Unmeasured Gains and Losses from Technology

Paul Krugman had an interesting exchange in a recent Substack with Erik Brynjolfsson, one of the world’s leading experts on AI. At one point they turned to benefits that are not counted in GDP as unmeasured sources of productivity gains.

Specifically, they talked about all the material that we can now get for free off the web. There is a huge amount of music, video material, video games, reference material (e.g. Wikipedia) and other items we can now get free over the web. These things used to be costly for us to purchase, but the gains from having free access are mostly not picked up in our measures of GDP and productivity.

I will add to their list that the ability to work from home and telemedicine are also great sources of gain and not picked up in GDP or productivity. Working from home allows people to save thousands of dollars a year in commuting costs and hundreds of hours of time.

Telemedicine also saves people from having to make trips to see healthcare professionals. This is an especially big deal, since the people making the appointments are often in bad health, which means the trip for an in-person visit can be difficult to navigate.

While all these are clearly positives from new technology, there are some aspects that go the other way, which are not picked up in our price indices. While we don’t count the gains from all the free stuff we can get off the web, the cost of Internet service is also not counted in the Consumer Price Index (CPI).

Just to be clear, the CPI does measure the increase or decrease in costs of Internet service, but it does not pick up the fact that most of us now find it pretty much a necessity and often have to pay $80 or more a month for this service.

It’s the same story for smartphones. They are a great convenience, but the cost of buying one and paying the monthly service fee does not enter as an increase in the cost of living measured by the CPI.

This actually is a much broader problem. If we need a car because the public transit system serves us poorly, that also is an unmeasured increase in the cost of living. Similarly, if fear of crime causes us to install a security system, or move to a new neighborhood, that also is an unmeasured increase in the cost of living.

Also, health issues can lead to unmeasured increases or decreases in the cost of living. The development of effect effective vaccines against polio, measles, and other diseases was a huge unmeasured reduction in the cost of living, whereas deteriorating health status due to pollutants, poor nutrition or addictive drugs, or a pandemic would translate into a higher cost of living.

Krugman and Brynjolfsson are undoubtedly aware of these broader issues, however that may not have been apparent in their brief exchange in Paul’s Substack. As a veteran of the debates over the CPI in 1990s, when many politicians wanted to claim a major overstatement of inflation as a rationale for pruning the annual Social Security cost-of-living adjustment, I remember many economists just wanting to look at the unmeasured gains from technology.

The reality is that the full picture is complicated, and we can’t just look at the positives if we move away from the CPI as a price index and say instead that we want to fully measure the changes in the cost of living. We have to look at negative factors as well, and in some cases they may outweigh the unmeasured gains from technology.