The surprising fragility of the infrastructure

Among the most persistent of fallacies is the belief that conscious design is superior to natural order. The socialist case against market economics long rested on the twin convictions that competition is more fiction than fact and capitalism both poorly coordinated and intrinsically wasteful.

 Among the most persistent of fallacies is the belief that conscious design is superior to natural order. The socialist case against market economics long rested on the twin convictions that competition is more fiction than fact and capitalism both poorly coordinated and intrinsically wasteful. When output is dominated by private corporations whose character is identical to that of the public bureaucracies, and the products and services of different companies seem indistinguishable, it is easy to conclude that planning and purpose and scale and organization are more influential in material success than competition for capital, labor and customers. An obvious symptom of the current turn of the intellectual tide away from market economics is the revival of this brand of thinking among the policymakers and regulators charged with rebuilding the financial systems of the world. Their faith in collective rather than individual endeavour is nowhere more explicit than in their enthusiasm to attenuate or even replace the commercial provision of securities services by so-called financial market infrastructures.

Laying responsibility for the safety and stability of the global financial system on clearing and settlement mechanisms (CSMs), central securities depositories (CSDs), central counterparty clearing houses (CCPs) and trade repositories is one of those decisions, like commemorating the victims of war, in which effect determines cause. The cataclysm that preceded the decision puts the choice beyond dispute. A banking industry that had to be rescued with trillions of dollars of public money suffers by comparison with financial market infrastructures that continued throughout the crisis of 2007-08 to make payments, deliver securities against cash and unwind the positions of failed investment banks. The invalidity of comparing banks (which exist to take risk) with infrastructures (which exist to mitigate risk) cannot disturb an accelerating official preference for collectively owned and managed utilities that stretches back to at least the formation in 1973 of the Depository Trust Company (DTC) to deal with the repeated paper-based settlement crises of the 1960s and early 1970s.

Since the publication of the influential Group of Thirty (G30) report of March 1989, the number of CSDs in the world has increased from 28 to nearly 150. CCPs, which in 1989 were confined to the commodities markets and the financial futures and options markets that grew out of them, now span the equity, fixed income, securities financing and swap markets too. Their number has grown to 83. One reason why there are so many CSDs and CCPs is that securities settlement and clearing utilities have followed the same pattern as payments utilities and remained largely domestic. At the last count, the global financial services industry was moving cash across 217 high-value payments systems, and 173 low equivalents. Two dozen trade repositories are now in various states of development. It seems that, just as every serious country must have an airline and a football team, so it must have a payments system, a CSD, a CCP and a trade repository. The clearance, settlement and recording of cash payments and securities and derivatives transactions is now increasingly dependent on a global collection of 650-odd infrastructural utilities.

Predictably, the conventional wisdom holds that this is far too many, and reflects nothing more sophisticated than the protectionist attitudes of local banks, broker-dealers, fund managers and investors and the local regulators whose behavior they dictate. Consolidation of market infrastructure is not merely predicted; it is prescribed. This also substitutes conviction for thought. In its belief that scale creates the efficiency that begets yet larger scale, rendering competition impossible to sustain, financial market infrastructure is the last bastion of 20th century socialist economics. The argument soon becomes self-fulfilling. Consolidation creates monopolies and quasi-monopolies that then attempt to profit from them by reducing output and raising prices.

It is not long before it is argued that allowing private interests to profit from monopoly is antithetical to the public interest, and that there is therefore no longer a case for leaving such important entities in private ownership. The Depository Trust and Clearing Corporation (DTCC), itself the product of a merger between a CSD and a CCP, argues openly that the commercial imperative to make a profit is ill-suited to market infrastructures charged with reducing risk. It believes they should be mutually owned and governed only. Principle 2 of the CPSS-IOSCO principles for financial market infrastructures agrees that market infrastructures should take into account the “legitimate interests” of the public as well as the banks and broker-dealers that make use of their services.

The irony, as the balance of the 24 CPSS-IOSCO principles makes plain, is that a global financial system reliant on a consolidating set of financial market infrastructures—however they are owned and governed—is not less dangerous to stability, but more so. The Basel III capital adequacy regime is proceeding on the basis that the financial crisis proved banks owned insufficient equity capital, and that at some point between now and 2019, systemically important banks must report genuine equity capital equivalent to 12% of risk-weighted assets. Despite the emphasis of the CPSS-IOSCO principles on disclosure of information to users and the public, so that they can make informed assessments of the risk they incur when using financial market infrastructures, meaningful information about the levels of capital CSDs and CCPs hold as a proportion of their real or notional liabilities is worryingly hard to find. In February this year, the Payments Risk Committee of the Federal Reserve published a detailed set of questions for CCPs covering initial margin, default funds, eligible collateral, investment balances, counterparty risk management, default procedures and other matters with the express goal of helping banks “conduct necessary due diligence and manage the risks they face as participants in such financial market infrastructures.” Bizarrely, amid dozens of detailed questions on collateral and collateral procedures, the question of equity capital was not brought up at all.

Nor is information about the equity capital of financial market infrastructures always that easy to find and assess in the financial reports of the major CCPs, particularly in the United States, where a string of mergers has bloated balance sheets with billions of dollars of goodwill. What is clear is that market infrastructures rarely have capital-to-assets ratios as generous as those of a bank, and frequently much less. LCH.Clearnet, the largest CCP in Europe, held at the end of last year equity capital equivalent to just 0.09% of its notional liabilities. The equity capital on the balance sheet of the CME Group, the largest CCP in North America, totalled $21.4 billion at the end of last year. It is an organization whose trading volumes had a notional value in 2012 of $806 trillion. At Euroclear, shareholders’ funds of €3.2 billion are equivalent to 15.5% of total assets, which is a little better than most banks, but still a trifle by comparison with average overnight client cash deposits of €15.8 billion, daily client collateral outstandings of €700.6 billion, client securities worth €23.0 trillion and settled transactions worth €541.6 trillion. At the end of last year, Clearstream had €11.1 trillion in custody and daily client collateral outstandings of €570.3 billion. Unlike Euroclear, but like CME, Clearstream is part of a publicly listed company. Four out of five euros on the balance sheet of the parent company, Deutsche Börse Group, are the gross value of the cash and securities posted as margin by the clearing members to Eurex Clearing. Deutsche Börse is supporting assets of €216.5 billion on equity of just €3.2 billion.

The traditional view is that market infrastructures do not need equity capital because they transfer cash or securities only if clients actually make them available: they take no principal risk. Indeed, the €178 billion of cash and securities owed to clients of Eurex Clearing on the liability side of the Deutsche Börse balance sheet is offset by exactly the same amount on the asset side of the balance sheet. Likewise, the customer cash deposits with Euroclear are almost perfectly matched with the cash holdings and short-term loans on the other side of the balance sheet. In effect, both CSDs and CCPs run matched books. On this view, CCP numbers only look alarmingly large because they collect initial and variation margin and default fund contributions from clearing members, and accounting rules do not permit the liabilities assumed to be netted down against the assets received.

This would be more reassuring if CSDs and CCPs were not operating in such an unstable competitive environment. CSDs are now competing with each other and with custodian banks for asset servicing business and access to commercial bank money. Custodian banks are competing with CSDs for issuance and collateral management business and access to central bank money. CCPs are competing with each other, especially across borders. They are also for the first time being forced by regulators to support transactions by client types (fund managers) in an asset class (swaps) of which they have limited experience.

Market infrastructure is in a bastardized condition, suspended uneasily between being forced take responsibility for financial stability and being forced to compete on price and margin and collateral segregation terms, which is scarcely compatible with stability. Inevitably, the primary loyalty of the management of a CCP in particular is not to such nebulous concepts as financial stability or the public interest, but to the clearing members that bring them business, and which wish to transact as much business as they can in exchange for as little in the way of capital and collateral commitments as possible. Though they want buy-side business, CCPs are ultimately answerable to the sell-side. Even in the extraordinary circumstances of today, it is commonplace for clearing brokers to press for economies in collateral payments, on grounds assets are uncorrelated, shorts are offset by longs and futures by swaps. To win their business, CCPs will have to compete for business on margin netting and collateral eligibility criteria and compromise on buy-side asset segregation. Clearing members are already lobbying domestic regulators, particularly in the United States, for the looser rules that will offer them a competitive advantage. Unsurprisingly, clearing brokers prefer covering exposures with initial margin (which is also contributed by buy-side clients) to default fund contributions (which are contributed by clearing brokers only, and carry a capital cost).

At bottom, margin calls are driven by the limited equity capital of the CCPs. Regulators have tacitly acknowledged this by specifying how much equity capital they believe financial market infrastructures should hold. Principle 15 of the CPSS-IOSCO principles recommends that they maintain equity sufficient to fund six months of operating expenses. The capital requirements set by the European Banking Authority (EBA) insist that a CCP hold equity capital sufficient to cover 12 months’ operational expenses plus the estimated value of all operational, legal, credit, counterparty and market risks not already covered by specific resources such as client collateral. The EBA also insisted CCPs bear some responsibility for losses before calling on clearing members and required them to set aside capital equivalent to 25% of the guarantee fund before losses were mutualized among members. The initial reaction of the CCPs to raising more capital and putting it at risk was to argue that it would only encourage their clearing members to take greater risks. The argument was effective enough to persuade the EBA to cut the proportion of capital at risk from half the value of the guarantee fund to a quarter. Nevertheless, LCH.Clearnet had to raise another €320 million in capital this spring. Eurex Clearing, which received a capital injection of €110 million in January, says it must raise another €150 million before it complies.

These are apt measures of how thinly capitalised the CCPs actually are, even by comparison with banks. The explanation for that is worryingly familiar to that once advanced by custodian banks for cash collateral reinvestment pools: the risks are fully collateralized. Regulators are obviously paying attention. The CPSS-IOSCO principles include specific recommendations on the management of credit, liquidity and collateral risks.

Principle 4 insists that infrastructures cover credit risk—even CSDs are prone to incur overnight credit risk—by a mix of collateralization and money set aside for the purpose. International CCPs are expected to put enough money aside to cover the default of the two largest participants in terms of aggregate credit exposures, and even smaller CCPs enough to cover the default of their largest participant. Principles 5 and 6 advise conservatism on collateral calculation methodologies, eligibility, haircuts, concentrations, offsets and cross-margining, rehypothecation and reinvestment. Principle 7 suggests financial market infrastructures maintain sufficient cash, liquid collateral and credit lines to settle their obligations on time in every currency they handle. In the case of payments systems, of course, this can mean in real-time, gross. CPSS-IOSCO reckons the minimum amount of liquidity necessary to cover this risk is a sum equivalent to the largest single obligation to any one counterparty. For most CSDs and CCPs, that implies a reduction in the maturity of short-term assets and an increase in contractually agreed credit lines and commercial paper facilities.

All of which makes it sound as if regulators have understood the consequences of their passion for infrastructure. In extremis, however, neat specifications of the sequence at which resources are put at risk and clever distinctions between capital, collateral, cash, near-cash and credit lines will be meaningless. In a serious event of default, credit lines will disappear. Collateral will have to be converted to cash and, if the default is large enough, the only distinction between equity capital and collateral will be the sequence in which they are put at the disposal of the non-defaulting counterparty. It is at this point that the true absurdity of the official faith in market infrastructures will become obvious. Utilities clearing and settling transactions worth hundreds of trillions of dollars a year have—even after netting—a gargantuan appetite for liquidity. When the Payments Risk Committee of the Federal Reserve measured the value of gross payments of cash in U.S. dollars alone, it came up with a figure of $15 trillion a day. CSMs, CSDs and CCPs generate massive daily demands for cash in multiple currencies to make payments, settle cash and derivatives transactions and margin positions. Current credit and commercial paper facilities between CSDs and CCPs and the banking system are nowhere near as large as this sum. It follows that in a crisis, cash will have to be raised, if it can be raised at all outside the central banks, against cash and securities collateral. This is why it is rightly said that, by transferring risk from banks to market infrastructures, official policy has not removed or even mitigated risk but merely repackaged it as liquidity risk in the markets for collateral—especially of the kind eligible at a central bank.

Indeed, the observation, in the seventh of the 24 principles of CPSS-IOSCO, that an infrastructure “should not assume the availability of emergency central bank credit as a part of its liquidity plan” is a more than usually pathetic triumph of hope over experience. After all, the only continuously available liquidity that has prevented the entire global financial system from imploding for six consecutive years is that available at the central banks. The sovereign debt of Greece and Portugal cannot be found on the eligibility schedules of the CCPs, and the long-term debt of Italy and Spain will cost a haircut of a tenth or more at LCH.Clearnet, but the European Central Bank will take any and all of it. In other words, there is a real market for sovereign risk and an unreal one. Some deluded observers look forward to the reversal of quantitative easing on grounds that the release of trillions of dollars of government bonds currently sitting on the balance sheets of the central banks will be released to alleviate the widely-anticipated shortage of eligible collateral occasioned by the regulatory enthusiasm for the collateralization of risks through financial market infrastructures. The panic that engulfs the market every time that possibility is even mentioned by a central banker offers a truer foretaste of what will actually happen.

And what will happen will be grimly familiar. A speculative bubble, created on this occasion by the leverage injected by a lack of discipline on counter-party selection and competition on margin terms fuelled by competing CCPs, will deflate. Some speculative clients of the clearing brokers will be ruined. The repo and reverse repo markets will close to the clearing brokers, which will then struggle to meet margin calls. Elaborate regulations designed to maintain a clear distinction between client and proprietary assets will prove to be meaningless as the clearing brokers ransack client accounts for eligible collateral. Because market infrastructures manage their risks with exactly the same techniques that failed the banks in 2007-08, the margin methodologies of the financial market infrastructures will prove inadequate to the scale of the defaults that occur. Financial infrastructures will fail. Even the final, irrevocable, un-rewindable and immediate payment offered by the CSMs will turn out to be not a guarantee of anything at all but merely a legal construct of value to the litigation departments of leading firms of lawyers only. Infrastructures will fail, just as banks failed in 2007-08. The crowning irony of the belief in market infrastructure as the route to a more stable financial system is that it will be the central banks that will rescue failed financial market infrastructures, just as it was the central banks that rescued failed banks five years ago. Perhaps at that point, the regulators will finally recognize that the instability of the financial system lies not in the structure of the system but in its nature.