By Wenbo Zhu, JDI Student Fellow, Queen’s University

Some technological advancements are skill-complementing, meaning that they tend to increase the productivity and demand for skilled workers. Other technological advancements are skill-replacing, meaning that they tend to reduce the demand for skilled workers and raise the productivity and demand for unskilled workers. Electronic computers are typically considered a prime example of skill-complementing technologies, whereas assembly lines and the use of interchangeable parts in the manufacturing industry are classic examples of skill-replacing technologies.

Disentangling the impacts of each type of technology is important for understanding of the impact of technological changes on labor markets. However, there is little work trying to disentangle the competing effects of the two types of technological advances. The main issue is that neither skill-complementing nor skill-replacing technologies can be easily measured at an aggregate level.

To this end, in a recent paper, I develop a framework that interprets a skill-complementing technology as the flow of a series of product innovations and interprets a skill-replacing technology as the flow of a series of process innovations. With this interpretation, I combine two interesting European datasets (Community Innovation Survey and the Structure of Earnings Survey), and identify a large positive impact of the skill-complementing technology relative to the skill-replacing technology on the skill premium (defined by the wage ratio between the college educated and high school graduates).

The assumption that is central to this interpretation, is that a new product, either a good or a service, when first introduced, is often complicated and tends to require skilled workers to implement (product innovations). Over time, as firms accumulate production experiences, they learn to develop new production methods and/or arrangements, which break down complicated production processes into less skill-requiring and more manageable pieces (process innovations). The importance of these process innovations is that they allow firms to replace expensive skilled workers with less-expensive unskilled ones, without compromising productivity.

The financial industry provides some useful examples of the “product-then-process” innovation pattern hypothesized above. The mortgage loan service, for example, when first introduced, requires banking officers and financial specialists to accomplish. Over the last 20 years or so, through the development and application of the information technologies and the re-organization of the banking industry, more and more people gather mortgage information and apply loans through online banking services or even mobile financial services and then finalize their transactions with less-specialized bank tellers. As a result, in the process of the mortgage loans service, skilled banking officers and financial specialists sitting in branches have been gradually replaced by a combination of the online banking interface and more and more less-specialized bank tellers.

My main empirical finding is that, a 1% increase in the skill-complementing technology relative to the skill-replacing technology would raise the skill premium by a range from 0.05% to 0.16%. To put the result in context, if the skill premium is 2, which means college educated workers earn twice as much as high school graduated workers, then after the 1% increase in the skill-complementing technology relative to the skill-replacing technology, the skill premium becomes something in the range of 2.1 to 2.32.

In contrast, I find that a 1% increase in the supply of skills would reduce the skill premium by a range from 0.06% to 0.09%. The empirical finding overall is remarkable as the relative labour supply is arguably the single most important factor on the supply side of the labour market, and yet the measure of the relative intensity of the two technologies generates a demand-side impact on the skill premium at least as large. My sample includes most European countries from 2000 to 2010 at the industry level. The Figure illustrates the positive correlation observed between the skill-complementing technology relative to the skill-replacing technology and the skill premium.

This result has direct policy implications, as the labour demand side is usually more responsive to policy interventions than the labour supply side. In particular, reducing the cost associated with process innovations that firms face, by encouraging collaborations of developing process innovations for instance, can be effective in reducing the skill premium.