A month or so ago I wrote about that Mismeasurement may be at the heart of the productivity paradox. The paradox is that despite AI and robots replacing jobs productivity doesn’t seem to be rising and my post drew on data showing that microprocessor prices may have been incorrectly calculated in productivity indices to argue that mismeasurement of productivity may be at the heart of the paradox.
My colleague Matt Bradley subsequently passed me a paper titled Moore’s Law at 50 by Bret Swanson which has some additional data points which give credence to this argument:
- In The Mumbo-Jumbo of ‘Middle-Class Economics’ Alan Reynolds argues that whilst real incomes have stagnated since 1968 real consumption has tripled. The major differences between the consumption and income measures come from realised capital gains, dramatic increases in public and private benefits (e.g. health insurance and food stamps), the reduction in middle class taxes or the impact of retirement savings. There is a related point that income is often calculated on a household basis but because average households now have fewer people in them a decline in household income doesn’t imply a decline in per capital income.
- We get a similar picture if we use the personal consumption expenditures (PCE) deflator rather than the consumer price index (CPI) deflator to adjust for inflation. Using CPI median incomes rose just 4.2% from 1967 to 2014, whilst using PCE they rose 33.0%. (Neither of these indices takes account of the impact of benefits, taxes and savings mentioned in the previous paragraph.
- New inventions have brought significant benefit to society but are impossible to capture in comparative income measures – e.g. life saving drugs, new surgeries or sites like Wikipedia.
The evidence is mounting…