Perky, productive robots, or nothing more than a few new smartphone apps? Cascading innovation, or just a few tweaks? Economists and technologists are debating what the future holds.
Pessimists like Robert Gordon of Northwestern University see decades of slow growth ahead, with little scope for big leaps forward. The optimists, among them Erik Brynjolfsson and Andrew McAfee of the Massachusetts Institute of Technology, expect new technological glories. Both sides are more wrong than right.
Everyone is wrong when the wrangling is numerical. Arguments based on GDP and productivity growth are too circular to resolve anything. A main cause of any slowdown in reported productivity numbers is a judgment that innovations are becoming less valuable. So a reported slowdown cannot logically be used to support the argument that technology is advancing more sluggishly.
The problem is that productivity measures depend upon what economists call hedonic adjustments. Consider a new model car that costs 2 percent more than the vehicle it replaces in a price index. The price change is clear, but the two cars aren’t identical. The new model will, at a minimum, have fancier electronics.
The statisticians have to put a value on the differences. They could decide the additional technology is just a gimmick with no effect on the real value of the car. In that case, car price inflation is 2 percent. If the same number of cars is sold this year as last, then car-related real GDP growth will be zero. And if the same number of hours went into building each car, then labour productivity would be unchanged.