Executive summary
- Question
- How much additional density still improves residual land value once construction, affordability, entitlement, OpEx, and absorption are counted?
- Site
- 344 14th Street, Mission District, San Francisco — UMU / 58-X infill near 16th Street BART
- Proposed program
- 108-unit studio-heavy rental benchmark · 1,449 SF retail · Type IIIA over IA · density-bonus height to 69 ft
- Critical constraint
- Hard cost: a $50/SF increase (~$525 → $575) wipes roughly $5.1M of residual land value
- Financial result
- Stabilized NOI $3.30M · value $70.1M at 4.70% exit · residual +$3.06M after required development profit · yield on cost 5.17%
- Conclusion
- Density paid in the modeled bonus scheme vs by-right, but only while the project stays in a wood mid-rise cost regime; unit mix and entitlement remain open variables
Thesis. Additional density creates value only while its added revenue outweighs the incremental construction, affordability, entitlement, operating, and absorption burden required to deliver it. The question is not whether density helps. It is how much density still pays.
The site
344 14th Street is a deep infill site between Stevenson and Woodward Streets near the western edge of the Mission. The parcel is zoned UMU / 58-X and sits within walking distance of 16th Street BART. The problem is how much additional density still improves residual land value once the full cost stack is counted.
What the market supported
The market analysis pointed clearly toward rental rather than for-sale housing. Approximately 77.8% of Mission households rent; 44.4% of residents are age 25 to 44; and median household income is about $161,000. Multifamily vacancy was 3.3%, while the submarket had added virtually no new rental supply since 2023.
Near-term competition was limited on the ground: only 304 competing units under construction, against another 1,522 units proposed. Low vacancy and limited deliveries supported new rental product; the larger proposed pipeline still created future absorption risk.
From that evidence, the class study used roughly 100-110 units as a prudent near-term program scale: studio/1BR-weighted rental, plus one small neighborhood-serving retail bay. That band is not a demonstrated maximum level of demand. It is the conservative scale the available market evidence supported for first-pass underwriting. RealPage institutional comps averaged about $3,339 studio / $3,992 one-bedroom / $5,290 two-bedroom.
Testing unit economics
RealPage observations from ten institutional rental properties showed rent increasing less than proportionally with unit size. A log-log specification estimated a size elasticity of 0.64; the pooled model produced an R² of 0.72, while the building fixed-effects specification produced an adjusted R² of 0.89. That supports the rent-per-square-foot economics of smaller units. It does not prove studios are optimal.
Market conclusion: studio/1BR-weighted rental. Class design: a studio-heavy benchmark (74 studios, 10 one-bedrooms, 24 two-bedrooms) used to test smaller-unit strategy economics.
Benchmark scheme
Type IIIA wood over Type IA podium, no required parking. Staying mid-rise wood was part of the strategy: added density is most valuable while the project can avoid a substantially more expensive construction regime.
- 108 rental units (74 studios, 10 one-bedrooms, 24 two-bedrooms)
- 18 BMR units (15 at 55% AMI, 3 at 80% AMI)
- 1,449 SF ground-floor retail
- Six stories / 69 ft via a density-bonus height waiver above the 58-X limit · ~84,185 design GFA · ~79% efficiency
Ground floor: retail bay, walk-up loft studios, family units at grade, and an interior courtyard. Nine ground-floor studios are walk-up lofts with stoops on 14th.
More exhibits (program, unit mix, elevation, rear axon)
Economics
Under the benchmark underwriting, the 108-unit density-bonus scheme produced the following outputs. Read them as linked layers: income → value → land residual → cost sensitivity → return implication.
Stabilized NOI
| Stabilized NOI | $3.30M |
|---|
Net operating income after modeled vacancy and operating expenses at stabilization.
Stabilized value
| Exit cap | 4.70% |
|---|---|
| Stabilized value | $70.1M |
Residual land value
| Hard cost basis | $525 / SF |
|---|---|
| Residual land value after required development profit | +$3.06M |
| Static yield on cost | 5.17% |
Residual land value is the amount the model can support for land after accounting for development costs and the required development-profit hurdle. Yield on cost is the static income return implication of the same underwriting—not a levered IRR claim.
Under the modeled assumptions, the by-right case generated a negative residual. It had fewer revenue-generating units and a less favorable modeled unit mix, while many fixed development costs remained. The density-bonus scheme added units, supported a more revenue-efficient program, and spread many shared development costs across a larger revenue base while remaining within a wood-frame mid-rise strategy.
Hard-cost sensitivity
Construction cost remains one of the strongest findings. Holding a 4.70% exit cap:
| Hard cost | Residual land value |
|---|---|
| $525/SF | +$3.06M |
| $550/SF | +$0.52M |
| $575/SF | -$2.02M |
A $50/SF increase in hard cost eliminates roughly $5.1M of residual land value.
Incremental density is most valuable when it can be added without a disproportionate jump in construction cost. The implication is not that maximum density always wins. Density is highly valuable up to the point where its marginal cost — construction, additional affordability, entitlement complexity, operating cost, and absorption risk — begins to exceed its marginal revenue.
Conclusion
What the analysis showed. On the modeled assumptions, the density-bonus 108-unit benchmark produced a positive residual land value (+$3.06M) and a $70.1M stabilized value, while the by-right case did not clear the land residual hurdle. Studio / 1BR weighting was supported by submarket rents and by a hedonic size elasticity below 1.0. Hard cost dominated sensitivity: modest $/SF moves flipped residual from positive to negative.
What would need to change to make the project more compelling. A next pass should treat density, unit mix, rear-yard strategy, height, turnover-adjusted OpEx, delivery-year rents, and bonus capacity as variables—not fixed class assumptions—and look for the strongest risk-adjusted residual while staying inside a wood mid-rise cost regime if possible.
What this taught about density and value. Density is not automatically valuable. It pays while incremental revenue outruns the full incremental cost stack, and it collapses quickly when construction cost steps up faster than rent can follow.
What remains open
108-unit benchmark · positive modeled residual · strong hard-cost sensitivity
Rental · studio/1BR weighting · conservative ~100-110-unit first-pass scale
Unit mix · density · rear-yard strategy · height · turnover-adjusted OpEx · delivery-year rents · bonus capacity
The current study does not assume that the existing BMR mix earns the full available bonus, that a rear-yard waiver will be granted, or that a density-bonus height waiver to 69 ft is assured. A 58-foot envelope also does not resolve all setback constraints on its own.
UC Berkeley academic case study (2026). Historical / educational note, not an offering, appraisal, solicitation, or advice to buy or develop any property. Market figures and underwriting are as of the study period and may not reflect current conditions. Comp rents from RealPage institutional observations; treat as study inputs, not a live rent forecast.