Automation and artificial intelligence are transforming the global economy in ways that central banks are still struggling to understand, Bank of Canada Governor Stephen Poloz said Thursday. That rapid change can be difficult to capture in official measures of the economy, he noted, and ends up sending conflicting signals that complicate monetary policy decisions.
“So far the statisticians are scrambling in many directions to try and capture all this,” Mr. Poloz said during a talk at the Federal Reserve in San Francisco, the heart of Silicon Valley.
A new Industrial Revolution is under way, driven by the shift to a digital economy fuelled by big data, artificial intelligence and machine learning, Mr. Poloz said. Like past technological transformations, the transition to a digital economy is likely to be bumpy, with winners and losers.
It has also been complicated by political tensions such as the ongoing U.S.-China trade tensions. “For instance, the Trump trade war is reducing investment virtually everywhere, and will permanently reduce the level and growth rate of global economic potential if it is sustained,” he said.
Mr. Poloz argued the shift is likely to prove to be a benefit to global economic growth in the long run. But it creates more uncertainty for the world’s central bankers, largely because current estimates of GDP growth struggle to accurately capture the signals coming from brand-new technology-driven industries. That raises the risks that official statistics underestimate just how much the economic output is expanding.
Estimates in Canada are that the amount of economic activity generated by data-driven services not captured in official measures of the economy amounts to as much as two percentage points of GDP, he said.
For instance, Statistics Canada doesn’t capture Canadian purchases on Amazon as part of retail sales because the U.S.-based e-commerce has no bricks and mortar stores in Canada. Estimates of business investment in particular are “very poorly measured,” Mr. Poloz said.
And official Statistics Canada measures of the economy are unlikely to be revised for years to come. “They’re the best statistical agency in the world, but it doesn’t mean they can get all this right, right away,” he said.
Mr. Poloz compared the dilemma central bankers’ face today from automation and machine learning to the dot-com bubble of the late 1990s and early 2000s.
The U.S. Federal Reserve has faced criticism for holding interest rates low for too long after the dot-com bubble burst, even as the U.S. economy was growing. That easy-money policy helped encourage speculation in the U.S. housing market, leading to the 2008 financial crisis.
The Fed at the time faced mixed signals – sluggish inflation and strong economic growth – though it took years for economists to fully appreciate what was happening.
Central banks are facing similar challenges today, he said, where the positive economic effects of automation may not be easy to detect for many years. "There is a good possibility that we will experience something very similar over the next decade, as the fourth Industrial Revolution unfolds.”
Mr. Poloz pointed to the trend of firms shifting many business functions into cloud computing services and hiring large numbers of technology workers to develop software and other services for clients. Both practices are recorded as expenses even when they should more accurately be considered business investment.
The Bank of Canada is alone among major central banks in holding its key overnight rate steady for the past year amid a global economic slowdown fuelled by concerns over the ongoing U.S.-China trade war. The Fed has cut interest rates three times this year. Canada decided against a proactive rate cut last month to guard against slumping global growth, leaving Canada with the highest official interest rates among advanced economies.
While Mr. Poloz didn’t offer any hints at the future path of Canada’s monetary policy in his speech, he argued that the current situation calls for a similar policy to the one that former Fed chair Alan Greenspan followed in the years before and after the dot-com bubble.
Central bank policy should allow the economy to continue to grow to make it easier on workers who are displaced by new technology – such as truck drivers, financial advisers, and agricultural workers – he said. But at the same time, central banks must closely watch the way that monetary policy encourages speculative financial bubbles.
“Such a framework is especially useful when financial vulnerabilities are already high, which is clearly the case for Canada’s household sector,” Mr. Poloz wrote, adding that allowing such risks to build “may prove to be an unacceptable trade-off.”
He also warned that the effects of the dramatic changes to the economy from automation could last years, or perhaps be felt more gradually. “I don’t know and that’s that the point,” he said. “So the point is to build your policy around that uncertainty.”
While the positive effects from new technology will eventually spread to all parts of the economy, in the short-term those changes are likely to displace large numbers of workers, some of whom may never find a new job. That’s a problem not only for central bankers, but for politicians, he said. “It does mean that we have to be not just having a good monetary policy, but making sure we have the right kind of safety net-type policies that give people time to adjust and adapt to this.”
Mr. Poloz was speaking in the U.S., where President Donald Trump has lashed out at Fed chair Jerome Powell for refusing to slash interest rates to zero.
While he didn’t directly address Mr. Trump’s actions, in a paper published online by the Bank of Canada shortly before his speech, Mr. Poloz acknowledged that it may be difficult to convince politicians and the broader public that the economy may be growing in ways that even economists don’t fully understand. “The situation is bound to attract considerable debate, during which the credibility of central banks will be tested,” he wrote.