"Video Killed the Radio Star": Why Hasn’t Technology Destroyed All Jobs?

Destruction of jobs by machines is an old issue that has appeared several times in history beginning with the famous Luddite resistance to the introduction of textile machinery in England early in the Industrial Revolution. It arose again during the 1960s when President Lyndon Johnson appointed the Bowen Commission to study the technology/job loss nexus (Autor). In 1979 The Buggles reflected back on that period with their famous song titled “Video Killed the Radio Star”, which included the lines “they took the credit for your second symphony, rewritten by machine on new technology”. The issue of the impact of technological change on jobs has resurfaced recently due to the new wave of automation. Recent interest has also been stimulated by the extraordinarily high unemployment during the Great Recession, and the slow growth in average wages since then.

Fast food chains in the U.S. are rolling out self-service kiosks (Wong). The four largest German airports have been testing automatic border control machines since 2014 (Spiegel). A hotel staffed entirely by robots opened its doors in Japan a few weeks ago (NY Daily News). These are just a few examples of how technological innovation is changing the economic landscape and more is yet to come. New technologies have undoubtedly destroyed specific jobs, and they have altered the distribution of employment for jobs with particular skills.

It is somewhat surprising that job loss should remain a sensitive issue in the U.S. at a time of full employment. The current unemployment rate in the United States of 5.3% is well below the long-term average unemployment rate of 5.8% for the last 30 years. However, unemployment continues to be a serious issue on the other side of the Atlantic – the seasonally adjusted unemployment rate for the euro area in May-June 2015 stood at 11.1% (Eurostat), clearly much above the natural rate.

There has been much discussion about the “polarization” of the work force, in which jobs requiring middle skills are disappearing at the same time jobs requiring high skills and low skills are expanding (Autor). Why is it that certain kinds of jobs are more amenable to automation than others? More generally, since technological innovation is a powerful and relentless force, why haven’t all jobs been destroyed by automation?

Routine Jobs and Automation

The main concern about job loss does not seem to be about insufficient total employment, but rather the kind of jobs and the average wages being paid. In a series of papers, David Autor (2015) has examined the effects of technical innovation on jobs. He distinguishes between routine jobs that are relatively easy to automate, and non-routine jobs that are not. Routine jobs can be automated because they are well-understood and computers can be programmed to instruct robots to perform a well-defined task. The routine job might be cognitive, such as alphabetizing a list of words, or it might be manual, such as moving heavy objects from one specific location to another. A task is routine if it is repetitive and can be codified. Some authors (Cortes et al) have attempted to identify specific routine jobs that are vulnerable to automation.

Conversely, non-routine jobs are not so easy to automate. They require judgment, discretion, and sometimes creativity. They involve new situations, and they are less repetitive than routine jobs. Artists, composers, inventers, and entrepreneurs try to create something new. Managers face and must respond to new situations. Some non-routine jobs are cognitive and require some formal education, but others are manual jobs that require judgment, but not necessarily a high level of formal education. Care-givers for people with physical or mental handicaps fit this category. The existence of non-routine skills provides an important limit to automation. On the subject on non-routine or tacit skills, Autor refers to the work of Michael Polanyi. Note that people possessing non-routine skills do not necessarily earn high wages. They are protected from automation, but if many people offer these skills, wages will be low. The wage for a specific skill depends on supplies and demands for each skill. Recently earnings of high-skilled workers in non-routine jobs have increased rapidly, as the supply of these skills has been rather unresponsive to higher earnings.  At the same time, earnings of workers with low-skill non-routine jobs have been stable or falling.

Polarization of Jobs

Will the trend toward polarization of employment and increasing inequality of earnings  continue? Autor notes that the rate of growth of high-skilled non-routine jobs has already slowed, which would mitigate further polarization of employment. Also the supply response to high earnings in high skill jobs places a limit on salaries. It takes time and money to acquire high skills, but adding to the supply of people with these skills limits earnings for incumbent workers. Reducing barriers to entering well-paid occupations would contribute to greater equality in earnings across occupations.

The recent increase in inequality of income has come primarily from an increase in the inequality of labor earnings, rather than an increase in the share of income going to owners of capital (Mulas-Granados). Increases in the salaries of the highest paid workers have been an important influence on inequality. Furthermore, the traditional distinction between workers who receive all their income from wages and salaries and capitalists, who receive all their income from financial assets and other forms of property has become increasingly blurred. The typical worker today owns substantial wealth in the form of housing as well as financial assets from pension funds.

 Supply of High Skill Non-Routine Jobs

Economic efficiency would be enhanced if the supply of specific skills responded quickly to changes in demand that might be stimulated by innovation. Conversely, an extremely inefficient response to an innovation would be a caste system that preserved jobs for members of a specific caste. Historically, guilds and some labor unions have benefitted members by blocking entry into lucrative occupations, but they have imposed larger losses on non-members. Occupational licensing and barriers to immigration have similar effects by increasing earnings in high wage occupations and increasing earnings inequality. Increasing the number of H1-B visas would be one way to add to the supply of science, technology, engineering, and mathematics (STEM) workers in the U.S. and reduce earnings inequality.

Access to training of workers in high-paid occupations is important, and training could come from many types of institutions, not just four year colleges. In addition to formal schools, apprenticeships, and other forms of on- the- job training could be important sources of new skills. However, minimum wages could interfere with profitable on-the-job training, and an explicit exemption or lower minimum wage for trainees would encourage employers to provide more training within businesses.

Emphasis on increasing the number of college graduates has become fashionable with some politicians, but it could be misguided. The average college graduate continues to earn more (1.6 times as much) than the average high school graduate, but not all college graduates are alike. “ Over the entire career, the highest-earning majors will earn about two-and-a-half times what the lowest-earning majors will earn, a range from over $2 million for some engineering majors to about $800,000 for early childhood education” (Hamilton Project). The difference between the highest and lowest earning majors is greater than the difference between college and high school graduates. Politicians continue to offer greater subsidies (loans, grants, lower tuition) to college students, but producing more graduates with skills in low demand is not economically efficient, and it can be frustrating for graduates and their parents if graduates with certain majors cannot get a job or can only earn a low salary.

In a period of rapid innovation, keeping workers’ skills current is an important task. Community colleges can be a useful source of skills, but they must avoid the problem of teaching skills that are obsolete by the time a student looks for a job. Certain for-profit schools have been criticized recently for poor performance, but there is no reason why education should be monopolized by not-for-profit institutions. Acquisition of skills within schools and on-the-job are both important, and in a dynamic economy skill acquisition and maintenance require lifelong education.

Conclusion

Anxiety about job destruction by machines has increased recently, but innovations that persist provide net benefits to society as a whole. The new technology substitutes for some workers, but by raising the productivity of other workers, it increases total production and is also complementary to labor.” Technology eliminates jobs, not work” (Bowen). Routine jobs are most vulnerable to automation, but non-routine jobs are protected from automation. A new song along the lines of “checkout machine killed the receptionist” remains to be written, but it will still be written by human artists, and thus its title will be more appealing than what we just proposed. The effect of technical innovation on earnings inequality depends on how responsive workers are in supplying skills in higher paid occupations.

Having a flexible labor market with rules and institutions that minimize barriers to acquisition of skills and entry to high paid occupations would mitigate earnings inequality. Policies that facilitate acquisition of skills in high wage occupations would improve the average quality of jobs, increase the national income, and reduce earnings inequality. Real world economies are more than zero-sum games, and policies that contribute to growth in total income should dominate those that merely attempt to redistribute a fixed total income. However, education and training relative to the demand for specific skills remains the most important single factor determining the employment and earnings outlook of an individual.

References

Autor, David. 2015. “Why Are There Still So Many Jobs?” Journal of Economic Perspectives, Summer.

Bowen, Harold R. (Chairman). 1966. Report of the National Commission on Technology, Automation and Economic Progress: Volume I.” Washington: U.S. Government Printing Office.

Cortes, Matias, Nir Jaimovich, Christopher Nekarda, and Henry Siu, 2014. “The Who and How of Disappearing Jobs”. VoxEU, October 2.

Hamilton Project. 2014. “Major Decision: What Graduates Earn Over Their Lifetimes”. September.

Mulas-Granados.Carlos. 2015. “Growing Wage Dispersion Increases Growing Wage Inequality”. EconoMonitor, June 15.

NY Daily News. 2015. “Japan opens hotel run entirely by robots.” July 20.

Spiegel. 2014. “EasyPass-System: Fluggäste können elektronische Grenzkontrolle nutzen.” June 17.

Wong, Venessa. 2015. “America, Meet McDonald’s Self-Service Kiosks”. BuzzFeed. July 28.