August 21, 2015
In a blogpost Paul Krugman picked up on a discussion by Rex Nutting of the Carter presidency. Nutting points to many of the positive accomplishments of the Carter years, including the fact that, by many measures, the economy actually performed quite well.
Krugman picks up on this theme and uses a chart of median family income to show that the typical family was actually better off in 1981 when Carter left the White House than they had been in 1977 when he took office. Krugman argues that the problem for Carter’s re-election prospects was that income was declining in the last years of his presidency, which is what people had in their minds when they went to vote.
While this point is undoubtedly accurate, there is another complication when we try to get a sense of people’s perceptions when they went to vote in November of 1980. The measure of inflation that is used to derive real median income in Krugman’s chart is the CPI-U-RS. This applies the methodology that we use today to construct what the CPI should have been in prior years. This gives a very different and much lower measure of inflation than the CPI that the Bureau of Labor Statistics was using at the time.
Here is how the two compare for 1978-1981.
CPI CPI-U-RS
1978 9.0% 7.8%
1979 13.3% 10.7%
1980 12.5% 10.7%
The cumulative difference for these three years is 5.6 percentage points. (Yes, this is just adding and I should be compounding, but let’s keep this simple.) This means that folks going to vote in 1980 would have been seeing in the data a 5.6 percent greater drop in real income by 1980 than what Krugman has in his chart. The question is whether this error in the data would have affected people’s perceptions of their well-being or whether we should only care about what we might think of now as the “true” rate of inflation.
I would argue for the importance of the errors in the data. First, none of us really have a clear idea of the true rate of inflation. It’s based on a basket of goods and services that none of us literally buy. There is a big weight for large purchases, like cars, that we may buy at five year intervals, or even longer. Also, the prices are quality adjusted. Is the typical person’s assessment of the rate of the quality improvement in a cell phone or computer the same as the BLS’s assessment? It’s very likely that if she pays more for a car or computer than for her last purchase, she sees that as a price increase, even if BLS has determined that the quality adjusted price has fallen.
On the other side, back in the late 1970s many contracts were legally tied to the CPI. This meant that workers had reason to know the inflation rate shown by the CPI since it would determine their pay increase that year. This was often true of rents as well. As a result, if the BLS said the rate of inflation was 13.3 percent in 1979, it is likely that many people thought the inflation rate was 13.3 percent, even though our methodology now tells us that the rate of inflation was actually just 10.7 percent.
There is more to this story of mis-perceived inflation—could mis-measured inflation lead to actual inflation? I’d argue yes, but we’ll leave that one for another day. For today, I’ll just say that it was not only Paul Volcker’s Fed that doomed Jimmy Carter’s re-election prospects, but also the mistakes made by the folks at BLS.
Comments