The great skills divide

There are lessons to be learned from international cases of labor market polarization, and how Indonesia can address its issues in the context of increased adoption of technology.

The great skills divide AFP Photo/Adek Berry

Indonesia, much like any other country in our time, must come to grips with the sweeping winds of creative destruction in the form of technological development. Technology has experienced an increasing rate of newfound use-case applications, such as those found in Industry 4.0 and beyond. This rapid pace of technological development poses serious questions about an already polarized labor force, characterized by a decrease in middle-skill jobs relative to low-skill and high-skill jobs, especially one that is fraught with structural and human capital issues.

The need to develop a multipronged approach and institute multifaceted reforms is pressing for policymakers. If the current situation is left to fester, Indonesia may very well find itself having wasted its demographic dividend, languishing in the bottom of the global value chain and experiencing increased inequality.

It takes two: labor and technology

The relationship between labor and technology is not governed by dichotomies of black and white or good and bad. The reality, in terms of technological development, is much subtler; the impacts, however, are far more overt. A job is made up of a series of tasks that are performed by the laborer. These tasks range in complexity with respect to cognitive ability required and routine-intensity. A routine task is one that requires certain conditions to be met and which conditions must be executed in an exact manner. A non-routine task, on the other hand, is riddled with the nuances of complex and infinite conditions, for the laborer not only to perceive but upon which also to act.

Business success can include more people and communities benefiting from that success.

Creative destruction has almost always taken the shape of technology replacing repetitive human labor – it is in the very nature of computing to execute programs that are based on a set of operable conditions. Since the time of the Luddites, technological development has improved at a rate consistent with Moore’s law, all the while the real cost of computing has declined precipitously. Early modern computing was purposefully built to handle rudimentary tasks that heavily depended on human input.

Clerks and bookkeepers became more productive in sorting and inputting information. However, especially in the decades leading up to the millennium and after, computing powers have widened
to include data input, storage and acting upon information. These technological developments, in a cruel twist of fate, make redundant some of the very jobs that early technological iterations created.

And so the relationship between labor and technological development starts to become apparent. For the sake of simplicity, the pattern in which the interplay occurs is as follows: technological development seeks to answer the inefficiencies that are present in the labor force. Because earlier iterations of technology were heavily dependent on human operation and input, this increased demand for routine inputting jobs. As computing power increased its share by taking over more inputting and processing tasks, routine labor was not necessarily deemed redundant; rather, it became more susceptible to automation.

According to a series of papers published by economists at the Massachusetts Institute of Technology (Acemoglu, Autor, et al), firms will tend to no longer require those who perform routine tasks. As such, the increasing relative demand for labor capable of performing nonroutine and cognitive tasks prompts laborers to go through a sinkor- swim moment. Those who are capable of upgrading their skill sets will be entitled to higher wages as they perform more productive tasks, while those who cannot may be left structurally unemployed. The job title very much stays the same; the tasks that comprise the job widely evolve. It is for this reason that, say, a biochemist is more productive than a manufacturing assembler. Technology is a complement to the former’s labor and thereby makes the former more productive, while acting as a substitute for the latter.

In effect, this is a polarization of the labor force. However, it is not only repetitive and noncognitive tasks that are under threat of being automated. Nonrepetitive and noncognitive tasks, such as driving, are also at risk of being automated as technology has gained perceptive and interpretive skills, thus enabling them to pick up on the intricacies of these tasks. Primarily, the tasks and, by transitive logic, the jobs of the future that will be relatively shielded from automation are those that are nonrepetitive and heavily cognitive.

Polarization in the West

Western Europe and the United States, with their considerable positions as early adopters of groundbreaking technology, are at the forefront of experiencing the waves of creative destruction brought on by technology. Below is a chart depicting the growth rates in several countries from 2000 to 2017 in the levels of employment based on skills. There seems to be a common theme in this data: middle-skilled employment growth either decreased unequivocally or grew at a slower rate than low- and high-skilled employment.

The U-shaped trend is quite ubiquitous among developed economies. However, each country has its own story to tell – a story that finds its roots mostly in labor market institutions and globalization. First, consider the outliers and what would seem to be contradictions to the idea of polarization: nations that have experienced sharp declines in low- and mid-skilled jobs coupled with dramatic increases in the number of high-skilled jobs. Interestingly enough, these countries are mostly Scandinavian ones.

Scandinavian labor forces have experienced high increases in high-skilled jobs, mostly due to the strength of their labor market institutions. A 2016 International Labor Organization (ILO) report estimated that around 84 percent of employed Danes, 89 percent of employed Fins and an astonishing 90 percent of Swedes were under some sort of collective bargaining scheme. This implies that low-skilled labor is comparatively more expensive than high-skilled labor. This is because not all of the additional marginal productivity from high-skilled labor is fully captured in wages when performing a cost-benefit analysis. So when given the chance to upgrade its talent pool, a company will do
so provided that the increase in productivity offsets the wage differential, hence effectively getting skilled labor at a relative discount. It must be added that Sweden’s moderate expansion in low-skilled labor, thereby creating the U-shape associated with polarization, is partially driven by the influx of relatively lowskilled immigrants that took place in the earlier part of the millennium.

Enrique Fernández-Macías, of the European Commission Joint Research Center in Seville, argued in a 2012 paper that Central European countries’ polarization occurred mostly due to a process of deregulation of the labor market that predominantly took hold in the late 1990s, prior to the formation of the European Union, as the region started integrating its value chains amid the backdrop of globalization. These policies were aimed at increasing employment figures. However, the set of policies achieved this through destandardization that created numerous temporary jobs. According to a 2014 study (Vergeer and Kleinknecht), such policies, while generating tangible gains in employment from a numerical sense, were expected to increase the “creation of low-productivity jobs, rather than skilled jobs.” This was partially due to the fact, as the two contended, that labor market flexibility, a measure of ease of firms’ decisionmaking, disincentivizes firms from undertaking necessary human capital development investments.

Polarization is a force that has no end in sight. The United States Bureau of Labor Statistics paints a bleak picture of things to come. The agency expects that by 2026, mid-skilled jobs, which are consequently mid-paying, that are routine-intensive are to decrease substantially. Meanwhile, strong growth is expected in high-skilled fields such as mathematics and computing, which are handsomely rewarding. The growth of American high-skilled labor is in tandem with an expected increase in healthcare-related jobs. While the health care industry may certainly employ and handsomely reward highly skilled laborers such as doctors, nurses and surgeons, the industry also heavily relies on a host of low-paying and low-skilled jobs that are more to do with maintenance and day-to-day hospital operations.

Therein lies the crux of labor market polarization brought on by creative destruction: as middling jobs are either lost due to automation or offshoring, those who cannot survive the sink-or-swim flashpoint are to be shifted to less productive parts of the economy – parts of the economy in which the business model is only economically and financially viable by keeping wages low. It is no wonder why the US labor statistics bureau forecast that jobs such as fast-food operators would increase by approximately 9 percent.

Where Indonesia stands

Indonesia’s economic development, from a numerical perspective for the two decades since the Asian financial crisis, has been impressive. However, it is clear that not all that glitters is gold. Indonesia’s growth story has mostly been one spurred onward by the uptick in commodity prices, mostly fueled by rising industrial demand from China and other emerging countries. This type of growth is one that is all too familiar in the Indonesian economic narrative. In addition to the commodities boom in the early part of the new millennium, Dutch disease occurred as well in the 1970s and 1980s with massive oil exports. While the uptick in commodity prices has been a boon to economic growth, it dissuaded
complacent policymakers from preparing for a post-commodity paradigm.

Traditionally, the economic growth narrative is that a society progresses from agriculture to manufacturing and then on to services. This is in the hope that the labor force will eventually move on to endeavors that are more productive when the technology is made available. However, in a combination of de-industrialization and an underdeveloped educational system brought on by Dutch disease, Indonesia has undergone a superficial structural economic change. All the talk of how Indonesia’s economy is inching ever closer to maturity due to the size of its services sector is misguided by simple nomenclature, and forgets to take into account the quality of the intersectoral shifts within the labor force.

In a post-commodity boom, national account data indicates that Indonesians were migrating from the most unproductive sector of the economy, agriculture, into the secondmost unproductive sector: low-end services. While economic productivity on a per laborer basis may improve, it is only marginal.

Despite its counterfactual nature, a decomposition growth analysis of the Indonesian economy from 2012 to 2018, using the World Bank Job Generation and Growth Decomposition tool, could provide

useful insights in understanding how the composition of the labor force has changed in a post-commodity boom Indonesia. Agriculture stands out among the industry classifications; it shows the largest gain in contribution to increased productivity and the largest decline in contribution to changes in employment uptick. This is reasonable as mentioned earlier: agriculture as an industry has experienced large declines in its share of employment. Additionally, the industry has had low-base productivity levels and is fairly labor intensive, and so with modest agrarian technological adoption its productivity grew significantly, though still being small. Agriculture contributed overwhelmingly to intersectoral shifts as a result of lower-than-economy-wide average productivity and a scaling back of jobs.

Trading, accommodation and food services and other low-value services represent the fastest growing sectors in the economy. Combined, they account for around one-fifth of the upward change in employment levels. However, because these sectors are more productive than agriculture and because the compositional shift away from agriculture and into these sectors was large enough, gross domestic product grew nevertheless. Which is why these two sectors accounted for around 20 percent of the GDP per capita growth in Indonesia during that period.

Employment-based on skills data from the ILO database, when combined with the
growth decomposition analysis above, shows a worrying picture of Indonesian labor. With high-skill and low-skill job growth outpacing that of middle-skill jobs, it can be said that Indonesia is experiencing a labor polarization of sorts.

Additionally, Indonesia’s problems are sure to compound when it finds itself having to deal with the tumult of impending creative destruction. Industry 4.0 boasts technological innovations that are not only capable of entering data, but also interpreting and acting upon them via a decentralized decision making process, mostly free of low- and middle-skilled human intervention. In short, the technological development is unlike any other. It’s not entirely certain if Indonesia’s labor force can survive the sink-or-swim moment, when it comes. This will result, as described earlier, in further polarization such as in Europe and the United States.

While Indonesia’s low-wage structure may deter investment and foster a penchant for developing labor-intensive business models in the coming years, there will come a tipping point. Eventually, wages will inflate cost structures and labor-saving technology will not only be economically feasible but necessary. As such, further polarization in Indonesia is inevitable. One would only need to refer to a 2017 study (Fonseca, Lima and Pereira) that describes labor polarization due to technological change and routinization in Portugal, a country with low capital spending and a relatively smaller share of college graduates compared to its peers. There are similarities between Indonesia’s and Portugal’s situations; there is certainly enough to see where Indonesia’s path lies.

Additionally, if the nature of jobs growth follows the same pattern whereby more high skill jobs are to be created as a result of the complementarity of cutting-edge technology and high-skill labor (Acemoglu and Autor), then it is reasonable to say that Indonesia could experience wage polarization. This is plausible given that in Indonesia the skill premium – the ratio of the wages of skilled to unskilled
workers – is significantly high and has worsened over time, as shown by an Australian National University-led group of Australian and Indonesian researchers in 2016.

Highly skilled workers displaced by technology may be able to transition to other high-paying white-collar jobs, while those without the necessary training are displaced and left structurally unemployed. From there, it obviously then follows that a difference in income flows can exacerbate an already unequal society, especially when the opportunity to upgrade via middle-skill jobs has disappeared and the legacy effects of work in family units perpetuate.

Regulatory solutions

Preparing Indonesia’s economy for the future and arresting labor market polarization are not two mutually exclusive policy goals. Polarization can mainly be solved by guiding the labor force through structural changes. Policymakers must see that there is no tradeoff to be made between protecting those at risk of structural unemployment and driving innovation – innovation that will boost the number of medium- and high-skilled jobs. This could very well be a case of having your cake and eating it too.

Thoughtful (de)regulation arises as a possible solution to labor market polarization. It has often been said that given increasing adoption of the latest technologies, Indonesia should pursue a policy that is conducive to labor or job growth along with advancements in technology. However, recent laissez-faire attitudes toward jobs growth in far-removed adjacencies of technology demonstrate that not all technology jobs are created equal. Much like the deregulation case in continental Europe, the creation of certain jobs in the “new age” Indonesian economy very much has the trappings of the newly created jobs in the post-commodity era.

Businesses are mostly turned away from operating in Indonesia, mainly due to its notoriously difficult and serpentine business environment.

Despite the legal classification of most of these jobs, their nature has very much remained the same, only now their frequency is dictated by optimizing software. These jobs face the same level of instability and insecurity as those in the informal sector. While the results of laissez-faire attitudes in the labor market are similar for Indonesia and continental Europe, the effect is surely to be worse for the former. This is because a neoliberal approach in the Indonesian labor market context, a country with low levels of human capital, can only mean increases in low-skilled jobs.

Simply put, such reforms are barely legally formalizing the informal sector of the economy, without the other added benefits of entering the formal realm of employment such as training and skills development. These semiformal jobs (traditionally informal jobs but only recently formalized) present a suitable substitute to skills upgrading, which requires time and effort. Additionally, formal jobs that do offer training are more likely to compensate poorly during and after the upgrading process when compared to these informal jobs.

While the notion that these people get shifted to less productive parts of the economy is debatable, considering the rupiah value of the services rendered compared to other industries, it does represent an opportunity cost economy-wide. These tech firms will eventually reach a critical mass that will
enable them to dictate wages as their cost structures see fit or adopt technologies to fully replace their labor, or they will be subject to regulations that limit prices, which in turn will harm their workers. This will result in an opportunity cost, but is also contradictory to some of the nation’s more strategic and idealistic aspirations and targets.

However, the increase in the number of semiformal jobs is partly due to the formal sector not being able to offer sought-after or lucrative jobs. This dearth is because few firms based in Indonesia engage in innovative endeavors. Businesses are mostly turned away from operating in Indonesia, mainly due to its notoriously difficult and serpentine business environment. Chief among the problems are
nationalistic economic policies that are not only promoted but also implemented. For example, consider the Indonesian local content requirements (LCR) that are levied on many innovative industries. While the policy goal is noteworthy in that it tries to encourage raw material and intermediary production in innovative industries, Indonesia fails to realize the stark reality of the labor market. Instituting LCR policies fails to understand that there are not enough skilled workers for any reasonable capital holder to invest in starting a supporting business here.

Protectionist attitudes are harming the ability of companies to be competitive and perform research. The best course of action is to consider which industries are really fully suited to temporary LCR policies. When industries are heavily dependent on foreign imports, they are more likely to depend on foreign talent, which can catalyze a spillover of technical know-how. This is especially true of firms that import and perform research and development in the country (Todo and Miyamoto, 2006.) It is through this spillover effect that local talent can hopefully gain the necessary skills and, with the right
encouragement and supporting institutions, eventually become entrepreneurial to start businesses that are not only economically viable but also present alternative solutions that are cost-effective compared to imported materials. Should the capabilities of local counterparts increase, they can possibly supplant existing foreign players in the domestic market and perhaps international markets.

In addition to a more thoughtful approach to local content requirements, an overarching and coherent intellectual property strategy is needed. Indonesia’s Ministry of Law and Human Rights recently released Regulation No 39/2018, which mandated compulsory licensing of innovative drugs. While implementing steps have not been carried out, it certainly sets back any progress gained initially by the postponement for five years or more of the local content requirements for manufacturers.

Such strategies are outdated and nearsighted. If the issue was about access to lifesaving drugs, then surely the solution is a proper and transparent pricing system underlined by a systematic approach to procurement. This is another case of faux nationalism coupled with poor policy-making that is counterproductive to Indonesia’s goal of creating a knowledge economy and introducing innovative drugs, and in the larger context innovations in general.

Beating a dead horse?

Unsurprisingly, education plays a role in addressing labor market polarization. Policymakers have made the right overtures in addressing the hollowing out of the labor market by undertaking new policies. However, public-private initiatives such as matching skills to job requirements, vocational programs, internships, link-and-match and the creation of new degree programs are mere stopgaps.

Assuming the current educational attainment of the labor force holds, these quick-win policies only target middle-skill jobs, catering to members of the work force that have high school-level vocational qualifications or a diploma or universitylevel education – roughly 20 percent of the working-age population. Even this estimate would be considered optimistic when considering the actual efficacy of the policies and conversion rate, given the quality of the graduates entering the work force. This
harrowing reality is reflected in a survey of adult skills conducted by the Organization for Economic Cooperation and Development (OECD), a group of mostly rich countries. The survey finds that the average Jakarta resident with a tertiary education scored closely in literacy to an average OECD citizen with less than an upper secondary education, but significantly less in numeracy to their OECD counterparts with a secondary education.

Developing policies that aim to stimulate productivity growth while reducing inequality cannot be inherently elitist and targeted; otherwise, the entire idea of an educational roadmap will devolve into a white elephant project. Reforms need to start much earlier than this phase. With oscillating scores
through each iteration of the test, falling ranks and a very small share of students classified as top performers, Indonesia’s Program for International Student Assessment (PISA) scores leave much to be desired. The PISA test is administered to 15-year-olds every three years. However, given the performance trend and the high levels of spending from the constitutionally mandated 20 percent of the national budget on education, it could be said that Indonesia’s investment in education is experiencing a negative marginal return of sorts on skills development in the formal education system.

Herein lies the issue of discussions regarding the higher order thinking competencies of Indonesian students. The trend of plateauing would seem to suggest that perhaps policy and investment focus should move to an earlier part of a child’s cognitive development.

Harvard University published a study in 2015 on the cognitive development of the human brain. In essence, it concluded that, from a neuroscience perspective, the moldability or plasticity of the human

What Indonesia faces is no easy task.

brain to take on skills that foster higher order thinking is at its best in the prenatal period up to the first five years. As the human mind ages, plasticity decreases and it becomes much more “wired.” With this, remedial actions that try to alter behavior and patterns built on shaky foundations are more costly, mentally and financially, than implementing nurturing environments and programs.

Research has indicated that enacting early childhood development policies that target those most at risk is more likely to generate more positive returns (externalities and beyond) than corrective measures aiming to provide workforce skills to teenagers and adults. Long-term studies on model programs have demonstrated a benefit-cost ratio of up to 17:1, while returning 18 percent on investment over 35 years. After years of policy neglect that has manifested itself in a measly 4.47 percent of the Indonesian Ministry of Education’s 2016 budget allocation, the need to invest in early childhood education and development programs is the most pressing. Such a program would complement President Joko Widodo’s anti-stunting program, as the nation aims to improve its human capital index score.

A program such as this will cost enormous amounts of financial and political capital, most of which decision makers would likely rather spend on policies that are conducive to electability or political support. Most of the problems that currently plague the existing educational system are surely to be found in the future implementation of an early development program, not least compounded by a lack of subject matter experts, a lack of coordination among stakeholders and continued politicization of curriculum. However, with objectives being missed within Indonesia’s existing school system, perhaps a novel approach is needed.

Looking forward

What Indonesia faces is no easy task. With the rapid pace of technological development and advancement, Indonesia’s window of opportunity to claim its rightful place in the global value chain is closing ever quicker. Having grown accustomed to the conditions that Dutch disease entailed,
especially the mentality of a “race to the bottom,” it now faces the difficult task of upgrading its labor force. There is hope there will be progress. Initiatives such as a recently established skills development fund by Indonesia’s national development planning ministry are a proper first step in addressing the skills gap.

However, more must be done in terms of integrated policy crafting and execution. As things stand, interministry initiatives seem more of a coincidental and convenient occurrence than a concerted effort and have only delivered incremental results. A sensible next step would be to create a skills development body akin to India’s National Skill Development Council – a body responsible for training labor in the informal sectors of the economy. Such a system embraces the need for increased collaboration between both the public and private sectors in the hope that the program will become much more demand-driven. However, with the ever lingering shadow of an emerging commodity boom and continued economic nationalism, it would be ill-advised to hold one’s breath.



"Giovanni Ugut is a senior associate at the Jakarta offices of a prominent global financial advisory firm. He completed a double degree in mathematics and economics at the University of California, San Diego."

You need to login to write a comment!