The digital transformation creates scope to make workflows and production processes more efficient. To quantify the impact of digital transformation on labour productivity, research often focuses on investments in digital technologies. This perspective neglects the fact that digital intermediate inputs, such as microchips or integrated software, also represent an important transmission channel for the efficiency gains from digitalisation. In a new study, Bundesbank researchers have examined how efficiency gains in the digital sectors have affected labour productivity in Germany, France and the United States.

Previous work on the impact of digitalisation on labour productivity has often focused on the role played by investments in digital goods. This includes, for example, the acquisition of computers for commercial purposes. Digital products, though, are used not only as capital goods but also represent important intermediate inputs, such as microchips and integrated software. In Germany, France and the United States, around half of the goods produced in the digital sectors serve as intermediate inputs. Besides the information and communication sector, the digital sectors also comprise parts of the manufacturing industry such as the “manufacture of computer, electronic and optical products”. The latter includes, for example, the production of graphics cards, microprocessors and the like.

Analysis of sectoral efficiency changes in a macroeconomic model with a production network

Our study takes this into account and examines how efficiency gains in the digital sectors have impacted on labour productivity in Germany, France and the United States via production linkages. Efficiency gains in a given branch of economic activity are measured here in terms of total factor productivity (TFP) growth rates. The latter reflect the part of sectoral output growth that is not attributable to changes in the inputs used. Estimates show that TFP growth in the digital sectors was exceptionally high in recent decades (see Figure 1).

To quantify the aggregate impact of these efficiency gains, we use a multi-sector dynamic general equilibrium model in which the economic sectors are interconnected through production linkages: products from one sector are not only used for consumption or investment but also serve as intermediate inputs for production in other sectors. The various intermediate inputs have only limited substitutability. Production flows between individual branches of economic activity are calibrated based on macroeconomic input-output tables. The latter provide information on production linkages in an economy.

Efficiency gains in the digital sectors are important but so, too, is their diffusion via digital intermediate inputs

Specifically, we start by feeding observed TFP growth rates of all economic sectors into the model. We then compare the simulated path of aggregate labour productivity with that of a counterfactual scenario. In the counterfactual scenario, it is assumed that there was no TFP growth in the digital goods-producing sectors of the economy during the period under review. The comparison reveals that TFP growth in the digital sectors had a considerable impact on aggregate labour productivity growth. Without the efficiency gains in these branches, US labour productivity growth would have been about 25 percentage points lower between 1996 and 2020, despite the comparatively small share of the digital sectors’ production in total gross value added. Similar results were found for Germany and France, where productivity growth would have been around 15 percentage points and 11 percentage points lower, respectively. Digital transformation has thus been a major driver of labour productivity in recent decades.

In a second scenario, we show that digital intermediate inputs are a key transmission channel for efficiency gains in the digital sector. Here, we assume that digital products can be used only for consumption or investment purposes, but not as intermediate inputs. Figure 2 shows that productivity growth would be significantly lower without digital intermediate inputs.

The simulation scenarios described above can also be conducted for other sectors of the economy. That exercise reveals that the non-digital sectors of manufacturing as well as the trade and transportation sectors have also made significant contributions to labour productivity growth. However, these economic sectors are significantly larger than the digital sector in terms of their share of total gross value added. Relative to its size, the digital sector has a disproportionally large impact on aggregate labour productivity growth.

Conclusion

The digital transformation has exerted a strong impact on aggregate labour productivity growth in recent decades. Our analyses based on a multi-sectoral macroeconomic model reveal that, absent the efficiency gains in the digital sectors, labour productivity growth in Germany, France and the United States between 1996 and 2020 would only have been about half as high. Production linkages play a key role in this regard. These ensure that the remarkable efficiency gains made in the digital sectors are passed on to end consumers via value chains. The extent to which this happens depends not only on the magnitude of the digital efficiency gains but, crucially, also on the relative importance of digital intermediate products. To increase the latter, continued efforts should be made to expand the digital infrastructure and digitalise businesses on a broad scale.

The authors
© Bert Bostelmann	© Nils Thies	© Nils Thies
Dr. Elisabeth Falck Economist at the Deutsche Bundesbank, Directorate General Economics	Dr. Oke Röhe Economist at the Deutsche Bundesbank, Directorate General Economics	Dr. Johannes Strobel Economist at the Deutsche Bundesbank, Directorate General Economics