Building the capital stack and closing the gap — Texas

What actually happens, and why the first fork is jurisdictional, not competitive

Structurally this mirrors California: structuring, soft-money assembly, and debt/equity procurement run in parallel over roughly the same 24–48-month window. Two Texas-specific facts reshape the sequencing inside that window.

In California the state credit competition is the thing that gates the bonds. In Texas there is no equivalent gate — a 2023 law (H.B. 1058) created a real, operating Texas State Housing Tax Credit, first claimable in 2026, but it is optional and non-gating: TDHCA does not withhold a bond allocation for failing to win it, the way CDLAC withholds bonds from a project that requested and was not scheduled to receive state credits. Requesting the Texas state credit is not the same risk position it is in California.

What gates a Texas deal instead is jurisdiction. TDHCA's own HOME loans — the larger half of its in-house multifamily soft-money apparatus — are categorically unavailable inside Houston, Dallas, Austin, San Antonio, or any other city that is its own federally designated HOME Participating Jurisdiction (PJ); those cities receive their own direct HUD HOME allocation and run their own programs instead. Whether the same wall applies to TDHCA's National Housing Trust Fund (NHTF) loans is genuinely unresolved — the 2026 NHTF NOFA does not repeat the HOME NOFA's explicit PJ-exclusion language, so NHTF eligibility inside a PJ city is an open question pending a direct read of 24 CFR Part 93, not a confirmed exclusion. A tool, or a developer, that does not know a site's PJ status before recommending a soft source will recommend a HOME program the deal is legally ineligible for — not a smaller allocation, a categorical exclusion — and should flag NHTF eligibility as unresolved rather than assuming either way.

The exit Texas didn't foreclose

This is the single most consequential structural divergence from California in this entire domain, so it belongs near the top rather than in a footnote. California's regulatory agreement runs 55 years with no meaningful early-exit path. Texas's default Extended Use Period is only the federal floor: a 15-year Compliance Period plus a 15-year extension, 30 years total — and Texas's own rules still define and administer a live Qualified Contract (QC) process, referenced directly from the current QAP's definitions and scoring sections.

A Texas owner who does not elect to extend past year 15 can, in principle, put the property through the statutory QC process. TDHCA additionally offers voluntary, scored extensions to 35, 40 or 45 years (worth 2, 3 or 4 competitive points) and a voluntary, scored Right of First Refusal commitment (1 point) — but neither is mandatory the way California's 55-year term and its ROFR requirement are for reserved-nonprofit deals.

One honest caveat: the QC rule's actual text — pricing formula, procedural hurdles TDHCA may have layered onto the federal floor price, and how often it is actually exercised — was not independently read this pass; only its existence and its live cross-reference from the current QAP are confirmed. Treat the mechanism as real and the mechanics as still to be verified before quoting a QC price to a client.

The practical consequence: market cap rates and comparable-sale data are more directly relevant to a Texas LIHTC deal's economics than to a California one, because a real, if formula-constrained, exit actually exists at or near year 15 for developments that don't elect extended affordability. That said, no credible LIHTC-specific Texas cap-rate series was found in freely available sources this pass — general Texas multifamily cap-rate ranges exist in secondary industry reporting, but restricted, income-capped LIHTC properties do not necessarily trade at the same rates as market-rate stock, and a tool should not print a number it cannot source to a primary or LIHTC-specific series.

Bonds: split three ways, and allocated by a lottery, not a score

Where CDLAC is a single state committee that both allocates bonds and runs a public-benefit tiebreaker for every 4% applicant statewide, Texas's private-activity-bond system is administered by the Texas Bond Review Board (BRB) and split among many separate issuers, allocated by an annual lottery rather than a score.

The bond test and the aggregate-basis cap — conditional, not automatic

TDHCA's own December 4, 2025 guidance memo on this test states plainly: "TDHCA does not currently have a requirement that limits the 25% test further." That is the cleanest confirmation available that Texas has not layered a CDLAC-style overlay on top of the federal floor.

What Texas has instead is a conditional cap, not an unconditional one. A single project's reservation in a program year may not exceed the greater of $50,000,000 or 1.70% of the available state ceiling. Separately, if statewide demand for QRRP bond reservations exceeds 55.75% of the state ceiling as of October 20 of the preceding year, every project's bond issuance is capped at 55% of the reasonably expected aggregate basis of the project and the land — but only in a year that trigger fires. In an undersubscribed year, no project-level aggregate-basis cap applies from this section at all. Compare CDLAC's now-unconditional 30% cap (40% by Executive Director exception): Texas's version only bites on demand, not by default.

One rulebook, and a DCR band instead of two separate rules

TDHCA is both the tax-credit-allocating agency and the state's primary soft-money lender, and its underwriting rules apply uniformly regardless of funding source. There is no second agency layering an independent cap on top the way HCD's Uniform Multifamily Regulations do in California — one rulebook, one Underwriter, for both the credit application and the Direct Loan.

The remedy for a DCR outside that band lands entirely on the financing structure, never on the credit or fee. Too low, the Underwriter reduces debt service in order — lower the Direct Loan's interest rate, then lengthen its amortization, then reduce its principal, then reduce the assumed permanent loan amount. Too high, the Underwriter raises debt service in the mirror order. This is a structurally different remedy than California's, where excess cash flow triggers a credit reduction under Section 10327(g)(6) — in Texas the lever is debt sizing, in California it is the credit award itself.

The deferred-fee-repayability test is the sharpest single contrast with California in this whole domain. California's equivalent concern is unverified practitioner folklore tied to an IRS revenue ruling, built at most as a soft warning. In Texas it is a codified, agency-administered pass/fail test the Underwriter actually applies using the same 30-year pro forma used for the DCR test — a tool built for Texas can and should implement it as a hard rule.

Pro forma escalators: a 30-year long-term pro forma, 2% annual income growth and 3% annual operating-expense growth (management fees instead scale directly with each year's effective gross income). If a Direct Loan is in the stack and the DCR would exceed 1.50 in any year of the loan term or federal affordability period, TDHCA will not recommend approval unless the applicant commits 25% of annual Cash Flow (after deferred-fee and cash-flow-loan payments) to a special reserve, or accepts a Direct Loan rate increase instead.

Developer fee: no dollar ceiling, but the credit caps do the job instead

There is no dollar ceiling anywhere in this formula — the headline contrast with California, which caps 9% fee at the lesser of a percentage formula and a hard $2.5M/$2.8M ceiling, with a further $6M cash-out cap on 4% deals. A 300-unit, $90M-TDC Texas deal can generate a percentage-based fee many multiples of what an equivalent California deal could ever book in eligible basis.

What actually constrains a large Texas 9% deal's fee is not a fee-cap section at all. It is the $6,000,000 per-Applicant/Developer/Affiliate/Guarantor aggregate credit limit per Application Round, and the per-Development cap of the lesser of 150% of the subregion's estimated credit amount or $2,000,000 — these cap the credit request, which in turn caps how large a 9% deal's basis (and its percentage-based fee) can realistically be, as a side effect of the credit-allocation cap rather than a developer-fee rule. A 4% bond deal, not subject to those credit caps, has essentially no upper bound on fee-in-basis beyond the percentage formula itself and what a lender and equity partner will underwrite.

The fee definition sweeps broadly: any fee to an affiliate or related party for work the Underwriter determines is typically completed by the developer, plus all developer general-and-administrative expense including travel, dining and courier costs, count as developer fee — the same catches-people-the-same-way spirit as California's expansive definition. And any fee claimed for ineligible costs above the percentage cap is excluded from Total Housing Development Costs entirely, not just from eligible basis — a harsher exclusion than California's.

Two structuring rules with no direct California analog. First, to be eligible for the Leveraging scoring points (worth 1–3 points), no more than 50% of developer fee may be deferred — a binary threshold, not California's sliding percentage haircut. Second, any Owner Contribution of capital or debt from a GP or MGP (not the syndication-equity partner) exceeding 5% of TDC is, with narrow exceptions, deemed added to Deferred Developer Fee for feasibility-testing purposes — a large GP capital contribution doesn't just sit quietly in the sources schedule, it gets folded into the same 15-year repayability test that governs deferred fee.

For identity-of-interest acquisitions in competitive HTC applications, cash-out to a related-party seller is prohibited outright, with narrow carve-outs for USDA existing developments and nonprofit/housing-authority structures. Any seller note must be cash-flow-contingent, carry no DCR requirement for payment eligibility, and sit subordinate to deferred developer fee — a stricter, blanket threshold rule than California's opt-in 80% cash-out haircut.

The Texas soft-money map: three genuinely separate pipelines

There is no single 'soft money' search in Texas — there are three pipelines with different eligibility rules, different administering levels of government, and no overlap: TDHCA's own Multifamily Direct Loan (MFDL) program for the balance of state; each Participating Jurisdiction city's own housing department inside its own boundary; and USDA Rural Development for genuinely rural sites, with no state agency in between.

Two agencies that look like plausible rural gap sources are not. The Texas Housing Trust Fund — $10,058,675 for the 2026–2027 biennium, state general revenue — funds only two single-family, owner-occupied programs (the Texas Bootstrap Loan Program and the Amy Young Barrier Removal Program); it has no multifamily use and no equivalent of HCD's Multifamily Housing Program. And the Texas Department of Agriculture runs TxCDBG, which funds rural public infrastructure and single-family rehab, not multifamily new construction — nothing in TDA's own program materials describes an LIHTC gap product. The real rural multifamily financing source is federal: USDA Rural Development's Section 515 direct-loan and Section 538 guaranteed-loan programs, paired with TDHCA's own statutory 5% USDA Set-Aside of the state's credit ceiling. Reported FY2026 national loan authority is roughly $50,000,000 for Section 515 and $400,000,000 for Section 538 — both from a secondary industry summary of the FY2026 appropriations act, not independently confirmed against a primary federal budget document.

Texas's own State Housing Tax Credit (H.B. 1058, 2023) is real and operating — first claimable in 2026 — but small and non-gating relative to California's. The minimum request is $3,000,000, and the fee treatment is a hard ceiling tied to the federal deal: total developer fee cannot exceed what was already underwritten at the federal-credit stage, so the state credit adds equity, not fee room. A $25,000,000 annual cap split evenly between 4% and 9% deals circulates in secondary law-firm reporting, but the QAP text itself does not state that dollar figure and the underlying Tax Code/Insurance Code provisions were not independently read this pass — treat the program's existence and mechanics above as settled and that specific cap as unconfirmed.

The calendar and the local gates: where deals actually die

Houston's Resolution of No Objection (RNO) process is the clearest illustration in this domain of a hard local gate sitting in front of the state deadline — implemented by a city council vote rather than a state regulation, but functionally similar in effect to CDLAC's state-credit-gates-the-bonds coupling in California.

The trap runs the opposite direction of what an applicant might intuitively pursue: HCD will not even review a Resolution request for a 4% deal until the applicant already holds a tax-exempt bond reservation. Seeking the city's blessing before the bond reservation is in hand gets a flat no, not a delay.

Unlike CDLAC's dollar-denominated tiebreaker, none of these four factors requires HUD FMR or AMI-targeting data — it is closer to a pure GIS problem, and the closest thing in this domain to an easy build if a parcel-level amenity-distance layer already exists.

The inputs nobody can source for you

Equity pricing is paywalled, same as California — CohnReznick's Housing Tax Credit Monitor and Novogradac's equity-pricing page are subscription products, and neither publishes a Texas-specific regional breakout in free secondary coverage. National secondary reporting places pricing broadly in the $0.80s per credit dollar in early 2026. Texas has no state-mandated minimum tax credit factor the way CTCAC publishes one — meaning there isn't even an agency number to fall back on as a soft hint. Carry price as an explicit sensitivity input, never a constant.

Building-cost estimation runs through the Marshall and Swift Residential Cost Handbook, a named but subscription-gated source that TDHCA's own rule directs its Underwriter to use. And the QAP's cost-per-square-foot scoring targets are re-indexed annually against the U.S. Census Bureau's Construction Price Index for Multifamily Housing Units Under Construction — meaning any figure encoded from this cycle's QAP is cycle-specific and must be versioned, not treated as a durable constant.

The Public Facility Corporation property-tax exemption is a real structuring lever — and Texas just made it dramatically less generous, in direct response to a well-publicized abuse pattern, while California's comparable welfare exemption has gone the other direction. A 2023 law tightened the required affordability depth, capped the exemption's term at 30 years from acquisition (60 for new construction) instead of the prior life-of-property exemption, and added notice, hearing, and compliance-reporting requirements. Every figure describing this change comes from law-firm client alerts, not a direct read of the codified Local Government Code and Tax Code text — present the mechanism (a modeled step-up date, a modeled affordability-depth requirement) without asserting a specific dollar savings until that text has actually been read.

And one live moving target: TDHCA's own Governing Board approved a draft new rule chapter for its Direct Loan program in January 2026. The Direct-Loan-specific citations used throughout the current QAP may relocate to that new chapter during the 2026 rule year — a genuine mid-cycle-about-to-change flag, the same category of risk the California guide flags for agency PDFs that get quietly replaced.