The Lewsearch Report · Issue 06
Pre-registered · resolves Sep 21, 2026

A Synthetic Forecast of August Household Spending, Frozen Before the Fed Print

Lewsearch forecasts that the median U.S. household will report monthly spending 4.9 percent higher than a year ago when the New York Fed publishes its August 2026 Household Spending Survey. The Fed's April reading was 4.8. The official number is due Monday, September 21.

Frozen September 15, 2026 · Resolves Monday, September 21, 2026 · n=1,000 synthetic household heads

The forecast below was frozen before the print. After September 21 nothing on this page changes except that a grade is added.

Forecast · August 2026

4.9

Median reported year-over-year change in monthly household spending. This is the graded number.

Last print · April 2026

4.8

The New York Fed's most recent official reading. The no-change comparator.

Panel · before calibration

4.2

The median answer from the 1,000 simulated respondents before Lewsearch's calibration step. Published for the record. Not graded.

The series so far

Apr 2025

4.5

Official

Aug 2025

4.1

Official

Dec 2025

4.9

Official

Apr 2026

4.8

Official

Aug 2026

4.9

Forecast

New York Fed, SCE Household Spending Survey: median reported year-over-year change in monthly household spending, rounded to one decimal. The August 2026 cell is Lewsearch's forecast, not a Fed number.

WIN rule

WIN if the official August 2026 reading lands within 0.4 points of 4.9 and 4.9 is closer to that reading than 4.8 is. Both conditions must hold. A tie is not closer.

  • Prints 4.8Miss. Unchanged from April: inside the band, but not closer than the last print.
  • Prints 5.0Win. 0.1 from the forecast, 0.2 from the last print.
  • Prints 5.4Miss. 4.9 is the nearer number, but it is outside the band.

Read plainly, the forecast is a call that the reading ticks up. Graded against the New York Fed's published August 2026 Household Spending Survey, expected Monday, September 21, 2026. If the Fed's posting date moves, the grade waits for the print; nothing else changes.

NY Fed SCE Household Spending Survey →

What we asked

How does your current monthly household spending compare with 12 months ago?

The Fed's headline is the median of that reported percent change, computed by interpolation. We put the question, in the Fed's terms, to 1,000 simulated U.S. household heads and published the calibrated median as the forecast. This is not a reproduction of the Survey of Consumer Expectations. It is a synthetic read of the same series.

Also frozen · not graded

The same panel answered the rest of the module. These forecasts were frozen with the headline and are published for the record. Only the headline is graded.

MeasureForecast · Aug 2026Last print · Apr 2026
Expected change in monthly household spending over the next 12 months (median, percent)
Overall4.83.4
Essential items5.85.1
Non-essential items1.71.8
Expected change by category over the next 12 months (median, percent)
Housing3.33.1
Utilities4.74.8
Food5.55.6
Clothing2.62.9
Transportation4.85.4
Medical care4.84.9
Recreation2.42.8
Education0.20.2
Likelihood of a large purchase in the next four months (average percent chance)
Vehicles9.89.6
Vacations28.531.0

One decimal. Last-print values are the April 2026 wave from the New York Fed's chart data.

How this was produced

Panel. 1,000 simulated U.S. household heads, weighted to Census household benchmarks. The same 1,000 answered every question on this page. No humans were interviewed.

Calibration. The panel's answers pass through Lewsearch's production calibration step. That step was frozen before this forecast was produced and was not changed after the number was seen. Earlier published waves of this series were available to it; the August 2026 wave was not, and has not been published by the Fed. The recipe is proprietary and is not published.

Record. Frozen September 15, 2026, before the print. The same numbers are in the aggregates file. We will not swap 4.9 after September 21.

Read before citing

This is a synthetic panel, not a probability sample. No humans were interviewed. 4.9 is the forecast and the only graded number. 4.2 is the panel's reading before calibration, published for the record. Neither will be swapped or revised after September 21.

Try it on your category

See the panel on a question you actually ask.

Book a demo. We will run a study on your audience and walk through how we publish the miss.

Revised September 15, 2026, before the print: series context and the ungraded table added; copy tightened. Forecast, last print, panel reading, and WIN rule unchanged.