Guide

Retail sales explained

U.S. retail sales is the Census Bureau’s monthly estimate of goods purchased at stores, online, and through other retail channels — the first hard dollar read on household spending each month. Unlike consumer confidence surveys, retail sales measures what consumers actually spent, not how they feel. The Advance Monthly Retail Trade Survey (MARTS) covers roughly two-thirds of total personal consumption in goods categories and feeds directly into BEA estimates for GDP. Markets react to the headline month-over-month change, but economists and portfolio managers often focus on the retail sales control group — a subset that maps cleanly to the personal consumption expenditures (PCE) component of GDP. This guide covers what the series includes and excludes, headline vs control group vs ex-auto/gas, category breakdown, seasonal adjustment and revision patterns, the link to GDP consumption, a Harbor Retail release-day read worked example, an indicator decision table, common pitfalls, and an investor checklist.

What retail sales measures

The Census Bureau surveys roughly 5,500 retail and food-service firms each month, representing firms that account for about 65% of national retail sales. Respondents report dollar sales for the reference month. Results are published as the Advance Monthly Sales for Retail and Food Services — usually around the 15th of the following month at 8:30 a.m. ET, alongside other “blockbuster” releases on the economic calendar.

Retail sales is nominal — not adjusted for inflation. A 0.5% monthly gain driven entirely by higher gasoline prices is not the same as 0.5% growth in real volume. Analysts cross-check with CPI subcomponents and BEA real PCE to separate price from quantity.

What is included

Merchandise at retail establishments — department stores, apparel, electronics, furniture, sporting goods, building materials, and nonstore retailers (e-commerce).
Motor vehicle and parts dealers — a large, volatile slice; often reported separately.
Gasoline stations — reflects price and volume; swings with crude oil.
Food services and drinking places — restaurants and bars; the only services category in the retail sales report (most services appear only in PCE).

What is excluded

Most services — healthcare, rent, insurance, education, and professional services are not in retail sales; they appear in BEA PCE instead.
Used vehicle sales between consumers — only dealer transactions count.
Spending abroad by U.S. residents — PCE includes some; retail sales is domestic retail establishment focus.

Because services dominate U.S. consumption (~two-thirds of PCE), retail sales alone is an incomplete picture of total household demand. It is still the earliest high-frequency goods read and moves equity sectors tied to discretionary merchandise.

Headline, control group, and ex-auto/gas

Financial media quotes the headline retail sales month-over-month percent change — total sales including autos, gasoline, and restaurants. Three common adjustments strip volatile or GDP-irrelevant pieces:

Ex-autos (or ex-motor vehicle and parts)

Auto sales are lumpy — incentive campaigns, supply-chain bottlenecks, and fleet orders create month-to-month noise. Retail sales excluding autos isolates underlying merchandise trends. Auto dealers themselves remain a key cyclical sector; watch both series.

Ex-gasoline

Gas station sales move with pump prices as much as with miles driven. When crude spikes, headline retail sales can rise even if consumers buy fewer gallons. Ex-gas versions remove this price illusion.

Retail sales control group

The control group excludes autos, gasoline, building materials, and food services. BEA economists designed this subset because it correlates with the goods portion of PCE that retail sales is meant to proxy for GDP forecasting. On release morning, sell-side economists often publish “control group” surprises — frequently more market-moving than headline for rates and growth trades.

A typical release headline might read: “Retail sales +0.3% m/m (consensus +0.4%); ex-autos +0.2%; control group +0.1%.” Read all three before drawing conclusions — strong autos with weak control group signals a different economy than broad-based goods strength.

Category breakdown — what to watch by sector

Census publishes sales by industry group. Investors map categories to public equities and to macro narratives:

Nonstore retailers — e-commerce and mail-order; structural share gainer vs brick-and-mortar; watch vs general merchandise and clothing stores.
Food and beverage stores — staples; relatively stable; spikes during pantry-loading events.
Clothing and accessories — discretionary; correlates with confidence and weather; back-to-school and holiday seasonality.
General merchandise — big-box retailers; middle-income spending barometer.
Furniture and home furnishings — interest-rate sensitive; ties to housing turnover (see housing starts).
Building material and garden equipment — excluded from control group but signals DIY and construction activity.
Electronics and appliance stores — lumpy around product cycles and promotions.
Sporting goods, hobby, musical instrument, and book stores — discretionary niche; useful for leisure spending reads.
Food services and drinking places — real-time services demand proxy inside an otherwise goods-heavy report.

Year-over-year comparisons matter for trend analysis; month-over-month seasonally adjusted changes dominate headline trading. Heat maps by category reveal whether weakness is broad or concentrated (e.g., only autos soft).

Seasonal adjustment, revisions, and data quirks

Published figures are seasonally adjusted annual rate (SAAR) for levels and seasonally adjusted month-over-month for changes — Census removes predictable patterns (holiday shopping, summer travel, tax-refund timing). Residual seasonality occasionally distorts January and December prints; compare three-month moving averages when single months look extreme.

Revision pattern

Advance — first estimate ~two weeks after month-end; based on incomplete sample.
Revised — published with the following month’s advance report; can flip the narrative.
Annual benchmark — once yearly, Census revises levels to match the Annual Retail Trade Survey; historical levels can shift.

Unlike GDP, retail sales revisions are smaller, but a +0.6% advance revised to +0.2% the next month still matters for growth tracking. Keep a revision log if you trade the release systematically.

Deflation and inflation

Nominal retail sales rising 4% year-over-year with 3% goods inflation implies roughly 1% real volume growth — back-of-envelope only; category-level CPI weights differ. BEA’s real PCE (released later) is the authoritative inflation-adjusted consumption series for GDP.

Retail sales vs GDP personal consumption

Personal consumption expenditures (PCE) is the largest component of U.S. GDP — roughly 68% of output. Retail sales covers a subset of PCE goods sold through retail channels plus food services. The mapping is approximate:

Control group retail sales — best single proxy for the goods slice BEA uses when forecasting the first PCE estimate.
Services PCE — healthcare, housing services, financial services, and recreation services have no retail sales equivalent; they arrive via other surveys and administrative data.
Timing — retail sales leads the monthly PCE goods detail; full PCE drops with the personal income and outlays report (usually end of month / early next month).
Revisions — PCE is revised across multiple GDP vintages; retail sales revisions are more contained but can still change the Q/Q consumption story at the margin.

Macro strategists build nowcasts — combining retail sales control group, imports, inventory data, and card spending aggregates to estimate GDP before the first official print. A soft control group two months in a row raises recession watchlist probability alongside output and labor indicators, but one month is rarely decisive alone.

How investors use retail sales release days

Retail sales prints at 8:30 a.m. ET. Markets trade the surprise versus Bloomberg/Reuters consensus and the component story — especially control group and revisions to prior months.

Typical asset reactions

Equities — strong control group supports consumer discretionary (retailers, restaurants, leisure); weak prints favor staples and prompt growth-scare rotations. Category detail moves individual names (home improvement vs apparel).
U.S. Treasuries — hot retail sales can sell off bonds (higher growth / inflation risk); weak sales support duration. Reaction size depends on whether the move confirms or contradicts prior payrolls and PMI data.
U.S. dollar — resilient consumption supports USD in growth-differential trades versus economies with softer retail data.
Commodities — gasoline category affects energy narrative; broad goods strength supports industrial metals demand expectations.
Fed expectations — sustained above-trend control group with sticky services inflation keeps “higher for longer” pricing; three consecutive soft months opens cut speculation — always cross-check with CPI and labor costs.

Position squaring ahead of the release is common; liquidity thins at 8:29 a.m. Long-term investors weight three-month control-group trends and real spending over single headline surprises.

Worked example: Harbor Retail release-day read

Harbor Retail operates 120 mid-market apparel and home-goods stores. On Census retail sales morning, the planning team runs a twenty-minute checklist before updating same-store sales guidance — complementing the confidence survey read from their monthly sentiment review:

Read control group first — if control group misses consensus while headline beats (autos/gas driven), Harbor treats it as neutral-to-negative for discretionary apparel — not a green light to raise guidance.
Check clothing and general merchandise lines — direct category comps for Harbor’s revenue mix; note year-over-year vs month-over-month direction.
Compare nonstore vs brick-and-mortar — if nonstore outpaces clothing stores by 300 bps again, Harbor accelerates ship-from-store and marketplace listings rather than new mall leases.
Read food services for context only — restaurant strength signals consumer willingness to spend on experiences; weak restaurants plus weak clothing is a double discretionary warning.
Note revisions to prior two months — an upward revision to last month’s control group changes the three-month trend Harbor uses for inventory buys.
Write one paragraph for the ops journal — e.g. “Advance retail sales headline +0.4% (cons +0.3%); control group −0.1% (cons +0.2%). Clothing stores −0.3% m/m. Revise August comp guidance from +2.5% to +1.8%; hold October inventory flat; marketing spend unchanged pending Michigan final.”

Harbor uses hard sales data to confirm or veto sentiment signals before committing to purchase orders — retail sales arrives weeks before their own monthly close and months before GDP revisions settle.

Indicator decision table

Question you have	Best indicator	Why
How much did consumers spend on goods last month?	Census retail sales (control group)	Earliest hard dollar goods read; maps to PCE goods
Will households cut spending before GDP?	Consumer confidence Expectations + retail sales trend	Sentiment leads; retail sales confirms magnitude
Total consumption including healthcare and rent?	BEA personal consumption expenditures	Full GDP consumption measure; released later
Are businesses cutting orders?	PMI New Orders	B2B channel; can lead retail by one to two months
Real volume vs inflation?	Real PCE + CPI goods components	Retail sales is nominal; must deflate for volume
Auto-specific demand?	Motor vehicle retail sales line + SAAR unit sales	Headline ex-autos hides dealer detail
Housing-related goods demand?	Furniture + building materials categories + housing starts	Rate-sensitive; lags home sales and permits
Official recession dating?	NBER coincident indicators	Retail sales alone does not date recessions

Common pitfalls

Trading headline while ignoring control group — autos and gas distort headline; GDP forecasters watch control group first.
Treating nominal sales as real growth — always ask how much of the move is prices, especially for gasoline and food.
Ignoring revisions — the story can change when prior months are revised in the same release.
Equating retail sales with total consumption — services dominate PCE; weak retail sales with strong healthcare spending still yields GDP growth.
Overreacting to one month — weather, holiday shifts, and tax-refund timing create noise; use three-month averages.
Missing seasonality artifacts — January and November/December are prone to residual seasonality surprises.
Comparing U.S. retail sales to eurozone retail without adjustment — different coverage, calendars, and deflation methods.
Forgetting food services is in the report but out of control group — restaurant strength won’t show in control group but still signals consumer health.

Investor checklist

On release morning (8:30 a.m. ET, ~15th of month), read headline, ex-autos, and control group versus consensus.
Scan revisions to the prior one to two months — recalculate three-month control-group annualized trend.
Review category table for breadth: how many major groups positive vs negative month-over-month.
Separate gasoline and auto contributions before updating GDP nowcasts.
Cross-check with prior confidence surveys — convergence strengthens the signal; divergence warrants caution.
Map category moves to sector weights (XRT retail ETF, restaurants, home improvement, autos).
Log surprises in your macro journal; note bond and equity reactions versus payroll week context.
When PCE releases later, compare BEA goods detail to your retail-sales-based forecast and note drift for next month.

Key takeaways

Retail sales is the Census Bureau’s monthly nominal estimate of goods sold at retail establishments plus food services — the earliest hard spending read.
The control group (ex autos, gas, building materials, food services) is the key GDP forecasting line; headline alone can mislead.
Most consumer services are excluded — pair retail sales with BEA PCE for total consumption.
Seasonal adjustment and monthly revisions add noise; trend three-month control-group changes for macro conclusions.
Category detail links to sector trades; confirm sentiment indicators with retail sales before sizing discretionary exposure.