In a recent article I explored whether investing in technology is inevitably going to harm the employment prospects of workers.  The article revolved around the fact that employment often rests upon productivity, and that not investing in technology is a sure fire way to harm productivity, which will then harm our employment prospects.

It’s a notion that has a lot of merit, but a recent paper reminds us, not all investments in technology are liable to have the same impact.  The authors argue that the benefits highlighted in my previous article do undoubtedly arrive when the technology we’re investing in provides clear productivity advantages over the existing human workforce.  Where problems (in terms of employment at least) come in, is when the technology only provides a slight boost to productivity.

“The presumption that all technologies increase (aggregate) labor demand simply because they raise productivity is wrong,” the researchers explain. “Some automation technologies may in fact reduce labor demand because they bring sizable displacement effects but modest productivity gains (especially when substituted workers were cheap to begin with and the automated technology is only marginally better than them).”

They conduct an empirical analysis of historical labor and technology trends to examine what the evidence from the labor market tells us in regards to past investments in technology.  They suggest that in both the 1990s and 2000s, organizations predominantly invested in technology that provided only a modest boost to productivity.  This, they argue, has contributed to a net decline in demand for labor, and a stagnation in wages.

Smarter investment

This chimes with other analyses of our current economic situation, which have equated our stagnating productivity and wage growth with a poor investment in technology, especially among smaller firms.  What this paper argues is that any investment has to actually deliver significant increases in productivity, and that arguing that technology investments are good in and of themselves is not enough.

What’s more, they argue that even if technology does raise productivity, it might not result in the kind of wage increases we normally associate with productivity growth.

“In fact, as we noted already, automation by itself always reduces the labor share in industry value added and tends to reduce the overall labor share in the economy (meaning that it leads to slower wage growth than productivity growth),” the authors say. “The reason why we have had rapid wage growth and stable labor shares in the past is a consequence of other technological changes that generated new tasks for labor and counterbalanced the effects of automation on the task content of production.”

The need for new tasks

As wage growth is predicated on people finding new tasks to perform, the researchers argue that the recent wave of technology has not resulted in new tasks being created. They argue that while the incentives for investing in technology have largely increased, those for creating new tasks for humans to perform have not.

Tax incentives often subsidize the investment in technology while at the same time dis-incentivizing the employment of labor. This has been exacerbated by the recent trend of technology companies growing very quickly with a relatively small workforce, with digital technologies allowing tasks to be either automated or performed at scale by a skeleton workforce.

“If the balance between automation and new tasks has shifted inefficiently and if indeed this is contributing to rapid automation, the absence of powerful reinstatement effects, and the slowdown of productivity growth, then there may be room for policy interventions to improve both job creation and productivity growth,” the authors argue.

The authors attempt to remind us that just as technology will not replace all human jobs, it’s also unlikely to be an instant route to productivity improvements either.  As is so often the case in life, the likely reality is more nuanced and sitting somewhere in between these two extremes.  It’s a level of nuance that is likely to be crucial if policy interventions are going to have a positive impact.