Worried about Canada’s productivity? Why too much productivity may soon be the problem

In a recent speech, Bank of Canada senior deputy governor Carolyn Rogers considered Canada’s productivity record a national “emergency” that’s jeopardizing our standard of living.
The solution seems quite simple: capital investment per unit of labour must rise — which implies greater labour-saving technologies embedded in new machinery and equipment. But the opposite has occurred: businesses investment per worker decreased 20 per cent from 2006 to 2021.
So why are Canadian businesses not making the necessary productivity-enhancing investments to increase their profits? To a large extent, government policies that kept real wages low — and profits high — created a strong disincentive for businesses to undertake risky investments to raise their profits.
In part, this disincentive may soon come to a halt. “We need Canadian businesses to invest in training and technology, not increasing their reliance on low-cost foreign labour,” Prime Minister Justin Trudeau recently stated in reference to incoming changes to the Temporary Foreign Workers program.
But even if governments cannot always be trusted, there is still room for optimism since the adoption of revolutionary artificial intelligence (AI) technology may be around the corner. “Every few decades, a new technology comes along that fundamentally redefines society,” says Joel Blit of the University of Waterloo. “Artificial intelligence is the technology of our age,” he adds. And the economics profession seems to agree.
Most economists are convinced that AI will lead to a surge in productivity and rapid economic growth. A Goldman Sachs study estimates that widespread adoption of generative AI could hike U.S. productivity by 1.5 per cent a year, thus raising annual productivity by about three per cent as in the decade before the turn of the millennium. In Canada, a Microsoft report finds that generative AI could boost productivity by eight per cent by 2030 — thus raising GDP by almost $200 billion a year.
But technological change also involves disruptions to the labour market. In the past, however, labour-saving technologies increased the ranks of the unemployed only in the short run for new products and new jobs were almost simultaneously created. This time, though, it may be different. “AI could end up destroying more jobs than it creates,” Bank of Canada governor Tiff Macklem warned us last month. 
Indeed, Goldman Sach researchers estimate that 300 million workers across the most advanced countries could be in danger of losing their jobs, raising unemployment to 35 per cent or so. In Canada, a recent StatCan study finds that widespread adoption of AI could result in the elimination of 30 per cent of existing jobs.
If so, will our socio-economic system survive a chronic situation of 30-35 per cent unemployment? Most likely not.
Let’s point out that technological advances do not directly lead to improvements in living standards. On the contrary, due to its negative impact on employment, productivity growth initially results in a regressive redistribution of income from workers to the owners of means of production. As employment eventually rises and labour markets become tight, workers regain bargaining power vis-à-vis their employers, and a redistribution of income in the opposite direction occurs. This way, both capital and labour eventually benefit from technological change.
But this time may be different. It appears the widespread adoption of AI may result in chronic high unemployment, thus preventing the progressive redistribution of income from capital to labour. The outcome, therefore, may be socially unsustainable: permanent high unemployment and greater income concentration.
So, the state must — and will — intervene to save the day and ensure the reproduction of the system. Indeed, the government will implement policies to de-escalate the situation, improving the distribution of income and reducing unemployment to manageable levels.
Introducing universal basic income (UBI) programs is one such policy proposed to help reduce income inequality. While expensive, UBI can be financed by taxing the windfall profits resulting from the widespread adoption of AI technologies.
But UBI cannot be the definite answer: it may facilitate the reproduction of the system in the short run, but likely not in the long run. Why not? Because it will create two “castes” in society: the “employed” and the “unemployed.” And these two castes will reproduce with little mobility across them, thus eroding society’s cohesion and making the situation unsustainable in the long run.
The most important issue, therefore, is to prevent unemployment from becoming entrenched at high levels. How? By reducing the workday.
The eight-hour workday was first introduced more than a century ago in Uruguay, in 1915. In the U.S., the eight-hour workday became law in 1938. Being the usual laggard, Canada waited until the 1960s to enact a similar law.
Since then, productivity has skyrocketed. In the U.S., it has increased by more than 150 per cent since 1938; in Canada, by about 110 per cent since 1960.
But wages have not kept pace with productivity growth — thus more profit has been squeezed out of every hour of a worker’s labour. So, reducing the workday seems long overdue.
Hence, we shouldn’t wait until unemployment becomes chronically high to start reducing the workday. We should start now to keep the unemployment and participation rates at “normal” levels as the adoption of AI proceeds.
And even if those who doubt the adoption of AI will result in chronic high unemployment were correct, UBI programs and a shorter workday should still be promoted as a matter of social justice.
Indeed, it’s time to start reversing the income concentration and greater labour exploitation of the neo-liberal era. That would be the sign of a caring society — a very Canadian thing.