A decade-old estimate gets a stress test
Back in 2017, Bank of England researchers Sebastian de-Ramon and Peter Straughan built a statistical model estimating how changes in bank capital ratios feed through into the price of credit. Their conclusion was that raising capital requirements does raise the cost of corporate borrowing somewhat, while doing almost nothing measurable to mortgage rates. That finding mattered because it fed directly into how the Prudential Regulation Authority weighs the costs and benefits of capital rules.
The obvious problem with a 2017 estimate is that almost everything about UK banking has changed since then: full implementation of post-crisis reforms, years of near-zero rates, and then the abrupt shock of the COVID-19 pandemic. Updated analysis from Federico D’Amario, Sebastian de-Ramon and William B. Francis reruns the same type of model with data running through 2024 — effectively asking whether the original conclusion survived contact with a decade of upheaval. It is a useful kind of study precisely because it is not trying to produce a dramatic new finding; it is trying to find out whether an old one still holds.
How a capital rule is supposed to travel through the system
Before looking at the numbers, it helps to trace the mechanism the model is actually measuring — because "banks pass on the cost" hides several distinct steps.
flowchart TD A[Capital ratio requirement rises] --> B[Bank adjusts balance sheet mix] B --> C[Pricing shifts more on corporate loans] C --> D[Corporate borrowing costs increase] D --> E[Investment and output effects]
When regulators require banks to hold more capital relative to their risk-weighted assets, a bank has several ways to comply: issue new equity, shrink its loan book, shift toward lower-risk assets, or reprice existing lending. In practice, banks tend to do some combination of all four, and the updated study confirms — consistent with earlier work by de-Ramon, Francis and Harris — that banks typically keep a voluntary buffer above the regulatory minimum, which itself shapes how quickly and how far they adjust.
The repricing step is where the corporate-versus-mortgage split emerges. Corporate loans generally carry higher regulatory risk weights and shorter maturities than residential mortgages, which makes them the more convenient lever for a bank trying to optimize its capital position quickly. That is a structural, mechanical reason for the asymmetry — not a claim about which borrowers regulators intend to burden.
What the updated numbers say
The headline result is that the corporate lending channel remains the dominant pathway, and it has not moved much. A permanent one-percentage-point increase in the risk-based capital ratio is now associated with an increase of roughly 8 to 10 basis points on corporate lending spreads — close to the 10 to 12 basis points found in the original 2017 study. A basis point is one hundredth of a percentage point, so this is a small number in absolute terms, but it is a persistent, structural adjustment rather than short-lived noise.
Mortgage spreads tell a different story: the long-run relationship between capital ratios and mortgage pricing remains statistically indistinguishable from zero, just as it was in 2017. That is a meaningfully different conclusion from "mortgage spreads rise a little" — the data cannot rule out that there is no long-run effect at all.
| Dimension | Corporate lending | Mortgage lending |
|---|---|---|
| Long-run pass-through per 1pp capital increase | ~7–10 basis points (updated); ~10–12 bps (2017) | Not statistically distinguishable from zero |
| Speed of adjustment | Adjusts relatively quickly toward new equilibrium | Adjusts more slowly, if at all |
| Effect of including COVID-period data | Adds short-run volatility, but long-run estimate holds | Adds short-run volatility, but long-run estimate holds |
| Robustness to added competition measures (Lerner index, Boone indicator, HHI) | Central result unchanged | Central result unchanged |
The researchers also tested whether their conclusions were an artifact of the extraordinary 2020–21 shock. Including COVID-period data does inject short-run volatility into the model, which is unsurprising given how unusual that period was economically. Following the approach used elsewhere in the economics literature for handling pandemic-era outliers, the authors relied primarily on estimates that either exclude the COVID period or explicitly control for it with time dummies — and in both versions, the long-run relationship held up. They further extended the model with newer measures of banking-sector competition, and none of it materially changed the central finding: the pass-through stays concentrated in corporate spreads.
Why corporate borrowers, and not mortgage holders
The asymmetry is not arbitrary. Corporate lending sits closer to the sharp edge of a bank’s capital-optimization decisions because it is riskier per pound lent, shorter in duration, and therefore easier to reprice without waiting out a 25-year mortgage book. But there is also something on the borrower side of the equation worth naming plainly: corporate borrowers, especially smaller and mid-sized firms, are typically more sensitive to financing costs than large diversified banks or the public bond markets a blue-chip company can tap instead. A firm that depends heavily on bank credit and has few alternative funding channels absorbs a repricing shock more directly than a household with a fixed-rate mortgage locked in for years. That is the practical reason the "corporate lending channel" is treated as a route through which capital policy can touch the real economy — not just a statistical curiosity, but a plausible channel from bank balance sheets to business investment.
What this evidence does not settle
It is worth being precise about the boundaries of what has actually been shown. The study does not tell us how much of that 8-to-10 basis point increase in corporate spreads is ultimately absorbed by shrinking bank or corporate profit margins, passed further downstream to the customers of those corporate borrowers, or offset by firms shifting toward other sources of funding. It is a UK-specific estimate built on UK regulatory history, and there is no evidence here that the same magnitude — or even the same direction of asymmetry — would show up identically in another country’s banking system or in a different part of the credit cycle. Nor does the paper break down which industries within corporate lending feel the effect most; "corporate borrowers" is a broad category covering firms with very different reliance on bank credit. None of this undermines the core finding, but it does mean the number should be read as a benchmark from one well-studied system, not a universal constant.
A modest cost is still a cost
The temptation with a finding like this is to pull it toward one of two extremes: either "capital requirements are basically free" or "capital requirements quietly punish borrowers." Neither reading survives contact with the actual numbers. A single-digit-to-low-double-digit basis point increase in corporate spreads, concentrated in one segment of lending and statistically absent in another, is neither trivial nor dramatic — it is exactly the kind of contained, imperfectly distributed cost that cost-benefit analysis is supposed to weigh against the benefit of a more resilient banking system. What this update mainly demonstrates is durability: an estimate built before the last decade of regulatory reform and economic shocks still describes the system reasonably well afterward. For anyone trying to evaluate popular claims about bank regulation, that stability is itself useful evidence — not proof that the debate is settled, but a sign that the effect being debated is real, specific, and considerably smaller than the loudest arguments on either side tend to suggest.
