Skip to main content

What do housing starts tell you about economic health?

Housing starts measure the number of new residential construction projects that broke ground in a given month. Released monthly by the Census Bureau, the data count single-family homes, townhomes, and apartment buildings. For economists and Fed policymakers, housing starts are a crucial leading indicator: they respond quickly to interest rate changes, tell you whether builders are confident about future demand, and directly measure construction employment—one of the fastest-hiring sectors in any recovery.

Housing is special in the economy. It's the largest asset most households own; it employs millions (construction workers, real estate agents, mortgage brokers, inspectors); it anchors consumer wealth and confidence (when house prices fall, consumers feel poorer and cut spending); and it's highly sensitive to interest rates (a 1 percent rise in mortgage rates kills affordability and immediately reduces housing demand). For these reasons, housing starts are a leading indicator economists watch obsessively.

Quick definition: Housing starts measure the number of new residential construction projects that began during a month, reported as an annualized rate (the rate, if sustained for a year, would equal that many starts), and serve as a gauge of builder confidence and housing demand.

Key takeaways​

  • Housing starts are reported as an annualized rate (e.g., "1.2 million starts" means the monthly count, extrapolated to a full year, would equal 1.2 million).
  • Starts respond quickly to interest rate changes and builder sentiment; a 50-basis-point rate hike often reduces starts 10–20 percent within two to three months.
  • Single-family starts dominate (about 75 percent of total); multifamily (apartments) are more cyclical and respond sharply to rent growth and migration patterns.
  • Building permits, released alongside starts, are a leading indicator to starts (permits often lead starts by one to two months).
  • Housing starts of 1.0–1.3 million annually are considered healthy for the U.S. population; below 0.8 million signals tight supply; above 1.8 million signals over-building.

How housing starts are measured and reported​

The Census Bureau conducts the Survey of Construction, sending questionnaires to about 10,000 residential construction firms monthly. The survey asks: How many new residential projects began in your area this month? From these survey responses, the Bureau projects to the national level and publishes three data points:

Total housing starts: All new residential units, including single-family homes, townhomes, and apartment buildings.

Single-family starts: Detached homes, typically owner-occupied. These are the most stable component because single-family home demand is driven by household formation (marriages, births, immigration) and is relatively steady year-to-year.

Multifamily starts: Buildings with five or more units (apartments, condos). Multifamily is more cyclical; it jumps when rents are rising (suggesting strong rental demand) and crashes when rents are falling or cap rates (investment returns for landlords) are compressing.

The data are reported as annualized rates. If the Census Bureau counts 100,000 housing starts in one month, the published number is "1.2 million," meaning that monthly rate extrapolated to 12 months would be 1.2 million. This annualization smooths monthly volatility and makes year-to-year comparisons cleaner, but it can be misleading: a single bad month doesn't actually mean starts fell 10 percent annualized; you have to look at three-month trends.

Why interest rates kill housing demand so quickly​

Housing is the most interest-rate-sensitive sector of the economy. Here's the math: a 30-year fixed mortgage at 3 percent on a $400,000 home costs about $1,686 per month (principal + interest). At 4 percent, the same home costs $1,909 per month—$223 more. At 5 percent, it's $2,147—a 27 percent increase in monthly cost.

Most home buyers have a monthly payment budget (based on their income and down payment). When interest rates rise by even 100 basis points (1 percentage point), that fixed budget buys a significantly cheaper house. In the 2022 rate-hike cycle, 30-year mortgage rates rose from 3.1 percent in January to 6.9 percent by October. The monthly payment on a median home jumped about 45 percent. The result: housing starts fell from 1.4 million annualized in December 2021 to 0.9 million by October 2022—a 36 percent collapse in nine months.

This is why the Federal Reserve watches housing starts like a hawk when it's hiking rates. If the Fed raises rates by 300 basis points (as it did from 2022 to 2023), housing starts will fall sharply within two to three months. Policymakers have to decide: is the pain (job losses in construction, falling home prices) worth the anti-inflation benefit? This is one of the hardest tradeoffs in monetary policy.

The lead from permits to starts​

Building permits are even more forward-looking than housing starts. A building permit is issued by local government before construction begins. The Census Bureau reports permits and starts together, and permits typically lead starts by one to two months. Strong permit growth predicts strong future start growth; weak permits predict weak future starts.

In 2021, permits jumped to 1.9 million annualized (highest in 15 years), signaling that builders expected strong demand ahead. Starts followed, reaching 1.6 million by late 2021. Then, as the Fed signaled rate hikes in early 2022, permits collapsed 30 percent by October 2022. Starts followed with a lag of one to two months.

Analysts use the permits-to-starts ratio to gauge builder caution. When permits are high relative to starts, builders are laying groundwork but haven't begun construction—a sign they're uncertain about the timeline or financing. When starts are close to permits, builders are moving fast. During the 2008 crisis, permits fell below starts (builders were completing projects begun before the crash but not starting new ones), a sign of severe capitulation.

Single-family versus multifamily: Two different stories​

Single-family starts are driven primarily by household formation (marriage, divorce, kids, immigration) and home prices. They're relatively stable and move slowly—shifts of ±10 percent year-to-year are normal. Single-family starts of 0.75–0.95 million annualized are considered healthy for a 130-million-person workforce.

Multifamily starts are driven by rent growth, cap rates (the return landlords expect), and commercial real estate cycle dynamics. Multifamily is boom-and-bust: when rents surge, developers start apartment buildings en masse. When rents level off, multifamily starts crash. The 2012–2022 period saw a multifamily boom—rents rose steadily due to demographic inflows (millennials, immigrants) and constrained supply, so multifamily starts surged to 400,000+ annualized by 2022. Then, in 2023, as rents stabilized and landlords couldn't raise rents anymore, multifamily starts collapsed 40+ percent.

Understanding this distinction is crucial. A headline decline in housing starts might be -5 percent, but if it's entirely driven by a -40 percent multifamily collapse while single-family held flat, the story is: "Apartment-building speculation is cooling, but household demand for homes is steady."

Employment and multiplier effects from housing construction​

Housing construction is one of the highest-multiplier sectors. Building a home directly employs: carpenters, electricians, plumbers, roofers, HVAC technicians, inspectors, and supervisors. It indirectly employs: truck drivers (moving materials), warehouse workers, equipment rental staff, and surveying engineers. A single new home start generates about 1.5–2.0 jobs in construction and related industries. Across the U.S., a million housing starts translates to roughly 1.5–2.0 million jobs directly and indirectly.

This is why housing is so important to the jobs market. During recessions, housing starts fall, construction employment crashes, and unemployment soars. During recoveries, housing leads: starts rise, construction hires heavily, and unemployment falls. The 2008 recession saw housing starts fall from 2.3 million to 0.4 million over two years—a 83 percent collapse. Construction employment fell 2.3 million. The subsequent 2009–2019 recovery saw housing starts climb back to 1.4 million, and construction employment recovered all the lost jobs by 2018.

Beyond employment, housing construction also drives demand for materials: lumber, copper, drywall, cement, appliances. These supplier industries employ millions more. When housing starts surge, lumber mills ramp production, steel mills get orders, and appliance factories see demand tick up. This is why housing starts are a leading indicator for the whole economy: they predict construction employment, materials demand, and manufacturing activity months ahead.

What causes swings in housing starts beyond interest rates​

While interest rates are the dominant driver, other factors matter:

Builder sentiment. The National Association of Home Builders publishes a Homebuilder Confidence Index monthly. When builders expect future demand to be strong and they expect to sell homes at profit, they start projects. When confidence collapses (due to recession fears, price declines, or margin pressure), starts fall. The index typically leads starts by one to three months.

Housing inventory and prices. When existing home inventory is low and prices are rising, builders see opportunity for profit and increase starts. When inventory is high and prices are falling, builders cut back. During 2020–2021, existing home inventory fell to historic lows (homeowners didn't want to sell in a low-rate environment) and prices surged. This spurred multifamily and single-family starts. By 2023, as existing home prices fell and inventory rose, builder interest in new construction cooled.

Supply-chain constraints. During 2021–2022, lumber prices surged 300 percent, gypsum board was scarce, and labor was hard to find. These supply shocks raised construction costs 15–25 percent, squeezing builder margins and killing profitability. Starts fell partly due to rate hikes, but also partly due to cost pressures and labor shortages. By 2023, as supply chains normalized and lumber prices fell 70 percent from their peak, cost pressures eased and builders' willingness to start projects improved.

Migration and demographic patterns. Population inflows to growing metros (Austin, Phoenix, Miami) drive local housing demand and construction. Outflows from declining cities (Detroit, Pittsburgh) reduce demand there. National housing starts are the sum of all these local patterns, so understanding where people are moving helps explain starts data.

Reading the monthly housing starts report​

The Census Bureau releases housing starts around the 17th of each month, covering the prior month. The release includes:

  • Total starts (monthly annualized rate). Reported in thousands (e.g., "1,200 starts" = 1.2 million annualized).
  • Month-over-month change (percent). Usually ±5 to ±10 percent is normal; beyond ±20 percent warrants attention.
  • Year-over-year change (percent). More stable than month-to-month; positive year-over-year for three consecutive months suggests trend strength.
  • Single-family and multifamily breakouts. See which category is driving changes.
  • Building permits (same metrics). Typically released same report or one day prior; always compare permits to starts.

For a quick read: if starts and permits are both above 1.2 million annualized and rising, housing demand is strong and construction employment should follow. If both are below 0.9 million and falling, housing is cooling and construction employment is at risk. If starts are rising but permits are flat or falling, builders are confident in near-term demand but uncertain about the future.

Real-world examples of housing starts turning points​

In mid-2006, housing starts peaked at 2.3 million annualized as the subprime mortgage bubble reached fever pitch. Builders couldn't start projects fast enough. By late 2006, leading indicators (mortgage applications, building permits) started falling. By mid-2007, starts were declining. By late 2008, starts had fallen to 0.55 million—a collapse of 76 percent in two years. This collapse preceded the full recession and was the first major warning sign that the economy was in serious trouble.

In early 2020, COVID caused a brief shock: starts fell from 1.4 million in February to 0.96 million in April (a 32 percent decline in two months). But then, as the Fed cut rates to zero and Congress passed stimulus, demand rebounded. Starts surged to 1.6 million by mid-2021—the highest in 15 years. This housing surge was a leading indicator of the post-COVID recovery; construction employment surged, materials demand climbed, and the economic rebound was underway.

Then, in 2022, as the Fed hiked rates from 0 to 4.25 percent (fastest in 40 years), housing starts fell from 1.4 million to 0.9 million by Q4 2022. Construction employment gains paused and housing prices began to fall. By early 2023, many forecasters predicted a housing crash and recession. However, as rate-hike expectations moderated in spring 2023 and the labor market remained strong, housing starts stabilized, suggesting the downturn would be modest, not severe.

Common mistakes when reading housing starts​

Confusing starts with sales. Housing starts measure new construction that broke ground; new home sales measure homes that closed. These are different. Starts are a leading indicator of future supply. Sales are a trailing indicator of demand. A gap between rising starts and flat sales suggests builders are ahead of demand and inventory is building.

Over-interpreting monthly volatility. Housing starts swing ±10 to ±15 percent month-to-month due to weather, permit processing delays, and random variation. A single bad month is not a trend. Look at three-month or six-month averages to identify trends.

Ignoring permits. If permits are flat but starts are rising, builders are using up their permit pipeline and future starts will eventually fall. If permits are surging but starts are flat, the pipeline is building and future starts should rise. Permits are the leading indicator to starts.

Forgetting the interest-rate sensitivity. The mortgage rate is the single most important variable predicting starts. A sudden 100-basis-point rate rise will depress starts 10–20 percent within two to three months. If you're reading housing starts, always check: what are mortgage rates doing?

Confusing starts with completions. A start is the beginning of construction; a completion is the end. Between start and completion, there's typically 6–12 months of building. A collapse in starts doesn't immediately reduce housing supply; it does so after the lag.

FAQ​

Why does the Census Bureau annualize housing starts instead of reporting them as a raw monthly count?​

Annualization makes month-to-month and year-to-year comparisons valid. Raw data would show huge seasonal swings (more starts in spring and summer, fewer in winter) that would obscure underlying trends. Annualization smooths these patterns and makes trends visible.

How does housing starts affect the GDP calculation?​

Housing starts don't directly increase GDP until construction is completed and the home is sold (a transaction, not included in GDP). What does increase GDP is the labor expended during construction (counted as value-added) and the goods consumed (lumber, steel, appliances). These show up in GDP as business investment (the construction itself) and goods consumption.

Can housing starts predict recessions?​

Housing starts are a strong recession indicator, not a predictor. A sudden collapse in starts (say, from 1.3 million to 0.9 million in six months) is almost always accompanied by or follows recession onset. However, a gradual decline in starts (0.5 percent per month for a year) might be normal supply-demand adjustment and not recession-related.

What happens to housing starts when the Fed raises interest rates but inflation is falling?​

This is the key scenario playing out in 2023–2024. The Fed raises rates primarily to fight inflation, but as inflation falls, the case for continued hikes weakens. If mortgage rates fall before the Fed actually cuts rates, housing demand can surge before the Fed acts. This is how "soft landing" scenarios unfold: Fed raises rates (housing starts fall), but then inflation falls fast enough that the Fed starts cutting (housing starts recover) before a true recession hits.

How did remote work affect housing starts in 2021–2022?​

Remote work enabled people to move away from expensive urban centers (San Francisco, New York, Boston) to cheaper, smaller cities and suburbs. This shifted demand geographically but didn't necessarily increase total national demand. Local housing starts surged in boomtowns (Austin, Miami, Phoenix) while declining in coastal cities. National starts were driven more by overall demand (low rates, stimulus, migration) than by remote work per se.

Summary​

Housing starts measure the number of new residential construction projects that began during a month, reported as an annualized rate, and serve as a leading indicator of builder confidence, construction employment, and economic momentum. The Census Bureau surveys construction firms monthly; starts include single-family homes (about 75 percent), multifamily apartments (about 25 percent), and mobile homes. Single-family starts are driven by household formation and are relatively stable (0.75–0.95 million annualized is healthy). Multifamily starts are cyclical and respond sharply to rent growth and investment returns for landlords. Interest rates are the dominant short-term driver: a 1 percent rise in mortgage rates reduces starts 10–20 percent within two to three months. Building permits, released alongside, lead starts by one to two months and gauge future builder intentions. Housing construction employment is high-multiplier (1.5–2.0 jobs per start) and drives demand for materials, making starts a reliable leading indicator of the broader economy. Sustained starts below 0.8 million signal tight supply; above 1.8 million signal over-building; 1.0–1.3 million is considered healthy.

Next​

→ Existing home sales explained