The Employment-Impact of Automation in Canada

Standard neoclassical models of labour demand predict that automation does not produce long-term increases in unemployment. Supporting evidence in Canada between 1970 and 2008 is explained by the reallocation of labour from industries with high levels of automation such as Manufacturing to industries with low levels of automation such as Retail and Wholesale Trade, and Business Services. Recent evidence indicates however that on-going technological advances are now driving labour automation in many industries, and will likely expand to others in the near future. As automation technologies grow more sophisticated and decline in cost, Canada faces the prospect of historically high levels of structural unemployment due to declining labour demand. The study utilizes previous work on occupation-level automation probabilities to estimate the impact of automation on Canadian employment over the next two decades. Findings suggest that up to 50% of Canadian jobs will be at high risk for automation over this time frame. Suggested policies for minimizing unemployment include a nationally coordinated education and skills-training framework, and greater publically-funded assistance for the creation and development of small- and medium-sized enterprises. Beyond the next two decades the study suggests that policies will need to shift in focus from employment to redistribution.