Send the valuer the repair backlog

Each December, external valuers estimate what our 196,400 flats are worth. Like other listed property companies that report under IFRS, we carry investment property at fair value, and the change in that value runs through the income statement. A movement of one per cent on a portfolio valued at about €20 billion is close to €200 million, more than a third of what the business earns from its operations in a year.

The valuers are qualified professionals and their judgement is their own. But the evidence on which they exercise it comes overwhelmingly from us, in a data tape that my team prepares. I have come to think that the contents of that tape are among the most consequential things a chief data officer in property controls, and that most landlords send valuers too little of what would lower a value.

What a residential valuer actually receives

A portfolio of this size cannot be inspected building by building each year. Valuers visit a sample, a few hundred buildings, and model the rest. For each valuation unit they need in-place rent, market rent, vacancy, lease data, floor areas, construction year, modernisation history and an assumption about the capital expenditure the buildings will need over the holding period. Then they discount the expected cash flows.

The early items on that list are well recorded, because they come from the rent roll, and the rent roll is audited. The last item is not. In many residential portfolios, ours included until 2023, the capital expenditure assumption was a standard amount per square metre adjusted for construction period and the date of the last modernisation. It was a reasonable convention. It was also blind to what we knew about particular buildings.

By 2023 we knew a good deal. Our riser programme had surveyed drinking-water and heating risers in 640 of the 1,840 buildings in its scope and found about 150 that needed replacement sooner than our plan assumed. We held roof surveys for 2,600 buildings. Our caretakers recorded damp, leaks and failing windows. None of it reached the valuers. They received construction year and the standard assumption, and for the buildings we knew to be in the worst condition that assumption was too low.

Bad news belongs in the data tape

In the summer of 2023 I proposed that the December tape include our condition evidence: the costed riser backlog for surveyed buildings, roof survey results and a summary of recorded defects, together with a statement of what we had not surveyed. The timing was awkward. Values across German residential property were falling as interest rates rose, and our own portfolio was heading for a decline of about nine per cent that year. Colleagues in finance and investor relations asked why we would hand valuers reasons to cut further in the worst year for a decade.

My answer had three parts. First, the evidence existed, and if a building later failed in a way our records had predicted, we would struggle to explain why the valuer had not been told. Second, auditors increasingly ask how the inputs to fair value are controlled, and a tape that omits known defects is a weak control. Third, condition evidence cuts both ways: some surveyed buildings were in better shape than the construction-year convention implied.

The chair of our audit committee asked to see the proposal, and the committee supported it. The valuers used the evidence. For the surveyed buildings, higher capital expenditure assumptions reduced value by about €95 million, partly offset by about €30 million where surveys showed better condition than convention assumed. In a year of large market-driven declines, analysts barely noticed. Inside the company it changed the conversation. The riser programme stopped being an engineering budget and became something the balance sheet recognised.

Condition data lowers values before it raises confidence. Every survey that finds a problem reduces a value before the repair is done, so a landlord that surveys more will, for a while, look worse than one that surveys less. Valuation conventions do not remove that perverse incentive. The honest response is to name it, and to make the survey programme visible, so that investors can distinguish a portfolio in poor condition from one that has started to measure its condition.

We now give valuers each building's condition evidence with its date and source, and we state the share of buildings for which we hold none. In the 2025 tape that share was 61 per cent. It is an unflattering figure. It is also the most useful single number in the tape, because it tells the valuers, the auditors and the board where our knowledge ends.

The tape has also changed what we survey. Once the board could see which buildings lacked evidence, it asked why the gaps fell where they did. The honest answer was that we had surveyed where engineers were already worried, so our evidence leaned towards bad buildings while the convention was applied mostly to buildings nobody had examined. Since 2025 a small part of each year's survey budget pays for a random sample of about 150 buildings outside any programme, so that the valuers and we can test whether construction-year assumptions hold where we know least. The first sample suggested that they understate repair needs in blocks from the early 1970s and overstate them for buildings modernised in the 1990s.

Questions an audit committee can ask

Audit committees of property companies spend a great deal of time on valuation, and rightly. Much of it concerns yields and market rents, matters of judgement on which committee members can only test reasonableness. I would encourage committees to spend some of that time on the data tape, where the questions have answers.

Which evidence about condition, defects and planned works did management hold that the valuers did not receive? Who decides what goes into the tape, and is that decision recorded? How many buildings did the valuers inspect, and did what they saw match our records? When a building suffers a major failure, had it been flagged in the tape?

None of those questions requires valuation expertise. They require a committee to treat the tape as a control, like any other input to the accounts. In my experience management teams answer them readily once asked, and the answers improve after the first year.

Fair value is an estimate, and no tape makes it certain. But an estimate made without evidence the owner possessed is not fair. It is merely convenient, and in residential property the inconvenient evidence is nearly always about pipes, roofs and damp: the things tenants notice first.