Terra insight · Market commentary

The Return of Price Discipline

1. Fifteen Years of Forgiveness

For fifteen years, the UK property market got away with a lot. Cheap debt covered a multitude of underwriting sins, and falling yields flattered even mediocre stock selection. Buy something reasonable, gear it, wait, and the market did the work. That regime has ended, and the numbers are unambiguous about it.

2. Leverage Has Stopped Adding Value

The arithmetic of gearing is simple: borrow below the yield on the asset and you enhance equity returns; borrow above it and you destroy them. For most of the last decade, the first condition held comfortably. It no longer does at the prime end of the market.

Prime City offices were quoted at a 5.25% net initial yield in Knight Frank’s June 2026 yield guide, and prime distribution at 5.25% [2]. Meanwhile senior commercial debt prices off a five-year SONIA swap that stood at 4.71% on 11 September 2026 [3], and once a realistic margin is added the all-in cost sits comfortably above 6%. Development finance is a different world again — 7.5–10% all-in from challenger banks and 9.5–12.5% for stretched senior, on Construction Capital’s May 2026 survey of the market [5].

The spread has not narrowed. It has inverted.

In some secondary and regional markets the arithmetic still works. A major regional city office at 6.50% or a regional shopping centre at 7.25% [2] gears positively against debt in the mid-sixes — thinly, but positively. Those opportunities are rarer than they were, and they deserve a harder question than whether the leverage is accretive: why does the asset trade at that yield in the first place? The market has priced something into it — a covenant, a location, a lease length, an obsolescence risk — and gearing does not remove that risk. It concentrates it in the equity.

The consequences are already visible in the loan books. Only 37% of UK commercial real estate loans now carry an interest cover ratio above 2x, down from 73% in 2017, and 13% sit below 1x — the income does not cover the interest [4]. At those levels, as Bayes Business School’s lending research puts it, losing a single tenant can turn into a payment default.

This is not a financing problem. It is a pricing problem wearing a financing costume.

3. The Equity Buyer Now Sets the Price

When debt is accretive, the levered buyer can always outbid. When debt is dilutive, that advantage reverses: the buyer with no borrowing requirement — the institution, the charity, the family office, the overseas equity buyer — can pay a price that the levered fund simply cannot justify to its investors.

If you are bidding with debt in the capital stack today, you are not competing on cost of capital. You are competing on what you can do to the asset.

4. Why Yields Still Have Work to Do

Here is the part the market has been slow to accept. On 11 September 2026 the ten-year gilt yielded 5.36%; the twenty- and thirty-year gilts sat at 5.85% and 5.91% [1]. The long end has been hovering near 6%, a level last seen in the 1990s.

Prime property, in other words, is being priced at or below the risk-free rate. Knight Frank’s own guide quoted the ten-year gilt at 4.97% on 4 June 2026 [2]; it has moved out roughly forty basis points since, and the prime yields have not followed. City offices at 5.25% against a 5.36% gilt is a negative nominal spread. West End core at 3.75–4.00% is more than a hundred and thirty basis points below it.

That cannot be right, because property is not risk-free and never has been. The investor who buys a building rather than a gilt takes on:

• Illiquidity. A gilt settles in two days. A building takes months to sell, and in a poor market may not sell at all at the carrying value.
• Obsolescence and capex. Gilts do not need a new roof, a plant replacement or an EPC upgrade. Buildings depreciate physically and functionally, and the regulatory bar keeps rising.
• Transaction costs. Roughly 6.8% to buy in England — the standard purchaser’s costs allowance in UK valuation practice, being stamp duty land tax with agent’s and legal fees on top. A gilt costs basis points.
• Tenant and income risk. A covenant can fail. A lease can end. A gilt coupon is a sovereign obligation.
• Valuation uncertainty. You know what a gilt is worth to the penny. A valuation is an opinion.

Each of those deserves compensation, and that compensation has to come from somewhere. It comes from yield. If the risk-free rate has repriced upwards and property yields have not followed, then either the risk premium has been compressed to nothing, or the market is assuming rental growth it has not yet seen. Knight Frank’s June commentary says as much from the other direction: returns “are likely to be driven more by income and rental growth than capital appreciation” [2].

5. You Cannot Underwrite a Yield You Don’t Control

Which brings us to the discipline point.

Between 2010 and 2021, a great many appraisals were rescued by the exit. Model in fifty basis points of compression on the way out and almost anything clears a hurdle rate. It worked, repeatedly, for long enough that it stopped feeling like an assumption and started feeling like a forecast.

It was never a forecast. Yield shift is a market variable. You do not control it, you cannot hedge it, and you have no informational edge in predicting it. Underwriting an exit yield tighter than your entry yield is not modelling — it is taking a position on interest rates and dressing it as an appraisal.

The honest approach is the conservative one: exit at or above entry, and make the deal work on the things you can influence — the price you pay, the rent you achieve, the cost you control, the lease you structure, the asset management you actually execute.

6. What This Means in Practice

Returns now have to be earned, not received. That changes what an appraisal is for. When yield shift was doing the heavy lifting, the model was a formality. When it isn’t, the model is the decision.

Which means the questions that matter have changed:

• What is the most I can pay and still hit my target return — and how does that answer move across profit on cost, IRR, leveraged and unleveraged?
• Does this deal still work if the exit yield is fifty basis points softer, not tighter?
• Where exactly does the cash go negative, and for how long?
• What does the debt actually do to my equity return — and is it doing anything at all?
• If rental growth comes in at zero, what am I left with?

None of those are answerable with a headline residual and a gut feel. They need a proper monthly cash flow, real lease-level assumptions, and sensitivities that move more than one variable and report more than one metric.

Terra solves the residual value against eleven target returns, runs sensitivities across several KPIs at once, and models leases letting by letting — see what you get.

7. Conclusion

Buying at the right price has always mattered. What has changed is that nothing else is coming along to bail you out if you get it wrong.

8. The Numbers, Dated

Every figure in the article, with the date it was read and the source it was read from. Not counted in the reading time above.

Gilts, base rate and swaps

Prime net initial yields

All from Knight Frank’s Prime Yield Guide dated June 2026 [2].

Debt

Investment senior debt is priced as the five-year swap plus a margin. Bayes reports that UK banks cut prime office margins by around 45 basis points, and debt funds by around 30, during a refinancing price war in early 2026 [4]; even so, all-in pricing lands in the region of 6.2–7.2%.

9. Sources

Sources 2, 3 and 5 were re-fetched on 14 September 2026 and every figure taken from them checked against the document; source 1 was re-fetched and showed the ten-year gilt at 5.39% that day, consistent with the 5.36% read on 11 September. The twenty- and thirty-year gilt figures and the loan-book statistics in source 4 are as read on 11 September 2026.

This article is commentary for general information, not investment, valuation or financial advice. Market figures are quoted as at the dates stated and will have moved since; any appraisal should be prepared on current data and in accordance with the applicable professional standards.