Greenwich Township School District (Warren County) — NJ

1. Snapshot

Suburb-Large K-8-only district in Warren County, NJ (Stewartsville, near the Pennsylvania border). 607 students across 2 schools — Greenwich School (PK-5, 397 students) and Stewartsville Middle School (6-8, 210 students). 8th-graders matriculate to Phillipsburg HS via send-receive relationship. SAIPE poverty 5.1% — low. Demographics 62% White / 15% Hispanic / 13% Black / 6% Asian / 3% Multiracial — modestly diverse for rural-fringe NJ. Per-pupil expenditure $30,042 (FY2020) — the highest in this brief set, driven largely by the small-district scale-disadvantage on overhead. Smallest dollar ask in this brief set ($6M), still failed at 63/37.

2. Why this was a hard sell — community context (ACS)

Despite being a non-unified K-8-only district, ACS data WAS available via get_community_profile — atypical for non-unified districts in NJ. This is a real data point for outreach.

This is the wealthiest, most educated, most homeowner-dominant district in the brief set after Wayne. $168K HHI, $432K homes, 58% Bachelor’s+, 93% owner-occupied — every economic indicator says voters could absorb a $6M bond easily. The 63/37 No vote is not a tax-base issue. It is almost certainly:

(a) Send-receive HS structural disconnect: K-8 districts in NJ that send to a separate HS (Phillipsburg in this case) have weaker political bonds with parents whose oldest kid is already gone — the “school spirit” coalition is thinner. (b) Tiny district, tiny turnout, organized opposition: 897 total votes cast (~20% turnout) means ~50 active opposition households could swing the result. (c) K-8 districts are notoriously underfunded for capital relative to NJ’s per-pupil norms — but the visible tax burden in Warren County (NJ’s high state property taxes + Phillipsburg HS share) is what voters feel, not the structural under-investment.

3. Peer comparison

Top peers identified via MCP (default weights + plantOps emphasis):

4. The gap story (what the data would have shown voters)

Greenwich Township is the rarest profile in the brief set: a high-spending district that is underspending on plant ops relative to peers. The data tells a “we’ve let buildings slip while protecting classroom” story that the bond could have leaned into.

• Plant operations spending: $1,322.43 / pupil vs national median $1,324 — Greenwich is exactly at national median. But versus peer median ($1,478), it is 11% below peer median. Of the 5 named peers, Greenwich has the lowest plant ops spend per pupil. This is the bond justification and it never made the materials we reviewed.
• Capital construction outlay (FY2020): $0 — zero across the entire district. Not “low” — zero.
• Per-pupil instruction $12,885 — above peer median ($12,599). Classroom is being funded above peers.
• Chronic absenteeism 6.1% — best in the peer set (peer median 9.5%). Suspension <1%, lowest in peer set. Behavioral data is excellent — no behavioral case for the bond.
• Teacher certification 95.4% — the lowest in the peer set (peer median 100.1%). Same red flag as Baraboo. New buildings won’t fix certification but failing buildings make recruitment harder.
• Counselor ratio 304 — at peer median (573 — actually better than peer median when looking at the full set, since several peers report 500+).
• No reported security FTE — neither Greenwich nor most of its peers report security staffing. Not a clear differentiator.
• 2 nurses across 2 schools — at peer median.

The structural pitch: Greenwich is a high-spending district that has chosen classroom over building. The bond was the chance to rebalance. The $6M ask was small enough that it should have passed in a $168K-HHI community — that it didn’t is a campaign-execution failure, not a tax-base failure.

5. Bond history (Ballotpedia + news)

• Mar 12, 2024: $6.0M bond, 63% No / 37% Yes (565 vs 332; ~20% turnout) — one confirmed failure
• No prior referendum documented in available coverage.

The Ballotpedia page for Greenwich Township SD Warren County elections was not accessible during research; confirmation of bond history pre-2024 would require a direct call to the district or county clerk.

6. What voters / opposition actually said

NJ Globe coverage was thin — Greenwich was one paragraph among 11 statewide special school referenda. No quoted opposition, no superintendent statement, no organized-opposition campaign documented. The absence of public debate is the signal: a 200-vote-margin loss in a 600-student district means roughly 100 households moved the result, and that was likely word-of-mouth opposition, not organized.

~20% turnout in a special school election is the political failure mode. In a district where 92.7% of households own homes, a special-election ask in March (off-cycle) drives a turnout pool that skews older, fixed-income, and tax-sensitive — exactly the wrong demographic to win a school bond from. The K-8 send-receive structure compounds this: parents whose oldest kid graduated to Phillipsburg HS have weakening reasons to show up.

7. What we could have told them

1. “Our plant operations spend is the lowest of any district in our peer comparison group — $1,322/student vs a peer median of $1,478. We spend less on keeping buildings standing than 5 of our 5 closest peers. The $6M ask is what closes that gap.” Single most defensible bond-justifying number, hiding in the data.
2. “Capital construction last year: $0 across 2 schools. Across 10 years, the same. The bond is what stops that from being the new normal.” Same data-quality argument as Vermilion.
3. “Our teacher certification is 95.4% — the lowest in our peer comparison group. Every other peer runs at or above 100%. We’re losing or failing to attract certified teachers, and one likely contributor is the buildings we’re asking them to teach in.” Defensible link from buildings to people.
4. NJ State Aid debt-service reimbursement on a $6M ask would have made the net local levy ~$3.6–4.2M. Across a 92.7% owner-occupied district of 607 households (approximate), that’s a per-household share that any honest tax-impact model would have made trivially small. The campaign appears not to have published one.
5. For the next attempt: move the vote to November general election, not a March special. Turnout would 3–4x, and at 70% Bachelor’s+ in this community, a higher-turnout vote should be more receptive to a building-condition argument backed by data.

8. FMX outreach hook

Greenwich Township is the smallest, simplest engagement in the brief set — 2 schools, named CFO equivalent (Scott Campbell), and a community wealthy enough to afford an FMX engagement even outside a bond cycle. The K-8 structure means lighter portfolio scope than HS districts. Lead with Scott Campbell (Business Administrator / Board Secretary): he’s the named decision-maker for finance and operations and the one defending the next ask. Opener: “$6M is the smallest ask in our failed-bond research set this year, and it still lost 63/37. Your plant ops spend is the lowest of any of your 5 closest peers — that’s the strongest pro-bond argument in your data, and it didn’t make the materials. With 92.7% owner-occupancy and 58% Bachelor’s+, you have a community that will respond to specific per-building condition data — they just need to see it. Berlin Township NJ in your peer cluster is the closest comp; 1 redacted FMX customer in Massachusetts is your other reference point. We can have your 2-building portfolio condition-scored and benchmarked in 30 days, in time for a November 2026 ballot move.” The November-election-strategy advice ties directly to the FMX engagement window.