The 2027 designations: who got picked, and why

The short version

1. Incumbency was the strongest predictor. Eligible tracts that had been OZ 1.0 zones were designated 42.6% of the time; tracts that had not were designated 19.2% of the time. Within states, former OZ tracts were picked 1.67× as often as each state's overall rate would predict, and the effect holds after controlling for everything else measured here (odds ratio 2.72). 45.2% of the new map is old map.
2. A deal track record came next. Tracts with a prior New Markets Tax Credit investment were designated 1.62× as often as expected. Together with incumbency, this points to states nominating where projects, sponsors and local capacity already existed.
3. Rural tracts were favored — modestly. Rural tracts were 33% of the eligible pool and 39.4% of designations (1.17× expected). The QROF rural bonus shifted selection at the margin, not wholesale.
4. Distress did not drive selection. Designation rates are flat across the EIG Distressed Communities Index: designated tracts sat at the 50.3th within-state percentile of distress, where 50 means no preference. CRA eligibility (lift 1.01) and persistent-poverty counties (0.91) made no difference either. Poverty rate tilted selection slightly upward (54th percentile).
5. States preferred mid-density, renter-heavy tracts over both the densest urban cores and the emptiest rural ones, and leaned toward tracts the Urban Institute rated likely to attract investment with larger impact (Tier 1, lift 1.28) over its low-probability Tier 3 (0.86).
6. Most of the choice is not explained by tract data. Everything above, combined in one model with state fixed effects, explains little of which tracts were picked (pseudo-R² 0.089). Local advocacy, project pipelines, political geography and county-level spreading — none of which the census measures — did most of the work.

Which traits raised the odds

Designations among eligible tracts with each trait, divided by the number expected if every state had picked at its own overall rate. 1.0 means no preference. Within-state comparison across 50 jurisdictions; the six that designated every eligible tract (VT, WY, AS, GU, MP, VI) are excluded because no selection took place.

The shape of the preference

Each eligible tract is ranked against the other eligible tracts in its own state, then grouped into tenths. Each panel shows the share of each tenth that was designated. A flat line at 26% means the trait made no difference.

Income and demographics

On income, the designated and passed-over tracts look much alike. Median family income relative to the area median shows a shallow U: the very lowest-income tracts and the near-threshold tracts were both picked a little more often than the middle. On education, unemployment and home values, there is essentially no difference.

On race and ethnicity, designated tracts had on average a somewhat smaller share of residents of color than the eligible tracts passed over in the same state (within-state percentile 48.1). Part of that comes from the rural tilt, since rural eligible tracts are whiter on average; but the association remains after controlling for rural status, density, income and the other traits (-2.7 percentage points per standard deviation). Neither the share of Black residents nor the share of Hispanic residents shows a consistent pattern on its own. This describes the selected set, not anyone's intent.

Medians pool all states and are for orientation; the percentile column is the within-state comparison.

All traits at once

A logistic regression of designation on every trait together, with a fixed effect for each state (so each tract is compared only with its own state's pool) and standard errors clustered by county. Continuous traits are scaled per standard deviation. The effect column is the average change in the probability of designation, in percentage points. n = 22,904.

The OZ 1.0 handoff

The 8,764 OZ 1.0 zones were drawn on 2010 census tracts. Mapped onto 2020 tracts, they cover 9,950 tracts. 6,882 of those are still low-income enough to be eligible; 2,974 were redesignated. That leaves 6,976 former zones — 70% — that will not be OZs after December 31, 2028, and that under Notice 2026-40 generally cannot take new capital after 2026 except under a working capital safe harbor plan adopted by December 31, 2026.

Nobody is documenting what happens to projects in those tracts. Each state page lists them by county. Transition rules →

The late additions

Treasury's October eligible list has 130 more tracts than the April Rev. Proc. 2026-14 appendix, and drops none. All 130 qualify under its missing-data rules — tracts without a usable ACS income or poverty figure that were shown eligible with supplemental data (new occupied housing, an income-distribution upper bound, or temporary disaster displacement). 15 of them were designated, 11.5% against 25.6% for everything else. Every designated tract is on Treasury's eligible list; there are no designations outside it. Background on off-list nominations →

What this means for the next ten years

• Expect OZ 2.0 capital to follow OZ 1.0's paths. States picked where deals had already happened. That makes early OZ 2.0 investment likelier — and makes it likelier the program repeats OZ 1.0's concentration, where 42% of capital went to 1% of tracts.
• The most distressed tracts are no better represented than the rest. Community-benefit terms, CDFI participation and public-sector gap financing will decide whether OZ 2.0 reaches them, not the map.
• For the 2036 round, the lesson is about pipelines. The tracts that won had a project history. A jurisdiction that wants designation next decade should start building a deal record now.

State by state

Every state designated its full allocation — 25% of its eligible tracts, or 25 tracts where that is larger. The columns compare the designated set with the state's full eligible pool. Click a state for its tract list, county table and the partners already active in its designated tracts.

Data and method

Download every eligible tract with its designation status and indicators (CSV).

• Designations and eligibility: Treasury Office of Tax Policy, 2027 QOZ designations and eligible LICs files, dated October 30, 2026. Poverty and income ratios are Treasury's own (ACS 2020–2024).
• Demographics: ACS 2020–2024 5-year Summary File (tables B01003, B03002, B15003, B23025, B25002, B25003, B25070, B25077); land area from the 2024 Census Gazetteer.
• OZ 1.0: IRS Notice 2018-48 and 2019-42, crosswalked to 2020 tracts with the Census 2020↔2010 tract relationship file; a 2020 tract counts as former OZ 1.0 if at least half its land area was in a 2018 zone. Connecticut's 2022 planning-region GEOIDs are remapped.
• Overlays: EIG Distressed Communities Index, FFIEC CRA income levels, Urban Institute investability tiers, CDFI Fund NMTC data, USDA ERS persistent poverty — all as used elsewhere on this site. These were joined to the April eligible list, so the 130 October additions have no overlay values.
• Scripts: ingest_designated_tracts.py, ingest_acs_tract_demographics.py, ingest_oz1_tracts.py, analyze_designations.py. Full detail in docs/designation-postmortem-2026-11.md.