Financial regulation, systemic risk, stability and AI
AI-enabled cyber attacks on the financial system are getting a lot of attention. But one ingredient is missing from the discussion, the double coincidence. Without it, policy risks false confidence and the wrong crisis preparations.
Does data lie to us when it captures the measurable but not the real? Yes, and it tells three lies. It omits, projects and flatters, misdirecting and leading us astray.
Artificial intelligence speeds up and intensifies financial crises and makes them harder to resolve. We need AI to meet the stability challenges posed by AI, only to find it is least reliable when needed the most. That problem is compounded because the authorities are likely to rely on the same models as the private sector while facing tighter restrictions on how they can use them. That…
Calls to simplify financial regulation are met with warnings that doing so would destabilise the financial system. Correct?
Can we measure extreme financial risk? Is financial stability only a question of technology and data? Many seem to think so. I disagree.
No jurisdiction can ignore stablecoins. They can be banned outright only in the most effectively authoritarian countries, or accommodated and even encouraged, which suits the United States and hub countries. For most countries, benign neglect is the best course of action, but countries where monetary sovereignty is already under threat are more vulnerable to stablecoin-linked risks.
Stablecoins are private digital money pegged to a fiat currency, overwhelmingly the US dollar. Dollar stablecoins reinforce American monetary and political hegemony, delivering fiscal and strategic benefits to the United States and hub jurisdictions while user countries absorb both the risks and the costs without corresponding control.
Stablecoins are private digital money designed to trade at par with a fiat currency. The promise of cheap and fast trading around the clock and across borders makes them run-optimised — faster, synchronised exits will compress stablecoin crises to minutes or hours. AI will exacerbate this.
Politics might be the single largest driver of financial crises, but impossible for the financial authorities to address — the fatal flaw in macroprudential policy.
"Same risk, same regulation" is one of those rare principles in finance that almost everybody seems to endorse. But is it as sensible as it sounds?
Authorities should proactively establish appropriate regulatory regimes for stablecoins rather than imposing outright bans. Since meaningful stablecoin exposure runs through regulated intermediaries, jurisdictions can monitor and control these assets to mitigate capital flight, banking disintermediation and financial instability while preserving beneficial innovation. With Ashutosh Jaiswal and…
The dark side of well meaning financial regulations is when they force banks to act against the system, like pouring petrol on a fire.
Increasingly stringent financial supervision paradoxically increases systemic fragility. The very attempt to safeguard stability can increase systemic risk by increasing complexity and synchronising behaviour. Rather than pursuing ever-tighter risk control, the authorities should focus on resilience through institutional diversification.
A data-driven risk management process can not capture endogenous risk, only exogenous risk.
The almost infinite complexity of the financial system is the main reason why it is so hard to keep it under control. And that complexity is due to everybody that works in the financial system having incentives to increase its complexity, including the regulators.
Artificial intelligence is transforming finance faster than the authorities can adapt. This column argues that while AI enhances the financial system’s efficiency, it also poses new, poorly understood risks. To stay relevant, the supervisory authorities must develop their own AI capabilities, use federated learning for cross-jurisdiction models, and design intervention facilities that can respond…
Warnings of an AI stock market bubble abound. Should investors and policymakers be concerned? Innovation-driven bubbles can yield real benefits when financed by equity rather than debt. Even if investors face large losses, policymakers have little reason for concern, unless it drives risk appetite elsewhere in the system and AI investment relies on bank credit.
Should supervision be about finding the optimal level of risk? It sometimes seems that way. A sensible idea in portfolio management, less so when applied to supervision.
There are plenty of warnings about a pending stock market crash, alongside a few hopeful voices. The only real uncertainty is when it will happen.
The common, perhaps even prevailing, culture among supervisors is to see undesirable outcomes as the result of insufficient resources, information and power. If only they knew all the positions and the associated risks, and were able to ensure the private sector had the right processes in place to control that risk, all would be well — Perfect supervision.
Artificial Intelligence (AI) adoption in finance is accelerating, giving private sector firms speed and agility that overwhelm human-centred supervision. Unless the authorities respond, they risk losing control of an AI-driven system. To remain effective, the supervisors must a) embed AI capabilities within core supervisory functions, b) favour domestic or collaborative AI engines to reduce…
The US dollar has been the reserve currency for the past century, never seriously challenged. Is that now under threat with Trump?
The America First thrust of the Trump tariffs caught many by surprise. But it shouldn't have. We have been there before — in (my) living memory.
The growing use of artificial intelligence (AI) poses difficult challenges for the financial authorities. AI allows private-sector firms to optimise against existing regulatory frameworks, helps those who want to damage the financial system, amplifies wrong-way risk and speeds up financial crises. It also gives the authorities new tools for executing their mandate. The authorities could become…
Systemic financial risk has both internal and external drivers. So, when we focus too strongly on preventing internal crises, such as the 2008 Global Financial Crisis, we tend to miss out on the more important external risks. Five external risk factors stand out: populism, debt-driven death spirals, manufactured tensions, artificial intelligence and geopolitics. To combat these threats, the…
The risk we measure is usually not the risk we care about. Why and does it matter?
Climate risk is not a systemic financial risk. Folding it into financial regulations will likely backfire, not mitigate climate damage and even increase systemic risk.
Financial institutions are rapidly embracing AI – but at what cost to financial stability? This column argues that AI introduces novel stability risks that the financial authorities may be unprepared for, raising the spectre of faster, more vicious financial crises. The authorities need to (1) establish internal AI expertise and AI systems, (2) make AI a core function of the financial stability…
Modern financial risk forecasting has much in common with scientific socialism, relying on pseudoscientific models that can create a false sense of security. Effective risk management requires acknowledging these limitations.
Financial crises usually inflict the most damage when banks suddenly shift from pursuing profits to survival. This column argues that such drastic behavioural changes render statistical analyses based on normal times ineffective. That is why we cannot predict the likelihood of crises, or what banks will do during those crises. Since this behaviour arises from a natural desire for…
Almost everyone insists that Expected Shortfall (ES) is superior to Value-at-Risk (VaR). But is that truly the case?
When measuring model risk with risk ratios, we find it highest when markets are in turmoil. In other words, model risk is highest, and the risk forecasting models least reliable, when they are needed the most.
What are the ideal properties of risk measures? Should we disregard VaR because it is not subadditive? Did the Basel Committee make a mistake by moving to ES? And what is the difference between theory and practice?
Only two variables predict financial crises: Credit growth and low risk. Everything else is the consequence of a crisis already happening. It is pointless to look at market prices, VIX, CDS and their ilk. They are only correlated with crises but don't predict them.
The tools my life runs on. What is impressive, and what is tolerated. All can be replaced, except one without which I would stop working.
Artificial intelligence can act to either stabilise the financial system or to increase the frequency and severity of financial crises. This second column in a two-part series argues that the way things turn out may depend on how the financial authorities choose to engage with AI. The authorities are at a considerable disadvantage because private-sector financial institutions have access to…
The rapid adoption of artificial intelligence is transforming the financial industry. This first of a two-column series argues that AI may either increase systemic financial risk or act to stabilise the system, depending on endogenous responses, strategic complementarities, the severity of events it faces, and the objectives it is given. Stress that might have taken days or weeks to unfold can now…
Financial crises are not complicated, and many claim to know why they happen and how to prevent them. Why then do they happen with such alarming frequency? This column argues that a key reason is failures in regulation, and proposes that the financial authorities instead adopt a fundamental concept from finance – diversification. By doing so, we can build a more resilient system that not only…
Artificial intelligence will both be of considerable help to the financial authorities and bring new challenges. This column argues the authorities risk irrelevance if they are reluctant and slow in engaging with AI, and discusses how the authorities might want to approach AI, where it can help, and what to watch out for.
Risk model hallucination happens when models are forced to forecast the likelihood of extreme events in cases where they have not been trained with such extreme outcomes. This is surprisingly common in applications such as financial regulations, pension funds, reinsurance, as well as the containment of market upheaval and financial crises. This column argues that measuring systemic financial risk…
As artificial intelligence makes inroads into the financial system, it exacerbates existing channels of instability and creates new ones. This column identifies several such channels, malicious and misinformed use, misalignment and the evasion of control, and finally risk monoculture and oligopolies. All arise when AI vulnerabilities interact with economic fragilities like strategic…
The use of artificial intelligence in the private sector is accelerating, and the financial authorities have no choice but to follow if they are to remain effective. Even when preferring prudence, their use of AI will probably grow by stealth. This column argues that although AI will bring considerable benefits, it also raises new challenges and can even destabilise the financial system.
Artificial intelligence is expected to be widely used by central banks as it brings considerable cost saving and efficiency benefits. However, as this column argues, it also raises difficult questions around which tasks can safely be outsourced to AI and what needs to stay in the hands of human decision makers. Senior decision makers will need to appreciate how AI advice differs from that produced…
Crypto-promoters and financial authorities are split on the future of cryptocurrencies. Should crypto join the mainstream or remain in the wilderness? Should the authorities control and extinguish crypto, ignore it, or embrace it? This column argues by adopting the best ideas from crypto, it can be leveraged to improve the efficiency of the financial system and reduce rent extraction.
The fallacy of composition in financial regulations is that if all the banks are prudent, keeping all their individual micro risks under control, the entire financial system is safe.
The financial system is infinitely complex, so the supervisors can only patrol a small part of it, why it is so hard to regulate finance effectively.
There are two main approaches to macroprudential regulations. The first is the current one of identifying all risks and using them to build buffers against shocks. Alternatively, shock absorption can be used to boost resilience. Which is better?
The financial regulators have recently taken an active interest in cryptocurrencies, more than a decade after their law enforcement counterparts did. So why are they doing this now, and what will the consequences be? Regulators feel compelled to respond due to political pressures and their actions may backfire.
The downfall of Silicon Valley Bank and Credit Suisse has exposed failures in how we regulate the financial system. This column argues that the problems we now see in the system have arisen because the financial authorities have been trying to do the impossible: maintain growth while keeping inflation under control and financial stability high. The best way forward would be to focus on shock…
The collapse of Silicon Valley Bank shows that banks still pose risks. Are they systemic? While it is unlikely that the failure of SVB will lead to a crisis, it shows us that the financial system is much more fragile than the public had been led to believe.