American productivity is finally on the rise

After a long decline through the 2000s and then several years of near-zero and even negative growth, the amount of stuff that Americans produce in the average work hour has been inching back upward since late 2016.

Productivity experienced its best growth in years during the fourth quarter of 2018, the Bureau of Labor Statistics reported last week, rising 1.8% compared with the same period last year.

What's driving it: Workers have been increasingly expensive and hard to find, forcing businesses to invest in equipment that makes their labor go further.

"Previous pickups have turned into a disappointment, so I definitely cannot say it's a true recovery," said Gad Levanon, chief economist for North America at the Conference Board. "But we are in a very tight labor market, so the pressure on companies is stronger than it was a few years ago. That's the only reason I think this time it's likely to be sustainable."

Why productivity matters

Productivity is important because it's the only way for the economy to grow without adding more people to the workforce. Since the US population is aging, making sure that younger people are able to produce more — without working around the clock — is essential to keep the wheels turning.

That's why economists were so concerned about the productivity slowdown in the 2010s, which affected many other developed countries, not just the United States. Although productivity briefly looked very strong immediately following the Great Recession, that was mostly a consequence of employers shedding everyone but their highest performers, and making them work harder in order to fill in for their laid-off colleagues.

After that, jobs returned quickly — but because wages were so low, employers had no incentive to invest in technology or better business processes in order to squeeze value from each working hour. Also, computers had already made supply chains a lot more efficient in the 2000s, so by the time employers started hiring again after 2011, most employees were already contributing as much as they could with the tools available to them.

"This recovery has been so rare in that it's been a job-rich and productivity-weak recovery," Jaana Remes, an economist with the think tank arm of the McKinsey & Company consulting group, said about the period following the Great Recession.

Weak productivity has been particularly visible in manufacturing: As the United States emerged from the recession, factories were producing far less than they had before the downturn, so there was plenty of room to make more stuff with existing equipment.

"You have a plant and the folks in there, and if you're not running it at 85% capacity, that immediately translates into lower output per worker," Remes said.

Technology as a factor

That started to change around 2016, however, as the unemployment rate continued to sink. Wages began rising, particularly in traditionally low-paid industries like retail and food service, and that changed the calculus for employers. It became harder to avoid investing in technology like self-checkout machines and online orders, which are expensive and may take time to generate returns.

Meanwhile, a new generation of machine learning and automation is just making its way into commercial use on a broad scale.

Take e-commerce, which improves the productivity of retail by delivering packages without the need for a customer-facing workforce. The warehousing and delivery company XPO Logistics just rolled out a new "labor productivity system" that allows managers to monitor and respond to workforce demands in real time, as well as organize warehouses in a way that allows them to fulfill orders more quickly.

The XPO system's algorithms can recognize that certain products are typically purchased together and slot those items next to each other, so that a floor worker — who already has a "co-bot" machine to help pick up and package orders — can spend less time on each one, says XPO Logistics Chief Information Officer Mario Harik.

"Technology's goal is to make our people more efficient in how they operate every day," said Harik, noting that the company has a technology budget of over $500 million a year, and that their automation and robotics make workers four times more productive. "This is what we find super exciting, and it's the reason we spend so much money on it."

The intangibles

It's also possible that technology has been improving productivity in ways that don't show up in official measurements, says Gregory Daco, chief US economist at Oxford Economics. That could mean that the slowdown wasn't as bad as it looked, and that the recovery is even better than it appears.

Intangible assets like a company's brand image, which legions of public relations professionals work to burnish, could be counted as productivity. Or if a company sells a smartphone that has unlimited data, maybe that should be counted as higher-value than its price would reflect.

"The digital economy's benefits to the overall economy might yet not be captured, either because of slow adoption or because our methodology isn't good enough to capture them," Daco said.

If workers are providing more value to their employers, of course, they should get paid more. There might be fewer short-order cooks churning out hamburgers, but the person who runs the burger-flipping machines should in theory be better trained and make more money.

That's why economists have long believed that productivity is key to wage growth. But the relationship has broken down over the past few decades, in part because the decline of unions and the rise of globalization has decreased the amount of bargaining power that workers have to demand higher wages.

Josh Bivens, director of research at the left-leaning Economic Policy Institute, argues that higher minimum wages, a stronger labor movement, and continued low interest rates would help productivity translate into better living standards.

"We need to make sure the unemployment rate stays really low for a really long time to give workers some leverage," Bivens says.