(Bloomberg Opinion) -- As the decade winds down and we assess how the economy did, two things stand out: productivity growth has been disappointing since the Great Recession and yet economic growth hasn't been all that bad, compounding at a 2.3% annual rate. But we've only been able to manage that level of growth because we had so many unemployed people available to go back to work. With labor slack continuing to diminish, maintaining this pace of economic growth in the 2020s will require a pick-up in productivity. The question is how to do it?

Economic growth is a function of two factors: more hours worked and how much more efficient, or productive, an economy is. Although the U.S. now has an aging population, a falling birthrate and less immigration than in the past, the level of job growth and hours worked during the past decade has been robust because of just how high unemployment was. In this environment, the fear that robots would displace workers proved unfounded.

It's now been 10 years since unemployment peaked at 10% in October 2009. Since that time, total employment in the U.S. has increased by 20 million. During that same period, the number of unemployed workers has fallen by 9.5 million, from 15.3 million to 5.8 million. In other words, about half of all job growth in the past decade has come from getting unemployed people back to work.

So future economic growth is probably going to have to come from somewhere other than the unemployed. Maybe a tight labor market can draw more people into the workforce -- we've seen some of that during the past few years -- but at some point that hits a limit, too. Ultimately, what we need is faster productivity growth. Fortunately, there's hope for that on a few different fronts.

The first is that as companies run out of cheap workers in industries such as retail, they finally have the incentive to invest in labor-saving robots and self-checkout machines. Maybe these technologies were advanced enough years ago to be put into production, but when there were plenty of unemployed workers available to hire, it was easier just to increase hiring. That's no longer the case as wages have risen both because of market pressures and state- and city-mandated higher minimum wages. During the past two years we've seen robust wage growth for retail workers, and self-checkout machines are popping up in more and more stores. Maybe this scenario unfolds in more industries, and in the 2020s we get a harmonious environment with tight labor markets, more automation, faster productivity growth and rising wages as workers shift into higher-paying and more-demanding jobs.

The second is the potential for large-scale government investment. Politics is cyclical, and the most recent election results suggest that after Donald Trump leaves office, whether that's in 2021 or 2025, Democrats will get a chance to govern. And whether it's a more centrist administration focused on infrastructure, or a leftist administration advocating for a Green New Deal, the Democratic Party sees using fiscal policy to invest as a key focus for its economic agenda rather than just tax cuts and deregulation. Tax cuts and deregulation did little to lift productivity growth, so perhaps a different course of action would lead to a different outcome.

Finally, there's the theory that workers in their 40s are the key to faster productivity growth. Workers in their 20s might be eager and inexperienced, while older workers are experienced but can have a harder time acquiring new skills. With the oldest members of the millennial generation turning 40 during the next couple of years, we're entering a period where this cohort will grow for more than a decade, perhaps providing a demographic tailwind to productivity growth.

Slow productivity growth is bad, and we're not fully sure how to increase it; if we did, policy makers would surely do it. But the good news is that there are multiple reasons productivity growth might accelerate in the coming years. If it does pick up, it will drive the faster economic growth we want. And in the process, we might just get some answers to the questions about why productivity growth has been so disappointing in this expansion.

To contact the author of this story: Conor Sen at csen9@bloomberg.net

To contact the editor responsible for this story: James Greiff at jgreiff@bloomberg.net

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.

Conor Sen is a Bloomberg Opinion columnist. He is a portfolio manager for New River Investments in Atlanta and has been a contributor to the Atlantic and Business Insider.

©2019 Bloomberg L.P.