In this week’s blog, NERI economist Paul Goldrick-Kelly summarises some key findings from the NERI's co-authored paper examining the effects of automation and the green transition on employment between 2008 and 2019. While the most gloomy predictions of mass job losses haven’t come to pass, workers within some sectors may face challenges.
In a forthcoming paper, Dr Lisa Wilson (NERI), Esmé Sheridan (UCD), Ian Walsh (UCD) and I investigate changes in employment in the Republic of Ireland between the crises of 2008 and our current pandemic crisis which began in early 2020. Specifically, we examine whether two major structural forces, in the form of automation/technological change and the low carbon transition shaped employment developments over the period.
Where automation is concerned, the evidence is mixed. Trends in aggregate employment and unemployment between 2007 and 2019 are not consistent with more pessimistic projections of mass unemployment from technological advance. Evidence of labour substitution (replacing workers with machines) is not evident in the data we have for hours worked, which recovered following the crisis period 2008-2012.
However, we have seen an apparent structural shift to increased part-time work and declines in labour market participation. This is driven in large part by declines in labour market participation on the part of men, although men’s participation rates still exceed those for women.
Despite disagreement about the employment effects of automation within the literature, there is a growing consensus that some types of jobs face higher risks than others. Low risk sectors appeared to fare best through the crisis. However, many sectors identified as high risk – from different theoretical perspectives – showed significant growth as the crisis ended. This suggests that advances in technology and their take up did not necessarily bias employment away from these jobs.
We have clearer evidence regarding skills and routine biased change. Employment in Ireland is increasingly skewed towards higher skilled work, away from low and middle skilled employment. This parallels the trend towards an increasingly educated workforce. Employment also seems to be biased towards non-routine abstract and service work, consistent with theories which predict greater job displacement where work has higher proportions of routine tasks. These routine tasks are more readily automated.
The available evidence suggests that a structural shift towards a green economy is more prospective than currently present. Data suggest that the Republic of Ireland may have begun a process of decoupling economic activity from greenhouse gas emissions, but this largely depends on how one assigns emissions occurring outside of Ireland’s borders. Territorial emissions recorded in the Republic of Ireland – rather than those attributed to Irish residents abroad (including airlines), or based upon consumption – fell between 2016 and 2019 while output and employment grew. However, emissions reductions have not occurred at necessary rates.
Green employment – as measured by more restrictive industry-based definitions – has grown significantly in Ireland. More expansive occupation or skills-based measures of green jobs would likely show a more significant employment profile. Emissions are sectorally concentrated, 90 per cent of emissions are tied to just over a fifth of overall employment.
Progress regarding emissions reductions is concentrated in the Electricity, gas, steam and air conditioning supply sector. While still a significant emitter, with high emissions per person employed, its reductions of more than 6 million tonnes of CO2 equivalent between 2008 and 2019. Agriculture, forestry and fishing remains the problem sector in territorial emissions terms – emissions grew by over 1 million tonnes of CO2 equivalent even as employment fell, increasing emissions per person employed.
A comparison with other comparably developed western European states suggests that our employment structure contributes to our laggard status in emissions terms. Hypothetical adjustments to our employment profile to match comparators keeping emissions per job constant shows significant emissions reductions (in many cases in the millions of tonnes), which have only grown over time. However, an alternative counterfactual, where employment remained constant but emissions per job were adjusted to match comparators also produced significant reductions in non-household territorial emissions in many cases – over 20 million tonnes of CO2 equivalent if emissions intensities matched those in Sweden in 2019.
The gloomiest predictions of the widespread net employment losses in employment due to automation and green transition processes are not apparent in the data or the prospective literature in the latter case.
However, these processes may affect some sectors and workers more than others. There is considerable scope for significant changes influenced by automation and the green transition is just beginning. While the green economy may be showing signs of growth, with some definitions implying a significant present scope, our current carbon situation and lack of progress suggests that these sectors will also face challenges.
Our analysis suggests that some sectors, such as Agriculture, forestry and fishing, face overlapping challenges from technological change and necessary emissions reductions. Further research should be undertaken to assess these sectorally specific transition difficulties and assemble appropriate responses.
At the same time, while these structural transitions may have sector or issue specific implications, they may show significant overlap in risk and difficulties. A common Just Transition approach incorporating social protection, activation support can ease difficulties faced by workers and their communities. A proactive approach to planning which incorporates workers and their representatives, alongside other civil society groups has also seen success elsewhere.
Our Working Paper can be accessed here and a related NERI InBrief is available here.