NAHB Housing Economics - HMI, Starts & Market Data

Builders do not start projects on vibes - they start them when the order book looks healthy, when lumber quotes are workable, and when buyer traffic clears a threshold. The National Association of Home Builders tracks all three, every month, and the result is the single most-watched sentiment gauge in residential construction: the NAHB/Wells Fargo Housing Market Index.

The flagship product is the composite index on a 0-100 scale, where any reading above 50 means more builders view conditions as good than poor. Underneath the headline sit three diffusion-style sub-indices, each one its own time series: Single-Family Sales: Present, Single-Family Sales: Next 6 Months, and Traffic of Prospective Buyers. The regional cut breaks the national number into Northeast, Midwest, South and West, and the regional series come with 3-month moving averages to smooth the sampling noise that single-month readings can carry.

The history is deep. Table 2 of the standard release backfills the national composite to January 1985 as a year-by-month grid, giving you more than four decades of sentiment cycles in one consistent series. Around the HMI itself, the housing-economics hub layers Census-derived starts and permits analysis, framing lumber price history, state and local indicators and forecasts - so the index is the door, and the rest of the data is the room behind it.

Datadory delivers the whole hub as one production panel: the composite, all three components, all four regions, the moving averages and the supporting analysis, all keyed to a single monthly date column so joins are trivial.

Three different cuts from the August 2026 research pass, rendered exactly as they land in your warehouse.

August 2026 key findings (current month).

month         | hmi | current_sales | next_6mo_sales | buyer_traffic | pct_cutting_prices | avg_price_reduction_pct | incentive_use_pct
2026-08       | 35  | 39            | 43             | 23            | 35                 | 6                       | 63

Table 2 - national HMI history (seasonally adjusted, year-by-month grid).

year | Jan | Feb | Mar | Apr | May | Jun
1985 | 50  | 58  | 54  | 49  | 51  | 54
1990 | 42  | 44  | 40  | 39  | 36  | 36

Table 1 - national and regional HMI (2009 recovery print).

series                       | start_year | Jan | Feb | Mar | Apr | May
Housing Market Index         | 2009       | 8   | 9   | 9   | 14  | 16
Single Family Sales: Present | 2009       | 6   | 7   | 8   | 12  | 14

Read the August 2026 row as a stress test: composite 35 with present sales at 39 and traffic at 23 means builders feel the pipeline behind them far more than the foot traffic in front. The 35 percent cutting prices, 6 percent average reduction and 63 percent incentive-use figures quantify the response - this is what a price-sensitive cycle looks like in builder-survey form.

Read the 2009 row as the cycle's floor: every series touched single digits before climbing. The 1990 row captures an earlier downturn before the long 1990s recovery, and the 1985 row shows the very first print in the series - useful anchors for any backtest that needs to validate against a known regime.

Six documented fields cover the headline series, regional breakdown and history grid. Definitions below are verified against the source card, not inferred.

Geography: the United States national reading plus the four Census regions - Northeast, Midwest, South, West. Every regional series comes paired with a 3-month moving average to smooth the sampling variability that single-month regional prints carry.

Temporal: national composite history runs from January 1985 to the present on a monthly cadence, more than forty years of uninterrupted monthly prints. Component history, regional history and regional moving averages ride the same monthly grid, so any cross-cut yields aligned rows.

Granularity: one row per month per series. Each row carries the composite (or component, or regional value) on the 0-100 scale, with the year-by-month history table reformatting that as a column-per-month grid for fast lookup. Behind the index itself, the broader hub layers housing starts, permits, framing lumber prices and state/local indicators at matching monthly cadence.

API, files, or your warehouse. Daily, weekly, or hourly.

• Cycle timing for homebuilder equities - investors and quants use the HMI as a coincident-to-leading signal for single-family housing equities, and the gap between present sales and next-six-month expectations tends to lead the cycle turns by one to two quarters.
• Builder pricing and incentive strategy - homebuilder executives benchmark their own price-reduction and incentive rates against the monthly NAHB survey, watching how the regional series diverge as one market weakens before another.
• Mortgage and credit risk modeling - the index tracks the demand side that mortgage pipelines, servicing portfolios and homebuilder loan books all hinge on, and the forty-year backfill gives modelers multiple downturns to train on.
• Building-products demand forecasting - windows, roofing, HVAC and insulation suppliers correlate shipments to the regional HMI's next-six-month expectation component, which is the best single proxy for forward build orders in the survey.
• Lumber and commodity desks - framing lumber prices sit in the same hub as the HMI, giving commodity traders a sentiment-plus-cost pair that the futures curve alone never offers.
• Housing-policy journalism and academic research - the index is the single most-cited number in housing-market research, and its regional cut makes it the cleanest way to study local shocks.

Investors and quants get a forty-year monthly sentiment series with sub-component and regional cuts - enough to build a proper factor model rather than rely on the headline number. Homebuilding executives and pricing strategists get a peer benchmark that tells them when the rest of the industry is cutting prices and how deep the average reduction is. Market researchers and consultants get a citable, trade-association-grade series behind every housing-market presentation. Data scientists and ML engineers get a clean, single-scale index on a fixed monthly grid - perfect as a target or feature with no schema drift. Policy analysts and journalists get the household-name number that anchors every housing-policy debate, with regional granularity to localize the story. Developers building housing dashboards get a free-license benchmark series to anchor product surfaces against.

Three things worth knowing upfront. First, the composite is a diffusion index, not a level: the 50 line is the breakeven between 'more builders see good conditions' and 'more builders see poor conditions', and 1995, 2000, 2009, 2014 and 2022 each show very different paths to and from that line. Treat directional changes and component divergences as the signal, not the absolute level.

Second, regional prints are noisy. The Northeast, Midwest, South and West series carry fewer respondents than the national composite, so single-month moves can be sampling artifacts - the 3-month moving averages that ship with the regional tables are the version to use for trend work, and the raw prints are the version to use for month-of-event studies.

Third, the supporting hub goes deeper than the HMI itself. The same housing-economics section also publishes framing lumber price history, Census-derived starts and permits analysis, state and local indicators and forecasts, so once you have the index series in your warehouse the next natural joins are lumber cost and starts/permits, not a third sentiment gauge.