Measuring and Reporting Property Level ESG Data

Owners collect property level energy data by mapping every meter and utility account to a stated building boundary, labeling each reading as measured or estimated, and recording the share of consumption that could not be obtained. Emissions come from multiplying each energy quantity by a factor in matching units, and the method, the energy types included and the estimated share travel with the result. Absolute totals and like-for-like comparisons answer different questions, so both are explained. Tenant-controlled consumption is the usual gap, especially under net leases.

Why tenant meters decide whether a total covers the building

A property total covers the whole building only when the tenant-paid accounts sit inside it, and under many leases those accounts belong to the tenant.

Electricity used by an industrial tenant on a triple net lease can be the tenant’s Scope 2 emissions and the landlord’s Scope 3 emissions. The kilowatt-hours are identical in both inventories. Only the reporting entity changes, so copying a tenant’s classification into an owner’s inventory misplaces the same consumption even when the meter reading is right.

Net lease structures also limit collection. Where tenants handle maintenance and daily property management, the owner has less access to property-level environmental metrics and less ability to enforce sustainability measures. Landlord-paid bills for such a property may show only common areas and vacant space.

The decision test is simple: does the figure cover owner-controlled accounts or the whole property? Either answer is acceptable if the label says which. A whole-property heading above an owner-only electricity total is an internal inconsistency, and accurate bill entry cannot repair it.

Why a 20,000 square foot building can still need a data process

Local benchmarking and reporting ordinances reach commercial buildings as small as 20,000 square feet, so a data process is no longer confined to large institutional assets.

Properties under 50,000 square feet make up more than 90% of the commercial building stock. The 20,000 square foot threshold sits well inside that band, which means a family holding a few small local buildings can face the same collection questions as a large owner.

Carbon rules add a cost to a bad number. In New York City, Local Law 97 fines owners that exceed a carbon emission threshold, so an emissions figure built on a wrong boundary or a wrong factor can matter in dollars as well as in disclosure.

Energy is also an operating line. In commercial office buildings it represents approximately one third of typical operating budgets, and low- or no-cost measures can cut building energy use by up to 30%. A measured baseline is what allows an owner to see whether any such cut occurred.

Building the collection record, account by account

A usable record starts with one row per meter or utility account, each tied to a property, an energy type, a unit and a period.

Step one is the account map. Every electricity, gas, steam or water account serving the building is listed with the party that pays it. Gaps show up here first, because an account with no row has no reading.

Step two is the status field. Each reading is tagged measured or estimated, and unavailable consumption is tagged separately. A completed total with no such tags cannot show how much of it rests on assumptions.

Step three is overlap review. A master meter that already includes a tenant’s submetered load makes the two readings non-additive, since adding both counts that load twice. A tenant load that sits on its own utility account outside the owner’s meters is the reverse case, and omitting it leaves a hole in a whole-property total.

Step four is a coverage statement with a denominator. Floor area coverage describes the area represented by collected records. Energy coverage describes the share of consumption represented. Equal area shares can hold very different energy, so converting one into the other needs a stated consumption assumption.

Disclosure follows from these tags. For elements outside the manager’s operational control, or where data could not be reliably collected, the degree of estimation is reported alongside the total.

“Disclosures and explanations should consider both absolute and like-for-like data.”

“Disclosures should clarify the degree to which estimated data was used in determining overall values for elements that are outside of the manager’s operational control, or for which data could not be reliably collected.”

Converting kilowatt-hours to tonnes without losing the audit trail

An emissions figure is energy multiplied by a factor, and the factor’s units must match the energy’s units.

Electricity in kilowatt-hours multiplied by kilograms per kilowatt-hour gives kilograms; dividing by 1000 gives tonnes. A factor quoted per another energy unit, such as per therm or per megajoule, needs a unit conversion before multiplication.

Each energy stream gets its own line because factors differ by fuel and by electricity supply. Emissions can be added across lines once all are in tonnes within one accounting boundary. Raw consumption in unlike units cannot be summed into a meaningful energy total.

The disclosed method names the energy and emissions types included and any normalization applied. With those items stated, a reader can rebuild the result from the meter records; without them, two owners reporting the same building can produce different numbers from the same bills.

Emissions can move because consumption moved, because the factor moved, or both. Recomputing the current period at the prior period’s factor isolates the consumption effect, and the worked example below shows the split.

Absolute totals and like-for-like totals answer different questions

Absolute reporting states the total for the whole portfolio in the period. Like-for-like reporting compares only properties held and measured on a consistent basis across the periods compared.

Explanations should cover both. An acquisition adds consumption to the absolute total, but reductions at properties previously held can outweigh that addition. A sale lowers it with no operating improvement at the properties kept. Like-for-like removes that population effect from the comparison.

A matching property list is not enough on its own. If tenant electricity is included for a building in one period and excluded in the next, the building has changed boundary inside a supposedly consistent set.

Normalization adds its own arithmetic. Energy per square foot divides consumption by area, so intensity can fall because the numerator shrank, because the denominator grew, or both. The area used has to match the consumption boundary. Dividing owner-only electricity by the whole building’s area mixes two boundaries in one ratio.

A reconciliation from the absolute movement to the like-for-like movement, with the coverage and estimation changes named, lets a reader see what operations contributed and what composition contributed.

Worked example: a 100000 square foot office building with missing tenant bills

The inputs are hypothetical and illustrative. A 100000 square foot multi-tenant office building has a prior period with full measured electricity of 1000000 kilowatt-hours and a factor of 0.4 kilograms per kilowatt-hour. Both periods are the same length, and the same building and area are used throughout.

Prior emissions are 1000000 multiplied by 0.4, which is 400000 kilograms. Dividing by 1000 gives 400 tonnes. Prior intensity is 1000000 divided by 100000, or 10 kilowatt-hours per square foot.

In the current period, owner-paid accounts measure 600000 kilowatt-hours. Tenant bills were not obtained, and the tenant load is estimated at 200000 kilowatt-hours. The current factor is 0.3 kilograms per kilowatt-hour. Two treatments of the same period follow.

Treatment A is the stressed case: tenant data treated as zero. Energy falls from 1000000 to 600000, a reduction of 40 percent. Emissions are 600000 multiplied by 0.3, which is 180000 kilograms, or 180 tonnes, and the reduction is 400 minus 180, divided by 400, or 55 percent. Intensity is 600000 divided by 100000, or 6.

Treatment B is the base case: tenant load estimated and labeled. Total energy is 600000 plus 200000, or 800000 kilowatt-hours, a reduction of 20 percent from the prior period. Emissions are 800000 multiplied by 0.3, which is 240000 kilograms, or 240 tonnes. The reduction is 400 minus 240, divided by 400, or 40 percent. Intensity is 800000 divided by 100000, or 8. Measured coverage of energy is 600000 divided by 800000, or 75 percent, and the estimated share is 200000 divided by 800000, or 25 percent.

Treating the gap as zero overstates the emissions reduction by 55 minus 40, or 15 percentage points, and doubles the apparent energy reduction from 20 percent to 40 percent.

Even Treatment B mixes two effects. At the prior factor, current consumption would emit 800000 multiplied by 0.4, or 320000 kilograms, which is 320 tonnes. The consumption effect is 400 minus 320, or 80 tonnes. The factor effect is 320 minus 240, or 80 tonnes. Together they equal the 160 tonne reduction from 400 minus 240.

Half the reported decline comes from the factor change, which no operating decision at the building produced. Of the consumption half, a quarter of the current energy figure is an estimate.

Pitfalls in property energy and emissions reporting

Treating unavailable tenant data as zero understates consumption by the omitted amount. Zero means no consumption inside the boundary; missing means unobtained. In the example above, the zero treatment turns a 40 percent emissions reduction into 55 percent. Correct handling keeps the gap visible and labels any fill-in as an estimate.

Adding a master meter to the submeters it already includes counts the same kilowatt-hours twice. Better meter precision leaves the duplication intact. The valid sum depends on the physical relationship between meters, which the account map has to record.

Reading an emissions decline as an operating improvement confuses consumption with the conversion factor. In the example, 80 of the 160 tonnes came from the factor. A constant-factor recomputation separates the two.

Using an area share as an energy share assumes every square foot uses equal energy. A data center suite and a storage suite of equal area do not, so 75 percent of area covered can sit far from 75 percent of energy covered.

Presenting an absolute portfolio change as a like-for-like improvement mixes composition with operations. A disposal lowers the total with no change at retained buildings, and a boundary change inside a retained building does the same quietly.

Calling an estimated-heavy total measured hides its dependence on assumptions. Decimal places do not change that, since a result computed exactly from an estimated input is still partly estimated. The share of estimated data, and why it was needed, belongs beside the number.

Checklist questions for a property energy and emissions report

Does the total state whether it covers owner-controlled accounts or the whole property? The label decides how every later figure is read.

Does every meter or utility account appear in a map with the party that pays it? An account missing from the map is a reading missing from the total.

Does each entry carry a unit and a period? Those two fields expose incompatible quantities and mismatched months before anything is added.

Is overlap between master meters and submeters resolved? A valid sum counts each physical kilowatt-hour once.

Is tenant electricity placed in both the coverage statement and the emissions scope classification? One question concerns what the building consumed and the other concerns whose inventory it enters.

Is each reading tagged measured, estimated or unavailable? A completed total is only as informative as its tags.

Does the emissions factor match the consumption unit, and is the factor source named in the method? A reader can then rebuild the tonnes from the meter records.

Does the comparison say whether it is absolute or like-for-like, and what normalization applies? Intensity depends on both the denominator and the property set.