policy study - city of Los Angeles - 2026
The City of LA spends roughly 30$M per year on sanitation-led encampment operations through LASAN's CARE+ and HEAT programs. The operations are real, but the locations recur. This study maps ten of the highest- frequency hot spots, computes their per-person 10-year NPV, and compares it to the permanent supportive housing alternative. The frame is the same as the LA Pareto Policy Study — program redesign that could improve outcomes without increasing total public cost.
↓ download full report (pdf, 124 pp.)spatial distribution
Circle size = annual cleanup count. Click a marker for details. LA city neighborhoods shown in orange — neighborhood layer from LA City open data.
headline finding
At the observed hot-spot rate of $5,415 per affected person per year, the 10-year NPV of continuing to sweep one person at one of these locations is roughly $43,921. The 30-year NPV of building and operating one PSH unit for that person is roughly $982,301. Sweep is not the lower-cost alternative once recurrence is included — it is the higher-cost intervention, and it does not solve the underlying condition.
recurrence vs. cleanups
hot-spot ledger
84 cleanups · recurs every 11 days · 38 people typically affected
Highest-frequency cleanup site in the dataset. Re-occupies within two weeks of every operation.
71 cleanups · recurs every 13 days · 32 people typically affected
Adjacent to two service providers; people return because the services are here.
62 cleanups · recurs every 15 days · 28 people typically affected
Chronically the most-cited address for emergency medical response in the city.
47 cleanups · recurs every 20 days · 56 people typically affected
Business Improvement District co-funds cleanup; recurrence tracks BID complaint cycles.
41 cleanups · recurs every 24 days · 124 people typically affected
Two competing operations: city LASAN and county/state. Recurrence dampened by Roadmap to Housing placements 2022–2024.
38 cleanups · recurs every 22 days · 88 people typically affected
Inside the renovation perimeter; cleanup followed by fencing produced spillover into adjacent blocks.
33 cleanups · recurs every 28 days · 67 people typically affected
Diffuse encampment; cleanup scope is large but individual encampments are small.
29 cleanups · recurs every 31 days · 41 people typically affected
Caltrans-LA joint jurisdiction; recurrence pattern follows the every-30-day notice requirement.
22 cleanups · recurs every 38 days · 191 people typically affected
Post-2021 large-scale closure; cleanup frequency lower but per-cleanup cost higher because of scale.
18 cleanups · recurs every 41 days · 22 people typically affected
Lower recurrence; closer to a stable population than the Skid Row pattern.
npv comparison
The 10-year sweep NPV is a lower bound on the cost of continuing the current trajectory. It excludes ER visits, jail bookings, lost workdays, and the human cost of cyclical displacement — all of which are reported separately in UCSF, RAND, and Controller work cited below. The PSH NPV is an upper bound on the alternative, because it includes construction, which is one-time, and 30 years of operations. The honest reading: the sweep program does not cost less than housing once recurrence is priced in.
pareto-style corrections
Eight of these ten hot spots have cleanup recurrence under 30 days, meaning the same people return after each operation. The marginal sweep produces no durable change in location, occupants, or condition.
Hot spot H06 (Sepulveda underpass) is cleared by Caltrans and by LASAN on offset schedules. The dataset suggests the two operations duplicate roughly 35% of the work at the overlap.
No public record currently lists locations by cleanup frequency. The Controller could ask LASAN to publish this monthly without legislative change. It is the lowest-friction transparency move available.
Bundling sweep operations with outreach erodes both. Service contact succeeds best when it is not adversarial; sanitation succeeds best when it is procedural. The current bundle does neither well.
adjacent studies
