What data source is HLC using to track total housing production?
HLC tabulated all metrics presented on this page from the Census Address Count Listing Files, a Census Bureau data product containing the number of residential addresses in each census block. This data is the best currently available source of information on housing deliveries. When a new home is created, local officials assign an address, which is then reported to the US Postal Service. The USPS adds the address to their inventory and shares this data with the Census Bureau twice per year. Using aerial photos, commercial data sources, and manual review, the Census Bureau assigns each address to a census block. The Census Bureau counts the total in each census block and subtracts demolitions and other loss. Starting in 2023, the Census Bureau began publishing this data each year. It is now the best currently available source of information about the net change in the housing stock. Documentation about the Census Bureau's address maintenance strategy can be found on the Geographic Support System webpage. More information on the benefits of the Address Count Listing files are detailed in a Niskanen Center report, New census data: A true picture of America's housing stock.
How is this data different from building permit numbers?
Until now, the best source of information about housing construction was the Census Building Permit Survey, which asks building inspectors to report on the number of new homes issued a building permit. That dataset provides a general picture of housing activity in a community, but it has shortcomings as described in a recent Boston Indicators analysis, The Surprising Lack of Good Permitting Data. Units appear in the data according to the month the permit was issued, even if it may be years before those units are ready for occupancy. The questions cover only ground-up construction, so inspectors do not report on commercial conversions, adaptive reuse, or internal ADUs. Nearly one-quarter of municipalities do not respond to the survey. As a result, the Building Permit Survey substantially underestimates housing change. In the decade after 2010, building permits underestimated actual housing growth by 37%. HLC compared address counts to Building Permit Survey data and found that the more than 34,000 units added in 2025 substantially exceeded the average number of permits reported during the prior four years by a factor of two. This finding reinforces concerns about using permits as a core housing metric, when they underestimate completions by such a huge margin.
What did HLC do to check the accuracy of the data?
HLC collected every version of the Address Counts (Decennial Census, July 2023, November 2023, July 2024, November 2024, July 2025, and November 2025) and calculated the change in housing units over each time period. We then identified census blocks in each region that had the greatest increase in units. These blocks were individually spot-checked by HLC staff and cross referenced with aerial photos and other sources to verify the unit count and timing of new housing developments. This review found a high level of accuracy with regard to unit count, though a small minority of developments may have been completed before or after they were added to the inventory. HLC staff also reviewed blocks with large decreases in housing units and found those were either attributable to a geographic correction or a reclassification from apartment to dormitory. HLC also consulted with staff in the state's largest municipalities to compare this information to the number of Certificates of Occupancy issued in 2025. None of these investigations identified major concerns or apparent sources of systematic error in the data.
Are you certain all these homes became ready for occupancy in 2025?
While the Address Count Listings provide an improved estimate of when homes enter the market, it's not yet perfect. The data measures the change in residential address counts between November 2024 and November 2025. It is not a precise estimate of new homes ready for occupancy during that time period. The timing of when addresses are created and how soon they reach the postal service may vary slightly from place to place or project to project. It is possible that some units added to the Listings last year were not actually available until after January 1, 2026, or that some of the units added to the Address Count Listings in 2025 were actually completed in 2024. In fact, this is a known issue in the City of Boston where the Address Count Listings include some units that received Certificates of Occupancy in 2024. HLC determined that the data is appropriate for overall progress measurement on multiyear goals, but year-over-year comparison should be conducted with caution. The relative impact of such timing issues will diminish as more and more years are collected.
What kinds of homes are being built? How many are affordable?
The Address Count Listings don't include detail on the exact location, type, size, or cost of new housing units, only the net change. More information is needed to understand progress toward the goals of A Home for Everyone. HLC is working to develop new data resources about various aspects of housing production, including an ADU Tracker. Forthcoming HLC data products about subsidized housing production and the MBTA Communities pipeline will help to provide a more well-rounded picture of new growth.
Some regions are well ahead of pace to meet their 10-year goal. Does that mean their crisis is over?
While some regions are projected to have little net increase in housing demand, they still have significant housing needs. Every region needs more affordable homes, more accessible homes, more sustainable homes, and more options for seniors. These goals can be accomplished through both new construction as well as smarter use of the existing housing stock. HLC is working in every region to help address the housing crisis from every angle, even in places that don't have lots of new demand.
Does this change anything related to 40B or the SHI?
No, this information does not affect the applicability or implementation of Chapter 40B, the state's Comprehensive Permit Law. Subsidized Housing Inventory calculations will continue to use the 2020 count of year round housing units, as required by statute. For more information about your city or town's SHI statistics, visit HLC's Subsidized Housing Inventory webpage.
Can I view the census block-level data somewhere?
Yes, the Census Bureau has published two map viewers: one showing detailed block-level data as of November 2025, and one with the total housing unit change from 2020 to 2025. Please note, the Housing Unit Change viewer covers a longer period than the Tracking Progress analysis of 2025 additions. Users should also be aware that the net change for each block includes geographic corrections (e.g., address moved from one block to another) or other data corrections (e.g., building reclassified from an apartment to a dorm) which may produce unexpected results.
What other caveats should I be aware of when using this information?
This data does not measure the addition of new homes to the pipeline. It is therefore a lagging indicator of the development environment and should not be used to draw conclusions about current or future development activity. Instead, it indicates homes that are currently entering the housing market, where residents can rent or purchase them and where they can help achieve housing abundance by 2035.
In spot-checking, there were a few edge cases in which one block reported a large increase and the adjacent block reported a large decrease, consistent with small changes in address geocoding for these projects. HLC therefore recommends caution in interpreting the block-level changes, but this issue poses no obstacles for interpretation at the municipal level and above, and it is still appropriate to use block-level data with appropriate quality assurance.
HLC is aware of a small number of blocks where the residential address count decreased in 2021 and 2022 vintages of the Address Count Listing Files due to the redesignation of housing units as Group Quarters for Census purposes. To HLC's knowledge, no major changes of this variety occurred in 2025, but HLC urges caution in interpreting negative results from early in the decade.
As with any new data product, the research community is exploring potential applications and limitations of the data. This list of limitations is not exhaustive and is subject to change.