A little less conversation, a little more data

23rd October 2025 

Key takeaways

Proper financial stability monitoring requires better data at global level.

• For IOSCO, threats to financial stability coming from asset management and connected sectors are a priority.
• However, the discussion so far has been very general and sometimes nebulous. The so-called “NBFI” is a too-diverse, too-heterogeneous term to have precise and effective regulatory discussions.
• Data accessible to regulators is fragmented, incomplete and suboptimal and would benefit from targeted reforms.

Without good data, not only we might have diminished visibility for supervisors but, as importantly, blunter and less targeted policy or macroprudential measures, which may damage investors unnecessarily.

Paraphrasing Elvis Presley on financial stability may seem unusual, but the topic is as exciting as rock and roll for many of us in the regulatory community. 

If the many crises in the last three decades have taught us anything, it is that financial stability risks can build up in any market and in any sector, be it commercial banking, investment banking, bond markets, asset management, equity markets, commodities or derivatives, just to name a few. It would be foolish to claim that some sub-sectors or some corners of the financial markets are immune to the build-up of vulnerabilities with potential systemic implications.

It is equally foolish to claim that all corners, vehicles and activities are equally exposed, at any given time, to the build-up of those vulnerabilities.

In recent years, the global regulatory community’s attention has once again been brought into sharp focus by a series of new market stresses, particularly but not only related to the Covid-19 pandemic. Policy makers have therefore addressed the delicate task of balancing ongoing support for economic growth with the need to monitor and mitigate potential systemic risks, which is a conversation that has become omnipresent.

IOSCO, founded in 1973, has been and continues to be an important voice in the global discussion around financial stability. One of our three core objectives is to prevent the buildup of systemic risk. Since the inception of the FSB in 2009, a fruitful regulatory dialogue was established across the entire regulatory community (securities, derivatives, asset management, insurance and banking) to identify vulnerabilities and to tackle them before they crystalize.

In this context, we have moved on from the old “shadow banking” and we have been recently using the NBFI acronym (Non-bank financial intermediation), which is not a precise concept either, for three main reasons. 

First, because it defines by exclusion (non-bank) a set of activities that are now as large as the banking sector. Globally, market-based finance amounted to close to $250 trillion in 2023, out of global financial assets estimated to be in the realm of $500 trillion.

Secondly, the NBFI definition is a too-diverse crowd, entailing many different risks. From traditional insurance to hedge funds; from plain-vanilla ultra-liquid ETFs to private unregulated investment trusts; from sovereign funds to family offices or retirement plans. Trying to find commonalities among such a diverse ecosystem would be similar to a congress of biologists trying to find remedies for “non-humans”, taking together alligators, mushrooms and sequoias. The chances of success of that exercise would be certainly slim.

Thirdly, because, as opposed to banks, only a small part of NBFI does actual intermediation. A family office or a sovereign fund, for instance, are not intermediating. They are simply investing their own money.

As the regulatory debate advances, we need to build a better taxonomy, a more precise language. One that allows a debate on specific types of vehicles, specific types of risks and specific geographic areas. Only through a more precise, focused and concrete discussion will we be able to advance meaningfully on the prevention of specific sources of systemic risk. 

Nonetheless, we are still using the “NBFI” all-encompassing category. The question is how to break down that concept in a manner that allows for the detection of the real pockets and sources of vulnerabilities.  How to break down the debate in a meaningful way, one that avoids putting sequoias and alligators in the same discussion.

The answer to that question is data.

We need to improve, collectively, the data that supervisors, regulators and international bodies receive on matters related to financial stability. Of course, this encompasses a very diverse range of actors and a heterogenous set of data. For example, data on bank’s balance sheets and exposures are more developed at global level and data on open ended funds are quite robust and comprehensive in many jurisdictions, although surprisingly scarce or lacking detail in many other countries. Data on exposures of hedge funds are a bit in the middle; partially present in some jurisdictions but absent in some important ones. At the end of the spectrum, data on private and unregulated vehicles and funds are nowadays the most difficult to obtain and interpret.

Let’s take leverage as an example. Leverage often takes center stage in this debate because of its prominence in recent instability episodes. Leverage affects both banks and investors equally. That is because banks are frequently at the center of the buildup of leveraged positions, be it as prime brokers, providers of credit or counterparties of derivatives.  This creates channels of contagion and contributes to the systemic nature of recent crises. However, leverage can be a positive force too. It increases market efficiency (by allowing arbitrage which corrects prices deviations), allows long term strategies that are essential for pension vehicles and increases market liquidity (and hence resilience). The problem is usually excessive leverage, non-visible leverage or, very frequently, too-abrupt deleveraging. This has been tackled head on by the FSB, of which IOSCO is an active contributor, with the FSB’s NBFI Leverage Recommendations published in July 2025. The message is clear: supervisors need better tools to monitor leverage; how it is spread, where it concentrates, which are the channels through which excessive leverage could destabilise the system. And they should have the tools and powers needed to intervene when the case arises. The only way to do this effectively is to have the right data. 

IOSCO has put a lot of effort to consolidate, at global level, the patchwork photograph that national data reporting regimes provide. We have published this year, for the first time, a global Dashboard on investment funds. It collects data from 39 jurisdictions, and captures 80% of the global assets under management, from 120,000 funds. Anyone can consult it through www.iosco.org and it provides figures of gross synthetic leverage and gross and net leverage by country, type of fund and strategy. It is an immensely rich database, which can throw some light to the more nuanced discussion as to which sectors, types of vehicles and even countries pose greater risks for financial stability. As a first try, is not perfect, of course. We will need to broaden the number of jurisdictions and vehicles covered, work on further metrics, improve the timeliness of publication, etc. Nonetheless, it marks a significant step in the direction of having more evidence-based regulatory discussions on key topics.

For example, if we look at the European region, we see that gross leverage in open ended funds (including interest rate and FX derivatives) was on average close to 1.4 times NAV in 2023. That same metric rises to 34 times NAV when we measure hedge funds. However, if we exclude interest rates and FX hedges or bets, European hedge funds show only 5x NAV leverage. The differences are even starker across countries: hedge funds in some countries in the European region showed 49x NAV leverage compared to Luxembourg hedge funds (AIFs in the EU jargon), which showed only 2x NAV. 

The conclusion is two-fold and clear: 1) data is the path to identify the risks that really matter and 2) there is no such thing as “global excessive leverage on NBFI”, as a kind of trend or global pattern. There are sector-specific, country-specific circumstances, which need to be addressed as precisely as possible, avoiding painting with a broad brush. That’s why the strong call by the FSB is for regulators and standard setters to develop within their jurisdictions the monitoring tools and the powers and competences needed to tackle those specific vulnerabilities.

However, we need to recognize that the current framework for reporting and supervision was built after the Great Financial Crisis and focused on the risks that materialized then, which are not identical to today’s key vulnerabilities. We need to review, at least at jurisdiction level, whether the reporting tools and the data sets at the disposal of supervisors are adequate for today’s risks or, even better, for tomorrow’s risks. For instance, data on leverage and exposures between systemic counterparties appear more relevant today than data on individual derivatives or repo transactions. Similarly, cross border exposures and collateralization have acquired new relevance with some market incidents and with the recent full deployment of the global system for collateralising bilateral derivatives exposures. But, besides specific improvements, there is value in having data sets with some commonalities that allow integration and aggregation at regional, sectoral or global levels. It takes us at IOSCO more than 9 months of hard work to integrate data from completely different reporting regimes, in order to obtain quality metrics and comparable data for our dashboard. And, in today’s markets, 9 months is way too long.

In IOSCO we are fully aware of the difficulties of bringing together national reporting regimes, to increase the chances of integrating those data, sharing them internationally (even on an aggregate manner, through key metrics) and building a global tool that allows the regulatory community to better understand and detect risks that could have global implications. It is clear that governments and parliaments are the ones with the legitimacy to determine which reporting regime suits their countries or regions best. But there is a broader benefit of being able to get timely and precise data about exposures that could destabilize global markets, damaging investors and citizens across borders. This is a goal worth pursuing.

The idea of having accurate indicators of potential vulnerabilities is pretty straight forward. However, it is not just the natural curiosity of supervisors which justifies sophisticated, precise and compatible reporting frameworks. Although it may seem paradoxical, it also benefits market participants, asset managers and their clients directly. The more we know, as a regulatory and supervisory community, the less invasive and costly will be any potential supervisory or macroprudential measure. The best way to avoid too-broad policy tools, or unnecessarily aggressive interventions is to have a proper and precise diagnostic based on reliable and timely data.

Reporting entails costs, no doubt. And the international mood in many regions is to lower the burden of regulation, not increasing it. We have seen political institutions in many regions calling for a cut on all reporting regimes, linked to the pro-growth agenda and it is imperative that regulators take into account those views. Some technology and data standardisation (like the use of LEIs, ISINs and other robust identifiers, which IOSCO has traditionally encouraged) can play an important role in reducing that cost and increasing the usability. The rise of AI can also offer chances to reduce the cost of reporting and also the cost of processing those reported data by regulators. But reported raw data, at international and national level, is still suboptimal to spot vulnerabilities and needs improvement.

On this particular topic, on this particular juncture, maintaining suboptimal and partial data sets or, even worse, rushing to cut reporting regimes related to financial stability risks could be a mistake. It might lead to poorer supervisory and risk-detection capabilities and, therefore, larger accumulation of vulnerabilities, deeper crises and blunter policy measures and interventions than needed. We need more, better and more harmonised and internationally comparable data to monitor systemic vulnerabilities.

Diagnostic image has a cost, but when someone has to undergo surgery, better to have the surgeon looking at a recent MRI than operating without any image assistance. The same applies to macro-prudential and micro-prudential interventions. In those instances, good data comes actually very cheap (also for the patient). So…a little less conversation and a little more action (at least on data), please.

Note: The views in this article are solely the author’s, not necessarily representing the views of IOSCO.

To comment, please go to the discussion forum