How will automation affect Canada’s post-COVID economy?

Coming out of the COVID-19 crisis, Canada is poised to implement more automation, according to a recent study conducted at the University of Waterloo.

Joel Blit, a professor in the economics department, highlights manufacturing and retail as the two sectors that will undergo transformational change in the post-COVID-19 economy in his report Automation and Reallocation: Will COVID-19 Usher in the Future of Work?

Blit says this change will be brought on by the 2020 economic recession, which was declared in May by the C.D. Howe Institute. While experts expect the recession to be short-lived, it will have a major impact on the workforce.

We talked to Blit about the rate of automation coming out of 2020, how many jobs may be affected by the increase, what the government should do to mitigate the impact – and why all of this is good for Canada in the long run.

Manufacturing AUTOMATION: Using data from the Labour Force Survey, you’ve determined there’s a link between recessions and a subsequent increase in automation due to the loss of routine jobs, like office support or parts assembly.

What got you thinking about this connection during the coronavirus crisis?

Joel Blit: When COVID hit, I think we all tried to figure out how we can contribute to the effort to overcome this crisis. So I started reading on the topic and saw that a lot of productivity increases seem to happen during recessions.

It seems that right now, there are two things happening. One is we’re going to be involved in a recession very much like all the previous ones. And in all the previous recessions since the beginning of the ICT (information and communications technology) revolution – so the last three – we’ve seen a lot of automation and reallocation happen towards the end.

Over the last 30 years, there has been a big decrease in the number of routine jobs per capita and a big increase in the number of non-routine jobs. But that drop in routine jobs hasn’t been a steady one. Fundamentally, they’re steady, and then a recession comes along and there’s a big drop, and those jobs never recover.

That’s consistent with the story that during recessions, firms are taking the opportunity to reorganize and re-examine their processes. Maybe they get rid of workers and then say if we automate, we don’t have to call them back.

The other thing that’s happening is what I call a reallocation. During recessions, there’s a lot of pressure on firms. The ones that are already more productive – which tend to be the ones that employ automation more strongly – tend to have lower costs, and so they survive.

The ones that are higher cost – that use more labour, less automation – very often will either see their revenues decrease or they’ll go bankrupt altogether. Then that market share is taken up by the bigger, more productive, more upgraded firms.

Whether it’s firms directly investing in automation, or whether it’s reallocation of market share resources from less automated firms, the outcome is the same.

MA: What about COVID-19 makes this different?

JB: Stacked on top of that story is there’s now a health incentive to automate as well. If you have a production line, and you have workers that are all next to each other, that’s a risk to the operation.

So you want to automate at least some of these processes, so that you get more individual-to-robot interactions. That not only protects workers, it also is going to mitigate some of the risky operations. Because you have these two stacked factors on top each other, I suspect that we’re going to see a massive amount of automation coming out of this recession.

The other point of my study was to try to get a sense of which industries [are] going to see the biggest automation impact. And it’s those industries that both have strong incentives to innovate because of health incentives, and that have lot of these routine jobs that could be automated. Manufacturing was one of those sectors that came up, and so was retail.

What all recessions have in common is that there’s less spending and firms have a harder time selling their products, so their revenues tend to decrease or stay flat. So they often need to cut costs or reinvent themselves. Here, it’s not that there was a decrease in consumer spending or a decrease in investment or anything of the sort. It was literally the government saying, “you need to shut down and you need to do it now.” In that sense it is different.

But the outcome is not going to be different in the sense that now that we have reopened, we’re not going to recover all of those routine jobs. We’ve already recovered a little more than half the jobs that we lost between March and April. But it’s the next 50 per cent that’s going to be very hard. The things in the world that had to come back have already come back. Now, I think we’re going to see that we are going to recover but it might take another year.

If a lot of people still don’t have jobs and the government pulls back supports, there’s going to be less consumer spending. If there’s still a lot of risk and uncertainty as to what demand is going to look like, or as to whether there’s going to be a second wave, there’s going to be less investment. Firms are not going to start investing a huge amount in increasing output.

MA: In your study, you mention that the Canada Emergency Wage Subsidy (CEWS) is stifling transformation in manufacturing and other sectors. Can you elaborate on why that is?

JB: The wage subsidy is really doing two things. The first thing is that it’s providing a subsidy to firms. I certainly do think that is warranted, because if there were no supports whatsoever, we would have good firms disappearing just because of this crisis that we’re in and we don’t want that. We want the good firms to survive.

But there is the risk that if you give too much support, the firms that are very unproductive are also going to survive. By supporting all firms, we’re going to get less of a shakeout in industry, which is part of what recessions are good for. We get to increase our productivity when the more productive firms expand and take up the market share that the less productive firms have left behind. By supporting all firms, we’re stopping this process of reallocation.

The second thing about the wage subsidy is that it’s un-leveling the playing field of the cost of different input factors for firms. Imagine you’re a firm and you’re [making a product]. You can use different types of inputs – labour, capital for technology, land, energy. And the firm gets to choose the optimal mix that they want to use of all these different factors of production.

Firms are often going to choose the mix of those things that is going to allow them to produce whatever amount at lowest cost. And what happens is that when you make the cost of one of these input factors less than the other artificially, then firms are going to find it more useful to use more of one and less of the rest.

Right now, these wage subsidies are making the cost of labour a quarter of the size. So, if at $20 per hour it makes sense to replace these workers with a robot, at $5 an hour it certainly does not. As a result, firms are now having the chance to continue using older technologies with fewer robots, fewer AI, fewer everything – and just use as many workers as possible because they’re not paying for it, the Canadian government is. Those two combined effects are really hurting the transformation that could be happening.

MA: In what ways could policy actually help to support manufacturing companies through this transition to more automation?

JB: First, there’s the question of how to support the companies themselves. Right now, we’re supporting them with a program that is pushing them towards hiring more people and not using technology.

The context of this, of course, is that over the last 20 years, Canada has been complaining that our productivity gap with the U.S. is increasing. With lower productivity, wages stagnate and eventually the standard of living of Canadians falls behind, which we’ve already seen. Now that we have this tremendous opportunity to actually increase productivity and make Canada richer, we’re actually putting in place policies that are slowing it down.

We could help firms in ways that are neutral – like with loans, because the firm can use that loan in any way they want. Or it could be a support that is actually positive towards technology. It could be something like any firm that invests in automation is going to get a partial subsidy or tax break from the government.

MA: What about the workers? How do we help them as these routine jobs disappear and they need to re-skill to become non-routine workers?

JB: AI and robotics [are] going to change the nature of work, and a lot of people are going to be hurt by this. It’s not just that people are going to lose their jobs, but that it’s going to generate inequality in the long run. The people who have skills that are complementary to the technologies are going to benefit tremendously – that’s already happened in the ICT revolution.

The first thing that we can do is to train these workers. For example, during the 2007-2008 Great Recession in the U.S., routine jobs were lost, but some of those were also upskilled. The nature of the job changed.

Instead of being someone working on the production line, now this person is overseeing the work of seven or eight or 10 robots. To do that, this person had to get extra training. Maybe they need a little bit of programming, maybe they need a little bit of operations management. A lot of these workers are going to become obsolete, so let’s help them retrain.

The other thing is to support them through the transition and perhaps even longer financially. Right now, we have the Canada Emergency Response Benefit (CERB) program. I think that’s a really good program; I think we should continue it. This is a historic opportunity to really rethink our social safety net and our welfare system.

Nobody has a certain place – there’s an increased solidarity and we all feel like we’re in this together. And there’s also an increased belief in government, and the role that they play in our economy and in making sure that all Canadians are doing well. Maybe the CERB, instead of getting rid of it, transitions to guaranteed basic income or universal basic income, to make sure that no one falls through the cracks.

MA: Your study estimates that 586,000 routine jobs may be lost during this recession even though new jobs will be created to replace them. What does that mean for workers now?

JB: It’s only in retrospect once new industries are created that it becomes richer and new jobs will be created. A [half] million jobs lost is going to hurt a lot of people. We can’t just ignore it.

For two reasons, number one because it takes time to create them. The jobs are lost and then the new ones might not be created for another five, 10, 15 years. A little-known fact is that at the beginning of the industrial revolution, it took 50 years until wages actually were permanently higher than they were relative to how they were the beginning of the industrial revolution.

The second thing is that even if jobs are created, there’s no guarantee that they’ll be as good as the current manufacturing jobs. In other words, these new technologies, when implemented, could have distributional consequences that could increase inequality. And I suspect that these will increase inequality. We have to put in place the right policies to make sure everyone benefits.

MA: And that inequality is because only certain groups have the skills, or have access to the education, required to do those jobs.

JB: Yes. We don’t know exactly what’s going to happen with AI and robotics, but the last revolution we’ve seen was the ICT revolution, and there’s a pretty big consensus as to what the distributional impacts were of that.

It displaced people in middle-income jobs – for example, bookkeepers – and other people that worked routine jobs. Computers completely replaced some workers, and those workers ended up [elsewhere] doing mostly service jobs that were not as highly paying. The people that were replaced, for the most part, made less money.

For people who were white-collar workers and who were already making more, the computer made them more productive. Generally speaking, rates go up. The ICT revolution has actually made some people have lower rates than they did before and the people whose skills were complementary to computers now make even more.

The same thing is going to happen with AI and robotics. It won’t be as clear cut that high-skilled workers are going to have complementary skills or that most skilled workers are going to be replaced. It’ll be a lot more nuanced, because we know that AI is going to replace a lot of work.

But the key question is, what kinds of skills are complimentary to what an AI algorithm does or what a robot does, and what types of skills are going to be replaced by them? That’s how you know who’s going to benefit.

MA: Cash flow is pretty low right now because of everything that’s happened with COVID-19. How does a company jump into this transformation in a timely manner given their current financial restraints?

JB: This is an excellent point of why the government needs to step in. Through the central bank, they are [already] pumping out so much money into the economy, so I suspect capital is going to be relatively accessible.

There could also be joint programs with the government’s help, like with loans that are specifically designed for technological upgrading. But a lot of firms are suffering. They don’t have the cash flow right at the time or maybe the cash flow to invest in the future.

Can we step in as a government, as a society, and make those investments for the future? It’s going to be hard, because right now they’re trying to protect jobs, and to invest in something that is actually going to get rid of jobs but make us better off in the long term is difficult.

With the smaller firms, it’s often a problem with knowledge. They’re often not aware how robots can be implemented, or how AI can be implemented. [We need to] partner these firms up with the knowledge that we’ve got at our universities to mobilize for the better of all of Canada.

At the University of Waterloo, for example, we have an AI Institute, and we have the RoboHub, both of which work closely with industry. When things get developed at the university, we want to make sure that firms are using that technology.

MA: What tends to hold Canadian companies back from innovation, in your view?

JB: We know that our firms tend not to do nearly as much R&D as firms in other countries. And they don’t tend to invest in IT.

We have a lot of small and medium-sized firms, more than the U.S. as a share of the economy. It seems that these firms might not have the knowledge, or the resources, or the talent to be able to innovate and automate and do things that’ll make more profit.

How can we help bridge that gap? Any of these programs that help Canadian firms, especially smaller ones, to partner with universities, or even work with each other in consortia, are fantastic ideas.

I’ll tell you one thing that isn’t holding companies back. We have cutting-edge technology being produced in Canada. If you look at the number of academic articles on AI or on other topics, it’s truly world class. It’s just that we are not taking that and turning it into opportunities.

Our firms should be encouraged to go global as much as possible. By being an international firm, you learn not just from the people that are located near you, but also from wherever you are overseas. If there’s a new development in semiconductors, or AI, in another country and you happen to have a presence there, you will learn about it much faster than if you’re a firm that’s only based in Canada. That’s actually some of the work that I did about five years ago – I showed empirically that is the case.

Encouraging more multinationals to come to Canada is a good idea because often they bring good ideas with them. The flip side of that argument is that when they do come to Canada, they compete with Canadian firms for Canadian talent.

It’s not 100 per cent clear cut, but certainly my research shows that when multinationals come to Canada or when they go to any country, they take with them knowledge and know-how. They can create a partnership with a university in Canada. It’s a two-way street, they both learn from each other.

A condensed version of this article originally appeared in the September 2020 issue of Manufacturing AUTOMATION. Read the digital edition here.