A taxonomy of productivity growth in Portugal

February 2, 2021

Technologies such as computers, robots, and artificial intelligence are changing how we work and interact, rendering routine jobs (which are also middle-wage jobs) obsolete due to automation and computerization of many tasks. In contrast, high-skilled workers that do complex abstract tasks benefit from this, as technology increases their productivity. There is evidence that several labor markets have polarized in the past two decades between high-skilled abstracted jobs and low paid jobs (with the simultaneous disappearance of routine middle-skilled jobs). This paper studies whether polarization has occurred within or across firms and how labor input changes have contributed to productivity growth in Portugal during the period from 2004 to 2009.

The paper develops a firm taxonomy based on their workers' tasks. Firms can be focused on one task - Abstract, Manual, or Routine - on a combination of tasks, or have no focus. The paper finds the abstract firms have increased their prevalence in the economy, while Routine firms declined (see Figure 1A). The results robustly show that Abstract firms are the most productive and manual firms the least (Figure 1B). Abstract firms accounted for the largest contribution to the Portuguese aggregate productivity growth (22%), while Routine and Manual firms contributed negatively with -0.6% and-1.5% respectively. Productivity growth in abstract firms increased, but was mostly propelled by increased market shares of the most productive incumbents and exit of the least productive firms.


The results suggest policymakers should develop innovation and knowledge capabilities to promote the creation and growth of competitive Abstract firms as well as education and training policies to reduce the skill mismatch occurring in labor markets with similar low productivity levels undergoing similar changes. Social welfare policies are also needed to tackle the high prevalence of long-term unemployment and the existence of large segments of the labor market where short duration and low-wage jobs prevail which will probably persist or be aggravated with the deepening of automation technologies.

Click here to go to the paper by Tiago Fonseca, Francisco Lima, and Sonia Pereira.