Refinancing a London office in 2026: leverage, capex reserves and the case for not maximising

Office refinancing is the part of the London commercial market where the gap between what a borrower wants and what an underwriter will write is widest at the moment. Understanding why makes the conversation shorter.

London is several office markets, not one

Greater London holds roughly 26,721,000 square metres of office floorspace. The distribution is heavily concentrated: the City accounts for around 7,740,000 square metres, Westminster around 5,780,000, Camden and Islington together around 2,294,000, Canary Wharf around 2,120,000 and Lambeth and Southwark together around 1,780,000.

Those particular floorspace figures date from 2001, so treat them as establishing the relative scale of the sub-markets rather than as a current stock count. The relative picture is still broadly right, and it is the relative picture that matters when you are working out whether a lender has appetite for your specific sub-market. For current stock figures, the sources to use are CoStar, Savills or the Greater London Authority's own office policy work.

Canary Wharf carries around 16 million square feet of office and retail space and roughly 105,000 workers. The City of London had 500,000-plus people working in it as of 2019, against 8,583 residents recorded in the 2021 census across 1.12 square miles.

What has changed in how underwriters read an office

Three things, and they compound.

Reversionary assumptions get a harder read. A valuer who would once have taken an agent's estimated rental value at face value now wants evidence of recent lettings at that level in that building or a directly comparable one.

Capital expenditure is priced in. Large occupiers expect a standard of fit-out and amenity that older stock does not deliver without spending. Lenders increasingly want to see money reserved for that spend rather than assumed to appear from cashflow.

Lease expiries inside the fixed-rate period get weighted heavily. A floor coming back in eighteen months on a five-year facility is a live risk to the underwriter, not a future one.

The structural answer: take less

The instinct on a refinance is to draw the maximum. On a London office in 2026 that instinct is usually wrong.

Consider a multi-let floor, 84% let across four tenants at a combined £610,000 a year, with one lease expiring in eighteen months. A borrower wanting 70% leverage on an expected £8M valuation is asking for £5.6M. Run at 60% on a £7.6M valuation, the facility is £4.6M, ICR clears at 168% stressed, and the difference funds a £250,000 capital expenditure reserve for the floor coming back.

Two things follow from taking the lower number. The ICR headroom gives the underwriter room to accept the near-term expiry without demanding an uncapped personal guarantee. And when the valuation lands below expectation, which on London office stock in 2026 it frequently does, the shortfall does not break the facility.

That is the entire argument for not running a London office refinance at maximum leverage, and it is worth more than 30 basis points on the rate.

Practical structuring points

Reserve the capex rather than promising it. Underwriters treat a ring-fenced reserve account as materially different from an undertaking to spend.

Get the tenancy schedule right. Every unit, tenant, passing rent, lease start and expiry, break dates, review dates and arrears position. On a multi-let London office this single document decides how quickly a desk can quote and how seriously it takes the case.

Go to desks that actually want offices. Taking a London office refinance to the full panel wastes three weeks, because a desk with no sector appetite will simply price itself out rather than decline quickly. Three genuine office lenders beats twelve names on a list.

Deal with the ERC arithmetic honestly. If the existing facility carries an early repayment charge, model the break-even precisely rather than assuming the better rate wins. Sometimes it does not.

Market temperature, correctly labelled

We use HM Land Registry price paid figures as a sentiment gauge only. This is residential sold data. It is not commercial transaction volume, not commercial pricing, and not evidence about office values.

With that caveat attached: in the 12 months to May 2026 the Central London sub-region (seven boroughs) showed a £650,000 median across 10,200 residential sales, down 1.9% year on year, against Greater London at £525,000 across 60,368 sales, down 0.9%. The sharpest fall of any London district was the E14 postcode area, down 15.9%. Westminster fell 9.8% and Kensington and Chelsea 11.3%.

None of that measures office values. What it does indicate is that sentiment in the central and Docklands markets has been weaker than in outer London, and valuers work inside the same sentiment. Build the refinance to survive a conservative report.

Send us the tenancy schedule, the existing facility terms and the ERC position, and we will model the leverage that actually clears rather than the leverage you would like.