Consumer confidence has been sky-high for some time now, with the major indices tracking various definitions of it at or just near highs not seen since the dot-com era. Economists place a lot of emphasis on confidence in all its forms, including that of consumers, and there is good reason for them to do so; or there was in the past.

spending and consumer sentiment used to track each other very closely and in terms of more general conditions there was a clear and obvious relationship; when they told whichever survey they were happy they also in the aggregate tended to act with fewer inhibitions about their purchasing behavior. The correlation was intuitive as well as mathematical.

Looking more closely at these sentiment indications, however, you get the sense that they like inflation expectations (market as well as survey-based) have become increasingly unanchored. There is a palpable dissonance more and more embedded in the numbers that hasn't been factored out of them by serious analysis (instead each index purveyor continues to assume by their regressions that consumers today are exactly like consumers ten, twenty, or thirty years ago).

Take, for instance, the University of Michigan's Surveys of Consumers. Its measure of confidence registered a sublime 100.7 in October. It was the highest since a brief spike in early 2004, closing in on levels that in the nineties used to be the norm.

Though the index is lower in both November and December, some of the reasons for that slight downtrend should give you pause. From the December estimate:

Tax reform was spontaneously mentioned by 29% of all respondents, with nearly an equal split between positive and negative impacts on economic prospects. Party affiliation was the dominant correlate of people's assessments of the tax legislation, with the long term economic outlook the most negatively affected.

If you are a Democrat, you are unhappy about the tax reform law; a Republican, quite pleased. It doesn't tell us what the tax cuts might actually do, instead the survey results reveal biases (that are always present) based on what people are told they will do.