Site sourcing and screening — New Jersey

What the two-week screen looks like in New Jersey

New Jersey doesn't carve out a separate phase called data acquisition either. An analyst gets a site and, over roughly the next two weeks, works through a checklist that looks structurally like every other state's — pull the parcel, read the zoning ordinance, check the hazards, price the deal — but almost every item on it resolves differently once you're inside New Jersey's own rulebook: N.J.A.C. 5:80-33, the Low-Income Housing Tax Credit Qualified Allocation Plan administered by the New Jersey Housing and Mortgage Finance Agency (NJHMFA). The 2026 QAP was adopted February 2, 2026 (58 N.J.R. 991(a)).

Two decisions come out of this phase, same as anywhere: whether to pursue the site, and what price goes into the option or contract. New Jersey's rule for the second one is more mechanical than a discretionary scoring discount — it caps the number outright. Under N.J.A.C. 5:80-33.12(c)2i, "the acquisition price and basis shall be limited to the lesser of the purchase price or the 'as is' appraised value of the building and/or land." For deals claiming acquisition credits specifically, N.J.A.C. 5:80-33.12(c)10 adds a second test: acquisition value is the lesser of appraised value or the price paid in the most recent arm's-length transaction within the past ten years, established by a title history, and the appraisal itself cannot be older than six months. Overpaying for land in New Jersey doesn't just weaken an underwriting memo — it is arithmetically removed from eligible basis before the application is even scored.

One statewide parcel layer, and a home-rule zoning map underneath it

New Jersey's parcel layer is the opposite structural story from a fragmented county-by-county build. NJGIN — the state's geographic information network — publishes one statewide parcel composite covering all 21 counties, available as a File Geodatabase (statewide or by county), county shapefiles, streaming web services, and a public-facing Property Explorer application. That's a genuine advantage over reconstructing county adapters one at a time.

The composite comes with real caveats baked into its own documentation, not discovered later by a user. The polygons "do not represent legal boundaries and should not be used for legal determinations" — they derive from that 2008–2014 normalization project, not survey data, and edge-matching between adjacent counties is incomplete in places. Owner names are redacted throughout per Daniel's Law, New Jersey's home-address privacy statute — a screening tool that expects to auto-populate a current owner's name for outreach will find that field empty for a meaningful share of parcels, and needs to treat the blank as a privacy redaction rather than a data gap to flag as missing.

Zoning underneath that parcel layer runs municipality by municipality. New Jersey has 564 municipalities — the same count DCA uses to rank its Municipal Revitalization Index — and each amends its own zoning ordinance independently, with no county zoning layer standing behind it. Two overlays can override the base zone entirely: the Highlands Water Protection and Planning Act (the Highlands Council is established at N.J.S.A. 13:20-4; its Transfer of Development Rights program at N.J.S.A. 13:20-13) governs a swath of northwestern New Jersey, and the Pinelands Protection Act (N.J.S.A. 13:18A-4) governs the Pinelands Area in the south. Both show up directly in the QAP's own scoring: a site inside a Pinelands Regional Growth Area, Village, or Town counts as a "smart growth area" outright, while a site inside a Designated Highlands Center, Highlands Redevelopment Area, or Highlands Development Credit Receiving Area satisfies the separate water-and-wastewater-infrastructure prong of the QAP's "ready to grow area" test — either route can unlock the two ready-to-grow points at N.J.A.C. 5:80-33.15(a)7.

NJHMFA points applicants to a real, open GIS layer for that determination: the New Jersey Smartgrowth Areas dataset, published on the state's open-data ArcGIS platform and keyed to the State Plan Policy Map's Planning Areas and Designated Centers. It is the closest thing New Jersey's screen has to an authoritative statewide land-use geography layer — open, queryable, and pointed to directly by the regulator rather than reconstructed by a third party.

Redevelopment-area status is worth flagging separately because it does double duty. A site inside a municipality's adopted redevelopment or rehabilitation area — defined by cross-reference to the Local Redevelopment and Housing Law, N.J.S.A. 40A:12A-3, or inside an approved neighborhood revitalization plan under the Neighborhood Revitalization State Tax Credit Act, N.J.S.A. 52:27D-491 — both satisfies one of the ready-to-grow criteria and can support a disposition and development agreement as a site-control instrument under N.J.A.C. 5:80-33.12(c)2i, including an eminent-domain path with the condemnor identified per N.J.S.A. 20:3-2. The redevelopment plan has to be adopted by the municipal governing body by the application deadline; one still moving through a planning board hasn't cleared the bar.

Site location is scored down to the land use — and the clock has to have already started

New Jersey scores site amenities with arithmetic, up front, on a schedule granular enough to be its own regulatory subject: N.J.A.C. 5:80-33.15(a)11 awards up to six points for proximity to twenty enumerated positive land uses — full-service grocery, hospital, public school, licensed day care, a one-stop career center, and fifteen more — each worth one or two points depending on category. Within a half-mile, each category is worth its listed value; a Family Cycle project can also claim a flat one point for any category located within three miles, and Age-Friendly Senior and Supportive Housing Cycle projects use a flat one-mile radius instead. Either way the application caps out at six points for positive land uses, and two instances of the same category — two supermarkets, say — don't stack.

The catch sits in one sentence easy to miss on a first read: to count at all, "structures must have building permits issued and be under construction." A grocery store that's been announced, permitted, or even fully entitled but hasn't broken ground earns nothing. A screen that geocodes against a general points-of-interest dataset — the kind that lists a planned store the day the press release runs — will overstate this score on exactly the parcels where the amenity story is doing the most work to sell the site.

The same section penalizes eight negative land uses — landfill, garbage dump, trash incinerator, nuclear power plant, oil or chemical refinery, an unremediated Superfund or toxic-waste site identified by EPA or DEP, jail or prison, and wastewater treatment facility — three points deducted per category within one mile, capped at six points lost.

Transit and jobs access sit in N.J.A.C. 5:80-33.15(a)14ii, and the category caps at five points total — a detail that matters because it isn't additive the way it first reads. The base transit tier runs from five points (fully inside a designated Transit Village) down to one (within a half-mile of privately-funded transit); a municipal jobs-to-housing ratio of 1.5 or higher (NJ Department of Labor and Workforce Development and ACS Table B25001 data) adds up to two more points, or one more for a ratio between 0.95 and 1.5 — but the combined total still can't exceed five. "Public transportation" itself carries a real definition, not just a distance: fixed fares, no seasonal interruption, and scheduled service at least once between 6:30 and 9:30 a.m. and once between 3:00 and 6:00 p.m., Monday through Friday. A stop running only seasonal or off-peak service doesn't qualify no matter how close it sits.

School quality is scored at (a)14iii, also capped at three points, not four. The base tier runs on the share of a school district's Grade 4 students scoring proficient or better on the NJSLA in both Math and English Language Arts, using the most recent New Jersey Department of Education data as of the deadline plus the prior year's: three points at 66 percent or higher, two points at 50 to 66 percent, one point at 40 to 50 percent. Participation in the Interdistrict Public School Choice Program starting no later than Grade 6 adds one more point — but only up to the same three-point ceiling, so it only moves the needle for a district already scoring below the top tier.

A separate, municipality-level distress score does its own scoring at (a)14iv: DCA's Municipal Revitalization Index (MRI) ranks all 564 municipalities from most distressed (rank 1) to least distressed (rank 564), and a project outside a Targeted Urban Municipality (TUM) in a municipality ranked 283 or higher earns three points, while one in a municipality ranked 282 or below earns two — plus a third if the project sits inside a court-approved municipal fair-share housing plan. And however a site scores, N.J.A.C. 5:80-33.4(c) caps how many winners can come from one place: no more than two Family Cycle awards per year to a municipality under 100,000 residents, three for one at or above that, and no more than three per year to any one developer or general partner group, using U.S. Census ACS Table DP05 population figures.

One more score belongs on a site-sourcing checklist even though it has nothing to do with the parcel: negative points under N.J.A.C. 5:80-33.15(a)15, 17, 18, and 19 attach to the development team. A general partner, voting member, developer, or related party with an uncorrected compliance violation, an unpaid NJHMFA monitoring fee, a missed annual certification, or a fair-housing finding at any other New Jersey LIHTC project costs this application 10 to 15 points and bars it from every set-aside outright. That's a diligence step that belongs at team formation, before a site is even under control, not something discovered after the application is filed.

The hazard the QAP itself never mentions

Search the full text of the 2026 QAP for "flood," "FEMA," or "floodplain" and it comes back empty. New Jersey's tax credit rulebook doesn't gate site eligibility on flood risk at all — not because the state doesn't regulate it, but because that regulation lives entirely outside N.J.A.C. 5:80-33. It sits instead in NJDEP's Flood Hazard Area Control Act rules, N.J.A.C. 7:13 (statutory authority at N.J.S.A. 58:16A-50 et seq.), last amended June 15, 2026 — a genuinely separate permitting regime a screen has to know to go looking for, because the tax credit application itself never points you there.

The sleeper number inside that rule is the riparian buffer, not the floodplain boundary. N.J.A.C. 7:13-4.1(c) sets a mandatory buffer measured landward from the top of bank: 300 feet along any Category One water and its upstream tributaries within the same HUC-14 watershed, 150 feet along trout production or maintenance waters or documented threatened-or-endangered-species habitat, and 50 feet along every other regulated water in the state. On a tight urban infill lot with even a minor stream at the edge, a 50-to-300-foot buffer can remove a large share of the site's buildable area — and it's invisible to any screen that checks FEMA flood zones and stops there, because the riparian zone and the FEMA floodplain are not the same geography.

What the QAP does require, at N.J.A.C. 5:80-33.12(c)4, is disclosure: known environmental conditions — wetlands, stream encroachment, steep-slope grading — and certification that DEP approvals have been obtained or at least applied for. A Phase I environmental site assessment (ASTM E1527-21 standard) isn't mandatory at application. But skipping it is a one-way bet: if the project is awarded credits and contamination surfaces later without a Phase I on file, the applicant is barred from applying for hardship credits to cover it. The QAP's own negative-land-use list at (a)11ii — unremediated Superfund or toxic-waste sites, among others — is proximity-based and self-reported, and functions as screening evidence for that disclosure, not as the DEP approval itself.

Where New Jersey's screen stops

Comparable-award research in New Jersey means several separate PDFs per cycle — an Applicant List, a Reservation List, a Ranking Chart, and a Points Chart — published back to 2009 for the competitive rounds, plus two running PDF logs of 4% credit activity. There's no consolidated, project-level file with parcel or census-tract keys the way a single downloadable award roster works in other states; building a comparable-awards map in New Jersey means parsing well over a decade of separately formatted PDFs by hand.

Income limits, maximum rents, and utility allowance charts are the same story: every year on record back to 2005 is a PDF, with no XLSX or CSV version ever published, and several years carry two or three separate versions branching on placed-in-service date.

Construction cost has the same hole in New Jersey it has everywhere: no free, authoritative, New Jersey-specific unit-level affordable-housing cost dataset exists. What NJHMFA publishes instead is a regulatory ceiling, not a cost estimate — the Family Cycle's total development cost caps ($385,000 per unit for one-to-four-story buildings, $427,500 for five or six stories, $461,250 for buildings over six stories, before enumerated exclusions like adaptive reuse or Passive House allowances). Treating that ceiling as a market cost benchmark is a real conflation risk: it's a cap on what NJHMFA will fund, not a forecast of what the building will actually cost to build.

One federal program worth flagging for New Jersey's more rural counties — Sussex, Warren, Cumberland, and similar — is USDA Rural Development. N.J.A.C. 5:80-33.12(c)12 requires a State Director approval letter confirming the RD loan is obligated, and because RD financing doesn't fund a developer fee, NJHMFA may cap the allocated credit amount to just enough to cover that fee — a real constraint on deal economics that only surfaces once RD financing is already on the table.

And the honest limit, same as anywhere: whether the seller actually extends the option, whether the planning board will fight the project or wave it through, whether a competing application is quietly being assembled in the next town over — none of that is a dataset, and screening shouldn't pretend otherwise. What it can and should nail down before the phone call is arithmetic: the site-selection score under (a)11 and (a)14, the municipality's de-concentration exposure, the riparian buffer eating into the buildable envelope, and whether the number on the LOI still clears once N.J.A.C. 5:80-33.12(c)2i caps acquisition basis at the lesser of price or appraised value.