Public Trust Lens · NYCHA Utilities
v2.0 · 2026
01 Case study · NYCHA utility dynamics

The bill residents don't see.

NYCHA tenants don't pay the electric, water, or gas bill directly — but every avoidable dollar still comes out of the system they depend on.

Total exposure
$3.87B
Combined utility charges
Estimated bills
16.0%
Not actually metered
Geometry mismatch
25.0%
Rows unmatched to development
Priority shortlist
51
Records needing explanation
02 The scale of the problem

If even a sliver of this were controllable

Across electricity, water, and gas, NYCHA's analyzed utility spend totals $3.87 billion. Move the slider below to see what a small efficiency gain — or a small accounting cleanup — could free up for repairs, cleaning, and resident services.

Charges by borough
Utility charges by category, 2010–2025
Stacked totals · interactive
Control scenario
→ Potential capacity unlocked
$116.1M
Roughly the cost of 2,300 deep apartment refurbishments at $50K each. Scenario, not a confirmed saving.

Caveat: these scenarios show opportunity costOpportunity costThe value of what else NYCHA could fund if utility spend were reduced — not a guaranteed saving., not recoverable dollars. Real savings depend on weather, tariffs, occupancy, and equipment.

03 Sixteen years of meters

$24M, every month, for sixteen years.

Monthly NYCHA utility charges have moved from a $13M floor in 2010 to peaks above $40M during winter cold snaps. Water enters the ledger in 2013, when the city first separates its meter records. The estimated-bill share rises and falls with operational turbulence — it never disappears.

Monthly utility charges, NYCHA-wide
Stacked by utility · $ per month
04 What residents actually hear

The bill is invisible, the friction is not.

Residents don't open a utility envelope every month — but they do live with the consequences of how that spending is managed. The hidden cost shows up as slower repairs, deferred maintenance, and explanations that arrive late or not at all.

Illustrative composite

When the heat goes out in February, nobody can tell me when it'll be back. The answer is always "ticket submitted."

— Composite of tenant association testimony
Illustrative composite

They say utilities are the biggest cost. Okay — so where does it go? Which building? That part nobody shows you.

— Composite of public meeting transcripts
Illustrative composite

If they overpaid, that's my elevator, my lock, my mailbox. It's not abstract money to me.

— Composite of resident advocate framing
05 The signal hunt

Where should explanation happen first?

Not every anomaly is waste — but every anomaly without an explanation chips away at trust. We ranked every development-utility pair by an interpretation-priority scoreInterpretation-priority scoreA composite of estimated bill share, exception rate, cost spread (p90/p10), and geometry match. It says "explain this one first.". The top ten are below.

Utility: Electric Water Gas Hover a row for detail
Rank Development Utility Estimated Cost spread Cost/unit trend Score

Source: NYCHA utility ledgers, 2010–2025. Score is heuristic, not a finding of fault.

06 Where the bills swing hardest

A meter that jumps is a story untold.

Cost-per-unit volatility (the ratio between a property's most expensive billing month and its cheapest) reveals where billing data behaves erratically. A high ratio doesn't prove waste — it proves the public can't easily explain what happened.

Low spread High spread (p90/p10)

Williamsburg, water

A 27× spread between top and bottom billing months — the single largest volatility flag in the dataset. Most plausibly a tariff or meter-scope artifact, but it warrants an explainer.

Patterson, gas

24% of bills are estimated. Combined with a 6.4× cost spread, this development should receive accounting attention before any operational claim.

The pattern, in plain English

Big buildings with old gas service tend to have the dirtiest billing data. Cleaning the data often comes before cleaning the cost.

07 Method & caveats

How the signal hunt works.

This isn't a savings claim — it's a triage layer. Six steps move from a million rows to a shortlist of records where institutional explanation should happen first.

Six steps

  1. Detect development-utility combinations with high interpretive friction.
  2. Separate resident opportunity cost from direct resident billing.
  3. Aggregate monthly records so one-off bills don't dominate the story.
  4. Compare estimated-bill flags, cost spread, and geography match quality.
  5. Shortlist records where the public explanation is weakest.
  6. Pressure-test the conclusion with caveats before assigning blame.

Known limitations

  • No normalization for apartment count, square footage, occupancy, building age, or weather.
  • Cost-per-unit spreads can reflect tariffs, fees, or commodity pricing — not just operations.
  • Water records include many FHA properties that don't map to standard NYCHA development polygons.
  • The score is an interpretation-priority heuristic, not a statistical model.
  • Scenario savings communicate scale, not recoverable dollars.