"What does it cost, and whose number wins if mine looks off?"
Four passes, and ADOH underwrites the deal itself — twice
Cost estimation follows the same four-pass ladder used everywhere else in the industry — a rough screening number, a priced concept set, a locked application budget, and a final reconciliation — but what makes Arizona different is who signs off on the locked number, and how many times it gets checked again before the deal is done.
| Pass | Who produces it | Precision | Timing |
|---|---|---|---|
| Napkin / screening estimate | Developer's own model, in Excel | ±30–40% | Minutes to hours |
| Concept estimate | Architect's SD set priced by GC preconstruction or a third-party estimator | ±15–20% | Weeks after schematic design |
| Application budget | Entered into ADOH's own Underwriting Workbook (Costs and Eligible Basis tab); ADOH runs a full underwriting pass before issuing a binding reservation | Locked at reservation | Prior to reservation |
| 10% test technical review | ADOH re-certifies that sources, uses, and equity interest and ownership have not changed since reservation | Verification only, not a re-price | Prior to acceptance of the 10% test |
| Form 8609 underwriting | ADOH conducts a second full underwriting pass at 8609 request — Arizona's placed-in-service-equivalent true-up | Final | At submission of Form 8609 documents |
The practical difference from a state like California is where the legally binding number gets set. CTCAC treats the application-stage figure as a developer-and-consultant product that CTCAC checks; ADOH's Underwriting Standards commit the agency itself to "conduct a full underwriting process a minimum of two (2) times: prior to issuing a binding reservation; at submission of documents requesting a Form 8609," plus one technical review at the 10% test. A budget that survives the first full underwriting can still fail the second.
Who is in the room tracks California and Texas closely — developer project manager, architect, GC preconstruction estimator, LIHTC consultant. A labor-compliance consultant is not part of the standard cast, for the same structural reason it usually is not in Texas: nothing about a bare LIHTC award pulls one in. What does pull one in is covered later in this phase, and in Arizona it is a shorter list than in either of the other two states.
No published cost ceiling — Arizona benchmarks a budget against the room, not a table
Neither CTCAC's per-county threshold basis limit tables nor TDHCA's per-square-foot scoring thresholds exist in Arizona. ADOH's cost-reasonableness test for new construction is comparative and after-the-fact: ADOH will "determine applications show development budget amounts outside the standard deviation among applications submitted," require those applicants to explain the variance, and treat an inability to explain it as grounds for disqualification.
That is a materially different planning problem than a fixed ceiling. A cost model built for California or Texas can check a proposed budget against a published number months before the application is filed. An Arizona model cannot — the standard deviation that decides whether a budget needs defending does not exist until every application in that specific round has been submitted, and ADOH does not publish the resulting distribution afterward either. The only defensible move is to price close to the pattern of recently awarded deals and be ready to explain any real outlier, not to chase a number that does not exist yet.
| Development type | Maximum contingency | Base |
|---|---|---|
| New construction / adaptive reuse | 7.5%–10% | New construction hard-cost line items |
| Rehabilitation | 10% | Rehabilitation hard-cost line-items |
The one true basis adjustment in the Arizona QAP is flat and discretionary rather than itemized. ADOH "will allow for a statewide thirty percent (30%) increase in eligible basis for 9% LIHTC projects which demonstrate a financial need," approved only "to the level needed to realize projects' feasibility" — the same statutory authority CTCAC and TDHCA both draw on for federally designated difficult development areas (I.R.C. §42(d)(5)(B)), but applied by ADOH statewide on a needs basis rather than confined to a mapped area. There is no menu underneath it — no separate bump for parking, elevators, Type I versus Type III construction, or energy certification the way California's §10327(c)(5)(A)–(F) menu works. A project either qualifies for the flat 30% on a feasibility showing, or it gets no basis boost at all.
Developer fee, contractor fee, and architect fee — three independent caps, not one blended number
The figures below come from ADOH's 2024-2025 Qualified Allocation Plan — the most recent QAP whose full text could be confirmed directly against the agency's own document rather than a secondary summary. ADOH has since adopted a 2026-2027 QAP; developer fee percentages, the deferred-fee threshold, and the contractor fee caps are exactly the kind of figure ADOH revises between two-year cycles, so confirm the current PDF before underwriting to any specific number in this section.
| Units | 4% LIHTC | 9% LIHTC | Percent allowed |
|---|---|---|---|
| 1–30 | 1–30 | 1–30 | 19% |
| 31–60 | 31–60 | 31–60 | 17% |
| 61–90 (4%) / 61+ (9%) | 61–90 | 61+ | 16% |
| 91+ | 91+ | N/A | 19% — 4% LIHTC only |
Minimum developer fee is $200,000; totals are inclusive of any consulting fees. Total Eligible Basis is the Eligible Basis in Cost Line Item Sections I–IV on the Uses tab of ADOH's Underwriting Workbook.
Deferral works differently by credit type. On a 9% deal, ADOH will not approve more than half the developer fee as deferred, and — independently — any amount over $2,750,000 must be deferred regardless of that 50% test; the deferred balance has to be repayable from project cash flow within 15 years at 0% interest, a real DSCR constraint rather than a formality. On a 4% deal the $2,750,000 threshold does not apply at all, and the fee is simply locked in at Form 8609.
| Line item | Maximum |
|---|---|
| General Requirements | 6% |
| Builder's Overhead | 4% |
| Builder's Profit | 5% |
ADOH may allow an increase in General Requirements "to accommodate an increase in costs for small Rural/Balance of State projects" — the closest thing the QAP has to a geographic cost adjustment, and it is discretionary and limited to that one line item.
The three caps are not interchangeable with a single blended ceiling. CTCAC caps builder OH&P plus general requirements together at 14% of the cost of construction; ADOH caps General Requirements, Overhead and Profit as three separate line items that only sum to 15% if a developer happens to hit every ceiling simultaneously. A model that nets these into one 15% number can let a single line item silently exceed its own individual cap while the blended total still looks compliant.
| Units | Per unit |
|---|---|
| 1–30 | $9,000 |
| 31–60 | $8,000 |
| 61+ | $7,000 |
Why California's "four-to-five-story cliff" has no Arizona equivalent
California's most consequential construction-cost interaction is a coincidence of three independent rule systems that all break at the same line: the prevailing-wage schedule switches from residential to general commercial at five stories (8 CCR §16001(d)), Type V-A construction is code-capped at four stories (2025 CBC Table 504.4), and crossing that line typically forces structured parking. In Arizona, only one of those three legs exists at all, and it is not read off a single statewide table.
The building-code leg is real but jurisdiction-specific — the same structural gap Texas has. ADOH's own Mandatory Design Standards require only that "all projects MUST meet or exceed the most recent local building codes"; there is no Arizona equivalent of one statewide code table. Each city or town adopts its own building code edition by ordinance under A.R.S. §9-802 (which lets a municipality enact "the provisions of a code or public record theretofore in existence without setting forth the provisions"), and counties do the same for unincorporated land. A story-count or construction-type cost model built for an Arizona site has to confirm the applicable code edition city by city, exactly as in Texas — not read it off one statewide table the way California's CBC Table 504.4 allows.
The wage-schedule leg does not exist in Arizona at any story count, for the reason covered in full in the next section: there is no prevailing-wage schedule at all for a state or local public-works contract to switch onto.
The basis-boost leg is also absent. Unlike CTCAC's §10327(c)(5)(A) menu, which separately rewards Type I and Type III construction, elevators, and structured parking, ADOH's QAP has no construction-type-specific basis adjustment of any kind — the only 9% boost is the flat, discretionary 30% financial-need increase described above, and it does not vary with story count, construction type, or parking configuration.
The physical cost of going from four stories of wood frame to five stories of something heavier — structured parking, a different structural system, an elevator — is exactly as real in Arizona as in California. What is missing is the compounding legal machinery. An Arizona cost model should price that jump as a pure construction-cost question, checked against whatever the local code and the GC's own pricing say, not as a regulatory-cliff question with its own citation trail.
The labor package: Arizona affirmatively bars prevailing wage, and just said so again
LIHTC alone triggers nothing in Arizona, exactly as in California and Texas — it is a tax credit, not a direct federal subsidy. What is different is that Arizona does not merely lack a state prevailing-wage law by omission; it has an affirmative, voter-enacted statute that bars any political subdivision from imposing one, and the state's courts have now upheld that bar twice.
Arizona's Little Davis-Bacon Act dates to 1933 (House Bill 37, codified at A.R.S. §34-322), requiring prevailing wages on public-works contracts over $1,000. The Arizona Court of Appeals declared it unconstitutional in 1979 for delegating too much legislative power to labor unions (Industrial Comm'n v. C&D Pipeline, Inc., 125 Ariz. 64, 67 (App. 1979)). In 1984, voters went further and buried it by referendum: Proposition 300 declared that "the rates of wages paid under public works contracts... is [a matter] of statewide concern," repealed §34-322 outright, and amended §34-321 so that no public-works contract may require paying "the prevailing rate of wages." Courts now call A.R.S. §34-321(B) the Prevailing Wage Prohibition.
A 2006 ballot measure created a wrinkle that briefly looked like a loophole. Proposition 202, the "Raise the Minimum Wage Act for Working Arizonans," let a "county, city, or town... regulate minimum wages... within its geographic boundaries" — the Local Permission Provision, now A.R.S. §23-364(I). In 2023 the Phoenix City Council enacted, then quickly repealed, a "Prevailing Wage Ordinance for City Projects" covering contracts of at least $250,000. Two months later the Arizona Attorney General opined that a city may regulate wages under §23-364(I) (Ariz. Att'y Gen. Op. I23-004, June 15, 2023). Phoenix then passed a new ordinance requiring prevailing wages — pegged to federal Davis-Bacon rates — on contracts over $4 million; Tucson passed a parallel ordinance (No. 12066) the same day, covering contracts of $2 million or more.
Contractor associations sued, and won the argument the Cities were betting against. The Maricopa County Superior Court granted summary judgment for the plaintiffs — Associated Minority Contractors of Arizona, the Arizona Chapter of the Associated General Contractors of America, and the Arizona Builders Alliance — holding that a prevailing wage is not a minimum wage, and declared both ordinances unlawful and enjoined them. The Arizona Court of Appeals, Division One, affirmed on February 27, 2026 (Associated Minority Contractors of Arizona, et al. v. City of Phoenix, et al., No. 1 CA-CV 24-0658): the 2006 Act's use of "minimum wages" does not, in context, reach prevailing wages, so the Local Permission Provision never repealed the 1984 Prevailing Wage Prohibition. As of that decision the Phoenix and Tucson ordinances remain void, subject only to further review by the Arizona Supreme Court.
The practical consequence for a feasibility model runs opposite to California's. In California, a land donation, an impact-fee waiver, or a residual-receipts loan can flip an entire deal into prevailing wage under Labor Code §1720(b) — a capital-stack property a screener has to test source by source. In Arizona, no city or county action can do that at all: the Prevailing Wage Prohibition bars it categorically, and no exemption analysis is needed because there is no rule to be exempt from. A state/local labor-standards trigger matrix — the kind that matters enormously in California and matters narrowly in Texas — is simply not a live question in Arizona.
Federal Davis-Bacon is the one labor-standards layer that still applies, exactly as in every other state, and it turns entirely on which federal program is actually in the stack.
| Program | Trigger |
|---|---|
| HOME | 12 or more HOME-assisted units (24 CFR §92.354) |
| CDBG | Rehabilitation of residential property with 8 or more units |
| Project-based Section 8 | New construction or substantial rehab at 9 or more assisted units, agreement executed before construction begins |
| Public Housing (1937 Act) | No unit threshold |
| NAHASDA | No unit threshold; $2,000 contract threshold |
HUD's Factors of Labor Standards Applicability page is the same federal baseline that applies in every state; nothing about it is Arizona-specific.
Where this actually shows up in an Arizona deal is worth flagging precisely because ADOH's own QAP scoring rewards it. Section V.D.6, "Below Market Loans and Local Support," awards up to 20 points for local gap financing — including CDBG, HOME from a non-ADOH source, and NAHASDA — subject to a minimum loan of $500,000 in the Metro area or $100,000 in Rural/Balance of State and Tribal areas. A developer chasing those points is, in the same move, opting into whatever federal Davis-Bacon string attaches to that specific source. The QAP's own scoring incentive and the real labor-cost trigger point at exactly the same funding sources.
NAHASDA's no-threshold trigger is not theoretical in Arizona. ADOH's FY2024 award round included two tribal-set-aside projects — TOKA Homes VI on the Tohono O'odham Nation and Yavapai-Apache Homes IX — both built as scattered-site single-family homes rather than apartment buildings, a construction-type detail worth carrying into any tribal-deal cost model.
What the FY2024 numbers actually show
Arizona's public cost-benchmark literature is thinner than California's and even thinner than Texas's — there is no Terner-style econometric study of a prevailing-wage cost premium, and the reason is not a research gap. It is that, per the section above, there is essentially no Arizona prevailing-wage population left to study: no state or local mandate can attach to a privately built LIHTC deal at all.
Dividing ADOH's own reported project-cost total by its own reported unit counts gives a rough per-unit benchmark — a calculation performed here from the report's project-level table, not a figure ADOH states directly: roughly $403,700 per LIHTC unit ($1,711,134,196 ÷ 4,239 LIHTC units), or roughly $395,500 per total unit counting the market-rate units in mixed-income projects too ($1,711,134,196 ÷ 4,326 total units). Either way, that sits far below California's roughly $708,000 per unit in 2023 — consistent with a state that carries no prevailing-wage cost layer at all.
Rural relief is the closest thing Arizona has to a regional cost adjustment, and it is narrow. Section VI.B.9 lets ADOH allow a higher General Requirements percentage specifically "to accommodate an increase in costs for small Rural/Balance of State projects" — discretionary, limited to one line item, and nothing like the published per-region coefficients Terner Center has built for California.
Rehabilitation cost runs in the opposite direction from a cost-containment test, and it is worth flagging because a model built on the California or Texas assumption that higher cost is always the risk will get this backwards. Section IV.B.2 makes an average of at least $25,000 per unit in rehabilitation hard costs a threshold-eligibility floor — miss it, and the application is not scored at all. Section V.C.2 then awards up to 60 additional points for pushing that same per-unit figure past $50,000, on a sliding scale verified by a third-party Capital Needs Assessment. In Arizona rehab, spending too little is the risk that gets penalized.
The order to run this in
| Step | Action | Why |
|---|---|---|
| 1 | Fix the unit mix against Exhibit A's minimum/maximum Residential Floor Area table by bedroom count | Arizona has no per-county cost limit to size the unit against instead — RFA is the first real constraint on the hard-cost baseline |
| 2 | Confirm which city's or county's building code edition governs the site before pricing story count or construction type | There is no statewide table to read this off; it has to be checked jurisdiction by jurisdiction |
| 3 | Decide the capital stack, and check each gap-financing source against the federal Davis-Bacon trigger table — not against any state or local prevailing-wage test | No such state or local test exists to check against; A.R.S. §34-321(B) forecloses it categorically |
| 4 | Size developer fee and contractor fee against the QAP's stacked percentage caps, and pressure-test the deferred-fee repayment schedule if a 9% fee exceeds half its own total or $2,750,000 | These are independent caps, not one blended number, and the deferred-fee test is a real 15-year cash-flow constraint |
| 5 | Only then compare the resulting budget against recent awards, and hold contingency for ADOH's after-the-fact outlier review | There is no published ceiling to check against in advance — the standard-deviation test only exists once the round closes |
The result is a different shape of problem than California's. CTCAC's boost menu makes cost, construction type and the labor decision a small joint-optimization problem worth iterating by hand before the application is filed. Arizona's caps are mostly independent and mostly fixed in advance; the one genuinely unresolved variable is the after-the-fact comparison to that round's own applicant pool, which cannot be optimized against because it does not exist yet. The one true certainty — and the one that most separates Arizona from every other state in this corpus — is that the largest single cost variable in a California underwrite, prevailing wage, is not a variable in Arizona at all. It is a settled, twice-litigated zero.
Where this goes wrong
- Assuming a published cost ceiling exists to check a budget against, the way CTCAC's threshold basis limits or TDHCA's per-square-foot thresholds work. ADOH publishes neither — the only cost test is whether a budget looks "outside the standard deviation among applications submitted" in that specific round, a number nobody outside ADOH can see in advance.
- Budgeting contingency for a state or local prevailing-wage risk that cannot legally attach. Unlike California, no city or county land donation, fee waiver, or soft loan can trigger a prevailing-wage mandate on an Arizona LIHTC deal — A.R.S. §34-321(B) bars it outright, a bar the Court of Appeals reaffirmed on February 27, 2026. Padding a budget against a risk that has been affirmatively and repeatedly foreclosed wastes real underwriting capacity that belongs elsewhere.
- Confusing the two different $25,000-per-unit rehabilitation numbers. Section IV.B.2's average is a threshold-eligibility floor — miss it and the application is not scored. Section V.C.2's $26,001-to-$50,001-plus ladder is a separate 60-point scoring item on top of it. Treating them as the same requirement, or assuming a higher rehab number is a red flag the way it would be in a cost-containment state, gets the incentive backwards.
- Applying one flat developer-fee percentage across the whole deal. The percentage of Total Eligible Basis moves non-monotonically with unit count — 19% at 1-30 units, down to 17% at 31-60, down again to 16% at 61-90 (4%) or 61+ (9%), then back up to 19% for 4% LIHTC deals at 91 or more units. A model built around one number will misprice any deal near those breakpoints.
- Missing the $2,750,000 deferred-fee cliff on a 9% deal. Any developer fee amount above $2,750,000 must be deferred regardless of the separate rule capping deferral at half the total fee, and the deferred balance has to be repayable from project cash flow within 15 years at 0% interest — a real, testable constraint on the pro forma, not a formality.
- Netting General Requirements, Builder's Overhead and Builder's Profit into one 15% cap. ADOH caps the three as independent line items — 6%, 4%, and 5% — that only sum to 15% if every one of them is maxed out simultaneously. A model that blends them can let one line item silently breach its own individual ceiling while the combined total still looks compliant.
- Assuming a single statewide building code table, the way California's CBC Table 504.4 works. ADOH's own Mandatory Design Standards require only meeting "the most recent local building codes" — the applicable code edition and any story or construction-type cap that follows from it has to be confirmed city by city, or county by county on unincorporated land.
- Building a construction-type cost adjustment on California's basis-boost menu. ADOH's QAP has no Type I, Type III, or parking-specific basis adjustment of any kind — the only 9% boost is a flat, discretionary 30% tied to demonstrated financial need, and it does not vary with story count or construction type.
- Modeling a tribal-set-aside deal as multifamily by default. Both of Arizona's FY2024 tribal-land LIHTC awards — TOKA Homes VI and Yavapai-Apache Homes IX — were scattered-site single-family construction, not apartment buildings. A per-square-foot assumption built for multifamily will misprice a reservation deal.
- Chasing the Below Market Loans and Local Support scoring points without pricing what rides along with them. CDBG, HOME from a non-ADOH source, and NAHASDA gap financing are worth up to 20 points under Section V.D.6, and each one carries its own federal Davis-Bacon threshold. The QAP's own scoring incentive points straight at the funding sources that can bring a real labor-standards cost, even though no state law requires one.
- Treating the 10% test technical review as a second full underwriting. ADOH only re-certifies that sources, uses, and ownership have not changed at the 10% test; the real second full underwriting happens at Form 8609. A budget that survives the technical review has not yet been through the check that actually re-prices the deal.
- Carrying 2024-2025 QAP figures forward into a 2026-2027 deal without checking the current document. ADOH has since adopted a new QAP cycle, and developer fee percentages, the deferred-fee threshold, and contractor fee caps are exactly the kind of number that moves between cycles — confirm the current PDF rather than assuming any figure in this phase carried forward unchanged.
- HUD
- LIHTC
- State QAPs
- IRS § 42
- Housing Finance Agencies