Property companies wearing pensioners as armour

When Aviva and Homes England announced one of the first investments backed by the new National Housing Bank earlier this year — an initial £100m to build family rental homes on brownfield sites in Liverpool and Manchester, with ambitions of 3,300 homes — the framing was familiar and comfortable. Patient institutional capital, deployed into the real economy, supporting the government's 1.5 million homes target. Pension Insurance Corporation's Habiko joint venture with Muse and Homes England had already established the template: the insurer forward-funds development and, in PIC's own words, will "ultimately own the homes" as long-term steward.

Own the homes. It is worth pausing on that phrase, because it marks a quiet but structural shift in what an annuity fund is. The traditional model had insurers lending against housing — to housing associations, against long leases, through income strips — holding fixed contractual cash flows with a legal claim senior to equity. The new model has them owning the bricks and collecting the rent. The difference is not cosmetic. A bondholder in a housing association holds credit risk. A freeholder of three thousand rental homes holds landlord risk: voids, maintenance, local market decline, and — above all — rent regulation. The question this article asks is not whether that risk is priced. It is who ends up bearing it, who booked the profit on it, and what happens to housing policy once the answer is "pensioners."

The demand problem

Start with why the industry wants this at all, because the causality is routinely reported backwards. The bulk purchase annuity boom has handed UK insurers back-books that need vast quantities of long-dated, illiquidity-premium-bearing assets, and the sterling corporate bond market cannot manufacture enough of them. The industry's ask of government — planning reform, Homes England land, the National Housing Bank's balance sheet — amounts to this: use public powers to de-risk the origination phase of housing, then hand us the stabilised income streams to hold against annuity liabilities.

The state absorbs the genuinely equity-like part of the risk — planning, development, lease-up. The insurer arrives at stabilised yield, with a quasi-public counterparty structure wrapped around the deal. And under the Solvency UK reforms, which widened matching adjustment eligibility from fixed cash flows to "highly predictable" ones, rental income can now creep towards annuity-backing status. That phrase — highly predictable — is doing enormous load-bearing work, and the matching adjustment converts it into something more valuable still: day-one recognition. The illiquidity premium on a rental estate is capitalised into shareholder profit at the point of purchase, decades before the evidence on whether the cash flows were in fact predictable can possibly arrive.

None of this is scandalous in isolation. It is the ordinary machinery of annuity investing, applied to a new asset class. The housing programme is not being supported by the annuity model; it is being shaped to produce the asset the annuity model requires. That distinction matters for everything that follows.

The mortality mismatch

A BPA book is a run-off machine. Payments peak within the first two decades and then decline as pensioners die, with a thin tail stretching out sixty years but a duration of perhaps twelve to fifteen. Direct ownership of housing is a perpetuity. Bolting a perpetual asset onto an amortising liability means the asset outlives, by construction, the thing it supposedly exists to secure.

So what happens to the rent when the annuitants are dead? For a while the question is deferred: each year's new BPA tranches need backing assets, and old cohorts fund new ones. But BPA is a finite extraction industry — there is roughly £1.4trn of defined benefit liability to transfer and then the seam is mined out. The theoretical answer is that the insurer sells down the estate as liabilities amortise. Notice what that concedes: the matching adjustment benefit was booked on hold-to-maturity economics, on the premise that market values do not matter — and the run-off endgame quietly reintroduces the premise that they do.

The practical answer is simpler and more important. Nothing compels disposal. As a closed book runs off, assets in excess of remaining liabilities are released as surplus — to shareholders. If the rental estate still exists when the last pension is paid, it does not revert to anyone's estate, or to the public purse that de-risked its construction. It is simply the insurer's property, free and clear, its rent now dividend feedstock rather than pension funding. Indeed, because a cash-flow-matched portfolio only "uses" the portion of rental income falling within the liability horizon, the perpetual residual was economically shareholder property from day one. The annuitants were never owners. They were a financing vehicle with a mortality table attached.

Follow the sequence end to end. Defined benefit members' accumulated capital purchases the housing. The state subsidises and de-risks the purchase. A generation priced out of ownership services the yield for decades. The yield pays the pensions. And when the pension liability extinguishes, the freeholds remain. The permanent product of the entire arrangement is the transfer of a slice of the national housing stock onto insurance company balance sheets.

The wrong worry

The instinctive objections to this picture tend to miss. Institutional ownership will not dominate the housing stock: even heroic growth leaves it in low single digits of 25 million dwellings, dwarfed by buy-to-let. Nor is the eventual sell-down a crash risk: disposals, if they ever happen, would drip out over half a century from the market's archetypal non-forced seller. And insurers owning housing is not even new — Prudential, Legal & General and their peers were among Britain's largest residential landlords in the mid-twentieth century, and exited precisely because rent regulation and political risk made housing a terrible insurance asset.

What has changed is not insurer appetite but the machinery pulling them back in — and the concentration that matters is in the flow, not the stock. If the state's delivery model for 1.5 million homes runs through forward-funding partnerships, the marginal home built in England is increasingly an institutionally-owned rental by construction. One does not need to own 30% of the market to set rents at the margin in the segments — new-build, urban, family rental — where the pricing power lives.

Germany ran this experiment first. Vonovia and Deutsche Wohnen consolidated hundreds of thousands of Berlin flats, and in 2021 Berliners voted in a referendum to expropriate them. That is the demonstrated end-state of institutional residential concentration: not quiet accumulation, but political rupture. Which raises the question of why UK insurers, with their own institutional memory of exiting residential, believe this time is different.

The answer is that this time, it is.

The lock-in

A property company can be regulated, taxed, competed against, and — as Berlin showed — credibly threatened with expropriation. An annuity fund cannot, because behind it stand the pensioners.

Consider a future government proposing rent controls at a moment when millions of annuities are backed by matching adjustment portfolios of rental cash flows whose "high predictability" was the basis for profits recognised years earlier. Rent regulation is no longer housing policy; it is a solvency event. The Prudential Regulation Authority, whose statutory objective is policyholder protection, is now structurally on the side of high rents — not through capture, not through lobbying, but in entirely good faith, because weakening the rent roll genuinely does threaten the pensions. The insurers will never need to argue against tenant protections. The prudential regulator will do it for them.

This is the mechanism that ownership-share arithmetic obscures. The state is engineering a constituency of pensioners whose retirement security is contractually adverse to the interests of the renters funding it, with the regulator positioned as enforcer. Once annuity security depends on rental income, no future government can cheapen housing without triggering a solvency question of its own creation. Housing costs will have been converted into a protected regulatory asset class.

That is the ratchet, and it clarifies what these firms are becoming. Not property companies — property companies are politically exposed. These are property companies wearing pensioners as armour.

The question worth asking

There is a legitimate version of every element here. Annuitants are owed their income; insurers must match it; the country must build. No actor in this arrangement is behaving indefensibly, and that is precisely what makes the emergent result so difficult to contest: intergenerational rent extraction, laundered through pension promises, with the profit capitalised on day one and the residual risk parked where it can never be politically touched.

So the question for policymakers is not whether insurers should invest in housing. It is this: if the state originates de-risked rental assets for annuity books — supplying the land, the planning, the development guarantee and the delivery vehicle — why does the illiquidity premium accrue to insurer shareholders rather than to the annuitants whose capital is deployed, or the taxpayer who absorbed the risk? And who, when the last pension of the BPA era is paid sometime late this century, was supposed to inherit the houses?

The industry describes bulk annuities as securing pensions to their final payment. That is true, and incomplete. The pensions end. The rents do not. The question of who inherits the difference was settled in the deal documents — and nobody asked the pensioners, the renters, or the public whose bank financed it.