from the San Francisco Fed

— this post authored by Pete Klenow and Huiyu Li

When products disappear from the market with no substitutes from the same manufacturer, they may have been replaced by cheaper or better products from a different manufacturer. Official measurements typically approximate price changes from such creative destruction using price changes for products that were not replaced. This can lead to overstating inflation and, in turn, understating economic growth. A recent estimate suggests that around 0.6 percentage point of growth is missed per year. The bias has not increased over time, however, so it does not explain the slowdown in productivity growth.

A gloomy view of the economy suggests that the measured slowdown in U.S. productivity growth since 2004 reflects a true slowdown in innovation and productivity gains. From 1995 to 2004 businesses boosted the average growth rate to nearly 3.75 percentage points annually by adopting information technology (IT) to make their operations more efficient (Byrne, Fernald, and Reinsdorf 2017). Since then, measured productivity growth has slowed to around 1.75 percentage points per year. Perhaps it has become more difficult to achieve the same rapid rate of improvement as in the past. But has productivity growth truly slowed down?

To answer this question, one needs to keep in mind that measured productivity growth is designed to capture growth in market activities. Thus, it may not fully capture the growth in people’s economic welfare because it misses out on important dimensions such as increasing lifespans and rising home production. So, even if the measurement is correct, a slowdown in measured productivity growth does not necessarily reflect a slowdown in welfare growth. For example, many recent IT innovations involve nonmarket activities such as time spent on social media and time saved from shopping online. Although these areas may improve welfare, they have not historically been covered by productivity measurements, so ignoring them cannot directly contribute to any growing understatement of market-sector growth.

This Economic Letter focuses on measuring growth from innovation in parts of the economy that have traditionally been within the scope of productivity measurement. Past research has found that measurement problems in the IT sector associated with market production and offshoring activities cannot explain much of the growth slowdown (see Aghion et al. 2017 for references). In this Letter, we consider whether errors in measuring innovation outside the IT sector can explain the substantial slowdown.

Quality adjustment in inflation measurement

As a first step toward measuring growth, the Bureau of Economic Analysis adds up the dollar value of goods and services sold, which equals “nominal output.” However, due to inflation – a general increase in overall price levels – one dollar today does not have the same purchasing power as one dollar three decades ago. This means the agencies measuring these data need to subtract the inflation rate from growth in nominal output to arrive at growth in real output. Hence, a 1% overstatement of inflation translates into a 1% understatement of measured real output growth. This translates into a 1% understatement of productivity growth as well, because productivity growth is the difference between real output growth and the growth in capital and labor inputs used in production.