The Bank of Canada (BoC) official who put the productivity “crisis” in the spotlight, returned this week with a familiar fix: invest in artificial intelligence (AI). Deputy Governor Carolyn Rogers urged businesses to adopt AI, comparing its productivity potential to the adoption of the internet. BMO chief economist Douglas Porter dumped cold water on those expectations, revealing the crisis isn’t recent, but rather the norm—and the internet adoption didn’t deliver the long-term boosts that many assume. He expects AI will help, but suggests the payoff will be far smaller than the hype.
Canadian Productivity Crisis 101: Wages Rise, But They Buy Less
Productivity is the difference between earning more and affording more. When workers don’t produce more, rising wages become inflationary (whether captured in CPI or not). That’s a big reason younger Canadians—and their peers in Europe—feel worse off, despite wage growth nominally outpacing inflation. Wages are up, but the basket is smaller, and the “security” that previous generations bought with much less income, remains out of reach.
Canada’s productivity problem isn’t new. Growth has been largely population-driven, even before it became a hot-button issue in the 2020s. Most growth has come from adding more people, rather than through innovation or productivity, as one would expect from a country with world-renowned universities. However, this deficit of productivity only became a public concern after BoC Deputy Governor Rogers called it a “crisis” two years ago, an alarm she rang again yesterday.
“This is an echo of her famous ‘break the glass, productivity emergency’ speech from two years ago. Since that speech, some serious upward revisions have left productivity in a less dismal spot,” explained BMO economist Douglas Porter.
Even so, BMO estimates output per hour since the pandemic has grown just under 0.7% annually. It’s weaker than the 1.0% long-run trend, which Porter noted is “already a bit disappointing.”
BoC Doubles Down On AI As The Fix For Productivity Crisis
Deputy Governor Carolyn Rogers pitched AI as the cure for Canada’s productivity shortfall. Her latest speech urged small and medium businesses to invest in AI technologies, comparing it to the adoption of the internet. This follows BoC Governor Tiff Macklem making a similar speech to business leaders last week, albeit a little less elegantly.
It’s a logical pitch, but has anyone actually done the math on the impact in Canada? The internet undeniably transformed business, but did the country turn those gains into a net productivity boost?
BoC Pitches AI As Internet-Scale Productivity Gains, but Canada’s Gains Were Temporary
Since productivity data is only readily available back to the 1980s, BMO used real GDP per capita as a long-run proxy. “And that shows quite vividly the underlying slowdown over the decades, only briefly interrupted by the burst in productivity in the late 1990s/early 2000s,” explained Porter. “That bump was spurred by the widespread adaptation of the internet.”
The “internet moment” we’re all hoping to re-live was a short-lived surge. After companies digitized, productivity resumed its downward trend. Porter hopes AI can replicate some of that shift, but remains mildly skeptical: “The hope is that AI can generate something akin to that surge—here’s predicting a much milder version of that episode over the next 10 years.”
BMO doesn’t directly spell it out, but there are two things hard to ignore. The first is that Canada’s internet-era boost didn’t lead to a permanent acceleration of productivity. It faded as the economy pivoted towards cheap credit and asset inflation. That’s when capital increasingly chased financial returns instead of productivity, producing the housing drag we’ve all come to know and love.
Second, AI presents a big opportunity for productivity gains, but it’s unclear if those are net gains. The risk to growth is whether AI scales faster than workers and institutions can adapt. If it does, the productivity gains concentrate in a winner-takes-all economy, while displacement spreads. In the medium run, it risks being a net drag as the productivity gains are offset by a drag on total economic output as workers are displaced for retraining. As Canada sees long-term unemployment rise to 90s levels ahead of this transformation, there’s a very real threat of productivity churn.
What’s interesting about these mass layoffs are people don’t realize it’s coming for the middle manager, white collared jobs first.
It’s going to take longer for AI to replace plumbers than it takes to replace anyone at Service Ontario.
The BoC ceased to be of any value in 2018. As anyone can tell you, bankers, politicians haven’t a clue how to create economic growth. In the last decade, not only did we see our standard of living collapse more than it did during the Great Depression, but outside of the trans Mountain Pipeline, not a penny spent was useful.
What should be a concern for Canadians is we somehow ended up with the same losers who wasted $2-3Tr on bad investments, corruption and waste to ‘invest’ even more? The job losses are just going to accumulate, and given the pressure on sme sectors, we wont actually know how bad it is.
I was at the supermarket yesterday and cucumbers were $5 each? Now this is clearly not just inflation. Canada has chosen to exclude the USA for short term political reasons, and apparently the BoC, and media thinks that all of this is still worth it? Remember that the main beneficiaries of the tariffs protecting banks, media, groceries, food producers, and so on are the same people telling us fighting the USA is our only choice?
If we cancelled sly management, allowed free trade in all foods, including dairy, poultry, beef and so on, how would that not lower prices? How about competition for our banks who operate in the USA with margins less than 25% of what ours are? They can apparently compete with US banks, just not in Canada?
Can the AI replace the politicians and bankers first?
I know CRA and bank layoffs last year were tied to AI, but I also know that noone believes the systems that replaced these people are fit for task.
AI is really just hastening the enshitification of everything that this point.
We have seen that the media, banks and politicians are actively choosing a path that is bad for everyone else to protect their positions and revenue. Carney has been a complete washout as PM, failing to accomplish anything in a year, with the exception of seeing Canada excluded from trade talks once again in DC? In addition, he is wasting hundreds of Billions on stuff that doesnt have any impact on the lives of regular Canadians? Does buying a bunch of weapons from external companies help Canada? Do we need subs, fighters, and so on, while we cant feed our kids?
So the answer is we picked the worst choice of the 5 for PM, and frankly, should have known better. As long as a liberal PM is in Ottawa, things will continue to slide.
Why is the Bank of Canada pushing so hard to accelerate job losses? Do they really need lower rates that bad? It’s a really bizarre, non-industry supported take.
Why isn’t every other central bank discussing this?
I’m sure it has nothing to do with this.
https://bam.brookfield.com/press-releases/brookfield-launches-100-billion-ai-infrastructure-program
The problem here is pretending that productivity, real gdp per capita, or cpi or nominal gdp per capita are in any way accurate. Consider, since Canada’s CPI is an outlier, using interest as a floating not fixed cost, despite it not being that, has understated Canada’s CPI for years by 15-20% or more. Remember if we compound a 15% understatement of CPI by 20 years, it makes any calulation of ‘real’ gdp nonsense. Add to that the major issues with cpi methodology to begin with, and none of this is accurate.
So if the data is wrong, who cares what it says? With respect to productivity, we have seen a net decline in the west since the 1970s relative to emerging markets. In 1950, Canada employed far far more people producing goods for use by people, not services of questionable value. To suggest that AI can ‘fix’ this is nonsense. AI can further replace clerical, sales and administration staff, manufacturing, and so on. Just as tractors and combine mechanized farming, a single farmer can now produce 1000x what they could in 1860. The problem is there are 10 people for every one there was in 1860, and almost none of them is producing anything of value. So if we can use AI to replace all these people, that might drive higher profits, but wont raise incomes, wages or standard of living.
Canada has huge potential for resource extraction, farming, technical innovation. The problem is, the govt seems focused on keeping investment where it isnt needed. Had ottawa decided to spend 90B building pipelines, upgrading rail for commercial purposes, ports in BC or Hudsons bay, that would drive productivity. Building a fast train from toronto to montreal through Ottawa wont.
Other studies have shown more than 95% of AI adopters so far have little tangible benefit so far. Maybe, these gains are more in the minds of the tech bros of Silicon Valley who are trying to raise trillions wherever they can.
You are bang on. Open the Canadian economy to foreign capital and do not punish sectors that are essential, where we have tradition and skills and lots of research.
Energy and mining, and rebuild the GRID!