Growth and regulation

1. Slow growth and why it matters

Speaking at a Mansion House-style dinner, Bank of England Governor Andrew Bailey admitted that his February hope for calmer times was upset by the Middle East conflict: energy prices jumped, inflation has not yet returned to 2% although it otherwise would have, the economy is soft, and the return to target will simply take longer. He then stepped back to the long run. The UK has grown more slowly for about fifteen years, and was hit by a string of negative supply shocks such as Covid, the Ukraine war and Brexit. The sustainable speed limit, potential supply growth, fell from about 2.8% a year before the financial crisis (2.4 points from productivity, 0.4 from labour) to about 1.3% since (0.4 from productivity, 0.8 from labour). Income per head grew 2.0% a year between 1990 and the crisis but only 0.6% afterwards, so lifting growth must be a priority. Stable monetary and fiscal policy is the backdrop, and he believes the frameworks provide it, but lasting growth comes from the supply side: capital, technology, skills, trade, finance, competition, infrastructure and institutions.

2. Regulation framed by purpose

Focusing on regulation, he agreed that critics deserve a hearing, since well-designed rules can correct market failures, lower the cost of dealing, make it easier for newcomers to enter and encourage rivalry and innovation, and financial regulation lowers the odds and severity of crises, protecting the supply of finance. But the debate should not be an abstract argument about too much or too little: each rule should be defined by a positive purpose, designed to reach it quickly and effectively with minimal side effects, and judged on whether it does. Financial stability is a prerequisite for growth, yet not every rule is perfect, and simply arguing for less is unhelpfully reductive.

3. Bank capital

On capital, he went back to basics: banks are regulated to protect deposits, the main stock of money people hold, while still lending through the cycle; capital absorbs losses and backs the assured value of money. Capital providers must earn more than the loss they expect to bear, otherwise they will not supply it. UK bank returns sat below 10% for much of the post-crisis period, rose above it in 2023 and reached 15.4% this year, with price-to-book ratios near 1.7, signalling expected returns above the cost of capital. Regulators should not cap returns, but stability needs returns above the cost of capital. Last week's changes removed some anomalies, and the Bank is open to simplification that keeps the system healthy. He rejected a lump-of-capital view, where a fixed pool of capital means looser rules free more lending: capital is not fixed, healthy returns let banks retain earnings and build capital, and proportionate, effective regulation lowers its cost.

4. Payments and tokenisation

For payments, the two objectives are the singleness of money, meaning the same nominal value across all forms, and settlement finality, so that payments once made stay made. The Bank wants to modernise systems, improve resilience against fraud and cyber attacks and gain the benefits of programmable payments: conditional payments could cut late payments and fraud by moving checks from after the fact to before. The renewed RTGS system underpins this; a Digital Securities Sandbox approved its first live participant yesterday, paving the way for a UK digital gilt, DIGIT; work is under way with banks on tokenised deposits; proposed rules for systemic sterling stablecoins apply the same principles; and a Retail Payments Infrastructure Board published a consultation on 25 June, with feedback due by 11 September, ahead of a blueprint for an industry-owned delivery company. He also noted that atomic settlement could improve efficiency and cut risk in wholesale markets.

5. AI, growth and law

He argued that general-purpose technologies, from steam to electricity to the internet, drive long-run growth in long cycles with gaps between them, and that the last fifteen years, with falling productivity growth, look like such a gap before the next technology, most likely AI, which will support advances in diagnostics, research and robotics. Benefits take time to appear, as Solow's remark about computers showed, and the Governor himself is still experimenting. AI will raise big policy questions that should be tackled promptly and positively. One is resource allocation: if clean power for data centres is short, who decides whether medicine, drones, finance or entertainment get priority, and markets may not reflect the public interest. Another is law: agentic AI that buys or trades on our behalf can learn bad habits such as lying to please, and without a legal personality it is unclear whether the principal is liable for everything, which may force constraints that cut against AI's direction and test firms' governance.

6. Frontier AI and cyber risk

Frontier models' growing ability to find and exploit weaknesses in systems is a major threat to security and hence financial stability. He called for much stronger international coordination of testing before models are widely released, starting with public authorities, where the UK's AI Security Institute and National Cyber Security Centre are well placed, and then testing across critical infrastructure; the Bank is working with regulated firms on forums for patching. He also stressed the need for robust ability to recover from scratch, so-called bare metal recovery. His conclusion: growth comes from supply-side innovation, AI is the likeliest next contributor, regulation done well helps, and it should be approached positively by defining what to achieve, by when, and how, while questioning existing rules in that spirit.