Key takeaways
- A meter reading and a resident charge are different records.
- Allocation formulas need a version and an effective date.
- Reconciliation should close at the invoice, not the estimate.
Source record
10 cited sources
Last verified
2026-09-18
In this article
When a rental portfolio pays a utility and then recovers part of that cost, the reliability of the recovery depends on records that connect three things: what the utility billed, what the meter or allocation formula produced, and what the resident was charged. This research note describes a reconciliation model for metered and allocated utilities. It is an operational method, not a legal opinion, and it does not decide which recovery structures are permitted.
The model is deliberately narrow. It asks whether the paper trail can show that a charge was computed and posted according to a rule that existed at the time.
Methodology and scope
On September 18, 2026, we reviewed ten public sources from energy, water, housing, and evidence authorities. We used the EIA residential energy consumption survey, the EIA Today in Energy series, the EPA WaterSense program page, the DOE Energy Saver site, the EPA indoor air quality page, the HUD housing choice voucher program page, HUD user fair market rent data, HUD user income limit data, NIST information quality standards, and the eCFR Part 100 fair housing regulations.
We then modeled four reconciliation layers and applied them to hypothetical portfolios: a master-metered building with a ratio allocation, a submetered building with estimated reads, a water and sewer split across unit counts, and a common-area charge recovered through a formula that changed mid-year. No utility bill, meter, or resident ledger was reviewed.
Four layers
The first layer is the utility invoice. This is the provider's bill for the account, with its own billing period, usage, rate components, taxes, and adjustments. The invoice is the anchor record. Any resident charge should be traceable to it, directly or through a documented formula.
The second layer is the measurement. For a true submeter, this is the unit-level reading and the read date. For an allocation, this is the input data, such as occupied unit counts, square footage, or a ratio method. The EPA WaterSense program page and the EIA residential consumption survey are useful context for why measurement quality matters, though neither prescribes a billing formula.
The third layer is the calculation. Here the file shows the formula version, the effective date, the inputs actually used, and the resulting amount. A formula without a version cannot be reconciled after it changes. NIST information quality standards support retaining the rule and its source alongside the result.
The fourth layer is the posting. This is the resident-facing charge on the ledger, with the billing period it covers and the date it was posted. The posting is the layer residents see, and it is the layer most likely to be questioned.
Analysis: where reconciliations fail
The most common failure is a missing invoice link. The resident is charged an amount, but nothing in the file connects that amount to the provider's invoice for the corresponding period. Without that link, the charge cannot be verified or corrected.
A second failure is the versionless formula. Allocations are often described in prose that changes quietly. When the formula is not versioned, a reviewer cannot tell whether the amount charged used the rule that applied at the time. The eCFR Part 100 fair housing regulations and the HUD housing choice voucher page show that housing programs place real weight on documented policies, which reinforces the value of version control even outside those programs.
A third failure is the estimated read that is never reconciled. Estimated readings are sometimes necessary, but an estimate that is never trued up against an actual reading leaves a gap that can accumulate. The record should show when the estimate was scheduled for reconciliation.
A fourth failure is the mixed-period charge. When a billing period does not align with a tenancy period, the allocation method should say how the period is split and why. A charge that silently spans two periods is hard to defend even when the arithmetic is sound.
Allocation methods and their evidence
Different recovery structures demand different evidence. A true submeter is the most direct, because the unit reading is the measurement, but it still requires a matching read date, a rate basis, and a link to the invoice. A ratio allocation based on occupancy depends on an accurate unit count for the period and a stable formula. A square-footage allocation depends on a current area schedule and a clear treatment of common areas.
Each method carries a specific failure mode. Submeters drift or fail and may be estimated for a period. Occupancy counts change during the month as units turn over. Area schedules become stale after renovations. The reconciliation record should name which method applies, which input data was used, and any known limitation. NIST information quality standards support documenting inputs and their sources so results can be reproduced.
Common-area and shared-system charges
Many portfolios recover a share of common-area electricity, water for irrigation, or shared laundry utilities. These charges need an explicit basis. A formula such as a fixed percentage or an equal split should be written down, versioned, and tied to the invoice. The file should also show how the common-area portion was separated from the resident portion, because a charge that quietly blends the two cannot be verified.
The EPA WaterSense program page and the DOE Energy Saver site provide general consumption context that helps teams sense-check whether a shared charge is plausible, but they do not establish a billing rule. The rule comes from the applicable policy and law, which should be recorded with its effective date.
A reconciliation workpaper that travels with the charge
The most durable practice is a one-page workpaper per billing period. It lists the provider invoice reference, the billing period, the total account usage and cost, the method and formula version, the input data, the resident-level result, and the posting date. The workpaper is short, but it lets any later reviewer reproduce the charge without hunting through several systems. When a formula changes, a new workpaper version begins, and the old version stays with the periods it governed.
Inference boundaries
A reconciled charge does not prove that the recovery structure is lawful or that the amount is reasonable. A gap does not prove that a resident was overcharged. The model tests whether the calculation can be reproduced, not whether it should have been made. Legal and program-specific rules, including rules attached to subsidized housing, must be applied by a qualified decision maker.
Limitations
The model excludes jurisdiction-specific utility billing rules, landlord-tenant limits, local rate structures, and program requirements. It was not tested against live bills. It does not address energy-efficiency upgrades or rate-optimization decisions, which are separate operating questions. It also does not evaluate whether a particular meter or sensor is accurate; that is a calibration question for the equipment provider rather than a records question.
Practical measurement
A portfolio can sample billing periods and count how many reconcile across all four layers, reporting each layer separately so the missing link is visible. Tracking the share of charges based on estimates, and the age of unreconciled estimates, adds useful context. Owners who want to review the wider method set can browse the research archive or ask for a portfolio review, and readers interested in a related billing-quality check can compare the note on moisture and mold incident records.