In an NYT Upshot column, Neil Irwin correctly points out that the Trump administration is very optimistic about GDP growth, it then adds that it shouldn't be. Irwin is certainly right about the optimistic part but should be more cautious in declaring the optimism wrong.

Irwin breaks down the factors that contribute to growth, labor, capital, and multifactor productivity growth. Labor growth is likely to be slow for the simple reason that the baby boom generation is moving into its sixties and seventies. They are now retiring in large numbers. This will seriously dampen the pace of labor force growth. While there is some room for more people to come into the labor market, even getting back to the peak prime-age (ages 25 to 54) employment to population ratios of 2000 would only around 0.3 percentage points to rate of labor force growth. We could have more immigrant workers, but that doesn't seem likely in the current political environment.

There could be an uptick in investment, which is the ostensible rationale for the big corporate tax cut. Irwin rightly is skeptical on this prospect. Historically, investment has not been very responsive to tax rates or the after-tax rate of profit, which in theory should be the key factor. The latter was already at post-World War II highs even before the tax cut, while the investment share of GDP has been mediocre.

The key issue is multifactor productivity. This is the increase in productivity that is the result of better organizing workplaces and introducing new technologies. Multifactor productivity growth has been very weak in recent years. It has averaged just over 0.4 percent a year since 2005. That compares to 1.6 percent annually from 1995 to 2005.

The Congressional Budget Office (CBO) in its projections essentially assumes that this slow pace of productivity growth continues. While that may prove correct, it is important to note that CBO both missed the slowdown in 2006 or the pick-up in 1995. (Multifactor growth had averaged less than 0.6 percent annually from 1987 to 1995 [the current series begins in 1987.]) In other words, CBO is not good at picking up turning points.

There are two reasons to believe that we could see an upturn. The first is the tightening of the labor market. This is putting upward pressure on wages, especially for those at the bottom end of the labor market. Usual weekly earnings for workers at the 10th percentile of the wage distribution (a worker who earns more than 10 percent of workers and less than 90 percent) have risen an average of 3.3 percent annually over the last three years. For workers at the 25th percentile (a worker who earns more than 25 percent of workers and less than 75 percent), they have risen an average of 3.6 percent.

Higher pay will have the effect of eliminating many of the least productive jobs, thereby raising average productivity. Walmart is likely to have many fewer greeters if they have to pay them $15 an hour than if they pay them $7.25 an hour. Eliminating the least productive jobs raises average productivity.

The other reason to expect an upturn in productivity growth is the robots-taking-all-the-jobs story. While this concern is hugely exaggerated, as I often point out, there are major new innovations in robotics and artificial intelligence. It is hard to believe that these will not have any major impact on the labor market and the economy. This is in effect what the growth pessimists are saying.

It shouldn't take a huge leap to think that job-killing robots can bring us back to something like the multifactor productivity growth rates we saw between 1995 and 2005 or in the long Golden Age from 1947 to 1973. If they do, then the Trump administration's 3.0 percent growth figure is very plausible.

I'm not predicting that multifactor productivity growth will necessarily jump back to the 1995 to 2005 pace. I just don't see any basis for dismissing this possibility. If I had to bet higher or lower on the pace over the next decade compared to the post-2005 pace, I would definitely say higher. We'll see whether that proves right.