As the International Monetary Fund and World Bank meet in Bangkok, the IMF’s Global Financial Stability report repeats the warning it has issued several times already: leverage is building in the non-bank financial sector. This sector has become the place where an apparently manageable market stress may easily turn into something systemic.
The IMF analysis has been precise every time and the alarm is very genuine. But the issue is not the IMF diagnostic power, which has grown considerably over the years. It is the lack of a credible mechanism for fast, internationally coordinated, preventive action.
This problem became concrete last April. According to the Bank of England’s July Financial Stability Report, after the Middle East conflict broke out in late February, hedge funds’ net borrowing in the UK gilt repurchase agreement market fell by around 40%, from £100bn in December to mid-April. The market absorbed it, and positions have since rebuilt to around £85bn. Yet the Bank treated the episode as a warning, not a reassurance. It was a small tremor in a market everyone now agrees can produce much larger ones.
Failure by design
The absence of a coordinating mechanism traces back to a choice made after the 2008 financial crisis. The major economies faced a fork in the road: strengthen the IMF’s mandate to address systemic spillovers, or preserve national policy sovereignty behind a softer coordination architecture. They chose the latter.
The 2012 Integrated Surveillance Decision, which still frames IMF surveillance today, integrated bilateral and multilateral surveillance without changing members’ legal obligations. As I argued in a background paper for the IMF’s Independent Evaluation Office, the ISD was built to accommodate member sovereignty, not to override it, and the International Monetary and Financial Committee, which might have evolved into a genuine ministerial forum for real-time coordination, remained advisory. The G20 self-elevated as the premier forum for international economic co-operation precisely because it created no binding obligations.
This situation reflects what I call a fundamental asymmetry: the mismatch between nationally anchored policy-making and globally transmitted risk. The Fund’s own instruments were built to discipline members individually, not to govern the transmission channels between them. So surveillance can name a risk correctly and still have no purchase on the cross-border mechanics that make it systemic. The ISD’s surveillance instruments and the Fund’s actual capacity to influence member policy were left pulling in different directions by design, not by oversight.
An inadequate toolkit
The risks now topping every financial stability report are not new in kind; they are only faster. The ISD framework already strains against the slow-moving spillovers it was built to manage: surveillance can name an exchange rate misalignment or a fiscal imbalance, but it rarely reshapes one before it corrects on its own timeline. Now picture NBFI leverage spirals and repo fragility, a sovereign-bond trade unwinding in days, born not in a finance ministry but simultaneously in a US hedge fund, a UK pension fund, a European repo chain and a Japanese carry trade, crossing regulatory perimeters as it goes. What is the ISD framework meant to do with that?
The gap is visible in two separate, unilateral responses. The BoE has gone beyond backstopping gilt market dysfunction and is advancing minimum-haircut requirements and wider central clearing for repo trades, with detailed proposals due in early 2027. The US Securities and Exchange Commission, on its side, requires that eligible Treasury repo be centrally cleared from June 2027. Both institutions are filling a gap on their own, and on their own timetables. The IMF, positioned to coordinate a common standard, can only publish a paper about it.
Each national response addresses the same underlying vulnerability, leverage building through under-priced repo haircuts, but neither is required to consult the other before it acts. This is the fundamental asymmetry in action: the textbook outcome of national regulators responding separately to a problem that isn’t national at all.
What the Fund could do without a new mandate
Closing this gap does not require rewriting the ISD or reopening the question the IMFC has deliberately kept unaddressed for years. It requires the Fund to use tools it already has, through the monetary and capital markets department that already produces the GFSR twice a year and tracks NBFI leverage across jurisdictions with enough granularity to know where haircuts, margining practices and clearing timetables diverge.
There is a working precedent for turning that comparative knowledge into pressure without compulsion. After 2008, the G20 asked the IMF and Financial Stability Board to run the Data Gaps Initiative, a multi-year, non-binding programme that set common reporting standards across dozens of jurisdictions and published each economy’s progress against those standards on a fixed schedule. No government was compelled to comply. Yet most did, because a public scorecard carries its own reputational cost and the initiative has now run through three phases with participation from all G20 members.
A technical working group could be established under MCM, drawing in the BoE, the SEC, and the European Central Bank alongside the FSB. The group would benchmark haircut floors and clear deadlines twice a year. The GFSR would report periodically and make divergence visible before it hardens into arbitrage. This solution would avoid binding rules and be consistent with the ISD.
The Fund has a second lever it currently under-uses. The Financial Sector Assessment Program already benchmarks national banking supervision against a standard – the Basel Core Principles – which carries no legal force of its own. There is no reason the same benchmarking cannot extend to repo-market haircuts and NBFI leverage limits. Article IV consultations could then flag a jurisdiction’s rules as a source of cross-border spillover risk, exactly as they already do for exchange rate misalignments.
Neither tool closes the gap that Bangkok will once again make visible. However, the G20 Common Framework for sovereign debt shows that IMF-facilitated coordination without a mandate to compel is not automatically toothless when the alternative is apparent incoherence. MCM could convene that forum without waiting for the ISD to be rewritten.
What Bangkok can decide
The IMFC’s communiqués keep calling for resilience, vigilance and co-operation on non-bank financial intermediation. But the IMFC was never given authority to convene anything more binding. What Bangkok can decide is to task MCM with the standing, published benchmarking exercise. That is not a minor step, given how easily another communiqué could repeat the same call for vigilance and change nothing.
If governments are serious about financial stability, they need more than better surveillance and two separate sets of national rules moving at two separate speeds. A standing cross-border playbook for NBFI stress does not require a new legal mandate. It requires Bangkok to ask the Fund to use the tools it already has, rather than letting national regulators write their own fire code for the same building.
Interested in this topic? Subscribe to OMFIF’s newsletter for more.