The simple version

The Conference Board publishes a Consumer Confidence Index. The University of Michigan publishes an Index of Consumer Sentiment. Both are monthly, both are widely quoted, and they come from organizations with no connection to each other.

Their most recent published readings are 90.8 and 51.0. A reader could be forgiven for concluding that one group of households is far gloomier than the other. That is not what those numbers say.

An index number is not a quantity of anything. It is a comparison to a base period the publisher chose, and the two publishers did not choose the same one.

The numbers

• The Conference Board Consumer Confidence Index most recently read 90.8 for July 2026, down 1.4 points from an upwardly revised 92.2 in June. The Conference Board prints the base alongside it as 1985 equals 100 (The Conference Board)
• Within that index, the Present Situation component read 114.9 and the Expectations component read 74.7 in July, both on the same scale as the headline (The Conference Board)
• The University of Michigan Index of Consumer Sentiment most recently read 51.0 for August 2026, a preliminary figure, against 55.2 in July and 58.2 in August 2025 (University of Michigan)
• Michigan publishes those readings with percent changes attached, down 7.6% from the prior month and down 12.4% from a year earlier, which is the form the comparison is meant to take (University of Michigan)
• Michigan's published results tables do not carry a base period next to the numbers, so a reader working from the release alone has the level without the reference point it is measured against (University of Michigan)
• The Conference Board publishes its index at 10 a.m. Eastern on the last Tuesday of every month, which is August 25 this month (The Conference Board)
• Michigan's next release is Friday, August 28, at 10 a.m. Eastern, for final August data. This article makes no claim about what either release will show (University of Michigan)
• Neither organization publishes the other's index, so each is the primary source for its own numbers and its own method and for nothing else (definition)

What a base period does

An index starts by picking a moment and calling it 100. Every later reading says how the measured thing compares to that moment. The choice of moment is arbitrary in the sense that it carries no meaning of its own, and it is decisive in the sense that it sets every number that follows.

Move the base period and every value in the series moves with it, without a single survey answer changing. A reading of 90 against one base and a reading of 50 against another can describe households in identical moods. The numbers are answers to different questions.

This is why the Conference Board prints 1985 equals 100 directly beside its headline figure. That notation is not decoration. It is the instruction for how to read the number, and it is the piece that gets dropped first when a figure travels into a headline.

Michigan's results tables take a different approach and lead with percent changes from the prior month and the prior year. That format sidesteps the problem, because a percent change against the same series carries its own reference point.

What each number can be compared to

The rule that falls out of this is narrow and firm. Each index is comparable to its own history and to nothing else.

The Conference Board reading of 90.8 in July is meaningful set against 92.2 in June, because both sit on the same scale and the same method produced them. Michigan's 51.0 in August is meaningful against its own 55.2 in July for the same reason.

What cannot be done is the subtraction between them. There is no defensible arithmetic that turns 90.8 and 51.0 into a statement about which set of households feels worse, and any story performing that comparison has produced a number with no referent.

Direction is the one thing that does travel between them. If both indexes move the same way in a month, that agreement is informative, and if they move opposite ways, that disagreement is informative. The levels stay in their own lanes.

The Real Cost lens on reading an index at all

The habit generalizes well past these two surveys, because most published economic numbers are indexes rather than counts.

• When a figure is an index, find the base period before doing anything else with it, because without one the level is unreadable
• Compare an index only to earlier values of the same index, produced the same way, and treat any cross-index subtraction as a mistake
• Percent change is the safer unit when comparing across series, because it carries its own reference point
• The Consumer Price Index works this way too, and so does almost every measure with the word index in its name

None of that is a criticism of either organization. Both publish their numbers, their components, and their methods, and the Conference Board prints its base period in the same sentence as its headline.

The information needed to read these figures correctly is public. It is the retelling that loses it.

What this means

Two confidence readings land this week, three days apart. The useful question for each is not what the level is but which direction it moved against its own prior reading, and whether the two moved together.

The transferable habit is to treat every index level as incomplete until you know its base period. A number without its reference point is not a small problem with the number. It is the absence of the thing that made it a measurement.

What this is NOT

This is not a prediction of either survey's reading, of consumer spending, or of the economy, and no forecast figure appears here. This is not a claim that American households are confident or worried, which is a judgment this article does not make, and the readings cited are reported as published values rather than as evidence for any conclusion about households. This is not a claim that either index is more accurate, more reliable, or more useful than the other; they are produced by separate organizations with separate methods, and each publishes its own documentation. This is not a description of anything either organization has said about the other. Michigan's August figure is preliminary and subject to revision at its final release. National survey averages describe no individual household. This is not advice about spending, saving, or any financial decision, and it is not advice about any security or fund. This is not investment or financial advice of any kind.

Sources

• The Conference Board, Consumer Confidence Survey: https://www.conference-board.org/topics/consumer-confidence/
• University of Michigan, Surveys of Consumers: https://data.sca.isr.umich.edu/
• University of Michigan, Surveys of Consumers overview: https://www.sca.isr.umich.edu/
• U.S. Bureau of Labor Statistics, Consumer Price Index, for an index that publishes its base period: https://www.bls.gov/cpi/