# Commuter vs. Resident Analysis — Redlands & Surrounding Cities

**Project:** Redlands Health strategy (personal side project; not a class assignment)
**Question:** Are Redlands and its neighbors *local/self-contained* communities (people live and work in the same city) or *commuter* communities (jobs filled by people who live elsewhere, and/or residents who commute out)? The purpose is to separate **employment-based covered lives** (people who *work* at a Redlands employer) from **residence-based lives** (people who *live* in the service area), because network efficiency — time/distance, where people actually seek longitudinal care — is governed by where people **live**, not where they work.

**Prepared:** June 2026
**Companion data file:** `commute_flows_by_city.csv` (same folder)

---

## 1. Data sources and method (every figure is labeled to source + year)

Two independent public datasets, used together so each cross-checks the other:

1. **U.S. Census LEHD LODES8 (Origin–Destination, Workplace, Residence Area Characteristics), data year 2022, segment S000 (all jobs), job type JT01 (primary jobs).**
   - Flat files downloaded directly from `https://lehd.ces.census.gov/data/lodes/LODES8/ca/` (od/, wac/, rac/) plus the 2020-block geographic crosswalk `ca_xwalk.csv.gz`.
   - Census blocks were aggregated to **place** geography (incorporated city / CDP) using the crosswalk's `stplc` field. This is the same aggregation OnTheMap's "Area Profile" performs; I used the flat files because the Census ACS API now requires a registered key and the public OnTheMap web API is not an openly documented JSON endpoint (see §8, Limitations).
   - LODES gives the **count-based** flow picture: jobs located in a city, how many of those jobs are held by people who also live there (live+work), in-commuters, employed residents, residents who work locally, out-commuters, and the **jobs–housing ratio**.
   - LODES caveats (Census-stated): based on UI-covered employment (excludes most federal, military, self-employed, and informal jobs); a single "primary job" per worker; counts are noise-infused for confidentiality. Treat all LODES figures as **(est)**.

2. **American Community Survey (ACS) 2024 5-year estimates (2020–2024), tables B08303 (travel time), B08013 (aggregate travel minutes), B08007 (place of work: in-/out-of-county), B08008 (worked in vs. outside place of residence), B08301 (means of transportation incl. work-from-home).**
   - Pulled at **place** level via the Census Reporter API mirror of the Census API (`api.censusreporter.org`, release `acs2024_5yr`), which serves the official ACS estimates without an API key. For Redlands I also pulled the **ACS 2024 1-year** as a freshness cross-check.
   - ACS gives the **resident-worker** picture: how long residents commute and whether they work inside or outside their own city/county. **ACS place-of-work tables (B08007/B08008) count only workers who did NOT work from home**; "% worked in place of residence" is therefore a share of *commuters*, not of all employed residents. All ACS figures carry sampling margins of error; treat as **(est)**.

**Important interpretive note on the two "work locally" numbers.** LODES "% of residents who work locally" (19.0% for Redlands) is *residents-working-in-city ÷ all employed residents*. ACS "% worked in place of residence" (39.2% for Redlands) is *residents-working-in-city ÷ commuters who didn't WFH*, and ACS place-of-work coding is self-reported and coarser. They measure related but different things; both are reported below and in the CSV, each labeled to its source. The **LODES count-based figures are the backbone of the strategic conclusions**; ACS is used for commute time and the in-/out-of-county split.

**Place FIPS codes used (authoritative, from `national_place2020.txt`):** Redlands 0659962 · Yucaipa 0687042 · Loma Linda 0642370 · Mentone CDP 0646884 · Calimesa 0609864 · San Bernardino 0665000 · Highland 0633588 · Colton 0614890 · Grand Terrace 0630658 · Beaumont 0604758 · Banning 0603820. (Cherry Valley CDP 0612916 was in scope but is in the Beaumont/Calimesa orbit; it is not separately tabulated below because it adds little to the service-area read — noted in §8.)

---

## 2. Headline table (LODES8 2022 jobs/flows + ACS 2024-5yr commute)

All counts **(est)**. Jobs-housing ratio = jobs in city ÷ employed residents (>1 = net in-commute employment importer; <1 = bedroom community/out-commute).

| City | Jobs in city | Live+work | Live+work % of jobs | Inflow (work-in/live-out) | Inflow % | Employed residents | Residents working locally % | Out-commuters | Out-commute % | Jobs-housing ratio | ACS mean commute (min) | Verdict |
|---|--:|--:|--:|--:|--:|--:|--:|--:|--:|--:|--:|---|
| **Redlands** | 39,161 | 5,836 | **14.9%** | **33,325** | **85.1%** | 30,755 | 19.0% | 24,919 | 81.0% | **1.26** | 24.9 | **In-commute employment center** |
| Loma Linda | 22,175 | 2,157 | 9.7% | 20,018 | 90.3% | 10,238 | 21.1% | 8,081 | 78.9% | **2.15** | 19.6 | In-commute employment center (extreme) |
| San Bernardino | 119,270 | 17,511 | 14.7% | 101,759 | 85.3% | 81,766 | 21.4% | 64,255 | 78.6% | **1.45** | 29.7 | In-commute employment center |
| Colton | 24,230 | 1,325 | 5.5% | 22,905 | 94.5% | 22,722 | 5.8% | 21,397 | 94.2% | 1.06 | 27.5 | Balanced count, high churn (pass-through) |
| Banning | 4,858 | 1,008 | 20.7% | 3,850 | 79.3% | 10,455 | 9.6% | 9,447 | 90.4% | 0.46 | 32.6 | Out-commute bedroom community |
| Beaumont | 8,189 | 1,510 | 18.4% | 6,679 | 81.6% | 20,995 | 7.2% | 19,485 | 92.8% | 0.39 | 35.7 | Out-commute bedroom community |
| Yucaipa | 7,931 | 2,388 | 30.1% | 5,543 | 69.9% | 22,052 | 10.8% | 19,664 | 89.2% | 0.36 | 32.6 | Out-commute bedroom community |
| Mentone CDP | 1,324 | 119 | 9.0% | 1,205 | 91.0% | 3,963 | 3.0% | 3,844 | 97.0% | 0.33 | 26.9 | Out-commute bedroom community |
| Grand Terrace | 2,325 | 150 | 6.5% | 2,175 | 93.5% | 6,003 | 2.5% | 5,853 | 97.5% | 0.38 | 25.6 | Out-commute bedroom community |
| Highland | 6,027 | 798 | 13.2% | 5,229 | 86.8% | 23,037 | 3.5% | 22,239 | 96.5% | 0.26 | 30.4 | Out-commute bedroom community (extreme) |
| Calimesa | 2,048 | 129 | 6.3% | 1,919 | 93.7% | 4,090 | 3.2% | 3,961 | 96.8% | 0.50 | 35.8 | Out-commute bedroom community |

*Source: jobs/flows = LEHD LODES8, JT01/S000, 2022 (est); mean commute = ACS 2024 5-yr, B08013÷B08303 (est).*

**No city in the study area is "self-contained."** Even the strongest live+work city (Yucaipa, 30.1% of its few local jobs held by residents) sends 89% of its own employed residents *out* to work. The dominant pattern across the entire eastern San Bernardino Valley is **separation of home and work** — exactly the condition under which an employment-based covered-lives count overstates the residentially-efficient market.

---

## 3. Redlands in depth

### 3a. As a workplace (the in-commute story)
- **39,161 primary jobs** are located in Redlands (LODES8 2022, est). This squares with the City's published figures: the FY2025 ACFR / EDD cite ~37,000 city employees, and the top-10 employers alone (Redlands USD 2,433; ESRI 2,416; Redlands Community Hospital 1,847; Optum/Beaver 1,000; City 539; Amazon 526; U. of Redlands 471; LLU Behavioral 458; etc.) total ~10,400. LODES' 39,161 also captures the thousands of smaller employers the ACFR doesn't itemize.
- Of those 39,161 jobs, **only 5,836 (14.9%) are held by people who also live in Redlands.** The other **33,325 (85.1%) are filled by in-commuters** who live outside the city.
- **Jobs-housing ratio 1.26** → Redlands is a **net employment importer**: it has 1.26 jobs for every employed resident. It pulls workers in.

### 3b. As a place to live (the out-commute story)
- **30,755 employed residents** live in Redlands (LODES8 2022, est).
- Only **5,836 of them (19.0%) work in Redlands.** **24,919 (81.0%) commute out** to jobs in other cities.
- ACS agrees in direction: among Redlands commuters who didn't WFH, **39.2%** report working in Redlands and **60.8%** work elsewhere (ACS 2024 5-yr, B08008, est); **82.9% work within San Bernardino County, 16.9% commute to another county** (B08007, est). Mean commute **24.9 min** (5-yr) / **24.3 min** (2024 1-yr) — moderate and the **shortest of all the bedroom communities**, reflecting that many residents who *do* stay close work in adjacent San Bernardino/Loma Linda.
- **Work-from-home 13.8%** of Redlands workers (ACS 2024 5-yr, B08301, est) — the highest in the study set, consistent with ESRI's software/professional workforce. WFH residents are, for care-access purposes, *resident lives*.

### 3c. Where the in-commuters live (top home locations of people who WORK in Redlands)
LODES8 2022 origin-destination, est. Total in-commuters 33,325.

| Home location of Redlands worker | Workers | % of in-commuters |
|---|--:|--:|
| San Bernardino city | 2,928 | 8.8% |
| Yucaipa | 2,238 | 6.7% |
| Highland | 1,746 | 5.2% |
| Riverside | 1,665 | 5.0% |
| Moreno Valley | 1,616 | 4.8% |
| Unincorporated San Bernardino County (incl. Citrus Plaza/Mountain Grove "donut hole", Mentone-adjacent pockets) | 1,348 | 4.0% |
| Beaumont | 1,071 | 3.2% |
| Fontana | 991 | 3.0% |
| Los Angeles | 896 | 2.7% |
| Loma Linda | 861 | 2.6% |
| Rialto | 851 | 2.6% |
| Colton | 814 | 2.4% |
| Mentone CDP | 590 | 1.8% |
| Rancho Cucamonga | 579 | 1.7% |
| Jurupa Valley | 527 | 1.6% |
| **Top-15 subtotal** | **18,721** | **56.2%** |
| Dispersed tail (all other places, each individually small) | 14,604 | 43.8% |

**Read:** the in-commute workforce is **geographically dispersed and reaches well past the efficient service area.** Grouping the top-15 by drive-efficiency to Redlands/RCH:
- From **near / efficient service-area cities** (San Bernardino, Yucaipa, Highland, Loma Linda, Colton, Mentone, Grand Terrace, and the immediately-adjacent unincorporated donut hole): **~10,525 workers ≈ 31.6% of inflow.**
- From **named-but-distant** cities in the top-15 (Riverside, Moreno Valley, Beaumont, Fontana, Los Angeles, Rialto, Rancho Cucamonga, Jurupa Valley): **~8,196 ≈ 24.6%.**
- From the **long dispersed tail** beyond the top-15: **~14,604 ≈ 43.8%**, almost all of it outside any RCH-centered narrow network.

So **roughly two-thirds of the people who work at Redlands employers live somewhere that is *not* efficient to serve with a Redlands-anchored longitudinal network.**

### 3d. Where the out-commuters work (top work destinations of Redlands RESIDENTS)
LODES8 2022, est. Total out-commuters 24,919.

| Work location of Redlands resident | Workers | % of out-commuters |
|---|--:|--:|
| San Bernardino city | 4,266 | 17.1% |
| Loma Linda | 2,371 | 9.5% |
| Riverside | 1,703 | 6.8% |
| Unincorporated San Bernardino County (donut hole etc.) | 1,236 | 5.0% |
| Los Angeles | 1,191 | 4.8% |
| Moreno Valley | 750 | 3.0% |
| Colton | 722 | 2.9% |
| Fontana | 698 | 2.8% |
| Rialto | 594 | 2.4% |
| Ontario | 569 | 2.3% |
| Yucaipa | 519 | 2.1% |
| Rancho Cucamonga | 498 | 2.0% |
| San Diego | 420 | 1.7% |
| Highland | 324 | 1.3% |
| Irvine | 300 | 1.2% |

**Read:** Redlands residents who leave town for work overwhelmingly stay close — **San Bernardino (17%) + Loma Linda (9.5%) + the donut hole (5%) + Colton/Highland/Yucaipa ≈ 35%+ work within ~15 minutes of home.** Their insurance comes from out-of-Redlands employers, but **their bodies are at home in Redlands evenings, weekends, and for any non-occupational care.** These are *efficient* lives to own.

---

## 4. Verdict per city

- **Redlands — net in-commute employment center that is also a desirable residential city.** JH 1.26; 85% of its jobs filled from outside; but also 81% of its own workforce commutes out. It is simultaneously a *daytime job magnet* (ESRI, the hospital, the district, Optum, U. of Redlands, Amazon) **and** a *residential community* whose own workers scatter to jobs all over the Inland Empire. The two populations — daytime workers and resident workers — are largely **different people in different places.**
- **Loma Linda — extreme in-commute employment center (JH 2.15).** The medical center / university / VA complex imports >20,000 workers for ~10,000 employed residents. Shortest commute (19.6 min) because so many residents work in-town at LLU. A clinical-employment island.
- **San Bernardino — large in-commute employment center (JH 1.45)** but with the region's longer commutes; the county-seat job base.
- **Colton — pass-through / balanced churn (JH 1.06):** roughly equal jobs and residents, but only ~5–6% live+work — people pour *through* it both directions (logistics corridor).
- **Yucaipa, Highland, Mentone, Grand Terrace, Calimesa, Beaumont, Banning — out-commute bedroom communities (JH 0.26–0.50).** They generate few local jobs and export 89–98% of their employed residents. **Highland (0.26)** and **Mentone (0.33)** are the purest dormitories. These residents are insured by employers elsewhere but **live, sleep, and seek routine care at home** — prime residence-based lives.

---

## 5. Strategic implication for Redlands Health

The covered-lives opportunity must be split along the home/work seam, because **a provider network's value to a member is a function of where that member lives.**

### (a) The employer-based TAM is heavily diluted by non-resident workers.
If Redlands Health sizes "covered lives" by counting people insured through a **Redlands employer**, the denominator is anchored to the **~39,000 jobs in Redlands** — but **85% of those job-holders live outside Redlands**, and **~two-thirds live outside the efficient (≈15-minute) service area entirely** (§3c). A worker who commutes in from Moreno Valley, Riverside, Fontana, or Los Angeles will **not** choose a Redlands PCP, will **not** route their family's care through RCH, and will use a Redlands-anchored narrow network only for whatever care happens *at or near the worksite during the workday*. Folding these lives into a provider-sponsored plan or RCH-centered narrow network at full value **overstates the addressable membership** and would seed the plan with members who leak out for most of their care.

### (b) The mirror: resident out-commuters are *efficient* lives that the employer count misses.
The ~24,900 Redlands residents who commute out (§3d) are insured by **out-of-area employers**, so they never appear in a "Redlands-employer" covered-lives count — yet they are **exactly the lives Redlands Health is best positioned to serve**: they live in town, their dependents are in town, and even most of their *own* daytime work is within ~15 minutes (San Bernardino/Loma Linda/Colton). For longitudinal primary care, value-based attribution, the CIN, and a resident-oriented health plan, **these are higher-value lives than many of the in-commuting workers** — they're just reached through a *residence* lens (geography, payer relationships, brand) rather than an *employer* lens.

### (c) Segment the market into two distinct lines of business.

**RESIDENT LIVES — "own the relationship."** Everyone who *lives* in the efficient service area (Redlands + adjacent bedroom communities: Yucaipa, Highland, Mentone, Loma Linda residents, the donut hole, Grand Terrace, Calimesa). For this segment the strategy is **longitudinal and total-cost-of-care**: employed PCP panels and a clinically-integrated network, value-based/ACO and capitated arrangements, a provider-sponsored or co-branded health **plan**, chronic-disease and prevention programs, and brand loyalty. These lives are sticky because home doesn't move with a job change. **This is where the JH<1 bedroom communities are an asset, not a liability** — they are dense reservoirs of resident lives whose care defaults home.

**DAYTIME WORKER LIVES — "serve episodically / at the worksite."** The ~33,000 people who *work* in Redlands but live elsewhere (and the analogous daytime populations in Loma Linda and San Bernardino). For this segment the strategy is **point-of-care and convenience, payer-agnostic**: occupational health and employer-sponsored **near-site / on-site clinics** (ESRI, Amazon, the District, the hospital itself, the donut-hole DCs), urgent care and imaging near the employment corridors, ED and ASC capacity, and **Centers of Excellence** (ortho/spine, cardiac, oncology) plus **direct-to-employer bundles** that *any* payer will route to on price/quality regardless of where the patient lives. You do **not** try to make these workers' families members of an RCH narrow network — you capture the **episodes** that occur during the workday and the high-acuity referrals that travel for reputation.

---

## 6. How to account for this in the covered-lives model (practical recommendation)

1. **Apply a "residence haircut" to the employer-based TAM** for any *network / plan / value-based* sizing. From §3c, only ~**32%** of Redlands' in-commute workers (≈10,500 of 33,325), plus the **14.9%** who already live+work, originate inside the efficient service area. A defensible planning rule:
   - **Network/plan-eligible share of Redlands-employer lives ≈ live+work (14.9%) + the near-resident slice of inflow (~32% of the 85.1% inflow ≈ 27 pts) ≈ ~40–45% of employer-based lives.** Size the narrow-network / provider-sponsored-plan TAM on roughly **0.40–0.45 × employer covered lives**, not 1.0×. Treat the remaining ~55–60% as **out-of-area worker lives** that will not anchor longitudinally to RCH.
   - Stress-test with a tighter (Redlands-only ≈ residence share) and looser (full Inland-Empire near-ring) boundary; the haircut sits between ~0.40 and ~0.55.

2. **Build a separate "daytime / worksite services" revenue line** for the in-commuter worker population. Size it on the **full daytime job base (~39,000 in Redlands; ~22,000 Loma Linda; ~119,000 San Bernardino)**, but model it as **episodic and B2B** — occupational health PMPM/PEPM, near-site clinic contracts, urgent-care/imaging volume, ED/ASC throughput, and bundled COE referrals — **not** as plan membership. Its economics are visit/contract-based and payer-agnostic, so it does not depend on owning the member's insurance.

3. **Add resident out-commuters back in.** The ~24,900 Redlands residents (and the tens of thousands more in the surrounding JH<1 bedroom communities) who are insured by out-of-area employers are **invisible to an employer-based count but fully addressable on a residence basis.** For plan/CIN/value-based planning, build the resident-lives TAM from **ACS/Census household population of the service area**, *not* from the local employer roster, and pursue them through payer contracts, PCP access, and brand — independent of who their employer is.

4. **Use residence, not workplace, as the unit of geography for anything longitudinal**; use workplace geography only for the occupational-health / worksite-clinic / COE line. The single most important modeling error to avoid is letting the impressive **39,000-job** workplace number set the size of a *plan* whose members must, by the data, mostly live somewhere else.

---

## 7. One-line takeaways

- **Redlands: 14.9% live+work, 85.1% in-commute, jobs-housing ratio 1.26, mean commute ~24.9 min — a net in-commute employment center sitting inside a ring of out-commute bedroom communities.**
- **~Two-thirds of the people who work in Redlands live outside the efficient service area; ~81% of Redlands' own residents commute out (but mostly stay within ~15 minutes).**
- **Strategic headline: apply a ~0.40–0.45× residence haircut to the employer-based covered-lives TAM for network/plan planning, own the RESIDENT lives (including the bedroom-community dormitories) longitudinally, and serve the in-commuting DAYTIME workforce episodically through a separate, payer-agnostic worksite/occupational-health/COE line.**

---

## 8. Limitations & honest flags

- **No fabricated figures.** Every number traces to LODES8 2022 or ACS 2024 estimates as labeled. All are estimates (LODES is noise-infused and UI-covered-employment-only; ACS carries sampling error). I did not retrieve margins of error into the CSV for brevity, but they are available from the same ACS tables if a confidence range is needed.
- **OnTheMap web API / Census ACS API:** the public OnTheMap `/api/` root returns 404 (no documented open JSON endpoint), and the Census ACS API now **requires a registered key even for single-place calls** (returns HTTP 302 → `missing_key.html`). I therefore used (a) the **LODES flat files** (which reproduce exactly what OnTheMap's Area Profile computes — jobs, inflow/outflow, live+work, home/work distributions) and (b) the **Census Reporter** mirror for keyless ACS. If you want the figures re-confirmed straight from OnTheMap's UI or the keyed Census API, the place FIPS codes in §1 are the lookup keys.
- **"Place" vs. reality of the donut hole.** Redlands is encircled by an unincorporated county island (Citrus Plaza / Mountain Grove / the Burlington & Amazon DCs). Those jobs are **outside** the Redlands place boundary, so the 39,161 Redlands jobs figure **understates** the true Redlands-area daytime employment by several thousand. In §3c the donut hole shows up as "Unincorporated San Bernardino County" (1,348 in-commuters into the *city*); the DC jobs themselves are not in the city total. If the strategic service area is drawn around the *hospital* rather than the *city line*, add the donut-hole employment back in — it only strengthens the in-commute-employment-center conclusion.
- **Mentone & Cherry Valley:** Mentone CDP (0646884) was fully tabulated. Cherry Valley CDP (0612916) was in scope but omitted from the per-city table; it behaves like Calimesa/Beaumont (a Riverside-County out-commute dormitory) and adds nothing to the service-area conclusion. It can be added on request.
- **ACS WFH vs. place-of-work base:** ACS place-of-work percentages exclude work-from-home, so they are shares of commuters, not of all employed residents; the LODES count-based percentages (the basis for the strategic conclusions) do not have this issue. Both are reported and labeled so the reader is not misled by the ~39% (ACS) vs. ~19% (LODES) "work locally" figures, which have different denominators.