Show investors which figures were measured

Investors in Japanese listed property now read environmental figures almost as closely as occupancy and rent. J-REITs and property groups publish energy use, water and emissions for their buildings, answer annual benchmark questionnaires and, increasingly, prepare for the sustainability disclosure standards Japan has adopted for listed companies. The figures appear in tidy tables, property by property, and they look like measurements. Many of them are not.

The reason is structural rather than dishonest. A building owner controls the energy it buys for common areas and for tenants it supplies directly. Where a tenant buys its own electricity, as single-tenant logistics warehouses and some shops often do, the owner sees nothing unless the lease gives it a right to the data. To produce a complete table, owners fill the gap with estimates, usually floor area multiplied by a published average intensity for the building type. The estimate sits in the same column as the meter readings, and a reader cannot tell them apart.

An estimate that looked like a measurement

I learnt this in 2019, while seconded to the asset manager of our group's J-REIT. I was preparing energy data for the annual sustainability report, which showed energy intensity covering every one of the REIT's 38 properties. When my team traced each figure back to its source, we found that 41 per cent of the floor area had no meter data at all. The estimates had been made carefully, in a working file, by people who were trying to complete a questionnaire. Nobody had decided to mislead anyone. But the published table presented a portfolio as fully measured when two-fifths of it was arithmetic.

Some of that arithmetic was also probably wrong in a direction that mattered. Two of the warehouses ran chilled storage, and their real consumption was likely to be several times the average for dry warehouses. An investor comparing the REIT with a peer that measured more of its portfolio would have drawn the wrong conclusion.

I recommended restating the previous three years and reporting measured coverage separately. The head of investor relations and the asset manager's president preferred to wait two years, until meters could be installed, and present the change then. Their argument deserved respect: the REIT's benchmark grade would fall, and an equity offering was being prepared. The investment corporation's board chose to restate. Our 2019 report disclosed measured coverage of 59 per cent and labelled every estimate, the benchmark grade fell by one level the following year, and the offering went ahead with some uncomfortable questions on the roadshow.

Coverage belongs beside the figure

The principle I took from that year is that coverage is a disclosure in its own right. Every environmental figure we supply now carries two companions: the share of floor area or consumption it covers, and a method label of metered, tenant-reported or estimated. Estimated figures are published, because a portfolio total that leaves out a third of the buildings would mislead in another direction, but they appear in their own column, and the method of estimation is stated.

This makes the tables longer and less flattering. A reader can see at once that one warehouse is estimated, and may reasonably ask why. That is the point. Investors are not asking owners to be perfect; they are asking to know how much weight each number will bear. A figure with 88 per cent measured coverage and a stated estimate for the rest is more useful than a figure that implies 100 per cent and cannot say which parts are real.

Labels also change how data is checked. Metered figures from our own supply can be reconciled with utility invoices every month, and we do that. Tenant-reported figures are harder: we compare each month with the same month a year earlier and with the logger readings where we have them, and we ask the tenant about any movement larger than a fifth. In the first year of tenant reporting, three warehouses sent figures in the wrong units, one of them a thousandfold too small. None of those errors would have been visible in a table that mixed every source together, and all three were caught because the method label told us which figures needed a second look.

Boards have their own reason to care. As sustainability disclosures move into statutory filings, directors will be signing off figures that rest on data their companies do not fully control. The question for an audit committee is not only whether the totals add up, but which parts were measured, which were reported by tenants and which were calculated, and whether each has been checked as rigorously as its weight in the total requires. Without method labels at the source, nobody can answer that question at the end.

The tenant's meter and the tenant's consent

Raising measured coverage depends on tenants. After the restatement we approached the 15 tenants in the REIT's portfolio who bought their own electricity and asked each to add a data clause to its lease: the tenant would share monthly consumption, and we would install and maintain the meters or loggers at our cost. Fourteen agreed over three years, most at lease renewal. We spent ¥86 million on equipment. One tenant declined, for reasons of commercial confidentiality that I understand, and its building remains labelled as estimated. Measured coverage reached 81 per cent by 2022 and 88 per cent in the year to March 2026.

Consent has a second dimension in residential buildings. Our parent group's rental housing produces energy data household by household, and consumption patterns can reveal when a flat is occupied, so we treat them as personal information under Japan's privacy law. We therefore report residential energy by building only, never by flat, and we do not hold household data longer than billing requires. The disclosure is less granular than it could be, and I think it is right that way.

There is a commercial cost to all of this. Our tables look less complete than those of owners who estimate without saying so, and in the short run benchmark grades can reward completeness over candour. Over six years I have watched that cost shrink as investors learnt to ask about coverage. I would still pay it if it did not shrink. A disclosure that cannot say how it was measured is a liability waiting for the day someone checks.