The official numbers and the lived experience have been telling different stories for months, and the gap is now wide enough to measure in both directions. Inflation, by the Bureau of Labor Statistics' count, is cooling: consumer prices rose 3.4 percent over the year through July, down from 3.5 percent in June and well off the year's 4.2 percent peak in May, and June actually brought the first monthly price decline since May 2020. The same week, the Chicago Fed's National Activity Index slipped negative for the tenth time in twelve months.
Meanwhile, a July survey by J.D. Power finds nearly 80 percent of consumers saying their daily expenses have changed, with groceries and gas leading the list, and roughly 30 percent reporting harder choices: selling possessions to cover bills, missing rent, mortgage, or utility payments, skipping prescriptions. American Banker's Paul Vigna, who does his household's grocery shopping, put the feeling plainly: the total always seems higher than it used to be. The two stories are not in conflict. They are about different variables. The statistics describe the rate. The households are paying the level.
The rate is a flow; the burden is a stock
Inflation, as measured, is a flow: the speed at which prices are rising. A cooling inflation rate means prices are rising more slowly, not that they are falling. A 3.4 percent annual rate is a promise that next year's grocery bill will be about 3.4 percent higher, and it is delivered on top of every increase of the past five years, none of which is ever unwound. Disinflation is not a refund. It is a deceleration, and the accumulated price level, the stock that households actually pay at the register, stays where the years of high readings put it.
The distinction sounds like a technicality until it explains behavior, and then it explains almost everything. A household is not experiencing the year-over-year percentage change. It is experiencing a checkout total that is a fifth higher than it used to be, and the total does not care that the monthly change was only 0.1 percent. The Fed watches the flow because the flow is what policy can steer. The household pays the stock because the stock is what the store charges. The two parties to the inflation conversation have been talking past each other, and the surveys are the sound of the household's side of the conversation.
This is also why the relief from cooling inflation arrives so slowly and feels so thin. Suppose inflation settles at 2 percent and stays there. Prices never come back down. The pain of the past five years is locked in permanently, and the only thing the household can hope for is that wages, which lag, eventually catch up to the new level. The best case the policy framework offers is not restoration. It is forgetting, and forgetting takes years, which is roughly the horizon the surveys keep pushing out.
Cooling is not relief, and the surveys keep saying so
The J.D. Power series has tracked the mismatch all year. In February, 65 percent of consumers said price increases were outpacing their incomes, and more than 85 percent said they had adjusted their daily lives in response, with a third turning down the thermostat or buying less food and 16 percent postponing or skipping medical care. In April, 87 percent said they expected prices for essentials to keep rising. By May, the firm classified around seven in ten consumers as financially unhealthy, with 81 percent having changed day-to-day spending habits and 30 percent cutting back on groceries or skipping meals.
The consistency is the finding. These are not readings of inflation expectations in the technical sense, and the surveys are samples, as Vigna is careful to note. They are readings of the stock: how the accumulated price level is landing on households whose incomes have not kept pace with it. When three in ten respondents report selling belongings or skipping prescriptions, the survey is not measuring a forecast. It is measuring a cash-flow emergency, distributed across a population that the aggregate data describes as coping. The extreme choices are the stock's bill coming due, one household at a time.
The extreme choices are the part the index cannot see
There is a reason the worst of this shows up in surveys rather than in the consumer price index. The index measures prices. It does not measure the second-order responses to prices: the prescription not filled, the mortgage payment postponed, the television sold. Those responses are real economic events with real costs, the untreated condition, the late fee, the future expense, and they are invisible to the price statistics until they surface somewhere else, in delinquency data, in health outcomes, in the momentum gauges. The economy does not only pay higher prices. It pays the price of not paying them, and the second cost is the one the surveys are picking up.
None of this requires a recession call. The Chicago Fed's index sits at negative 0.08, far above the negative 0.70 reading the bank treats as a historical recession signal, and the index's own components are mixed: production and sales held up in July, and employment improved, while personal consumption and housing weakened sharply. But the direction of travel is the survey's direction. Growth below trend, ten negative months out of twelve, a consumer sector worn down not by the rate of change but by the level it has been paying for years. The momentum gauge is a flow gauge too, which is why it can read mild while the household balance sheets read acute.
The economy pays twice for the level
The extreme choices deserve their own ledger, because each one is a cost transferred, not avoided. A skipped prescription is not a saving. It is a health expense moved forward, usually with interest, arriving later as a worse condition and a bigger bill, paid by the same household or the system around it. A missed rent payment becomes a late fee, then an arrears, then an eviction risk. Selling a possession converts an asset into cash at the worst moment to sell, and the cash buys a month of groceries that would have cost a fraction of the asset's worth in better times. None of this appears in the consumer price index, and none of it is cheap.
The macroeconomy absorbs the second bill the same way the household does. The Chicago Fed's July reading shows personal consumption and housing swinging from a positive contribution in June to the sharpest drag in the index, and consumption is roughly two-thirds of economic activity. When households postpone, the weakness shows up in the very gauges the policymakers watch, just later, at a lag that makes the cause easy to forget. The price level is a tax on households that collects invisibly. The surveys are the collection notices, and the momentum gauges are the receipts.
What policy can and cannot do about a level
It is worth being precise about the policy frontier, because the frontier explains why the frustration has no quick fix. The Federal Reserve's instruments steer the flow. Interest rates can slow price increases or speed them up, and the July report, with headline inflation at 3.4 percent and core at 2.5 percent, shows the steering working. What the instruments cannot do is subtract from the stock. No rate decision returns a grocery bill to its old level, and no central banker has ever proposed one that could. The level is history. Policy can only decide how much more history gets added, and the households paying the accumulated prices are not confused about the difference, even if the distinction rarely makes the press release.
The September rate decision now in focus will be about the flow, as it must be, and the households filling out the surveys will experience it, if at all, as a slightly slower climb. The two conversations are not going to merge. The Fed will keep reporting on the variable it controls, and the surveys will keep reporting on the one the households pay, and the gap between them will remain the best single explanation of why the economy can look cooled off and worn down at the same time. The moment the two variables converge will be the moment the story changes. The indicators will say so when it happens, and the indicators are not saying it yet.
The two variables will converge eventually, one way or the other
There are two ways this ends. One is the slow way: inflation stays near target, wages keep rising, and the price level gradually recedes into the background of household budgets, the way past inflations always did. The other is the fast way: the accumulated strain breaks something, spending cracks, and the level is reconciled with incomes all at once, at a cost nobody chose. Every indicator currently available is a bet on the first path, and the surveys are the reason the bet is not free.
The honest sentence about the moment is that the policy problem is being solved while the household problem is not. The Fed's variable is coming down. The household's variable never does. Prices rose, they will not fall back, and the only question left is whether incomes catch the level before the level catches the spending. The people filling out the surveys have been answering that question all year, and their answer is that they are still running behind the checkout total.
Primary sources
- American Banker's Bank Notes column by Paul Vigna for the J.D. Power July survey findings, the 30 percent extreme-choices figure, the Chicago Fed index reading, and the columnist's account of his own grocery shopping.
- The Hill's coverage of the J.D. Power survey series for the February and April survey figures and the 4,000-respondent methodology.
- Trading Economics' report on the July National Activity Index for the index level, the component contributions, and the recession threshold, and the July CPI report via the Bureau of Labor Statistics for the annual rate, monthly changes, and core reading.