HUD’s 2015 fair housing rule tied local assessments to its data and tool, so withdrawing the tool delayed them
The rule published July 16, 2015 required jurisdictions and housing agencies to plan with HUD-provided data in an HUD-prescribed tool. When HUD withdrew the local government tool in May 2018, it said the rule extended local governments’ first-assessment deadlines.
Looking Back revisits past weeks one at a time, up to the most recent, and asks what it meant for the banks, health systems and public agencies we serve. The series was written in 2026, with the benefit of hindsight, and each article says which week it looks back on.
Two-minute summaryThe dek and the first sentence of each section, quoted from the article
The rule published July 16, 2015 required jurisdictions and housing agencies to plan with HUD-provided data in an HUD-prescribed tool. When HUD withdrew the local government tool in May 2018, it said the rule extended local governments’ first-assessment deadlines.
- The week The rule implemented a duty in the Fair Housing Act.
- The data: what HUD would provide and what participants had to add A third input was “local knowledge.”
- The tool: how the Assessment Tool shaped the analysis The rule required an assessment to be submitted “in a manner and form prescribed by HUD.” [1] Its content section opened: “Using the Assessment Tool prescribed by HUD, each program participant shall conduct an AFH” for the purpose of examining its programs, jurisdiction and region. [1] In our reading, a large share of the rule’s working detail lived in the tool, not the regulation.
- The review: how HUD would test an assessment against the data The rule limited what acceptance meant.
- The schedule: deadlines tied to the tool The rule staggered first assessments.
- How the rule reached different agencies Entitlement jurisdictions with consolidated plan cycles starting in 2017 were in the first group. [1] For a large city with its own planning and data staff, in our reading, the rule’s main demand was integration.
- Objections, and what the record shows Commenters said HUD data had been inaccurate and out of date in the past. [2] HUD’s answer was the local data and local knowledge channels, and qualitative explanation. [1] In our reading, the answer was sound in design and demanding in practice.
- Checks a program director or data lead could run These checks are our suggestions, not requirements of the 2015 rule or the current regulation.
- With hindsight The withdrawal notice listed seven categories of problems.
The week
On July 16, 2015, the Department of Housing and Urban Development published its Affirmatively Furthering Fair Housing final rule in the Federal Register, at 80 FR 42272. [1] The rule gave an effective date of August 17, 2015. [1] It amended HUD’s regulations in 24 CFR parts 5, 91, 92, 570, 574, 576 and 903. [1]
The rule implemented a duty in the Fair Housing Act. Section 808 requires executive agencies to administer housing and urban development programs in a manner that affirmatively furthers fair housing. [1] HUD’s program participants certify that they will do so. [1]
Before the rule, participants met that duty by preparing an analysis of impediments to fair housing choice. [1] The rule said that analysis was generally not submitted to or reviewed by HUD. [1] It said program participants, civil rights advocates, other stakeholders and the Government Accountability Office had all told HUD the approach was not as effective as originally envisioned. [1]
The final rule replaced the analysis with an Assessment of Fair Housing. [1] Participants would complete it using an Assessment Tool prescribed by HUD and data provided by HUD. [1] HUD would review each assessment and accept or not accept it. [1] Its goals would then feed the consolidated plan or the public housing agency plan. [1]
The rule reached two groups. The first was jurisdictions and insular areas that submit consolidated plans for the Community Development Block Grant, Emergency Solutions Grants, HOME and Housing Opportunities for Persons With AIDS programs. [1] The second was public housing agencies receiving assistance under sections 8 or 9 of the United States Housing Act of 1937. [1]
Data sat at the center of the design. HUD committed to provide states, local governments, public housing agencies and the public, to the fullest extent possible, with local and regional data. [1] The data would cover integrated and segregated living patterns, racially or ethnically concentrated areas of poverty, the location of certain publicly supported housing, access to opportunity and disproportionate housing needs. [1]
The rule said it does not mandate specific outcomes for the planning process. [1] It estimated compliance costs to program participants of $25 million annually, and resource costs to HUD of $9 million annually. [1] It said HUD received over 1,000 public comments, including duplicate mass mailings, resulting in approximately 885 unique public submissions. [1]
The same issue of the Federal Register carried a second HUD document. It was a 30-day notice under the Paperwork Reduction Act seeking comment on the Assessment Tool, with comments due August 17, 2015. [2]
The notice described a Geospatial Mapping Tool, which HUD called the Data Tool, with interactive maps and exportable tables. [2] It said the Data Tool contained the same data released on September 26, 2014, with some minor changes, now accessible through an interactive web-based interface. [2] It listed further changes HUD anticipated before the tool’s final release. [2]
The notice also recorded commenters’ concerns about the data. It said some commenters stated that HUD data had in the past been inaccurate and out of date, and that the HUD-provided data was unwieldy and difficult to understand. [2]
It estimated the burden. The template was primarily designed for about 1,194 entitlement jurisdictions, with an estimated 240 hours each, and for public housing agencies joining them. [2] The table listed approximately 1,314 public housing agencies at 120 hours each, and a total of 444,240 burden hours. [2]
For a city’s planning department, a county’s community development office or a housing authority, the week’s two documents set out a new way of working. The rest of this piece is about what it asked of their data and their systems.
HUD’s 2015 fair housing rule is remembered for its politics and for its reversals. For a city’s chief data officer, a county planning director or a housing authority’s policy lead, its design deserves attention for a different reason. It built a federal planning duty on federal data, delivered through a federal tool.
Our argument is that the rule made HUD’s data and its Assessment Tool the working parts of every local plan. Participants had to use HUD-provided data, supplement it with local data and knowledge, and answer the questions in a tool HUD prescribed. [1] HUD’s review then tested whether goals were consistent with the data. [1] In our reading, that design made the plans only as dependable as the federal data and tool behind them. The 2018 withdrawal of the local government tool showed what happened when the tool failed. [4]
That framing matters to anyone who builds a planning or reporting process on a shared federal dataset. It also explains the rule’s later history, which turned on the tool as much as on policy. [3] [4] [5]
We take the rule in four parts:
- The data: what HUD would provide, and what participants had to add.
- The tool: how the Assessment Tool shaped the analysis.
- The review: how HUD would test an assessment against the data.
- The schedule: how deadlines were tied to the tool’s availability.
The parts do not overlap. The first is the evidence, the second the form, the third the check and the fourth the timing. After them we look at how the rule reached different kinds of agencies, take the objections in order and end with checks a program director or data lead could run.
The data: what HUD would provide and what participants had to add
Two kinds of data
The final rule added definitions that did not appear in the proposal. [1] “HUD-provided data” meant HUD-provided metrics, statistics and other quantified information required to be used with the Assessment Tool. [1] The rule said that data would be provided to participants and posted on HUD’s website for the public. [1]
“Local data” meant metrics, statistics and other quantified information relevant to the participant’s areas of analysis. [1] To count, it had to be found through a reasonable amount of search, be readily available at little or no cost and be necessary to complete the assessment. [1] It was also subject to a determination of statistical validity by HUD. [1]
A third input was “local knowledge.” The rule defined it as information the participant provides about its areas of analysis that is relevant, known or becomes known to it, and necessary to complete the assessment. [1]
The required use
The regulatory text was direct. Participants “will use HUD-provided data,” and supplement it as needed with local data and local knowledge, as guided by the Assessment Tool. [1] The analysis had to address segregation and integration, racially or ethnically concentrated areas of poverty, disparities in access to opportunity and disproportionate housing needs. [1]
In our reading, that made HUD’s dataset the baseline for every plan in scope. A participant could add to it. It could not skip it. The same national data would anchor a large city’s assessment and a small entitlement community’s.
Statistical validity as a gate
The “statistical validity” phrase drew questions. HUD explained that it meant HUD may decline to accept local data it has determined is not valid. [1] HUD said it did not mean it would apply a rigorous statistical validity test to all local data. [1]
That answer left the standard unwritten. In our reading, a local agency adding its own data took on a documentation task. It needed to be able to explain where a local dataset came from, how it was built and why it was reliable enough for HUD to accept.
Public data, public scrutiny
The rule made the baseline data public. HUD-provided data would be posted on HUD’s website for availability to all of the public. [1] Commenters asked HUD to make the data tool available to the public, and to require participants to post the data they used before public hearings. [1]
The community participation section pointed the same way. It said participants should use communications designed to reach the broadest audience, which may include making copies of each document available on the Internet and the participant’s official website. [1] The assessment had to summarize public comments, including those not accepted and the reasons. [1]
In our reading, public data changed the audit trail. An advocate, a developer or a neighboring jurisdiction could open the same HUD maps and compare them with the assessment. A finding that did not match the public data would need an explanation on the record. That raised the value of documenting every departure from HUD’s figures.
What HUD said about the data’s limits
HUD was candid about the limits of its approach. It said the rule and the Assessment Tool rely heavily on census tracts to identify areas of concentration and opportunity. [1] It said there are well-known limitations to census tracts among researchers. [1]
HUD gave examples. A tract with relatively high poverty may sit within a larger area experiencing significant economic improvement. [1] Tracts are not always synonymous with neighborhoods as understood locally, and their relevance varies between central cities and rural areas. [1] HUD said participants should take nearby tracts into account and should expand on the data through qualitative discussion. [1]
Commenters raised other limits. They said the American Community Survey has high margins of error at the tract level. [1] HUD agreed that a single numeric threshold would not work, revised the definition of disproportionate housing needs and removed the proposed 10 percent threshold. [1] In our reading, those exchanges showed that the data would need local interpretation everywhere it was used.
What commenters wanted added
Commenters asked HUD to provide data as raw files or tables, to let groups upload additional data and to add datasets such as mortgage lending data, foreclosures and fair housing complaints. [1] Others said the data HUD required was excessive and duplicated the consolidated plan. [1]
HUD said the data was not intended to be exhaustive. It described the data as a baseline that participants should supplement. [1] It said it would refine the data tool based on ongoing public input and research, and add data as more nationally uniform sources became available. [1]
A hypothetical county
Consider, as an illustration, an urban county preparing its first assessment. HUD’s tool flags a cluster of census tracts as a racially or ethnically concentrated area of poverty. The county’s own records show a new transit line and new housing under construction in two of those tracts.
Under the 2015 rule, the county would use HUD’s data, add its local data, and explain the difference. [1] HUD had said a tract with high poverty may sit within an improving area, and that participants should look at nearby tracts. [1] The county’s explanation would be its own analysis, kept as a record and open to public comment. [1]
The illustration shows the skill the rule assumed. Someone had to read federal data critically, set it beside local records and write down the reasoning. In our reading, that was a data analyst’s job as much as a planner’s.
Records that had to be kept
The rule required each participant to keep records sufficient for HUD to determine compliance. [1] For an assessment, those records included statistical data, studies and other diagnostic tools used, and any documents relating to the analysis. [1] In our reading, that made the data a record to preserve: the version of HUD’s data used, the local datasets added and the reasoning applied to both.
The tool: how the Assessment Tool shaped the analysis
A prescribed form
The rule required an assessment to be submitted “in a manner and form prescribed by HUD.” [1] Its content section opened: “Using the Assessment Tool prescribed by HUD, each program participant shall conduct an AFH” for the purpose of examining its programs, jurisdiction and region. [1] In our reading, a large share of the rule’s working detail lived in the tool, not the regulation.
HUD defended that choice. In response to commenters who wanted more detail in the regulatory text, HUD said “many of these items are better addressed in the Assessment Tool and as guidance.” [1] It said that would allow refinements on a timelier basis in response to public input and experience. [1]
In our reading, that choice moved the operative requirements into an information collection that could change on its own schedule. The regulation set the duty. The tool set what a participant actually filled in.
The Paperwork Reduction Act path
Because the tool was an information collection, it went through the Paperwork Reduction Act. The 2015 rule recounted that HUD had published a proposed tool for 60 days of comment in September 2014. [1] The notice published the same day as the rule opened the 30-day comment period. [2]
The rule said HUD had decided that the state and public housing agency tools would each go through the full 60-day and 30-day process. [1] In our reading, that meant separate tools, each with its own approval and its own clock. A participant’s obligations depended on which tool applied to it and when that tool was approved.
Community participation inside the tool
The tool also had to carry the public’s voice. Participants had to give the public reasonable opportunities to be involved in developing the assessment. [1] Consolidated plan participants would follow their citizen participation plans, and public housing agencies their own plan procedures. [1]
The assessment had to include a concise summary of the participation process, the comments received and efforts to broaden participation. [1] In our reading, that made the tool a meeting point for two kinds of evidence. Quantitative data came from HUD and local systems. Qualitative evidence came from residents, and had to be summarized beside it.
The Data Tool
The notice described the mapping application in some detail. It contained interactive maps and exportable tables, and provided data for the region based on the participant’s Core Based Statistical Area. [2] HUD said it would be posted online. [2] It said one table now included two transit-related indices. [2]
HUD listed changes it expected before final release. [2] They included consolidating redundant tables, improving the contrast and size of dots and icons on maps, and better access to data on publicly supported housing. [2] They also included the ability to export maps and tables for use in community participation. [2]
That list described software still in development. As we read it, participants were being asked to plan around an application whose features were changing, and whose outputs would become evidence in their submissions.
The region as a unit of analysis
The rule required analysis “within the jurisdiction and region” for each of its four core questions. [1] The Data Tool supplied regional data based on the participant’s Core Based Statistical Area. [2] HUD said the regional framing responded to commenters’ concerns about the area of analysis. [2]
On our reading, regional analysis was where federal data was most useful and local data hardest to match. A city could supply its own records. It could not easily supply comparable records for every neighboring jurisdiction. The federal dataset was the only consistent view of the region a participant was likely to have, which made its accuracy matter more.
Burden as HUD estimated it
The notice estimated 240 hours for each entitlement jurisdiction and 120 hours for each public housing agency. [2] It said those estimates were not net new costs, because the analysis of impediments the assessment replaced had its own costs. [2] It said costs for that analysis “varied substantially and often involved costs for hiring consultants and outside parties.” [2]
The rule recalled what HUD had said in its September 2014 notice on the tool: “With the data that HUD provides for use with the Assessment Tool supplemented by available local data and local knowledge, HUD does not anticipate the need for any program participant to turn to outside consultants to collect data and conduct the assessment.” [1] It then acknowledged that, in some cases, participants may hire consultants, as they had for the analysis of impediments, and said HUD believed most would not require them. [1] That expectation rested on the data and the tool working as intended.
As we read it, the hours estimate assumed the tool and data worked as designed. Every hour spent reconciling a map with local records, or explaining a tract that did not match a neighborhood, came on top.
What the tool asked of local systems
Participants needed to export HUD’s maps and tables into their own documents, and to place local data beside them. [2] They needed to publish drafts for community participation. [1] They needed to keep the materials as records. [1]
That combination called for ordinary but disciplined data management. A version-controlled copy of what HUD’s tool showed on the day it was used. A register of local datasets and their sources. A link from each finding in the assessment to the data behind it.
The review: how HUD would test an assessment against the data
Acceptance, not approval
HUD would review each assessment to decide whether the participant had met the requirements for analysis, assessment and goal setting. [1] In the rule’s words, an assessment “will be deemed accepted after 60 calendar days after the date that HUD receives the AFH,” unless HUD has said by then that it does not accept it. [1]
The rule limited what acceptance meant. It meant only that, for purposes of HUD program funding, the assessment met the required elements. [1] It did not mean the participant had complied with the Fair Housing Act or other civil rights laws. [1] HUD said it chose acceptance over approval to avoid creating the perception of a safe harbor. [1]
What acceptance did not prove
One commenter argued that deemed acceptance would create the impression of a safe harbor for jurisdictions that may be violating the Fair Housing Act. [1] The commenter asked HUD to replace it with an audit-type review. [1] HUD kept deemed acceptance and limited what acceptance meant. [1]
For an agency, that limit had a practical consequence. An accepted assessment showed HUD had found the required elements present, for funding purposes. [1] It did not settle any claim that the agency had failed its fair housing obligations. [1] In our reading, the agency’s own records of data, reasoning and participation remained its evidence, whatever HUD decided. A plan accepted by default, without substantive review, added nothing to that evidence.
Data as a ground for nonacceptance
The rule set out when HUD would not accept an assessment. One ground was that it was substantially incomplete. [1] That included failing to satisfy a required element. [1]
The rule spelled out one form of failure. It included an assessment in which priorities or goals were “materially inconsistent with the data or other evidence available to the program participant.” [1] In our reading, that sentence made the data the measuring stick for the plan. A goal that did not follow from the data could be a reason to send the assessment back.
Funding consequences
The review had teeth. Without an accepted assessment, HUD would disapprove a consolidated plan or a public housing agency plan, except where delayed submission was permitted. [1] The rule noted that failure to submit a consolidated plan by August 16 of the federal fiscal year results automatically in the loss of the CDBG funds the jurisdiction would otherwise receive. [1]
For a nonaccepted assessment, the rule set a resubmission period of “no less than 45 calendar days” from HUD’s written notice. [1] A revised assessment would be deemed accepted after 30 calendar days from HUD’s receipt, unless HUD said otherwise. [1]
Goals with metrics and milestones
The rule asked for more than findings. Participants had to identify significant contributing factors, prioritize them and justify the prioritization. [1] They had to give highest priority to factors that limit or deny fair housing choice or access to opportunity, or harm fair housing or civil rights compliance. [1]
Each goal had to name the contributing factors it addressed and how it related to them. [1] It also had to identify “the metrics and milestones for determining what fair housing results will be achieved.” [1] The rule gave an example: where segregation was a fair housing issue with a significant contributing factor, HUD would expect goals to reduce the segregation. [1]
In our reading, that made the assessment a measurement plan. A goal needed a baseline, a measure and a milestone. The baseline came from the data, and later progress would be reported against it in the next assessment. [1] That is where planning, data and performance management met. HUD’s January 2018 notice later said participants struggled with exactly this step. [3]
Revising for material change
An accepted assessment had to be revised after a material change. [1] The rule’s examples included presidentially declared disasters, significant demographic changes, new significant contributing factors, and civil rights findings or court orders. [1] In our reading, that kept the data question open after acceptance. A participant needed to notice when its facts had changed enough to require a revision.
What the review asked of HUD
In our reading, the review design put a large workload on HUD’s own staff. Each assessment had to be decided within the review period, or it would be deemed accepted. [1] HUD said it believed the time allotted, coupled with staggered deadlines, was sufficient. [1] The 2018 notices later described the staff resources nonacceptance decisions required. [3]
The schedule: deadlines tied to the tool
Staggered first submissions
The rule staggered first assessments. For consolidated plan participants not given a later date, the first assessment was due “no later than 270 calendar days prior to the start of” a program year. [1] That was the program year beginning on or after January 1, 2017, for which a new consolidated plan is due. [1] Participants with a fiscal year 2015 CDBG grant of $500,000 or less, and states and insular areas, would follow the program year beginning on or after January 1, 2018. [1]
Public housing agencies other than qualified agencies would follow their fiscal year beginning on or after January 1, 2018. [1] Qualified agencies, a defined group of smaller agencies, would follow the fiscal year beginning on or after January 1, 2019. [1] Later assessments were due, in the rule’s words, “195 calendar days before the start of the first year of the next 3 to 5-year cycle,” and no less frequently than once every five years, or at a time HUD and the participant agreed in writing. [1]
The tool as a trigger
One provision tied the clock to the tool. If a deadline would fall less than nine months from publication of the tool applicable to a participant, the deadline would be extended as specified in that tool’s publication. [1] The rule set the floor in its own terms: “a date that will not be less than 9 months from the date of publication of the Assessment Tool.” [1]
In our reading, that provision made each participant’s deadline depend on HUD’s own delivery. No tool, no clock. That protected participants from a tool published late. It also meant that withdrawing a tool would push the deadline back until a revised tool was published, which is what HUD’s May 2018 notice said had happened to local governments. [4]
Clocks HUD could move
The schedule was HUD’s to adjust. HUD’s January 2018 notice recounted an earlier notice, of October 24, 2016, for small CDBG grantees. [3] It moved them from the program year beginning on or after January 1, 2018 to the one beginning on or after January 1, 2019. [3] The January 2018 notice then moved all local government participants to their next submission date falling after October 31, 2020. [3]
It explained the date. The rule required at least nine months between an approved tool and a submission, and HUD expected to have a renewed tool approved by January 31, 2020. [3] In our reading, the deadline calculation ran through HUD’s own paperwork schedule. A local agency’s planning calendar depended on a federal approval it could not see or influence.
The analysis of impediments in the meantime
Until a participant was required to submit an assessment, it had to keep conducting the analysis of impediments under the requirements in effect before August 17, 2015. [1] That kept two regimes running at once. An agency could be maintaining the old analysis while preparing data for the new one.
Regional collaboration and its timing
The rule encouraged joint and regional assessments. [1] Collaborators within the same Core Based Statistical Area could cross state lines; others had to ask HUD to approve the collaboration. [1] A joint or regional assessment followed the lead entity’s deadline. [1] All collaborating participants were accountable for the analysis and joint goals, and each had to sign. [1]
Regional work multiplied the data question. Several agencies would rely on the same HUD data and the same tool output, and each would sign for the result.
How the rule reached different agencies
Large cities and urban counties
Entitlement jurisdictions with consolidated plan cycles starting in 2017 were in the first group. [1] For a large city with its own planning and data staff, in our reading, the rule’s main demand was integration. HUD’s maps had to sit beside local data on zoning, schools, transit and housing, and the assessment had to explain where they agreed and where they did not.
Smaller entitlement communities
Participants with CDBG grants of $500,000 or less received a later first deadline. [1] The notice said the cost of the earlier analysis of impediments varied substantially. [2] In our reading, smaller jurisdictions would depend more heavily on HUD’s data and tool, because they had the least capacity to supplement them.
States and insular areas
States followed a separate tool and a later deadline. [1] Commenters said state and rural participants needed a different level of data and analysis than metropolitan areas. [1] HUD said the level of analysis for different participants was best addressed in the tool, the data tool and guidance. [1]
Public housing agencies
Public housing agencies were in scope through their PHA plans. [1] The rule let an agency participate with a consolidated plan participant, participate with other agencies, or prepare its own assessment. [1] It encouraged memorandums of understanding to define each party’s role when agencies worked together. [1] The notice said HUD estimates that approximately a third of agencies may seek to join an entitlement jurisdiction. [2] For an agency preparing its own, the data about publicly supported housing and its residents was central.
Regional councils and consortia
Joint and regional collaborations would share one assessment and one lead entity. [1] The rule required participants intending to collaborate to notify HUD and give it a copy of their written agreement. [1] In our reading, the agreement was also a data governance document. It needed to say who held the data, who ran the tool and who kept the records.
Consultants and data vendors
The 2015 rule said HUD believed most participants would not require consultants, while acknowledging that some might hire them. [1] The 2018 withdrawal notice said many jurisdictions found it necessary to incur additional expense to hire consultants. [4] We read the gap as a procurement signal. Where a federal tool is hard to use, local agencies buy expertise to use it, and the cost moves from the federal estimate to local budgets.
A contract for that work, in our reading, needed to cover more than the report. It needed to leave the agency with the data, the tool outputs and the working files the recordkeeping section required. [1]
Local technology and data teams
The rule did not mention information technology departments. In our reading, it still reached them. Exporting maps and tables, publishing drafts online in accessible formats and keeping records of data used all ran through local systems. [1] [2]
The participation rules required compliance with section 504 of the Rehabilitation Act and the Americans with Disabilities Act where they applied. [1] A map-heavy assessment posted online, in our reading, had to be usable by people who could not see the maps. That was a publishing task, and it fell to whoever ran the jurisdiction’s website.
HUD itself
The rule also made HUD a data provider, a software publisher and a reviewer at once. It committed HUD to provide data to the fullest extent possible. [1] It committed HUD to review each assessment within a fixed period. [1] Those commitments became the points where the system later strained. [3] [4]
Objections, and what the record shows
“HUD’s data will be wrong for our community”
Commenters said HUD data had been inaccurate and out of date in the past. [2] HUD’s answer was the local data and local knowledge channels, and qualitative explanation. [1] In our reading, the answer was sound in design and demanding in practice. It worked only if participants had the capacity to produce and document local data.
“The tool is unfinished”
Commenters said releasing the tool was premature. [2] HUD published it for comment anyway, and listed changes still to come. [2] The 2018 withdrawal notice later called the local government tool inadequate for its purpose. [4] The objection, in our reading, proved closer to the mark than HUD allowed in 2015.
“Federal data will dictate local outcomes”
The rule said it did not mandate specific outcomes. [1] It also said goals materially inconsistent with the data could lead to nonacceptance. [1] In our reading, both statements were true at once. The rule did not set outcomes, but it did set the evidence that outcomes had to answer to.
“The burden is too high”
The notice estimated 444,240 burden hours across participants. [2] The rule estimated $25 million a year in participant costs. [1] The 2018 withdrawal notice said “many jurisdictions found it necessary to incur additional expense to hire consultants,” and called the tool unduly burdensome. [4] The 2015 rule had expected most participants not to need them. [1] In our reading, the objection held up.
“Local knowledge cannot be standardized”
The rule allowed local knowledge, defined broadly, to enter the analysis. [1] It also required HUD to judge whether goals were consistent with the data and other evidence. [1] In our reading, the tension was real. HUD needed consistency to review thousands of plans, and participants needed room to explain local conditions. The rule resolved it with written explanation, which is slow to produce and slow to review.
“HUD cannot review thousands of plans in time”
Commenters asked HUD to approve assessments rather than accept them, and objected to the deemed-acceptance provision. [1] HUD said the review period, with staggered deadlines, was sufficient. [1] The January 2018 notice later said nonacceptance decisions required significant staff resources, and that HUD would enhance its technical assistance. [3] We read the record as partly vindicating the objection. Review capacity, as well as participant capacity, shaped the rule’s pace.
“Census tracts are the wrong unit”
HUD conceded the limits of tracts and asked participants to interpret them with care. [1] It allowed block groups at a participant’s discretion, with more caveats. [1] In our reading, the unit question was never fully resolved. It was handed to participants, who had to explain it in their own words.
Checks a program director or data lead could run
These checks are our suggestions, not requirements of the 2015 rule or the current regulation. Agencies should confirm their fair housing obligations with counsel, especially since the regulation has changed several times.
- Federal dataset inventory: list every federal dataset or tool your planning or reporting relies on, and who owns the relationship with the provider.
- Snapshot discipline: keep a dated copy of any federal data or tool output used in a submission.
- Local data register: record the source, method and reliability of each local dataset added to federal data.
- Traceable findings: link each finding and goal to the data behind it, so a reviewer can follow the reasoning.
- Fallback plan: decide what you would do if a federal tool or dataset were withdrawn mid-cycle.
- Shared agreements: in regional work, write down who holds the data, who runs the tools and who keeps the records.
- Change triggers: define the events that would make a submitted plan out of date.
- Publication: plan how drafts and data will be published for community participation in accessible formats.
The 2015 rule put federal data and a federal tool at the center of local fair housing plans. [1] The checks above follow that line: know which federal inputs you depend on, keep what they showed, and be ready when they change.
With hindsight
Between October 2016 and December 2017, HUD received, reviewed and issued initial decisions on 49 assessments from local governments, according to its later notice. [4] A notice published January 5, 2018 extended the deadline for local government participants to their next submission date falling after October 31, 2020. [3] It said HUD had found, among the first 49 assessments that received a decision, that more than a third (35%) were initially non-accepted. [3]
The notice said participants struggled to set goals that could be expected to result in meaningful actions, and to develop metrics and milestones. [3] It said nonacceptance decisions required significant staff resources. [3] It noted that each of the 19 participants in the largest regional assessment met the standard for nonacceptance. [3]
On May 23, 2018, HUD withdrew the Local Government Assessment Tool. [4] The notice said HUD had become aware of significant deficiencies in the tool, and that it was inadequate to accomplish its purpose. [4] It said only 37% of the initial 49 submissions were acceptable on first submission, 35% were returned as unacceptable and 28% were accepted only after revisions. [4]
The withdrawal notice listed seven categories of problems. One was insufficient use of local data and knowledge. [4] It gave an example: one participant did not identify Superfund sites in its jurisdiction, which HUD maps did not show. [4] Another was a lack of regional analysis, although HUD provided regional data through its data and mapping tool. [4]
The notice applied the rule’s nine-month floor. With no approved local government tool, the first-assessment deadline was “extended to a date not less than 9 months following the future publication” of a revised and approved tool. [4] It said the public housing agency tool had been approved in January 2017 but was not yet available, “because the HUD data needed to make the Assessment Tool workable was not yet available.” [4] Their deadline was extended until a workable tool became available. [4]
HUD’s 2025 interim final rule later recounted that HUD published three notices on May 23, 2018. One withdrew the local government tool, a second directed participants that had not completed an assessment to continue conducting an analysis of impediments, and a third withdrew the January 2018 notice extending local governments’ deadlines. [5]
The rule itself then changed several times. HUD’s 2025 interim final rule recounts the history. A 2020 final rule repealed the 2015 rule; a 2021 interim final rule restored some definitions and certifications but not the assessment; a 2023 proposed rule was withdrawn on January 16, 2025. [5]
The 2025 interim final rule, published March 3, 2025 with an effective date of April 2, 2025, revised the regulation again. [5] It said a participant’s certification will be deemed sufficient if it took any action during the relevant period rationally related to promoting fair housing. [5] It does not reinstate the analysis of impediments or mandate any specific planning mechanism. [5]
As checked on November 16, 2026, the eCFR text of 24 CFR 5.150 and 5.151, up to date as of November 12, 2026, matches that approach and cites the 2025 rule. [6] The 2015 rule’s sections on the assessment, such as §§ 5.152 and 5.154, are marked reserved in the current text, which lists §§ 5.152–5.180 as reserved. [6]
What held up from 2015 is the diagnosis of dependence. HUD’s May 2018 notice attributed the high rate of unacceptable assessments, in large measure, to problems with its own tool, and its housing agency tool had been held back because the HUD data needed to make it workable was not available. [4] What did not hold up is the regime itself. [5]
In our view, the lasting point for agency data leaders is about inherited tools. A plan built on a provider’s dataset and form is exposed to that provider’s schedule, and needs its own record of what it used.
Spotted an error? Tell us and we will check it and correct the article where needed.