Timothy Taylor reports on a symposium on productivity trends. He quotes Robert Gordon,
I have often posed the following set of choices. Option A is to keep everything invented up until ten years ago, including laptops, Google, Amazon, and Wikipedia, while also keeping running water and indoor toilets. Option B is to keep everything invented up until yesterday, including Facebook, iphones, and ipads, but give up running water and indoor toilets; one must go outside to take care of one’s needs; one must carry all the water for cooking, cleaning, and bathing in buckets and pails. Often audiences laugh when confronted with the choice between A and B, because the answer seems so obvious.
I think that what this anecdote indicates is that measured productivity is bunk. Gordon’s anecdote suggests that people derive a lot of consumers’ surplus from modern water systems. But this consumers’ surplus does not show up in measures of productivity, either for one hundred years ago or for today.
I am becoming a productivity measurement pessimist. That is, I am becoming pessimistic that what we call “productivity” is anything more than a crude indicator of trends in living standards.
I can imagine coming up with an accurate measure of productivity in soybean output. However, it is difficult to imagine coming up with anything accurate for health care, where we have little idea about what generates value at the margin, or for education, we where have almost no idea at all.
Moreover, the value of many goods and services, including the Internet and modern water systems, is under-estimated because we do not measure consumers’ surplus. Going forward, suppose that researchers come up with a way to prevent or cure Alzheimer’s. The effect on consumers’ surplus would be quite large. The effect on measured productivity? To a first approximation, nil.
In assessing economic progress, productivity may be the best indicator we have. However, we take small differences in measured productivity growth rates way too seriously. On the one hand, it is correct to say that if you extrapolate a difference in productivity growth of 1 or 2 percentage points over thirty years, it accumulates to a big number. But I fear that it is quite possible that the error in measuring productivity growth can exceed 1 or 2 percentage points for thirty years or more. That is, I think it is quite possible to take two thirty-year periods and arrive at a very large estimate of the difference in the rate of growth of living standards that is entirely due to mis-measurement.