The impact of electricity on productivity: an industry-level approach to Swedish manufacturing, 1913–1938

Lilla salen Session 2: Consequences of Technological Change organized by Suvi Heikkuri


Svante Prado, Christopher Absell and Jesper Hamark


Technological change takes place spasmodically, as research demonstrates; a commonly shared view among economic historians is that a few sets of specific technologies, coined GPTs, have had the most far-reaching impact on economic development (Bresnahan & Trajtenberg 1995). Most authors agree that electricity deserves the epithet of a GPT. The introduction of electricity, and in particular the use of small electric motors, did away with the shafts, pulleys and belts that criss-crossed the factory hall. Instead of being propelled by a single source of motive power through line shafts and countershafts, each machine was run by a small electric motor. The growing use of unit drive facilitated a complete reconfiguration of the factory. The operation of the factory now required less manual workers. Besides reducing the demand for manual labour, unit drive entailed several indirect savings, such as increased flow of production, improved working environments, improved machine control and ease of plant expansion.

The transition from steam to electricity was rapid in the US, in particular during WWII. Much of the remarkably rapid growth of total factor productivity in manufacturing during the 1920s probably stemmed from the benefits that could be reaped from electrification (Field 2006). Because electrification occurred across the board of industries, we could perhaps expect that productivity advances across industries be similar. A way to test this notion is to deploy the Harberger’s distinction between yeast and mushroom-like patterns of productivity growth rates: he labelled an even pattern yeast-like and an uneven pattern mushroom-like. David and Wright (2003) argue that productivity growth rates became yeast-like in the 1920s and 1930s as the advantages of electricity came to fruition.

In Sweden, though, electrification proceeded even faster. Prado (2014) argues that the uniform productivity pattern that manifested itself in the first decade of the twentieth century owed to electrification. Before the turn of the century, productivity advances were largely mushroom-like because steam never had the potential to foster productivity growth rates across large swathes of the manufacturing sector. Prado’s investigation ends in 1912, however, coinciding with the end of the old version of the industrial statistics. To confirm the truth of his conjecture – that electricity was responsible for the unified productivity growth patterns in the first decade of the twentieth century – would require us to examine the interwar years, when electrification accelerated, first, and then came to an end when all manufacturing processes were run by electric motors in the end of the 1930s.

Therefore, in this paper we will attempt to continue the search for productivity advances and electrification in the interwar years. We do so by drawing on a sample of sixty industries in 1913–1950, combining the volume output series of Johansson (1985) and various issues of Board of Trade, and employment and other variables from the Swedish Industrial Statistics. We use the Harberger metaphor of yeast versus mushroom to identify whether electricity had the qualities of a true GPT. We then apply panel data econometrics to establish the impact of electricity on productivity advances.


No PDF available.