How far do asset managers create systemic risk?

The world’s financial authorities have been discussing whether asset managers should be designated as systematically important financial institutions, or SIFIs. This follows such designations for banks and insurance companies. For now, the authorities have concluded that such a designation is not recommended, but concerns about the systemic impact of asset managers are unlikely to go away.

This attention was prompted by the G20 meeting in November 2011 directing the Financial Stability Board (FSB) and the International Organization of Securities Commissions (IOSCO) to establish methodologies for identifying SIFIs.  This has been very controversial, as discussed by the Economist (2014). The FSB and ISOCO recently (2015) came out with their second consultation document, outlining how they aim to identify SIFIs. The proposals attracted a number of industry comments, with Blackrock (2015) just one example.

The exercise is quite problematic. The mandate to identify SIFIs is not accompanied by any instruction on what to do with institutions designated as such, even though it does not seem logical to separate out the identification and the remedy. The FSB and IOSCO do not seem to see eye to eye on this question, a division deepened by subsequent public statements by these and other public bodies.

The reason for this state of affairs is the fact that one really can only designate institutions as SIFIs if one has previously identified in what precise sense they may contribute to systemic risk and if one then can argue why the designation would mitigate systemic risk.

Based on current official documents, it is hard to see what could be the motivation for the SIFI designation beyond mere size, and even then, size as a contributor to systemic risk for an asset manager is not nearly as clear as it would be for a bank.

There may well be externalities implying that asset managers, and the funds managed by those managers, contribute to systemic risk. If no such externalities can be found, then intervention is ill advised and welfare-reducing. If such externalities can be identified, then any SIFI-related policy measures adopted must be designed to improve on the current hands-off policy by balancing benefits with costs.

The cost of acting when one should not, or not acting when one should (type I and type II errors) is very high and it is our view that strong policy tools and impacts should be much better understood before being put in use.  We propose a clear conceptual test for this, one that is based on the fundamental nature of asset management.

Existing research on systemic risk created by asset managers

In the banking world, an extensive body of academic and policy literature exists on how banks may contribute to systemic risk, enabling sophisticated discussions on why banks should be regulated, and then how.

In the asset management world, such formal analysis is almost completely absent and most resulting SIFI discussion appears to be based on anecdotal examples (often only Long-Term Capital Management) rather than hard analysis. With very few exceptions, discussions are mostly silent on why asset management is systemic in the first place, as if this was self-evident.

We maintain that there is no reason to believe in general that the nature of systemic risk created by banks is the same as systemic risk created by asset managers. Of course there are similarities, but such important banking contributions to systemic risk as direct interconnected balance sheet exposures, maturity mismatches, runs, leverage (up to 12 times Basel regulated banks) are largely absent in the mutual fund, exchange-traded fund, or pension fund part of the asset managing world (e.g. UCITS).¹ The largest funds are passive or tightly benchmarked managers using no leverage.

The current policy literature identifies a few main risks borrowed from banking, such as liquidity mismatches leading to runs, leverage, a fund in distress, and operational risks. We do not find any of these convincing without further elaboration and amplification. Operational risk is almost by definition non-systemic, and in one example we have – Bill Gross’ resignation from PIMCO – there were no serious consequences. In the absence of over-leverage, the effective closure of secondary markets or mis-selling, the connectivity of funds, even large ones, would lead to distress that would be qualitatively and quantitatively more subdued than a similar distress in the banking world.

What externalities created by asset managers may merit regulation?

To our mind there are several ways an asset manager could still contribute to systemic risk:

1. The abrupt withdrawal of an important manager in a particular market can create a dry up in liquidity that spills over into other markets and impose external costs on other investors through interlinkages. If the scale and connections are important enough, the spillovers can be systemic, especially if the markets are already under stress and liquidity is scarce.
2. Fire sale scenarios whereby various asset managers mechanically sell into a crashing market, directed by large withdrawals, by the need to keep mandates, and by pecuniary externalities, e.g. due to accounting, risk, or margin considerations. This then creates endogenous risk as discussed here on Vox (see Danielsson et al. 2009), and leads to increased correlations if funds liquidate across the spectrum, affecting asset classes that have otherwise no fundamental similarities. The reduced diversification in times where it matters most may leave investors disaffected.
3. Asset managers who rely on complicated derivatives positions or on high levels of leverage to achieve returns represent counterparty and model risk that only manifest in the worst state of the world.
4. Funds that advertise daily or short-term redemptions – ‘liquidity’ – but invest in illiquid or long-term assets may allow the concentrated build-up of large positions that would not have happened had it not been for the advertised free liquidity and maturity transformation services. Perhaps the best example of this is money market mutual funds that act as direct competitors to banks, though the redemption-at-par feature is not required and many other liquid alternatives or corporate bond funds perform liquidity and maturity transformations without guaranteeing par.

However, these are only conjectures and have not been examined either theoretically or empirically. Furthermore, these could just as well apply to the actions of end investors as well as their managers.

The crucial test for SIFI status

To us, the crucial test for SIFI status relies on the following replicability counterfactual:

‘Suppose a fund acts in a way that creates systemic risk. If the same investments had instead been made directly by end investors, would they have behaved differently, in a way that did not create systemic risk?’

If the answer is ‘no,’ perhaps:

• Retail investors are led to believe by fund managers that liquidity provisions or maturity transformations are free; or
• A fund permits leverage that retail investors could not establish on their own; or
• A fund manager has incentives (e.g. churn, leverage, take certain risks, or engage in sec lending chains) different from investors; or
• Retail investors may not have been able to take similar positions, e.g. it would not be possible to individually replicate defined benefit pensions or unit-linked insurance because such schemes rely on large numbers of investors.

In these situations, there may be a case for treating the asset manager as systemically significant, whether or not some of those SIFI costs are passed on.

Funds may end up creating an endogenous risk by adding net correlation to the market, thereby increasing procyclicality and endogenous risk. One example is a fund investor who is worried about the riskiness of some of the assets of the fund and as a consequence reduces her exposure, in turn causing the fund to liquidate a portion of all its investments across the board. If the investor had instead held the constituent assets, she would only have sold some assets. In this case the fund dis-investment would add more correlations to the market than the individual asset disposal.

If the answer however is ‘yes,’ then the ‘polluter pays principle’ applies to the final investors directly and not to the fund or the fund manager.

The reason is that if the measures are not borne directly by the final investors, then the fund will pass some of the costs to the investors. Depending on replicability, the latter may find it advantageous to invest directly into similar strategies, bypassing the regulations intended to make the system safer, so imposing frictional costs but achieving no benefit. The outcome may be a net addition to pro-cyclicality of the financial system.

Conclusion

Some policy authorities have taken firm steps towards designating asset managers as SIFIs. Some aspects of this are worrying.

Well-crafted financial regulations can be effective in mitigating the impact of externalities created by financial institutions, but can just as easily make the problem worse. This is especially likely to happen when it comes to macroprudential policy given its relative immaturity.

The authorities do not provide much in terms of identification of the externalities created by asset managers that might justify a SIFI designation, and seem to borrow much of the analysis inappropriately from banking regulations. In our view, there are externalities that might give rise to systemic concerns, and we propose a clear test for such identification.

In addition, the policy authorities are essentially silent on the question of what the SIFI designation should mean in practice, even though in any discussion on regulation, identification and remedy is intimately linked. Given the heavy reliance of the existing analysis on the banking literature, we are concerned that any eventual regulations might also draw on banking. That would be unfortunate because resilience comes from heterogeneity, not monoculture.

By ignoring the decisions of end investors that choose to allocate to asset managers in the first place, measures meant to encourage systemic stability at a fund or asset management level may lead to larger and less well-monitored build-ups of systemic risk through direct investing.

Instead of rushing ahead in designating asset managers as SIFIs, it would be better if the next steps were to increase our understanding of how asset managers contribute to systemic risk, identify the data necessary for making an assessment and develop the necessary theoretical frameworks. It is therefore fortunate that the authorities have decided not to proceed now with the SIFI designation of asset managers.

References

Blackrock (2015) “Comments on the Consultative document (2nd) assessment methodologies for identifying non-bank non-insurer global systemically important financial institutions”, Blackrock.

Danielsson, J, H S Shin and J P Zigrand (2009) “Modelling financial turmoil through endogenous risk”, VoxEU.org, 11 March.

FSB-IOSCO (2015) “Consultative document (2nd) assessment methodologies for identifying non-bank non-insurer global systemically important financial institutions”, Financial Stability Board.

The Economist (2014) “Assets or liabilities? Regulators worry that the asset-management industry may spawn the next financial crisis ”, 2 August.

Endnotes

[1] The Undertakings for Collective Investment in Transferable Securities EU directive.

This article is published in collaboration with VoxEU. Publication does not imply endorsement of views by the World Economic Forum.

To keep up with the Agenda subscribe to our weekly newsletter.

Authors: Jon Danielsson is a director of the ESRC funded Systemic Risk Centre at the London School of Economics. Dr Jean-Pierre Zigrand is Associate Professor of Finance and Co-Director of the Systemic Risk Centre at the London School of Economics.  

Image: U.S. one dollar bills blow near the Andalusian capital of Seville. REUTERS/Marcelo Del Pozo.