Robotics: Manufacturing A Technology Policy Blueprint
POLICY ANALYSIS, IMPLICATIONS & RECOMMENDATIONS
Introduction – The Robotics Revolution
When Chinese President Xi Jinping visited the University of Science and Technology of China in 2017, he was greeted by two humanoid-robots named “Jia Jia” and “Xiao Man” – “Hello Mr. President, we are pleased to be part of China’s great rejuvenation,” they said. Xi praised the engineers’ work, and stressed the importance of self-led robotics innovation. Miles away, in a factory in Guangdong province, nine robots now do the job of 150 full-time workers (Bland, 2016). Robotic arms hoist circuit boards, process before transporting them through a computer-linked camera for quality inspection. Across China’s manufacturing belts, thousands of factories like this are turning to automation in a major government-backed, robotics-driven industrial revolution.
Catalyzed by the government’s concerted efforts and incentives over the past years, China is moving full steam ahead in the field of robotics. According to the International Federation of Robotics (IFR) (2017), China is the biggest robot market in the world in terms of annual sales and operational stock, with a net worth of US$30 billion. The reason for China’s rise to dominance is the Robotics Industry Development Plan – a five-year technology and industrial policy designed to rapidly expand its robotics sector. China wants to be able to manufacture at least 100,000 industrial robots annually by 2020, with majority of robotic orders fulfilled by Chinese companies (Gonzalez, 2018).
China’s economic growth phenomenon is attributable to its manufacturing sector. Approximately 100 million people are employed in manufacturing, and the sector accounts for over 35% of China’s gross domestic product (GDP). In 1990, China contributed just 3% of global manufacturing output, today it produces almost a quarter (Knight, 2016). In recent years, however, the Chinese manufacturing engine has started to stall. Exports have fallen, wages have increased, productivity is increasingly stagnant, and the sector has been witnessing contraction. To the Chinese government, robotics appears to be an enticing technological solution.
This paper analyzes advanced robotics as a class of technology rapidly proliferation across the world, designs a policy analysis framework, evaluates policy implications (the key research question) in the context of China’s Robotics Industry Development Plan, and extrapolates/applies the analysis to design, develop and recommend a technology policy blueprint framework for the world.
Global Trend – March of the Machines
China is not alone in its robotics pursuit. Worldwide sales of industrial robots more than doubled from 112,000 in 2008 to close to 300,000 in 2016 (Statista, 2018), a new peak for the fourth year in a row. The number of industrial robots in the world grew by a compound annual growth rate (CAGR) of over 15% from 2010 to 2015. By 2020, over 2 million industrial robots are expected to be installed in factories around the world, and the robotics industry is expected to grow to US$150 billion. Figure 1 charts the actual and projected global stock of operational robots.
Fig. 1: Global stock of operational robots.
Source: chart illustrated by Author, data from IFR.
Around the world, many countries see robotics as a panacea to the productivity conundrum, and a key to unlocking manufacturing growth. The Japanese government established a “New Robot Strategy” to revitalize the country, which for some time has been grappling with an aging population and shrinking labour base. Manufacturing powerhouses Germany, United States, and South Korea all have robotics and automation programs in some form. Even Singapore – an island city-state with a small total land area of just 721.5 km2 – has a National Robotics Program.
Figure 2 shows the estimated annual sales and operational stock of industrial robots in selected countries.
Fig. 2: Estimated annual sales and operational stock of industrial robots in selected countries.
Source: illustrated by Author, data from IFR report.
There is no single agreed definition of robotics, although all definitions include an outcome of tasks completed without human intervention. Some definitions involve physical machines that move, maneuver and respond to commands, while others include the element of task completion by software without physical embodiments. The International Organization for Standardization (ISO) defines a robot as an automatically controlled, reprogrammable, multipurpose manipulator for use in industrial automation.
Fig. 3: Example of robotics supporting different manufacturing subsectors. Illustrated by Author.
Robotics is alluring as it is a horizontal class of technology able to support a myriad of manufacturing subsectors (Figure 3). Most of these robotics have taken the form of expensive, high-precision industrial machines, utilized to automate manufacturing processes in production lines and factory shop-floors. A pictorial example is shown in the cover page, where industrial robots are assisting with high-precision manufacturing and production.
Understanding China’s Robotics Industry Development Policy
The Robotics Industry Development Plan, announced in 2015, was launched jointly by the Ministries of Finance and Industry and the National Development and Reform Commission. The five-year plan aims to rapidly expand China’s industrial robotics sector – to enhance production capacity and strengthen competitiveness – ultimately propelling China to become a global leader in robotics. The key targets of the plan include:
i. Annual production of China-made industrial robots to reach 100,000 by year 2020;
ii. Annual sales of indigenous-branded robotics exceeding RMB 30 billion (~US$4.7 billion);
iii. Robotics density exceeding 150 units per 10,000 workers;
iv. To build at least three large, globally competitive robotics companies; and
v. Other technical objectives.
The plan also calls for building competitive advantages in 10 sub-areas of robotics, such as healthcare and surgical applications, domestic cleaning and precision welding. Policy levers under the plan include subsidies, low-interest loans, tax waivers and rent-free land (Bloomberg, 2017), which robot-makers and companies that automate would be eligible for. Other policy mechanisms in the plan include strategies to strengthen overall planning and resource integration, broaden financial support, foster conducive market environments, enhance talent flow, and expand international cooperation.
Designing a Technology Policy Analysis Framework
Public policy structures and programmatic initiatives to advanced science, technology and innovation (STI) have been receiving increased focus. With intensifying competition in the global economy, policy-makers have sought to leverage STI policies to enhance national economies. Additionally, new patterns and modalities of research and development, technology translation and commercialization, knowledge transfer and information exchanges have induced policy-makers to review their priorities and approaches. These forces, combined with budgetary pressure and demand for accountability, have resulted in an emphasis on performance – to capture/create more value from STI investments.
In this regard, organizations, governments and policy scholars have developed frameworks for policy analysis and evaluation. The OECD in 2014 released its framework for regulatory policy evaluation, to assist countries in in systematically evaluating policy design and implementation against the achievement of strategic objectives. Other policy evaluation models include frameworks from organizations such as European Environmental Agency (2017), Stanford University (2003), and from policy scholars including Hankivsky et al. (2014) and Wenzl, Naci & Mossialos (2017).
Synthesizing from the myriad of sources such international organizations, academic and industry, a framework is designed (Figure 4) to guide the analysis and evaluation of the robotics industry development plan.
Fig. 4: Policy analysis and evaluation framework. Source: designed by Author.
Policy Analysis and Implications
The analysis and evaluation framework designed consists of three dimensions (viz. economic, environmental and social) and three key evaluation parameters (viz. effectiveness, efficiency and equity). Policies, especially in the field of STI, are intertwined, with strong interlinkages between the economic, social and environmental dimensions. Thus, policy design, analysis and evaluation should not be conducted in silos. The effectiveness parameter analyzes whether the policy achieves its intended objectives without adversely affecting other areas; efficiency measures the manner through which the policy was implemented including cost-effectiveness; and equity evaluates the impact on different regions and demographic groups.
Fig. 5: 3×3 matrix evaluation of the dimensions and parameters. Source: designed by Author.
The Robotics Industry Development Policy is effective and efficient in the economic domain, but has significant gaps in environmental and social aspects (Figure 5) – these gaps have impacts that were not considered. The next subsections discuss the analysis in the various dimensions.
As the policy and overall plan is due for completion in 2020, this analysis/evaluation is done in medias res.
China’s obsession with robotics automation has its roots in a pressing economic problem. Since the 1980s, China became “the world’s factory” when central rulers opened its market to global trade. Low-cost labor attracted international enterprises, and breakneck economic growth soon followed. However, a growing middle class and aging population – exacerbated by the one-child policy – is starting to erode China’s competitiveness. China’s working-age population is also expected to decline significantly by 200 million before 2050. To central planners, robotics is the solution to the labor gap and productivity slowdown.
Against this backdrop, China’s robotics plan as an economic policy is efficient and effective – even at the halfway mark of implementation. In typical top-down Chinese fashion, the effects of centralized governmental support and the policy levers of financial and resource incentives (viz. subsidies, tax waivers, low-interest loans and land) are already evident. The Guangdong province alone have committed to spend RMB150 billion to equip manufacturers with industrial robots (Bland, 2016). Other provinces, such as Yunnan, has also signed partnerships with robotics companies to develop capabilities (Zaleski, 2017).
Fig. 6: Operational stock of industrial robots in China. Source: chart illustrated by Author, data from IFR database.
In 2016, a year after the robotics plan was announced, China installed 87,000 new industrial robots, an increase of more than 25% from 2015 and a record for any country (Figure 6). The IFR predicted that annual growth would continue at a 20% pace to 2020. Annual sales volume for China has reached the highest-level record for a single country, and sales increase of between 15% to 20% on average per year is within sight (IFR, 2017).
Backed by the government, Chinese robotics companies are also acquiring international robotics firms at a blazing pace to make up for any engineering and technology gaps. Chinese manufacturer Midea’s takeover of Germans robotics firm Kuka, for example, has sparked concerns among the European administration, especially Germany. The Kuka acquisition is just one in a long list of Chinese takeovers, others include Huachangda’s takeover of Swedish Robot System Products, and Chongqing Nanshang’s buyout of Michigan based HTI Cybernetics. China already has large enterprises able to rival major western technology firms – Alibaba is China’s answer to Amazon, Tencent to Facebook – and with an aggressive, government backed growth playbook, it would not be long before Chinese robotics companies dominate the global market.
However, despite relative success, two policy gaps are glaring. First, the industry-academia-government nexus is conspicuously missing from the plan. Partnerships and collaborations between the three domains have been a cornerstone in highly innovative, technology-based ecosystems. China’s top-down, centralized approach may be effective and efficient in driving (robotics) technology adoption and industrial growth, but the approach has limits. A significant amount of innovation in the west, such as in Silicon Valley and Silicon Fen, stemmed from bottom-up collaborations between industry, academia and government. The robotics plan and its associated policy levers may drive rapid short-term growth, but without a strong, collaborative R&D base, the incentives-fueled growth may not be sustainable. Second, the policy focused disproportionately on the “D of R&D” without considerable strategies to enhance the research landscape. Policy scholars have often commented that the quality of Chinese research falls short – majority of Chinese patents are minor novelties instead of genuine innovation, and China has relatively lesser high-impact publications (Qiu, 2014). This may be detrimental to the overall innovation ecosystem in the long run.
In terms of economic equity, a centralized technology policy of this magnitude raises fears of greater inequality, as benefits of productivity gains are significantly skewed towards owners of capital at the expense of workers. In China, as with everywhere else, automation will likely erode incomes of those with less skills (Orlik, 2017). The plan lacks policy levers for redistribution, or transition mechanisms to help rank-and-file and blue-collar workers (which make up a significant proportion of the workforce) to adapt to the technological change. Separately, high degree of centralized, top-down planning bears also other economic risks, as individual provinces may struggle to interpret central directives, and venture off the officially sanctioned pathway.
Proponents of trickle-down economics believe that a rising tide lifts all boats – they argued that the benefits of economic growth will make their way down the system, lifting income levels, reducing unemployment, and gradually eliminating poverty at the lower end of the socio-economic spectrum. In reality, the aphorism is not necessarily true, especially in China. China has one of the world’s highest levels of incomeinequality, with the richest 1% owning one-third of the country’s overall wealth, and the poorest just 1% (Wildau & Mitchell, 2016). The robotics plan, by focusing largely on driving economic growth, neglects the social side-effects of the policy, and will likely intensify any existing inequality. Therefore, the policy is weak in social effectiveness, efficiency and equity.
Fig. 7: China’s GDP per person based on region (2015).
Source: Economist, CEIC, World Bank
Figure 7 illustrates China’s GDP per person based on region. It is no coincidence that many regions with significantly higher GDP per person, such as Zhejiang, Guangdong and Tianjin, are manufacturing hubs. Zhejiang has significant manufacturing activities in metal products and textiles; Guangdong in electronics and petrochemicals; and Tianjin in pharmaceutical manufacturing. Regions with heavy manufacturing activities would benefit from the robotics plan, while the rural-poor areas may lack further behind.
The benefits derived from the policy/plan will likely be uneven even within manufacturing sub-sectors. Figure 8 plots manufacturing subsectors’ ability to automate vs deviation from average manufacturing wages. Evidently, high-tech manufacturing subsectors such as electronics, automotive and computing hardware, which possess higher ability to automate and enjoys higher average wages, will benefit from the robotics plan to a much larger extent than low-technology, conventional manufacturing subsectors such as textile, wood and apparel. Low-end manufacturing subsectors are being squeezed out, while higher value-add subsectors are enjoying financial incentives and reaping the benefits of the robotics plan.
Fig. 8: Manufacturing subsectors’ ability to automate. Illustrated by Author, data from BCG, US Labor Authority.
Another major social issue is the displacement of existing jobs, deskilling and further income erosion of workers with fewer skillsets and education level. The effects are already apparent. In Dongguan city, which launched a US$30 million annual fund to enhance productivity (based on the robotics plan), 87,000 workers were replaced by robots between 2014 and 2016. In Zhejiang province, two million workers lost their jobs between 2013 and 2015 (Chan, 2017). In many other manufacturing outfits, workers are increasingly channeled into dead-end jobs such as cleaning and polishing. McKinsey, a consultancy, estimated that over 100 million workers in China may need to switch jobs and learn new skills due to the robotics onslaught.
The social issues discussed above have profound impact on the Chinese society and economy. High inequality would impede China’s drive towards a consumer-driven economy – the rich enjoys high savings while the middle-class and the poor save almost nothing. By skewing income distribution more towards the rich and reducing social mobility, robotics automation risks further eroding China’s low consumption, and may even create social problems such as violence, crime and squatting, especially in poorer regions.
Fig. 9: China’s environmental problems by region.
Source: Woon (2010)
The robotics policy is also weak in terms of effectiveness, efficiency and equity in the environmental domain, as there is a severe lack of provisions for managing environmental impact. The relationship between the economy and the environment in China has always been tense. Rapid industrialization and economic development have taken place since the 1980s, creating huge pressure on the environment. Currently, a fifth of China’s arable land is poisoned with heavy metals, most of its groundwater is polluted, and many experts believe its air quality is a bigger killer than smoking (Gracie, 2018). Enterprises are also not doing their part for the environment – 10,000 companies in northern China alone failed to meet standards for controlling air pollution (Wong, 2018).
Many of China’s regions are already experiencing environmental problems such as severe air pollution, water scarcity, desertification, toxic land/water, and even acid rain. Figure 9 shows China’s environmental problems by regions. Many polluted areas in eastern China coincides with heavy industrialized regions with intense manufacturing activities.
The rise of the robots may become a problem not just because of the immediate social and economic cost, there are environmental repercussions as well. First, robotics are notably energy-hungry, and mass adoption would increase energy consumption significantly. Increasing the number of transistors on a microprocessor chip and reducing the size of processing units would also make them increasingly vulnerable to quantum instabilities, and thus requiring more energy to power. China’s energy consumption has already been experiencing rapid growth over the past decades, raising concerns on future adjustments of its energy utilization structure (Dong et al., 2016). A major national technology policy such as the robotics plan should consider energy requirements, and how these requirements would be met and managed. Second, the proliferation and mass adoption of robotics would inevitably lead to increasing amount of electronic waste. As computer parts shrink in size, they also become harder to recycle. Third, the process of disposing electronic waste and potential events of malfunction could release harmful chemicals and toxins into the atmosphere.
The impacts of environmental problems are significant and multifaceted – they include damage to human health, social conflict and economic losses (Fu et al., 2007). With China’s recent push for sustainable growth, and setting governmental targets for conserving energy (Kahn & Yardley, 2018), it is surprising that its robotics policy neglected to incorporate environmental considerations.
Comparing Best Practices
Table 1 depicts a summarized comparison of other notable national robotics policies/programs.
|Singapore||Japan||United States||South Korea|
|Policy||National Robotics Program||New Robot Strategy||National Robotics Initiative 2.0||Second Basic Plan for Intelligent Robot Development|
|Objectives||Develop a globally competitive robotics industry; enhance productivity and competitiveness of economic sectors; support public sector adoption of robotics||To make Japan into a robot innovation hub; to enhance utilization and deployment of robotics; to globally standardize Japan’s robotics technologies||Supports 4 main research thrusts (viz. scalability, customizability, lowering barriers to entry, and societal impact) to advanced goal of ubiquitous co-robots||Global recognition of domestic robot product quality and international standardization of domestic technology; industry expansion; and meet various societal needs and expand into global markets.|
|Key Elements||– Robotics for social and environmental needs, not just economic
– Building R&D capabilities and technologies
|– Develop domestic R&D capabilities
– Promote regulatory reforms and best practices
|– Encourage collaborations between academic, industry, non-profit and other organizations||– Extensive R&D projects
– Expand demand
– Creating open ecosystem
– Attracting investment
– Globalize standards
– Develop SMEs
– Broaden collaborations
|Strengths||– Public sector test-bed projects
– Public sector-led demand
|– Sectoral perspective (strategy for advancement of robotics for different sectors)||– Supported by multiple agencies such as NSF, Dept. of Agriculture, Dept. of Energy, and Dept. of Defense||– Large-scale simulation test-beds
– Apply robotics to other sectors
– Supply projects to generate demand
|Funding||US$340 million||US$30 million||US$50 million||US$316 million|
Table 1: Comparison of other notable national robotics policies and program.
The comparative analysis validates the analysis in the previous section of this paper, and further highlights several weaknesses in China’s robotics policy. First, apart from funding and resource policy levers, the specific modalities to achieve policy objectives are unclear. There is a lack of strategy to leverage the industry-academia-government nexus and synergies, and there are almost no policy levers to catalyze research, especially in academia (the levers are primarily directed at industry). Second, although it is acknowledged that the robotics plan is a STI policy with a primary focus on the economic domain, it neglects social and environmental implications and impacts almost entirely. China’s robotics plan seems like a “bulldozer” policy – a top-down, centralized approach razing its way through, leaving along the way questions and implications unanswered. Third, the centrally-planned policy seems to lack consultation, or any form of empirical analysis (e.g. cost-benefit analysis). There is no evidence of any stakeholder engagement or policy analysis. Last and more importantly, the policy lacks any form of evaluation mechanism, performance indicators or success measures for tracking, reporting, accountability and iterative improvement purposes.
Technology Policy Recommendations
Prior to proposing recommendations, discussions were conducted with policy-makers from United Kingdom, China and Singapore. Figure 10 illustrates a sunburst diagram summarizing the key points/issues raised or mentioned based on frequency, classified under the categories of economic, social and environmental.
Fig. 10: Sunburst diagram of key issues/points mentioned.
Source: generated by Author
The findings are consistent with the analysis presented earlier, where issues such as industry-academia-government partnerships, inequality, job loss, and energy consumption were mentioned. Additional points that were not covered in detail in the analysis include research commercialization, robotics infrastructure, and demand drivers. These are relevant points for robotics policy, but not critical gaps in China’s context.
Synthesizing the points above and the analysis in the previous sections of this paper, a two-part technology policy blueprint framework is designed and proposed. The first part (Figure 11) is a model for policy-maker to drive economic, environmental and social value capture/creation through technology (e.g. robotics). The second is a performance matrix (Figure 12), to track in medias res and/or ex post the key performance indicators of a technology policy.
Fig. 11: Value creation model for technology policy-makers to drive economic, social and environmental outcomes. Source: designed by Author.
The proposed value creation model is a template to drive economic, social and environmental outcomes with technology (and technology policy). With the industry-academia-government nexus at the heart of the model, appropriate innovation and commercialization mechanisms can be identified and leveraged. The technology policy and levers shouldbe constantly reviewed and scaled in an iterative development manner, to optimize economic, social and environmental outcomes.
Fig. 12: Performance matrix to track key performance indicators. Source: designed by Author
Every (technology) policy must be evaluated for impact and effectiveness. Figure 12 depicts a performance matrix to track and measure the performance of the policy (performance indicators from each category of economic, social, and environment can be customized to suit different policies based on context). In the case of China’s Robotics Industry Development Plan, key performance indicators to measure impact and effectiveness can include enterprise revenue growth (economic), employment (economic/social), income inequality (social), and energy conservation (environmental). These indicators should be tracked and monitored throughout the policy implementation period. Figure 12, for example, illustrates a policy that scores high on economic performance indicators, but low in the social and environmental domains – like China’s robotics plan.
This paper analyzed a robotics policy in China’s Robotics Industry Development Plan, analyzed gaps, and proposed a technology policy blueprint framework to aid policy-makers to identify and design effective levers and measure the impact and effectiveness of the policy. Utilizing the blueprint framework in the previous section for China’s robotics context, specific policy levers to plug the gaps can include the following:
Re-training or vocational training schemes/programs to manage structural transition of workers arising from mass adoption of robotics technologies (a good example would be Singapore’s SkillsFuture program);
Modalities to enhance R&D and robotics infrastructure around China, not just in manufacturing concentrated regions;
Programs to help small-and-medium-sized enterprises (SMEs) adopt and extract value from robotics, to ensure the benefits of the robotics policy are not only reaped by major firms;
Specific programs/initiatives to create demand and jobs in less wealthy rural areas; and
Technology translation mechanisms to help commercialize robotics research arising from academia, and/or help spur university spin-offs/spin-offs.