Garden vs. Mid-Rise vs. High-Rise: Multifamily Density Economics

Key Takeaways

Density Is the Master Variable

In multifamily underwriting, the density tier sets almost everything else. Once you decide whether a site is garden, wrap, mid-rise, or high-rise, you have implicitly committed to a construction type, a hard cost per square foot, a parking ratio, an amenity stack, a rent ceiling, a target NOI margin, a target exit cap, a levered IRR band, and a specific capital market that finances the deal. Pick the wrong tier for the site and every downstream number is misallocated: land basis is mis-sized, the construction type doesn't fit the cost basis, the rent premium isn't there, the parking ratio is wrong, the cap rate at exit doesn't compress to where the proforma assumed, and the lender you brought to the table isn't the one this deal should have gone to.

Density also determines who else is in the market with you. The deliveries you compete with for lease-up, the cap rates you exit into, and the LPs writing equity checks for the strategy are different across tiers. A garden deal in a Sun Belt secondary is competing with agency-funded merchant builders and private syndicators; a high-rise in a gateway market is competing with sovereign and core institutional capital on a 24-to-36-month construction timeline. The two are not interchangeable allocations.

The four density tiers used in this article borrow the vocabulary of NCREIF's property-type definitions(garden, low-rise, mid-rise, high-rise) but set their own story cutoffs: garden 1–4 stories, wrap 4–5, mid-rise 5–7, and high-rise 8+. NCREIF's own definitions differ: Garden follows the National Apartment Association definition (three stories or less, typically no elevators), and the non-garden clusters are Low-Rise (1–4 floors), Mid-Rise (5–10 floors), and High-Rise (more than 10 floors). Confirm NCREIF's current definitions before mapping a specific property to an NPI sub-type. The mid-rise tier here also splits between wood-frame wrap or podium construction and concrete construction at the same floor count. The article walks all four tiers in the institutional order, prices them at illustrative 2026 numbers, maps each to the right capital source, and ends with a 200-unit side-by-side.

A note on the ranges: hard cost, rent, NOI margin, exit cap, target IRR, parking ancillary, and debt-spread ranges in this article are illustrative 2026 underwriting ranges for a Sun Belt metro, not survey statistics. Where a figure comes from a published source, the source is linked in the text. Replace the ranges with local comps before underwriting a specific site.

Tier 1: Garden (2–4 Stories, Type V Wood-Frame, Surface Parking)

Garden-style multifamily is the workhorse of U.S. apartment supply. Two to four stories, all wood-frame (Type V-A or V-B per the International Building Code), spread across multiple buildings on 8–20+ acres of land, with surface parking and a modest amenity stack. Density runs 15–30 units per acre. Typical project size is 150–400 units across the site. NCREIF's Garden design class (the NAA definition: three stories or less, landscaped grounds, typically no elevators) overlaps the 1–3 story end of this tier; NCREIF classes elevator-served 1–4 story product as Low-Rise.

The construction-cost story is the cleanest of the four tiers. Wood-frame is the cheapest residential structural system in the U.S., and surface parking eliminates the structured-parking premium that podium and high-rise construction carry. This article uses $150–210 PSF of hard cost for garden in 2026, an illustrative range that sits above older garden cost tables because the 2026 overlay adds about $22 PSF to garden (see the cost-overlay figure below). As a loose cross-check, RSMeans' 2026 apartment cost guide puts the cost to build a 1–3 story apartment complex at $220–575 PSF; the same page puts hard costs at 70% to 80% of the budget, so its $220 floor implies about $154–176 PSF of hard cost (220 × 0.70 = 154; 220 × 0.80 = 176), which falls in the lower half of this range. Cost is also regional: Multifamily.loans' investor guide gives a 2023 national average multifamily construction cost of $350 PSF ($300–350 for secondary markets such as Phoenix and Orlando, $250–300 for tertiary and rural markets); the page does not say whether these are hard-cost-only, so they are not directly comparable with the ranges here. The headline is that garden works only if land basis stays low per unit (about $15K in the worked example), because the structure is the cheapest of the four tiers.

Garden rents are the lowest absolute PSF of the four tiers and the lowest absolute per-unit. In the 2026 Sun Belt, garden rent PSF generally runs $1.50–1.95 across the major metros (Houston, Phoenix, Dallas, Atlanta, Nashville, Tampa, Raleigh), which is $1,425–1,853 per unit at the 950 SF average unit used in the worked example below (1.50 × 950 = 1,425; 1.95 × 950 = 1,852.50). Garden trades on absolute rent affordability, not premium pricing. The renter is paying less per unit than they would in any other typology in the same metro, and the operator is delivering it with the lowest cost basis.

Parking and amenity are correspondingly modest. 1.5–2.0 spaces per unit, surface, uncovered, free to residents. The amenity stack is pool, fitness room, leasing office, sometimes a clubhouse and a dog park. No structured parking, no rooftop deck, no concierge, no co-working floor. The operating model is simple and the OpEx margin reflects it: NOI margins typically run 58–62% on stabilized garden in the Sun Belt.

Exit caps in the 2026 environment sit at 5.25–6.25% for institutional-quality garden core/core-plus in the Sun Belt; 6.0–7.0%+ for tertiary or value-add. Sun Belt garden sits wider on cap rate than gateway prime product, which trades at the tightest caps of any tier. Target levered IRR is11–14% on core/core-plus stabilized acquisitions with agency debt, stretching to14–18% on value-add with a credible interior-renovation and rent-roll-improvement program.

The capital source profile is what makes garden the deepest-funded tier in the U.S. multifamily market.Agency debt (Fannie Mae DUS and Freddie Mac Optigo) is dominant for stabilized garden acquisitions: DSCR Plus, Small Balance Loans, Green Up programs, and standard agency execution all favor garden underwriting because garden is the most standardized product type in agency underwriting. Bridge debt funds finance transitional and value-add garden deals; HUD 223(f) handles refinances after a value-add program completes. Equity ranges from regional sponsors and syndicators (the BiggerPockets register at the smaller end) through institutional LPs deploying via Sun Belt-focused private funds.

Garden works when land basis stays near or below roughly $15K per unit (the worked example below carries $3.0M of land on 8 acres, or $375K per acre, against a 200-unit plan), the site is large enough to spread 15–30 UPA, the metro supports professional single-family-to-multifamily rental flight (Sun Belt and secondary Midwest), and demographics support the rent band. It stops working on coastal urban infill where land is $80K+ per unit, in metros where parking ratios are regulated below 1.5 spaces per unit, and in submarkets where existing garden has aged into Class C and rent ceilings cap returns. Stabilized agency-financed garden in the 2026 Sun Belt is a low-cap, deep-debt, modest-IRR allocation that institutional capital buys for income and inflation protection, not for upside.

Tier 2: Wrap / Texas Donut (4–5 Stories, Type V over Type I Podium)

The wrap (or Texas donut) is the bridge between garden and mid-rise. Four to five stories of wood-frame apartments built around an above-grade concrete parking deck: the "donut" with the deck in the middle. Density runs 50–90 units per acre. Construction is Type V wood-frame for the residential floors over a Type I concrete parking structure; the prototype originated in Texas as a way to capture higher density on infill sites without paying the full structured-parking premium of a true podium. The prototype now ships in every Sun Belt metro and increasingly in secondary Midwest markets.

Cost basis runs $200–275 PSF in 2026; the lower end of the range captures suburban donuts in markets with low labor cost. The wood-frame residential floors are still cheap; the cost adder versus garden is the concrete parking structure and the deeper site excavation. Wrap rents PSF run roughly 1.2–1.4x garden in the same submarket because the higher density typically supports a more urban submarket placement; absolute per-unit rents typically come in at $1,500–1,950 in 2026 Sun Belt. Parking ratio is 1.2–1.5 spaces per unit, structured in the deck, often paid by residents as a separate ancillary at $50–125 per month per space.

Exit caps for institutional wrap in 2026 Sun Belt sit at 5.00–5.75%, about 25–50 bps tighter than the garden in the same metro because the urban-infill placement and structured-parking ancillary typically command modest cap compression. Target levered IRR is 13–16% on core-plus stabilized acquisitions; merchant-build development targets 17–21%.

The capital source for wrap is mixed and depends on stage. Agency debt finances stabilized wrap, sometimes at tighter pricing than equivalent garden if the property has a stabilized rent roll and 90-day trailing operations. Construction debt for ground-up wrap is dominated by regional banks and balance-sheet community bank lending, with HUD 221(d)(4) the underused higher-leverage, fixed-rate, non-recourse alternative for sponsors who can absorb the Davis-Bacon prevailing-wage requirement and a longer HUD approval timeline. The construction-debt market for wood-frame mid-rise has tightened as regional banks pulled back; life-co construction debt and dedicated debt funds have stepped into the gap with wider spreads (this article assumes SOFR + 300–450 bps). See our construction-loan deep-dive for the draw-schedule and interest-reserve mechanics that determine whether the cost basis actually pencils.

The 2026 overlay lands heavily on wrap. In the cost-overlay figure below, wrap carries about $28 PSF of insurance, labor, and steel adders. The fire-frame insurance surcharge falls on Type V construction throughout the residential floors, not on the Type I concrete podium. The data-center labor pull, which ConstructionDivereports is making limited electrical and mechanical craft capacity a gating factor on large projects, falls on trades that this article assumes wrap construction relies on. The pencil math has tightened.

Tier 3: Mid-Rise (5–7 Stories, Podium or Concrete)

The mid-rise tier covers two distinct construction types delivering similar density and rent profiles: the wood-frame podium (the iconic 5-over-1: five stories of wood-frame over a one-story Type I concrete podium for parking and ground-floor retail), and the concrete-or-mass-timber mid-rise (Type I-B at 5–7 stories, or Type IV-A/B/C mass timber). Both produce 80–120+ units per acre and serve urban infill submarkets that don't justify high-rise but warrant more than wrap density. Mixed-use ground-floor retail is common.

The 5-over-1 podium has been a dominant ground-up apartment delivery prototype in U.S. urban infill for the last fifteen years. Construction is Type III for five-story all-wood-frame buildings (exterior walls of noncombustible or fire-retardant-treated wood), or wood-frame (Type III-A or V-A, depending on jurisdiction and code edition) over a Type I podium for the podium variant. Wood framing is materially cheaper PSF than concrete or steel residential floor plates, which is what drove the prototype's economics: the savings on the residential floors offset the cost of the concrete podium and made 5-over-1 one of the highest-density wood-frame apartment prototypes the U.S. code allows.

Cost basis in 2026 runs $235–350 PSF across the mid-rise tier, with the lower end capturing suburban 5-over-1 ($235–265 PSF) and the upper end capturing urban podium with deeper structured parking and richer amenity ($300–350 PSF). The 2026 number reflects the cumulative effect of fire-frame insurance, steel and aluminum tariffs, and trades labor pressure on top of older cost levels. Wrap ($200–275 PSF) is a different prototype from podium; see the common-mistakes list below.

Concrete and mass-timber mid-rise (Type I-B and Type IV) run $280–380 PSF in 2026. Mass timber's pre-fab elements can shorten the vertical structure schedule relative to concrete on a comparable mid-rise envelope, which can partially offset any per-PSF premium through reduced interest carry and earlier rent commencement. The size of the premium and the schedule gain are project-specific, so take both from contractor pricing.

Rent PSF on mid-rise sits 30–60% above garden in the same metro (the worked example below uses $2.50 against $1.85, a 35% premium). 2026 Sun Belt urban infill mid-rise commands $1,800–2,400 per unit, $2.10–2.85 PSF. Parking ratio is 1.0–1.5 spaces per unit, structured, often paid as an ancillary at $100–200 per month per space. Amenity stack expands materially: pool deck, full fitness with peloton bikes, co-working lounge, rooftop terrace, package room, EV charging, dog wash, ground-floor retail in mixed-use. NOI margin runs 60–64% on stabilized urban infill mid-rise, slightly higher than garden because higher absolute rents dilute the fixed-cost base.

Exit caps run 4.75–5.50% in 2026 Sun Belt urban infill institutional; tighter to 4.50–5.00% in true gateway submarkets. Target levered IRR is 12–15% core-plus,15–20% merchant-build development. The merchant build/sell strategy behind the 2014–2022 supply wave was concentrated in this tier: merchant builders built and sold 5-over-1 podiums in the same Sun Belt urban submarkets, and the resulting supply is a headwind to near-term rent growth in the metros that absorbed the most podium deliveries.

Capital sources mix more than at any other tier. Stabilized mid-rise finances through agency debt (Fannie DUS, Freddie Optigo) once a 90-day trailing rent roll exists. Construction debt is balance-sheet bank (regional or money-center depending on size), with HUD 221(d)(4) as the higher-LTC alternative for sponsors who can navigate the process: our HUD 221(d)(4) deep-dive walks the underwriting and timeline mechanics. Bridge debt funds bridge transitional deals; insurance company permanent debt finances stabilized urban infill at competitive long-term fixed-rate execution. Equity is institutional LP commingled funds and merchant builders recycling equity through the build/sell cycle.

Tier 4: High-Rise (8+ Stories, Type I-A Concrete and Steel, Structured Parking)

High-rise multifamily is its own asset class. Eight or more stories in this article's tiering, Type I-A construction per the IBC (concrete columns and post-tensioned slabs is the most common U.S. high-rise residential system; steel with composite floors is the alternative on tall towers), 2–3 hour fire-resistive rating, no combustible structural elements. Density runs 140–300+ units per acre. Typical projects are 200–600 units in a single tower; signature towers reach 800–1,200 units. NCREIF's High-Rise cluster starts above 10 floors; 8–10 story product sits in NCREIF's Mid-Rise (5–10 floors), so this article's 8+ cutoff is broader.

Cost basis runs $400–675 PSF in 2026 across most institutional U.S. high-rise multifamily, with the $400–500 PSF range capturing typical Sun Belt urban core (Austin Domain, Charlotte SouthEnd, Nashville Gulch, Tampa Water Street). NYC, SF, Boston, Honolulu, and DC high-rise: $550–800+ PSF. For reference, RSMeans' 2026 guide puts 8–24 story apartment complexes at $250–700 PSF; the page's 70–80% hard-cost share implies roughly $175–560 PSF of hard cost (250 × 0.70 = 175; 700 × 0.80 = 560), which is below the top of this article's range; the top reflects gateway sites. ConstructionDive's 2026 outlook reports cement and concrete prices largely flat while steel and aluminum remain elevated on tariffs, with overall material costs forecast to inflate roughly 2% to 4% in 2026. This article sizes the steel-tariff adder on high-rise at an illustrative $22 PSF.

Construction speed is the slowest of the four tiers. Typical high-rise multifamily takes 24–36 months from groundbreaking to certificate of occupancy, with another 12–18 months for lease-up. The full development timeline from site control to stabilization is routinely 4–5 years on high-rise versus 30–36 months on mid-rise and 24–30 months on garden. The longer timeline compounds with the higher cost basis to produce materially larger interest reserves; the construction-loan interest-reserve math in our construction-loan article applies most acutely to this tier.

Rent PSF is the highest of the four tiers; the worked example below uses $3.75 against $2.50 for mid-rise, a 50% premium. In 2026, gateway high-rise routinely runs $3,000–5,000+ per unit ($3.75–6.50 PSF in places); Sun Belt urban high-rise runs $2,200–3,500 per unit ($2.75–4.50 PSF). Parking ratio is the lowest of the four tiers: 0.7–1.2 spaces per unit, structured(typically below-grade in gateway markets or wrapped above-grade in Sun Belt), paid as a meaningful ancillary at $150–300 per month per space in major markets. Amenity stack is full institutional: sky lounge, indoor pool, full fitness with spa and sauna, co-working floor, screening room, golf simulator, dog spa, 24/7 concierge, valet, rooftop bar (in mixed-use), retail on multiple floors.

NOI margin on high-rise typically runs 55–60%, lower than mid-rise because the amenity stack adds significant fixed OpEx (concierge headcount, pool deck maintenance, elevator service contracts on multiple high-speed elevators, façade cleaning, structural inspections, higher insurance for tall buildings) that absolute rent levels don't always overcome. Exit caps run 4.00–4.75% in gateway core (NYC, SF, LA, Boston, DC) for institutional high-rise; 4.50–5.25% in Sun Belt urban core. Target levered IRR is 10–13% core (stabilized acquisition), 16–22%development (merchant build), and 18–25%+ opportunistic (entitled-shell, partial-tower acquisitions).

The capital market for high-rise is the most constrained of any density tier in 2026. Agency caps and DSCR-driven sizing don't always support the high-rise cost basis on Sun Belt deals; gateway high-rise often exceeds agency's largest single-loan parameters. Construction debt is dominated by money-center balance sheet (JPMorgan, BofA, Wells Fargo, Citi, plus the Canadian banks at the upper end and the Japanese and Korean banks on select gateway deals), with a handful of life cos and dedicated debt funds active. HUD 221(d)(4) is technically available at high-rise scale but is rare in practice: the HUD process favors smaller transactions and the cost-basis threshold limits applicability. EB-5 capital appears in some high-rise capital stacks, particularly in NYC and Florida; mezzanine and preferred equity from real estate credit funds and bespoke balance-sheet allocators fill the gap between senior debt and common equity.

Common equity is dominated by institutional core funds (Blackstone, Brookfield, JPMorgan Asset Management, Ares), sovereign capital (Singapore GIC, Norges Bank, ADIA, CPP Investment Board), and the largest sponsors with permanent or near-permanent capital (Related, Tishman Speyer, Greystar's institutional vehicles, Equity Residential, AvalonBay). See our bank-debt recourse-vs-nonrecourse guide for the recourse mechanics that materially shape sponsor selection at this tier.

High-rise works on gateway urban core sites where land basis exceeds $200K per unit, rent achievable is at the top of the gateway range (the worked example below needs about $6.40 PSF), the capital partner has patient duration, and the merchant-build path has a deep pre-stabilization exit market or a build-to-core institutional hold. It doesn't work in Sun Belt sub-markets where land doesn't justify Type I-A premium over mid-rise, in markets where construction debt for high-rise is effectively closed, or in supply-wave-affected submarkets where the 24–36 month construction timeline crosses the lease-up cycle wrong.

Construction Type to Density Mapping

The relationship between density tier and IBC construction type is direct, though the boundary cases between Type III, Type V over Type I, and Type I-B can be ambiguous. The institutional mapping:

Construction-type to density-tier mapping. Story cutoffs follow this article's tiering; taller mass timber falls in the high-rise tier. UPA ranges are the author's working ranges.

The 2026 Construction Cost Overlay

The 2010–2022 cost ranges that older cost tables still cite are stale. Four cost pressures have moved through U.S. multifamily construction since 2023, and the cumulative effect is to lift cost in every tier and to move cost between the wood-frame and concrete tiers: insurance lands on wood, steel lands on concrete. The figure below sizes each adder by typology as an illustrative assumption. Reading the 2026 economics through 2022 cost data is a common error in underwriting briefs.

The fire-frame insurance surcharge. Builder's-risk coverage for wood-frame construction costs more than it did before 2023, and the increase falls on Type V construction; Type III mixed-construction is hit moderately, while Type I concrete and steel carry only a nominal adder ($1–2 PSF in the figure). This article assumes the surcharge adds $4–8 PSF on a $200 PSF wood-frame baseline, depending on construction duration and the specific carrier's underwriting. That is 2.0–4.0% of the wood-frame baseline ($4 / $200 = 2.0%; $8 / $200 = 4.0%). Insurance alone therefore raises wood-frame cost relative to concrete; steel and labor, below, push the other way.

The data-center labor pull. ConstructionDive's 2026 outlook reports that data center construction is expected to stay busy, with hyperscalers pushing to build billion-dollar facilities at pace, and that the limited capacity of qualified electrical and mechanical craft workers is becoming a gating factor as project scale grows. It also reports that labor costs are exerting far greater pressure on project budgets than materials. This article assumes multifamily competes for the same electrical, mechanical, and finish trades, so the labor adder is applied to every tier. The figure sizes it at $14–20 PSF across the tiers, with mid-rise wood ($18) above mid-rise concrete ($16) because wood-frame relies on framing, MEP, and finish trades.

The steel and aluminum tariff overhang. ConstructionDive's 2026 outlook reports that cement and concrete prices appear largely flat while steel and aluminum remain elevated on tariff-related impacts, with overall material costs forecast to inflate roughly 2% to 4% in 2026. The tariff overhang falls mainly on Type I-A high-rise and on steel-intensive Type I-B mid-rise. Wood-frame tiers carry only a small steel adder ($2–6 PSF in the figure: connector hardware, hold-downs, and the steel in any concrete podium or deck). Net effect in the illustration: the cost gap between high-rise and the wood-frame tiers stays wide, while mid-rise wood and mid-rise concrete each carry $32 PSF of total adders, so the dollar gap between them is unchanged.

The 5-over-1 stall. Fire-frame insurance, labor pull, and construction-debt market contraction have made the 5-over-1 prototype tougher to pencil, and the worked example below shows the arithmetic: at $245 PSF and a 5.00% exit cap, the 5-over-1 yields 4.16% on cost and loses $12.6M of value against cost. Where the math is this negative, starts tend to shift toward lower-density typologies.

The cost backdrop may be stabilizing rather than accelerating. ConstructionDive's 2026 outlook quotes a Wipfli partner forecasting material costs inflating roughly 2% to 4% in 2026, with labor exerting far greater pressure on project budgets than materials. The implication is that 2026 deliveries are entering a stabilization band for material costs, while labor remains the line to watch. The deals that pencil at 2026 cost basis are the ones moving forward; marginal deals that stalled in 2024–2025 need better rents or tighter exit caps to restart.

Worked Side-by-Side: 200 Units, Same Sun Belt Metro, Three Density Tiers

The cleanest way to see the density-tier decision is to model the same unit count and the same metro at three different prototypes. The table below works 200 units in an urban Sun Belt market (Austin, Charlotte, Nashville, Phoenix, or Tampa; the numbers are illustrative and not deal-specific) as garden, mid-rise (5-over-1 podium), and high-rise. Land basis, construction type, achievable rent, OpEx, and exit cap shift across the three columns. The shape is what matters: the absolute numbers will adjust for the specific metro and submarket.

200 units, same Sun Belt metro, three density prototypes. Illustrative figures; the shape (yield-on-cost compresses sharply as density rises; exit cap compresses modestly) is what matters for the prototype decision. NOI as a share of effective gross income (gross potential rent less vacancy, plus other income) is 59.8% / 60.5% / 54.8%.

The development margins are negative across all three tiers at 2026 cost basis and 2026 exit caps in this illustrative metro, which is a feature of the current environment, not the example. Cost basis has caught up to stabilized cap rates faster than rent growth has caught up to cost basis. The point isn't that all three tiers fail (real Sun Belt markets in 2026 vary significantly: Austin and Nashville pencil differently than Tampa or Charlotte; the metro choice matters as much as the tier choice). The point is that the shapeof the three columns reveals which tier is most defensive when cost basis runs ahead of rent growth.

Garden has the lowest absolute cost basis ($46.8M), the highest yield-on-cost (5.40%), and the smallest yield-on-cost-to-exit-cap gap (5.40% vs. 5.50% = −10 bps). The garden development margin in this illustration is essentially zero: the deal pencils to value-equals-cost in this metro. Mid-rise carries a meaningfully higher cost basis ($75.4M, 61% above garden), a 4.16% yield-on-cost, and a −84 bps gap to exit cap. The mid-rise pencils negative by $12.6M in this illustration. At the table's 8% vacancy, $0.50M of other income, $2.05M of OpEx, and 5.00% cap, break-even needs $5.78M of gross potential rent, about 13% above the $5.10M assumed, or roughly $2,410 per unit ($2.83 PSF) versus $2.50 PSF. High-rise carries a $185.8M cost basis (4x garden), a 2.22% yield-on-cost, and a −253 bps gap to exit cap. At the table's 8% vacancy, $0.90M of other income, $3.40M of OpEx, and 4.75% cap, break-even needs $8.82M of NOI and $12.31M of gross potential rent, about 71% above the $7.20M assumed, or roughly $5,130 per unit ($6.41 PSF) versus $3.75 PSF. That is the top of the gateway range, not a Sun Belt rent.

The institutional reading: in this illustration, garden is the only density tier that pencils close to break-even on development math at 2026 cost basis. Mid-rise pencils only where rent reaches roughly $2.83 PSF, the very top of the $2.10–2.85 PSF range for Sun Belt urban infill. High-rise pencils only in gateway-grade submarkets at rents near $6.40 PSF, well above the $3.75 PSF illustrated and above the $2.75–4.50 PSF Sun Belt urban high-rise range. Real metros vary, so rerun the table on local cost and rent comps; the shape of the three columns is the point.

Density Determines the Debt Market

The most under-appreciated implication of the density-tier decision is that it determines the debt market the deal will live in. Garden, wrap, mid-rise, and high-rise are not financed by the same lenders. The capital structure that fits one tier doesn't fit the next; the lender that loves your garden deal can't underwrite your high-rise, and vice versa. This is the bridge from the density decision to the capital-structure articles listed under Related Articles below.

Garden → agency dominant. Fannie Mae DUS and Freddie Mac Optigo are the deepest debt markets in U.S. multifamily, and they are most aggressive at the garden tier, where the performance record is longest and best documented and underwriting standards are most established. HUD 223(f) handles refinances after value-add programs complete. The garden tier is the only one where senior debt is genuinely abundant in 2026.

Wrap → bank balance sheet, with HUD 221(d)(4) as the higher-LTC alternative. Construction debt for wrap and Texas donut comes through regional and community banks, with HUD 221(d)(4) as the higher-LTC, non-recourse, fixed-rate alternative for sponsors who can absorb Davis-Bacon prevailing-wage and a longer HUD approval timeline. The HUD execution (covered in detail in our HUD 221(d)(4) deep-dive) is the underused alternative when bank construction-debt market constraints close standard execution. Stabilized takeout is agency, same as garden.

Mid-rise wood-frame → bank balance sheet or HUD 221(d)(4). Same pattern as wrap at higher cost basis. Mid-rise construction debt sits with regional banks for smaller deals and money-center banks for larger; HUD 221(d)(4) is the institutional alternative for the right sponsor. Bridge debt funds bridge transitional mid-rise deals. Insurance-company permanent debt is increasingly active at the mid-rise tier for stabilized urban infill on long-term fixed-rate execution. Pref equity and mezz fill the gap between senior debt (typically capped at 65–75% LTC on bank balance sheet for non-HUD execution) and the common-equity capitalization.

Mid-rise concrete and mass timber → same as wood mid-rise plus ESG capital. Same senior debt and subordinate stack as wood mid-rise, with a small but growing slice of institutional equity preferentially allocating to mass timber for ESG mandate compliance. Life-co construction debt has expanded share at the concrete and mass-timber mid-rise tier.

High-rise → constrained debt market. Construction debt is dominated by money-center balance sheet (JPMorgan, BofA, Wells Fargo, Citi at the U.S. end, with Canadian, Japanese, and Korean bank syndicates on gateway executions) and a handful of life cos and dedicated debt funds. HUD 221(d)(4) is technically available but rare at high-rise scale. EB-5 capital appears in some NYC and Florida high-rise stacks. Mezzanine and preferred equity from real estate credit funds and bespoke balance-sheet allocators fill the gap between senior debt and common equity. Our bank-debt recourse-vs-nonrecourse guide covers the recourse mechanics that materially shape sponsor selection at this tier: high-rise construction debt is routinely partial-recourse to creditworthy sponsors and rarely fully non-recourse outside HUD execution.

Common equity at the high-rise tier is dominated by institutional core funds (Blackstone, Brookfield, JPMorgan Asset Management, Ares), sovereign wealth (Singapore GIC, Norges Bank, ADIA, CPP Investment Board, Korea Investment Corporation), and the largest sponsors with permanent or near-permanent capital (Related, Tishman Speyer, Greystar's institutional vehicles, Equity Residential, AvalonBay, Camden Property Trust). The capital market is deep but selective: sovereigns underwrite gateway high-rise on 15–25 year hold horizons at IRR targets in the high single digits to low double digits; core funds underwrite on 10–15 year holds at low teens. The high-rise tier is the most institutional and the most concentrated at the top of the LP capital stack.

The Texas Donut Callout

The Texas donut (also called the wrap, or the "fat-five" in some markets) merits a dedicated section because it occupies a specific institutional niche between garden and podium that is easy to blur in a four-tier framework.

What the Texas Donut Is

The Texas donut is a four-to-five-story wood-frame apartment building constructed around an above-grade concrete parking deck: the "donut" with the parking deck as the hole in the middle. The residential units form a rectangular or oval ring around the deck on each floor. Stairs and elevators connect the ring to the parking levels. The prototype delivers 50–90 units per acre, places ~1.0–1.5 parking spaces per unit in structured deck (vs. 1.5–2.0 in surface garden), and fits on roughly a 1.5–3 acre infill site.

Origin and Spread

The prototype originated in Texas as developers looked for a way to capture higher density on urbanizing infill sites without paying the full structured-parking premium of a true podium. The above-grade deck wrapped by wood-frame residential delivers meaningful density at a lower cost basis than a true podium. The prototype spread through the Sun Belt and is now a common urban-infill apartment typology in Austin, Charlotte, Nashville, Phoenix, Dallas, Houston, Tampa, Raleigh, and increasingly Atlanta and Orlando. It also appears in secondary Midwest markets (Indianapolis, Columbus, Kansas City) and parts of the Mountain West (Denver, Salt Lake City, Boise, Reno).

2026 Economics

Cost basis runs $200–275 PSF in 2026. Density 50–90 UPA. Rent PSF roughly 1.2–1.4x equivalent garden in the same metro. Exit cap 5.00–5.75% Sun Belt. Target levered IRR 13–16% core-plus, 17–21% merchant-build. The prototype's structural appeal is the combination of urban-infill density with wood-frame cost economics, a position that mid-rise podium and high-rise can't replicate at the same basis. Its current vulnerability is the same as wood-frame mid-rise: the fire-frame insurance surcharge and the data-center labor pull are tightening the basis.

When the Texas Donut Wins

The Texas donut wins in submarkets that are urbanizing fast enough to support 50–90 UPA density but where rents don't yet support full podium cost basis: the in-between case where garden's surface parking doesn't fit the site and mid-rise podium overshoots the rent ceiling. Most of urban Sun Belt fits this pattern.

When You'd Build Down Rather Than Up

The default development heuristic from 2010 through 2022 was straightforward: if a site supports mid-rise, build mid-rise; if it supports high-rise, build high-rise. Higher density meant higher absolute rents, higher exit valuations, and a better merchant-build outcome. The 2026 math has shifted enough that the heuristic doesn't always hold. In a meaningful number of submarkets, the right answer on the same site is to build down rather than up: garden where the zoning supports mid-rise, mid-rise where it supports high-rise.

Insurance surcharges falling disproportionately on wood-frame raise its cost, but they don't erase the absolute cost advantage of wood-frame at lower densities: surface parking costs a small fraction of the $15,000–30,000 per space this article assumes for a structured deck. Structured-parking cost increases (steel rebar, post-tensioning labor) have raised the per-space cost of decks meaningfully, which falls on wrap and podium more than on garden.Supply-wave rent compression through 2024–2026 in the major Sun Belt urban infill submarkets has compressed the rent premium that mid-rise needs to justify its cost-basis premium over garden. Andthe agency-debt advantage of garden (75–80% LTV, 30–35 year amortization, IO on the front end, non-recourse with standard carve-outs) widens the levered IRR gap in garden's favor at lower yield-on-cost.

The combination can flip the prototype decision on a site that would have been mid-rise in 2018. A suburban infill parcel that would have penciled as mid-rise in 2018 can pencil more defensively in 2026 as garden at $175 PSF: the garden delivers fewer units on the same land, but at a cost basis that needs only about 1% more rent to break even in the worked example, and on agency debt that levers the equity check more efficiently. The choice depends on the specific submarket: in genuine urban-infill core (Austin downtown, Nashville Gulch, Charlotte SouthEnd at the right block), mid-rise still wins when the rent achievable is genuinely 30%+ above garden in the same submarket. In transitional submarkets (Charlotte University, Phoenix North, Dallas Frisco edge), the math can flip.

Practitioners who learned the 2018-vintage heuristic need to update it for the current cycle. The mid-rise default is no longer the right default.

NCREIF Building-Type Sub-Index Shape

NCREIF's NPI includes apartment properties and sub-types within each property type. This section is a qualitative reading of how garden, low-rise, mid-rise, and high-rise apartment product has behaved across cycles; it is not a computed result, so pull the current series from NCREIF before relying on any of it. What follows is a common expectation, not a result from NCREIF data:

• Garden is generally expected to be the more defensive sub-type: higher rent volatility (more concession activity, more turnover) but lower price volatility (less cap-rate compression and decompression) than high-rise, possibly helped by agency-debt cost advantage and a simpler operating model.

• Mid-Rise is generally expected to sit between the two, with more exposure than garden to urban-infill cap-rate and rent-growth swings. That is a profile the 2024–2026 environment is likely to test, as supply-wave rent compression and cost-basis inflation meet higher exit caps.

• High-Rise is expected to show the highest price volatility (most cap-rate sensitivity, most gateway-market beta) and lower NOI margin than mid-rise but the highest absolute rent levels, and it is the sub-type most sensitive to interest-rate moves.

The expectation is that density correlates with rent level and exit-cap sensitivity but not necessarily with total return: garden is expected to be the most defensive allocation mid-cycle, high-rise the most sensitive to cap-rate moves in either direction, and mid-rise generally expected to sit between the two. Treat these as hypotheses to test against the NCREIF series, not as findings.

From Density Choice to Underwrite

The density decision sets the prototype; the prototype sets the underwriting model. An acquisitions team screening stabilized multifamily across the four density tiers needs a model that handles agency-debt sizing for garden, balance-sheet construction debt sizing for wrap and mid-rise, HUD 221(d)(4) for the underused alternative, and money-center bank construction plus mezz/pref stacking for high-rise. The capital structure varies tier-by-tier; the underwriting model has to vary with it.

On the acquisitions side, the workflow is screen-many, underwrite-few. The pocket model that handles all four density tiers without rebuild (agency, balance-sheet, HUD, and money-center execution all in the same sheet) is what turns the density choice into a five-minute screening decision. On the development side, the workflow is site-specific: which prototype, at what density, with which capital stack, at what cost basis, with what rent and exit assumption. The pro forma that handles all four density tiers on a single site with a side-by-side comparison is what turns prototype selection from a six-week exercise into a one-day decision.

Seven Mistakes Practitioners Make

• Reading 2026 economics through 2022 cost data. Cost tables built before the 2023–2026 adders understate the 2026 basis. The overlay figure above sizes those adders at $22–43 PSF by tier (illustrative), and RSMeans' 2026 guide puts 1–3 story apartment complexes at $220 PSF and up (about $154–176 PSF of hard cost at the 70–80% hard-cost share on the same page, a loose cross-check that falls in the lower half of this article's garden range). Pull current local cost data before basing a development prototype decision on any published cost table.

• Underwriting mid-rise rents off 2018–2022 cap-rate compression. The 2014–2022 podium pencil math leaned on cap-rate compression at exit and steady rent growth. Neither is a safe assumption in 2026. In this article's example, the 5-over-1 at $245 PSF and a 5.00% exit cap yields 4.16% on cost and loses $12.6M of value against cost. The same prototype, very different underwrite.

• Defaulting to mid-rise because the zoning supports it. The "build the highest density the zoning allows" heuristic is wrong in 2026 in transitional submarkets. Garden at agency-debt economics and lower cost basis often delivers a more defensive levered IRR than mid-rise at the same site, even when zoning supports both.

• Bringing the wrong lender for the tier. Money-center bank construction lenders aren't going to underwrite a $35M garden deal at competitive terms; agency lenders aren't going to take the lease-up risk on a ground-up high-rise. The capital source map for the density tier has to be set before lender outreach starts. Garden → agency. Wrap and mid-rise → bank balance sheet or HUD 221(d)(4). High-rise → money-center bank plus mezz/pref/sovereign equity.

• Ignoring the fire-frame insurance line in wood-frame underwriting. Wood-frame builder's-risk premiums are a recurring cost line, not a 2023 anomaly. Underwriting wood-frame at 2022 insurance assumptions understates the carrying cost; this article assumes $4–8 PSF, or 2.0–4.0% of a $200 PSF wood-frame baseline.

• Treating wrap and podium as interchangeable. A 4-story wrap at 50–90 UPA and a 5-over-1 podium at 80–120 UPA are different prototypes with different cost basis, different parking economics, and different exit caps. This article prices wrap at $200–275 PSF and podium at $235–350 PSF; the gap reflects the structural and parking-deck differences.

• Underwriting high-rise to Sun Belt exit caps at gateway construction timelines. The 24–36 month construction period for high-rise plus 12–18 month lease-up makes the prototype most sensitive to mid-construction rate and rent shocks. Sun Belt high-rise in 2024 vintage routinely delivered into supply-wave compression and missed underwriting; the 2026 vintage that pencils is the one that underwrites to flat-to-expanding exit cap and conservative lease-up rent.

Do It in Apers

Related Articles

• Rental Property Pro Forma Calculator: the underwriting model that runs underneath every density-tier decision.
• Multifamily Development: Construction and Lease-Up: the development-side companion that walks construction execution and stabilization.
• Multifamily Underwriting Fundamentals and Rent Roll Analysis: the rent-roll mechanics that drive achievable rent PSF assumptions in this article.
• Construction Loans: Draw Schedules and Interest Reserves: the draw-schedule and capitalized-interest mechanics for any ground-up density tier.
• HUD 221(d)(4) Construction-to-Permanent Loans: the higher-LTC alternative that finances wrap and mid-rise outside the bank construction market.
• Bank Debt: Recourse vs. Non-Recourse and Covenants: the recourse mechanics that shape sponsor selection at the high-rise tier.

FAQ