At ten o'clock Tuesday morning, the Conference Board publishes its August read on consumer confidence. Buried inside that release is a line that has been flashing a recession warning for eighteen consecutive months without a recession bothering to show up.
The line is the Expectations Index. It captures how Americans feel about their income, business conditions, and the job market six months out. The Conference Board has said for years that a reading below 80 has historically preceded a recession within the following twelve months. July's reading came in at 74.7, unchanged from June. It hasn't printed above 80 since February 2025.
That's a year and a half of a well-documented leading indicator telling you the economy is about to break, while the economy kept expanding, corporate profits kept climbing, and the S&P 500 kept doing what it does. If you had repositioned your portfolio around that signal in February 2025, you'd have spent eighteen months paying for insurance against a fire that never started.
I'm not writing this to tell you the indicator is worthless. I'm writing it because the way most people consume this report is backwards, and the fix is genuinely simple once you understand what the survey is built out of.
Start with the architecture. The headline Consumer Confidence Index isn't one measurement. It's a weighted blend of two separate sub-indexes that measure completely different things, and blending them destroys most of the information. July's headline was 90.8, down 1.4 points from an upwardly revised 92.2 in June, and below the roughly 92.3 that forecasters expected. That number tells you almost nothing on its own.
Underneath it, the Present Situation Index came in at 114.9, down 3.6 points, its third straight monthly decline and its weakest reading since February 2021. The Expectations Index sat at 74.7. The distance between those two figures is roughly forty points, and that gap is the actual story.
Here's why the two halves behave so differently. The Present Situation Index asks people to describe conditions they're currently experiencing. Is business good or bad right now? Are jobs plentiful or hard to get right now? People are reasonably accurate reporters of their own circumstances. They know whether their hours got cut. They know whether the last three job applications went unanswered. That makes the Present Situation Index a coincident indicator. It confirms what's happening, and it confirms it with real fidelity.
The Expectations Index asks people to forecast. And people are terrible forecasters, especially about macroeconomics, especially when the question is filtered through whatever they watched on television the night before. Academic work on consumer sentiment has been pointing at this for a while, and the divergence got much wider after 2021. Sentiment readings collapsed and stayed collapsed while actual consumption held up. The survey stopped predicting behavior because respondents stopped answering the economic question and started answering a mood question.
The Conference Board itself noted that mentions of war, geopolitics, and conflict shifted through the July sample period, and that references to food and grocery prices got more frequent. That's not a growth forecast. That's a read on what was loud that month.
So the honest framing is this: the headline confidence number is a sentiment reading with a recession threshold attached to it that made sense in a different behavioral era. The Present Situation half still works. The Expectations half has become a proxy for how annoyed people are about prices and headlines.
Now here's the part worth your attention. There's a third number inside the same release that most coverage never mentions, and it's the one institutional desks actually pull.
It's called the labor differential. The Conference Board asks respondents whether jobs are currently plentiful or currently hard to get, and the differential is simply the first percentage minus the second. It's been narrowing steadily, and it's now near its tightest reading in more than five years, with the share of people describing jobs as hard to get running around levels last seen in early 2021.
The labor differential works when the Expectations Index doesn't, and the reason is structural. It isn't asking anyone to forecast anything. It's asking a few thousand households to report on something they've personally observed. That's the same reason the Present Situation Index has held its signal quality. Surveys that ask about observed experience stay useful. Surveys that ask about the future degrade into vibes.
The labor differential also has a documented relationship with the unemployment rate, and it tends to move first because it captures the friction people feel before that friction turns into a layoff statistic. When the differential compresses, it means the people who are already employed are noticing that the exits are getting narrower. That shows up in behavior well before it shows up in a jobs report.
There's a second reason to prefer it over the headline, and it's about how the sausage gets made. The Conference Board's survey runs on an online panel with a cutoff date roughly a week before publication, which means the headline you read Tuesday reflects responses gathered during the first three weeks of August. Preliminary results get revised as late responses come in. A headline that moves two points on preliminary data and then revises back is extremely common, and reacting to the preliminary print means reacting to a number that partially doesn't exist yet. Note that June's reading was revised upward before July's release, which is precisely why July's decline looked steeper than it was.
Contrast that with how a professional actually treats this release. Nobody on a macro desk builds a position off the headline confidence number. What they do is compare the labor differential against the JOLTS quits rate and continuing jobless claims, because when all three point the same direction you have corroboration from a survey, an employer report, and an administrative record. One indicator is an anecdote. Three independent measurement methods agreeing is a signal. The mistake retail investors make isn't reading the wrong indicator, it's reading one indicator in isolation and granting it more authority than a single data series can carry.
One more line from the July release deserves a mention, because it has direct consequences for your balance sheet. Roughly 61 percent of respondents expected higher interest rates over the next twelve months, essentially unchanged from the prior month. That's a large majority, and it happens to line up with what the bond market is pricing. The effective federal funds rate has been sitting at 3.63 percent, but the two-year Treasury has been trading around 4.19 percent and the ten-year around 4.71 percent, which is the market's way of saying it expects the next move to go up rather than down.
When most consumers and the bond market agree that rates are heading higher, the practical question stops being "what will the Fed do" and becomes "which of my liabilities is exposed to that." You can check the current Treasury curve yourself on the Federal Reserve's daily H.15 release, which posts every business day at 4:15pm and is free.
So what do you actually do with all of this on a Tuesday morning?
The first thing is to stop letting sentiment data touch your allocation. This sounds obvious and almost nobody does it. Write down, in advance and in plain language, the specific conditions that would cause you to change your portfolio. Not "if things look bad." Something you can verify: a change in your own income, a change in your time horizon, a change in your obligations, an allocation that has drifted more than a set number of percentage points from target. A survey of how strangers feel about the next six months isn't on that list, and if you make yourself write the list down, you'll notice it isn't.
The second thing is to use the labor differential as a personal risk gauge rather than a market one. This is where the data genuinely earns its place in your process. A compressing labor differential means the cost of losing your income just went up, because replacing it takes longer. That's not a signal to sell equities. It's a signal to extend your cash buffer. If you've been running three months of expenses in reserve, a tightening labor market is the argument for six. If you were planning to negotiate an exit, it's an argument for doing it while you still have leverage. If you carry variable-rate debt and a majority of the country expects rates to rise, that's the moment to price out a fixed alternative.
Notice what those decisions have in common. They're all about your own balance sheet, and none of them require you to be right about the direction of the S&P 500. That's the entire distinction between using macro data and reacting to it.
The third thing is to separate your observation window from your decision window. You can read every release. Read them all. But the decision cadence should be slower than the data cadence, deliberately and by design, because the data arrives monthly and your financial situation does not change monthly. Institutions handle this with an investment policy statement that specifies when rebalancing happens regardless of what the news said that week. You can do the same thing on one page.
The mechanics matter here more than the philosophy. If your decision window is quarterly but your accounts require you to log in and manually move money, you'll end up making decisions on whatever day you happen to log in, which will be a day the news was loud. That's how good intentions turn into reactive trading. Automation isn't about convenience, it's about removing the moment where sentiment gets a vote. Setting up recurring contributions and target allocations at a platform like M1 Finance puts the rebalance on a schedule instead of on a mood, which is the whole point.
You also need to be able to see your actual position before you can judge whether anything meaningful has changed. Most people can't answer basic questions about their own finances quickly, which is exactly why headlines feel so urgent. When you don't know your numbers, every number sounds important. Connecting your accounts to a free tracker like Empower gives you a running view of net worth, allocation drift, and cash flow, and it turns "should I be worried" into a question you can answer with arithmetic in about ninety seconds.
And if you're running a business alongside your portfolio, the same principle scales. The reports that matter should reach you on a fixed schedule without you going to fetch them. I run my own release calendar and portfolio alerts through Make, which pulls economic release dates and account thresholds into one place and pings me only when something crosses a line I set in advance. The point isn't the tool. The point is that the trigger is defined before the news arrives, not after.
Tuesday's number will move markets for about twenty minutes. Some outlet will write that confidence hit a multi-year low and that the recession signal is flashing again. It will be technically accurate and practically useless, because that signal has been flashing since February 2025 and the correct response then was to do nothing, which remains the correct response now.
The Present Situation Index and the labor differential are the two lines worth reading, and they're worth reading for what they say about your job and your cash buffer, not about your equity allocation. Everything else in that release is a mood ring with a press embargo.
I built a one-page reference that sorts the major monthly economic releases into three buckets: the ones that lead, the ones that merely confirm, and the ones that are noise dressed up as information. It lists what each report actually measures, which sub-component carries the signal, the release date and time, and the specific personal finance decision it should inform, if any. It's the filter I use to decide what's worth opening.
Reply to this email with the word SIGNAL and I'll send it over.
And if you're finding this useful, forward it to someone who's been talking about a recession since 2023. Three referrals unlocks the full Money Systems Lab playbook library. Ten unlocks lifetime premium access.
Taylor Voss
Money Systems Lab