Policy 4.0 - Why an inclusive industrial policy is needed to navigate the complex landscape of digitization in the manufacturing sector

That new digital technologies will transform the way we live, work and do business is common knowledge at this point. It is widely agreed that the New Digital Era will change the world of labor by making some jobs redundant and creating new, different jobs. It will challenge our welfare systems as well as the way that companies do business. And yet, the complexity of these challenges, and in particular those to the labor market, seems to be captured in a somewhat polarized fashion between technology advocates and prudent risk-averse contenders.[1] Measuring the risks and opportunities of digitization correctly and adequately preparing the policy landscape to meet the social and economic challenges to come is in no small share dependent on the unit of analysis: nations will face different challenges, but even within one country, there are different regional realities; some sectors may be affected strongly by automation and could lay off large amounts of workers while other industries may experience explosive growth and job creation; and the development of skill requirements are equally diverse depending on the industry. The reality of digitization is diverse, multifaceted, confusing. There is no one-size-fits-all solution, but an examination of the current landscape will illuminate potential remedies.

The velocity of digitization across industries varies widely depending on the cost vs. benefit of adopting new technologies and the lifespan of a company's products. Cyber-Physical Systems (CPS) and the Internet of Things (IoT) require a significant amount of investment in innovative and largely completely new industrial hardware––an investment that not all companies are able to uphold, especially SMEs, unless significantly subsidized. For large corporations, these transitions are feasible and, in many ways, are an extension of the steady production innovation that has already taken place over the past years.[2] The efficiency of these “smart factories”, however, not only lies in the technologies, but in the symbiosis of technology with highly skilled workers, both blue and white collar, that manage the production process and are equipped with the skill set to work with the hardware (production machinery) and the software (IoT).[3] High value-added industries and high-skill manufacturing will be the first to experience the transition to smart manufacturing and, in some cases, already have (like in many companies in the automotive industry[4]), while other industries, despite the expected profit margins, will incorporate smart manufacturing more slowly due to the long life cycle of their products (see, for example, the case of Airbus[5]).

For small and medium enterprises, Digital Assistance Systems offer a different opportunity to increase profit margins and efficiency. These computer-aided systems consist of software solutions that work with the existing hardware and, thus, require lower initial investments. Assistance Systems optimize the assembly of products and lead workers through this process step by step, thereby keeping the training and education effort required from workers at a minimum.[6]

Artificial Intelligence (AI) requires by far the highest investment in hardware and software. While commercial uses are spreading and becoming more and more advanced with products like Siri or Google Home, industrial use of AI is only slowly picking up in the main manufacturing sectors. And while research on this issue is progressing at a fast pace, in many industries the improvements in profit margins do not currently offset the initial investments and, once they do, will likely be applied first in the high value-added and high-tech industries.

The required skill levels through digitization will increase––and decrease. With the variations in the levels of digitization in production comes a variation in the skills required for the new jobs. Smart manufacturing requires a set of skills that is, if not entirely new, at least a significant expansion of the training of a large part of the workforce. A smart factory will require control jobs that are closely working with digital technology and familiarity with programming languages, a skill set that is in no way comparable to traditional industrial jobs.[7] The effects on the labor force are a reduction in the workforce and a more complex skill profile of the workers remaining at the plants.

For those SMEs that are more likely to make use of digital assistance systems in their assembly lines, skill profiles may actually decrease, giving job opportunities to low and medium skilled workers in medium-tech industries,[8] which could actually have very strong beneficial effects on the labor market and on the traditional working and middle class––that is, if these jobs are decently paid.

The 2017 McKinsey Report “Jobs Lost, Jobs Gained – Workforce Transitions in a Time of Automation” predicts a total of 30 percent of work hours being automated by 2030. It shows that occupations requiring only a high school degree (like most traditional manufacturing jobs in the United States) are more than 200 percent higher at risk of being automated than occupations requiring a college degree[9]. With privatized education and a continuously biased access to higher education, this transition bears the risk of turning existing inequalities into social cleavages based on class, gender, race and ethnicity. Similarly, the urban-rural divide will likely increase and leave more rural communities with the struggle to absorb job losses.[10]

Whether one sides with the technology advocates, which see the potential for job gains in mechanical engineering, machine learning and software development, or with the antagonists, which see the risks of drastic workforce downsizing––the jobs lost and the jobs created are very different, require intense additional training and education, and will be difficult to be matched ad hoc.[11]

To make the Digital Era a success industries must address not only the skills shortage but also the increasing skill misfit in supply and demand. Qualifications are one of the most prominent challenges when it comes to digitization. The changes in the required skills are not just a challenge for workers, but they also have strong impacts on societies, especially in developed countries where skills shortage and skills misfit are already common problems of the labor market (see figure below).

In most developed economies, industrial design and the manufacturing of high quality products require a large amount of highly qualified workers and engineers. At the same time, there is a continuing need for private and personal services (such as cleaning, laundry, maintenance, etc.) that require lower skills. On the other hand, medium skills are demanded to a much lesser extent because a large share of medium skill manufacturing has relocated to emerging economies.

Modeling the existing qualifications in Western European societies shows that they differ greatly in distribution from those needed[13]: Well established educational systems (and in some cases apprenticeship programs) means that a large share of the society has at least medium skills, and a relatively small share of the population has low skills. While this might be a good sign for educational systems, it also points to a problem of supply and demand: The overproduction of a medium skilled workforce means that there is a large share of this group struggling to find a job to match their qualifications––they are overqualified for the lower-paying low-skill jobs and they are not skilled enough to fill the skills shortage in the higher qualification jobs. With industrial work and design losing its appeal to younger generations in Western economies, the problem of skills shortages has arisen in higher-skilled jobs. In response, both companies and governments in Europe have taken measures to make industrial jobs more appealing, i.e. by giving out specific scholarships for STEM subjects and guaranteed job offers after successful apprenticeship programs.[14]

The right training and education for our current (and expected) manufacturing landscape is key to a smooth transition into digitized production. The skills misfit means that a fair share of workers are forced to work at jobs that actually do not match their own qualifications and that they are overqualified and underpaid. It also means that a large share of the existing human capital in society actually remains unused, even though the average qualification is fairly high. In Western societies, low qualification demand is likely to stay fairly stagnant. Low qualification manufacturing has relocated, while low qualification services can often not be outsourced or relocated. Medium skill manufacturing jobs, however, are prone to being automated, thereby being at higher risk for workforce downsizing and rationalization. Medium skilled services (web design, calculations, etc.) can easily be outsourced and provided from anywhere around the world for a fraction of the price via various platforms and will therefore also shrink local job opportunities in the developed world for the service sector.

The example of Western Europe shows the digitization side effects and challenges that need concrete policy solutions: In Europe, an even smaller share of medium skill jobs will be available for the relatively large share of workers with these skills. Some high skilled manufacturing can be done through smart manufacturing in the future, which will translate into job cuts, while in some SMEs, the process may be enhanced by digital assistance systems––the latter could actually generate jobs for medium skilled workers that do not currently work in jobs that fit their skills. The risk of automation eliminating medium-skilled jobs and the need for training and education is even higher in the United States, with a lower average qualification level of the workforce due to the regional discrepancies and urban-to-rural differences, among other things.

Employer-sponsored trainings in the US have decreased from 19.4 percent in 1996 to 11.2 percent in 2008 and government spending on labor policies has consistently decreased over the last 30 years, with the exception of the final years of Obama’s presidency[15]––a trend that should be turned around quickly and incentivized more than ever before in order to ensure that workers are not left behind. Skilled workers are an asset to any company––the regional supply of the much-needed talent is a deciding factor for many companies when they decide to open a new plant––and a skilled labor force presents a competitive advantage for nations. Strengthening the US labor force means strengthening the performance of the US in global comparison, resulting in an increased ability to attract new investments and increasing productivity.

The rise of the Digital Era will be a game-changing technological and economic transition for our societies––but it does not have to spell catastrophe. The McKinsey Report cited earlier highlighted a risk of automation for 30 percent of the hours worked. It may be seen as a discussion of semantics by some, but it is important to note that automation of a large share of hours worked does not have to translate into a 30 percent of workers being displaced. Given the regional variations and the bias towards minorities in the risk of automation, displacement should be avoided. Instead, tax reforms that draw from the increased profit margins due to the technology innovation, in combination with a reduction of weekly hours, should provide a viable strategy to relieve the stress of automation on workers. In 2017, full-time workers in the US worked 44h/week on average, with more than 40% of them working more than 50h/week on average.[16] Dropping to only 40h/week would significantly reduce the strain––and would still be at much higher levels than European countries with a high GDP/capita and competitive labor markets, like Germany (35h/week)[17] or the United Kingdom (37h/week)[18].

Decent wages and a strong middle class do not represent a normative point or an ideological argument. In an era in which economies rely on a strong consumption model with a low share of manufacturing and a rising share of services, ensuring a strong purchasing power through high wages is crucial – both for the services and the manufacturing sectors, especially with the current and upcoming transformations in the context of digitization.

The stress of digitization needs to be met with smart policy making that takes into consideration the necessity of reform in economic, social, education and welfare policy ––an action plan for the future of work. It is clear that the labor market will experience strong and very diverse effects of digitization that will require stakeholders to be strongly involved in the policy design process.

1. An inclusive industrial policy that captures the various facets of digitization and matches it with the adequate policy solutions is necessary. The policy reform needed in order to tackle digitization is not one that can be nestled under economic policy, nor is it only a social problem or one of education or welfare. Solutions need to involve experts from all of these policy fields and stakeholders.
2. Social partnership and parity models are more important than ever. Digitization is not a process that can be stopped and this is a lesson that needs to be accepted by all parties. But the success of this transition greatly depends on the willingness to compromise, to develop solutions that boost our economies and distribute wealth, one that develops our cities into centers of innovative human capital without the precarization of more rural communities.
3. Nations must develop suitable long-term training and education strategies and infrastructures. The skills that are needed now will vary from those needed in the decade to come––flexibility in education, both academic and non-academic (i.e. through apprenticeships) is therefore crucial. Where social partnership and industrial relations are weak, employer-sponsored trainings need to be incentivized and enforced by the adequate legislative framework in order to maintain the competitiveness of US labor.
4. Tax reforms that offset potential workforce reductions and redistribute from digitization beneficiaries to those at risk of automation are necessary to maintain social cohesion and avoid (regionalized) class, race and gender divides. New technologies will bring tremendous opportunity for growth to companies creating a demand for the development of human capital that supplies their needs. A redistribution of profit gains through tax reform could be a path into the future: Just Transition and training and education models, together with the exploration of unconventional policy ideas such as the (conditional) basic income in combination with the reduction of working hours would become more feasible in the process.
5. A systemic view on digitization is necessary to put metropolitan, regional and national realities into perspective––especially in the context of a globalized world. The discussion on digitization is largely led by developed nations in very specific contexts. But we must not forget that the developing world is facing drastically different preexisting inequalities and precarities. When production becomes cheaper for high-quality and high-profit products in the developed world, relocation into the developed world becomes a possibility. Developing nations that currently have the competitive advantage of low wages and few regulations are suddenly catapulted into a race to the bottom––a score that is very likely to be settled at the expense of workers and wages and could, long term cause many more problems for developing and developed nations alike, such as a wave of “digital refugees”.[19]

Digitization offers countless opportunities. But fair access to these opportunities needs to be ensured to maintain social cohesion and economic prosperity.

____________________________________

[1]?Barker, Rachel; Berube, Alan, 2018: Delivering shared prosperity for workers in a rapidly changing economy.

[2]?Bosch Software Solutions, 2015: Industrial Internet: Putting the Vision into Practice.

[3]?Eurofound, 2018: New Tasks in Old Jobs. Drivers of Change and Implications for Job Quality.

[4]?See for example the case of Mercedes-Benz.

[5]?Adrian Bridgwater: Airbus: engineering the future of intelligent factories. Internet of Business. 10 February 2017

[6]?Fraunhofer-Gesellschaft. "Better quality control with digital assistance systems." ScienceDaily. ScienceDaily, 6 July 2016.

[7]?Smith, Aaron; Anderson, Janna, 2014: AI, Robotics, and the Future of Jobs. Pew Research Center.

[8]?Ford, Martin, 2015: Rise of the Robots. Technology and the Threat of a Jobless Future. Basic Books: NY.

[9]?McKinsey Global Institute, December 2017: Jobs Lost, Jobs Gained – Workforce Transitions in a Time of Automation.

[10] Muro, Mark; Maxim, Roberto & Whiton, Jacob, 2019: Automation and Artificial Intelligence: How machines are affecting people and places.

[11] The Economist, 2015: The weaker sex. Blue-collar men in rich countries are in trouble.

[12] In adaptation of Sandulli, Francesco, 2016: Job Polarization and Innovation. The Contribution of Middle-Skilled Workers. Conference Paper.

[13] Sandulli, Francesco, 2016: Job Polarization and Innovation. The Contribution of Middle-Skilled Workers. Conference Paper

[14] Hirsch-Kreinsen, Hartmut, 2016: Digitization of industrial work: development paths and prospects. Journal for Labour Market Research vol 49/1.

Frey, Carl Benedikt; Osborne, Michael A. (2013): The Future of Employment: How Susceptible are Jobs to Computerization?

[15] Maxim, Robert; Muro, Mark, 2019: Automation and AI will disrupt the American labor force. Here’s how we can protect workers.

[16] Bureau of Labor Statistics, 2017.

[17] “Germans work less than the EU average”, Die Zeit, 20 August 2018.

[18] Office for National Statistics GB, 2019.

[19] Kelly, Ross: A “Digital Refugee” Crisis Could Emerge Sooner than Many Expect: CEOs Warn in Davos. 18 January 2017