← The Underwriting Desk
MultifamilySan Francisco, CA · 12 units · Built 1916

Edgewater Flats: 12-unit multifamily sample, San Francisco, CA

32.8% loss-to-lease ($151,608/yr) is concentrated in long-tenured, likely rent-controlled units that present minimal near-term reversion upside.

Units
12
Year built
1916
Asset class
Multifamily
Market
San Francisco, CA

Key findings

Scheduled Rent (Annual)
$311,016
Sum of 12 in-place unit rents × 12; no ancillary income, expenses, or vacancy applied.
Loss-to-Lease
$151,608/yr (32.8% of market)
Gap between in-place and market rents is material; reversion path is constrained by SF rent control.
Long-Tenure Units (20+ yrs)
3 units @ avg $1,098/mo in-place
Market rent avg for this cohort is $2,578/mo — a $1,480/unit/mo gap unlikely to close without vacancy.
Lease Rollover Exposure
0 leases expiring in 0–12 mo
No near-term contractual roll; all lease-end data is either expired, indeterminate, or beyond 13 months.
Rent Range (Min–Max)
$729 – $8,828/mo
Wide dispersion (mean $2,160, median $1,641) signals a mixed tenure and unit-type profile.
Occupancy
12/12 units occupied (100%)
Full occupancy is consistent with SF market tightness but removes any immediate mark-to-market lever.

Analysis

## Rent Position and Loss-to-Lease The 12-unit rent roll produces a scheduled rent of **$311,016/year** ($25,918/month). Against market rents, the portfolio carries **$151,608/year of loss-to-lease**, equal to 32.8% of market — a figure that is large in absolute terms but must be evaluated against the realistic reversion timeline given SF rent-control constraints. ## Tenure Band Analysis and Reversion Probability The loss-to-lease is not evenly distributed. The **3 units with 20+ years of tenure** average $1,098/month in-place against a $2,578/month market average — a per-unit gap of roughly $1,480/month. These units are almost certainly subject to SF Rent Ordinance annual allowable increases (historically 1–3%), meaning reversion to market is effectively gated on voluntary vacancy or owner-move-in proceedings, both of which carry legal and timeline risk. The **6 units in the 3-to-10-year cohort** average $1,796/month in-place versus $2,705/month market — a $909/unit/month gap that is also largely protected but represents a more moderate drag. The **2 units in the 10-to-20-year band** are the outlier: at an average in-place rent of $5,220/month against a $6,580.50 market average, these units are above the portfolio median and suggest larger-format or premium units where the absolute gap ($1,360/month combined) is real but the in-place rent is already at a level that limits distress. One unit has an unparseable move-in date, which prevents tenure classification and should be resolved in due diligence. ## Rollover and Cash Flow Stability The rollover table shows **zero leases expiring in the 0-to-12-month window**, which means there is no near-term contractual mechanism to reset rents to market. This cuts both ways: scheduled rent is stable in the short term, but the mark-to-market thesis depends entirely on natural turnover, which in a 100%-occupied, long-tenured SF building is historically low. The absence of any lease expiration data in the 0-to-13-month buckets suggests leases may be month-to-month (common post-expiration in SF) or that lease-end dates were not captured — either scenario warrants clarification. ## Underwriting Considerations The **$729/month minimum rent** and **$8,828/month maximum rent** reflect a 12x spread within a 12-unit building, which is atypical and suggests the rent roll may include a mix of deeply controlled legacy tenants and one or more recently vacated/re-leased units at or near market. The mean-median gap ($2,160 mean vs. $1,641 median) confirms the distribution is right-skewed by the high-rent unit(s). Underwriters should not model loss-to-lease burn-down on any accelerated schedule without vacancy assumptions that are legally and operationally supportable under SF law. Scheduled rent of $311,016 is the appropriate baseline; any upside scenario requires explicit turnover rate assumptions, not a simple mark-to-market adjustment.
Anonymized sample. Property name, address, and tenant identifiers are synthetic. Numeric figures jittered ±5% from a real source rent roll.