There is a remarkable continuity in the use of robots for certain tasks from the late 1980s to 2016. The only difference is that more of them are used for the same kind of tasks. The only task that has increased its share is material handling, which shows that industrial robots are primarily used for tasks which are difficult for human beings to perform.

Firms employ robots to automate specific tasks—many of them harmful to human health. The range of automatable tasks is increasing and will continue to increase through advances in vision and end-effector technologies.2 But this does not imply that jobs will be wiped out.

We now turn our attention to the Indian case. As mentioned earlier, the operational stock of industrial robots in India has seen a tremendous increase since 2000 (Figure 1). However, the rate of growth has been fluctuating (Table 3). The manufacturing sector’s share has been rising steadily and now accounts for about two-thirds of the operational stock. It is also interesting to note that the number of robots being used in both construction, and education/research and development has increased significantly. However, the industry-wise usage numbers are coloured by the large number of robots whose usage cannot be ascribed to any specific sector.

The manufacturing sector accounts for the lion’s share of delivered robots in India (Table 4, p 45). However, within manufacturing, most of the robot installations are in four industries: automotive, electrical and electronics; metal; chemical; rubber and plastics. In 2017, there was a 27% increase in the number of delivered robots compared to the previous year. On average, the number has increased by 64% per annum during the period under consideration.

An analysis of the industry-wise operational stock of industrial robots shows that robot use is highest in the automotive industry, followed by plastics, rubber and chemicals, and metal. In short, it is the growth of the automotive industry that accounts for most of the growth in robot installations in India. Thus, the pattern in India is very similar to the international pattern (Table 5, p 45).

Task-based installations: Robot usage in India is confined to two tasks, namely welding and soldering, and handling and machine tending (Table 6). Within the former, it is almost entirely concentrated in arc and spot welding. Material handling involving plastic moulding and machine tools accounts for the second largest share. This resembles the pattern observed historically even in developed countries.

This finding has deep implications for employment. Industrial robots have hitherto been used for tasks that are inhospitable for human labour and where a lot of precision is required. However, in order to understand the employment implications of robot use, one has to analyse the density of robots per unit of employment.

Density of robots: The density of robots in India is one of the lowest among robot using countries (Table 7). Density is an important indicator of the labour-displacing effect of industrial robot use. Figure 4 (p 47) provides estimates of robot density in two different industries: the manufacturing industry and the automotive industry. Although both industries are showing an increase in robot density, it is much higher in the automotive industry than the general manufacturing industry. Since the automotive industry in India is dominated by affiliates of multinational companies, with the parent companies having a long history of using industrial robots in various manufacturing operations, it is only natural that their affiliates in India will be using industrial robots (Table 8).

The robot densities in these state-of-the-art plants are significantly higher than the average for the Indian automotive industry. Even domestic automotive manufacturers such as Tata Motors are deploying industrial robots, albeit at a lower density level. In a single plant in Pune, Tata is said to have installed 100 robots. According to the company, automation increased turnover by 250% despite the production force being reduced by 20% during the same period.

Highly labour-intensive industries such as paper and wood products, textiles, non-metallic products, food products, metal products, and machinery are the least automated. The most automated industries, such as automotives, rubber and petroleum, basic metals and chemicals, are less labour-intensive. So the effect of automation has an insignificant effect on the quantum of overall employment in the manufacturing industry.

Production of industrial robots in India: Some of the world’s leading firms in factory automation, such as Fanuc, Kuka, Gudel, and ABB, have manufacturing and sales operations in India. One of them has even established a training academy in Pune to train engineering graduates in robotics. These operations could hasten the diffusion of industrial robots in non-traditional industries. Further, TAL Manufacturing Solutions, a subsidiary of Tata Motors, has launched its much-awaited TAL Brabo robot. It comes in two variants with payloads of 2 kg and 10 kg each, and is priced between ₹ 5,00,000 and ₹ 7,00,000. The robot was developed in-house with TAL Manufacturing Solutions doing the design and Tata Elxsi handling the styling and Tata AutoComp manufacturing some critical components of the robot. The TAL Brabo has apparently been developed to cater to micro, small, and medium enterprises, as well as for the use of large-scale manufacturers who require cost-competitive automated solutions in manufacturing. Conceptualised to complement a human workforce and perform dangerous, repetitive, high-volume, and time-consuming tasks, the robot can be deployed across industries. Having successfully tested the TAL Brabo in over 50 customer work streams, TAL Manufacturing is ready to supply these robots to several sectors, including automotive, light engineering, precision machining, electronics, software testing, plastics, logistics, education, and aerospace.

Implications of Automation on labour-intensive Industries

One of the most labour-intensive industries in India is the cotton textile industry—especially ready-made garments. The Sewbot technology, being developed by the American company Softwear Automation, aims to automate the entire clothes-making process. However, the technology is very highly-priced and its diffusion in the textile industry will take years to fructify. There are four processes that go into making an item of clothing: picking up the item, aligning it, sewing it, and disposing of it. Of these, only the sewing has so far been automated—by the sewing machine, which came in a long time ago. The other parts of the process are still done faster and cheaper by humans than by robots.

In fact, the most automated industry in India, the automotive industry, accounts for only about 10% of manufacturing employment in the country (Figures 5a and 5b). We have already seen that even within this industry, only certain tasks are automated.

However, automation technologies are improving fast, with significant developments forthcoming in the areas of AI and machine learning—especially a technique known as deep reinforcement learning—and robotics. Further, improvements in the following eight technologies will have a strong positive effect on faster diffusion of automation technologies: (i) computing performance; (ii) electromechanical design tools and numerically-controlled manufacturing tools; (iii) electrical energy storage; (iv) electronics power efficiency; (v) local wireless digital communications; (vi) internet; (vii) data storage; (viii) computation power.

Another significant factor is China’s increasing share of the robotics industry (Figure 6, p 48). China’s tech industry is shifting away from copying Western companies and has identified AI and machine learning as the next big areas of innovation. Chinese investors are now investing heavily in AI-focused start-ups and by pledging to invest about $15 billion by 2018, the Chinese government has signalled its desire to see the country’s AI industry blossom. The combination of these three factors could make industrial robots more intelligent and capable of performing tasks which were hitherto considered impossible. China’s entry into robotics could make robots much cheaper and make them affordable even to newer industries. Faster adoption of these new automation technologies could have a deleterious effect on employment in Indian manufacturing—especially in labour-intensive industries such as textiles and clothing.

In Conclusion

In this study, we have analysed the possible links between diffusion of automation technologies and employment. Automation technology is narrowly defined in terms of the highest form of automation—namely, the use of industrial robots—primarily because of the availability of reliable data on industrial robot use at the level of tasks within occupations. Analysis of the data shows that although robot density has increased, its usage is restricted to a few industries, with the automotive industry being the most important user. Within the automotive industry, the use of industrial robots is concentrated in certain tasks that are less labour-intensive. So, for the present, automation does not pose a threat to manufacturing employment. However, with rapid developments in technology, the situation could change. Therefore, there is a need for a policy on automation in an economy such as India’s, wherein labour is abundant.

Notes

1 See symposium on “Automation and Labour Markets,” Journal of Economic Perspectives,
Vol 29, No 3, Summer, 2015, https://www.aeaweb.org/issues/381. The three papers in the symposium are by Autor (2015), Mokyr et al (2015) and Pratt (2015).

2 In robotics, an end effector is the device at the end of a robotic arm, which is designed to interact with the environment.

References

Acemoglu, Daron and Pascual Restrepo (2017): “Robots and Jobs: Evidence from US Labor Markets,” National Bureau of Economic Research, Working Paper No 2328, www.nber.org/papers/w23285.

Arntz, Melanie, Terry Gregory and Ulrich Zierahn (2017): “Revisiting the Risk of Automation,” Economic Letters, Vol 159, pp 157–60.

Autor, David (2015): “Why Are There Still So Many Jobs? The History and Future of Workplace Automation,” Journal of Economic Perspectives, Vol 29, No 3, pp 3–30.

Boston Consulting Group (2015): “The Robotics Revolution: The Next Great Leap in Manufacturing,” .

Brynjolfsson, Erik and Andrew McAfee (2014):The Second Machine Age: Work, Progress and Prosperity in a Time of Brilliant Technologies, New York: W W Norton and Company.

Chang, Jay-Hee, Gary Rynjhart and Phu Huynh (2016): ASEAN in Transformation, Textiles, Clothing and Footwear, Refashioning the Future, International Labour Organization, http://www.ilo.org/actemp/publications/WCMS_579560/lang--en/index.htm.

Central Statistical Organization (2015): “Annual Survey of Industries 2014–15,” Ministry of Statistics and Planning, Government of India, Delhi.

Frey, Carl B and Michael A Osborne (2013): “The Future of Employment: How Susceptible Are Jobs to Computerisation?” working paper, Oxford Martin School, .

— (2017): “The Future of Employment: How Susceptible Are Jobs to Computerisation?” Technological Forecasting and Social Change, Vol 114, pp 254–80.

Flamm, Kenneth (1988): “The Changing Pattern of Industrial Robot Use,” The Impact of Technological Change on Employment and Economic Growth, Richard M Cyert and David C Mowery (eds), Cambridge: Ballinger Publishing House, pp 267–328.

Graetz, Georg and Guy Michaels (2017): “Robots at Work,” Discussion Paper 1335, Centre for Economic Performance, London School of Economics and Political Science, http://cep.lse.ac.uk/pubs/download/dp1335.pdf.

Hallward-Driemeier, Mary and Gauvrav Nayyar (2018): Trouble in the Making? The Future of Manufacturing-led Development, World Bank, Washington, DC.

IFR (2017): “World Robotics Industrial Robots, Online Database,” International Federation of Robotics, https://ifr.org/worldrobotics.

Intellectual Property Office (2014): “Eight Great Technologies: Robotics and autonomous Systems,” A Patent Overview, Newport: Intellectual Property Office, UK.

Mokyr, Joel, Chris Vickers and Nicolas L Ziebarth (2015): “The History of Technological Anxiety and the Future of Economic Growth: Is This Time Different?Journal of Economic Perspectives, Vol 29, No 3, pp 31–50.

Mckinsey Global Institute (2017): “A Future that Works: Automation, Employment and Productivity,” New York: McKinsey.

Pratt, Gill A (2015): “Is a Cambrian Explosion Coming for Robotics?” Journal of Economic Perspectives, Vol 29, No 3, pp 51–60.

Organisation for Economic Cooperation and Development (2016): “Automation and Independent Work in a Digital Economy,” Policy Brief on the Future of Work, OECD Publishing, https://www.oecd.org/els/emp/Policy%20brief%20-%20Automation%20and%20Independent%20Work%20in%20a%20Digital%20Economy.pdf.

Sen, Kunal and Deb Kusum Das (2014): “Where Have the Workers Gone? The Puzzle of Declining Labour Intensity in Organised Indian Manufacturing,” Department of Economics and Public Policy Working Paper No 36, Institute for Development Policy and Management, University of Manchester, http://hummedia.manchester.ac.uk/institutes/gdi/publications/workingpape....

Sirkin, Harold, Michael Zinser and Justin Rose (2015): “How Robots Will Redefine Competitiveness?” Boston Consulting Group, https://www.bcgperspectives.com/content/articles/lean-manufacturing-inno....

World Bank (2018): Jobless Growth, South Asia Economic Focus, World Bank, Washington, DC.

WEF (2016): “Five Million Jobs by 2020: The Real Challenge of the Fourth Industrial Revolution,” World Economic Forum, Published 18 January 2016, https://www.weforum.org/press/2016/01/five-million-jobs-by-2020-the-real....

Updated On : 25th Feb, 2019

Comments

(-) Hide

EPW looks forward to your comments. Please note that comments are moderated as per our comments policy. They may take some time to appear. A comment, if suitable, may be selected for publication in the Letters pages of EPW.

Back to Top