4
Regulation

4.1 REGULATION AND THE GLOBAL FINANCIAL CRISIS

Over the years, derivatives – especially over-the-counter (OTC) ones – have given institutional investors, corporates, sovereigns, and supranational organisations a flexible tool for hedging a large range of risks, and the market has grown to be very large. However, critics have accused the OTC derivatives market of triggering and amplifying the global financial crisis (GFC) and, accordingly, regulators have proposed a number of initiatives aimed at making it safer. The GFC brought about many changes to regulation as a result of experiences during the crisis, such as thinly-capitalised banks. Some of the regulatory policy has brought about alterations in existing regulatory rules, but the majority has implemented completely new requirements. Many of these changes have important impacts for the assessment of valuation adjustments as they relate to counterparty risk, funding, collateral, and capital.

Firstly, the changes in regulation can be seen as covering the strengthening of capital bases and the implementation of liquidity standards which are broadly covered by the Basel III regulation originally published in 2010 (Figure 4.1). These requirements are detailed and affect banks across many activities, not just derivatives and their associated valuation adjustments. The discussion below will focus mainly on aspects that are relevant to the scope of this book. Secondly, and specific to OTC derivatives and this topic, is the requirement to clear standardised OTC derivatives, which is backed up by the bilateral margin requirements for non-centrally-cleared OTC derivatives. These requirements are potentially applicable to all users of OTC derivatives, not merely banks. However, small users will generally be exempt.

Schematic illustration of an overview of regulatory changes since the GFC that are relevant for the consideration of xVA. — **Figure 4.1** Overview of regulatory changes since the GFC that are relevant for the consideration of xVA.

Regulatory initiatives, such as Basel III, are often global guidelines that need to be implemented into law by each region or country. Major implementations of new financial regulation include the Dodd–Frank Act in the US and the European Market Infrastructures Regulation (EMIR) in the European Union (EU). The two are similar – but not identical – in the way they treat OTC derivatives, particularly in terms of aspects such as reducing systemic risk and improving transparency.

4.2 CAPITAL REQUIREMENTS

4.2.1 Overview

The Basel II capital rules generally in force at the time of the GFC were seen as dramatically undercapitalising banks and promoting dangerous behaviour such as regulatory arbitrage prior to the GFC.¹ Defining the increased capital requirements is the job of the Basel Committee on Banking Supervision (BCBS),² which did this as part of Basel III (BCBS 2009b, 2010a). Basel III proposed rules to dramatically increase capital by a combination of re-parameterisation of existing methods and the addition of new requirements. Counterparty risk was a key focus for Basel III, with the view that the failure to adequately capture OTC derivatives exposure was a key factor in amplifying the GFC. In particular, the BCBS noted that only around one-third of the counterparty risk was capitalised prior to the crisis. The remaining two-thirds would be captured by a new credit value adjustment (CVA) capital charge to cover the mark-to-market (MTM) of the counterparty risk of a bank. This would capture the potential losses when credit spreads widened, even in the absence of any actual defaults. Notably, the only exemption (under Basel III at least) from the CVA capital charge for OTC derivatives trades was for those cleared with a central counterparty (CCP).³

The danger in calibrating risk models with relatively recent historical data is that benign and quiet periods tend to precede major crises. This means that risk measures and capital requirements are particularly low at the worst possible time. The higher leverage levels that such low-risk measures ultimately allow may increase the likelihood and severity of any crisis. This problem is typically known as procyclicality. This has led to the requirement to use stressed data to calibrate risk models. The use of the stressed period should reduce procyclicality problems by ensuring that capital does not become artificially low during quiet periods in financial markets.

Other small changes to Basel III rules have also increased counterparty risk capital requirements. For example, the margin period of risk (MPoR) used by advanced banks for computing capital requirements on collateralised/margined transactions was increased in certain situations, basically reducing the benefit from a bilateral margin agreement (see Section 7.2.3). The overall total effect of Basel III is to increase a bank's capital requirements for counterparty risk by a significant factor, which clearly incentivises a push towards central clearing. This push is potentially made more relevant by the nature of the CVA capital charge, which penalises higher-rated counterparties in particular,⁴ and by the changes to the rules for collateralised counterparties.

4.2.2 Capital Ratios

Capital acts as a buffer to absorb losses during turbulent periods and therefore contributes significantly to defining the creditworthiness of a bank. Banks strive for profits and will therefore naturally wish to hold the minimum amount of capital possible in order to maximise the amount of business they can do and risk they are able to take. There is clearly a balance to be found in defining the capital requirements for a bank: it must be high enough to contribute to a very low possibility of failure, and yet not so severe as to unfairly penalise the bank and have adverse consequences for its clients and the economy as a whole. Capital requirements govern the ratio of equity to debt, recorded on the liability side of a firm's balance sheet.

Regulatory capital is the minimum amount of capital a bank must hold as required by its regulator. This minimum capital requirement aims to ensure that banks do not take on excess leverage and become insolvent. The BCBS sets out a framework for regulatory capital standards (Pillar 1) which is generally adopted by all national regulators. This framework has evolved from Basel I in 1998 to Basel II in 2004 and finally Basel III, which has had a phased implementation since the GFC. If a bank does not meet the minimum capital requirements, then regulators may have the power to restrict dividends and remuneration policies.

Banks have also used the concept of economic capital, which is an internal measure of the capital they require. However, economic capital has become very much secondary to regulatory capital in evaluating the performance of a bank. This is mainly due to the impact of regulation increasing regulatory capital to levels where it almost certainly exceeds economic capital.

Intuitively, capital is often thought of as representing an unexpected loss (i.e. a loss that is more than expected) (Figure 4.2). The idea is that expected losses will be naturally considered and should be part of valuation (e.g. CVA is the expected loss with respect to counterparty risk). Unexpected losses, defined to some confidence level, should be absorbed by the capital of a bank. In some capital methodologies, unexpected losses are defined by a quantile of a model-generated distribution. These are generally referred to as internal model methods or approaches (IMMs or IMAs). However, some regulatory methodologies are more simple and formulaic and do not use internal models.

Regulatory capital requirements are normally expressed as a capital ratio, which is the percentage of a bank's capital to its risk-weighted assets (RWAs). The RWAs are defined by various methodologies across different areas and risk types. Since Basel II, it has been a requirement that the total capital ratio of a bank must be greater than 8%. There are, broadly speaking, two types of capital: Tier 1 – sometimes referred to as Common Equity Tier 1 (CET1) – and Tier 2 (there was previously a Tier 3 that has been abolished). CET1 is generally ‘going concern’ capital and consists of common equity and some other components such as retained earnings. Tier 2 capital is generally ‘gone concern’ capital and consists of components such as preference shares. Going concern capital reduces the chance of insolvency, whilst gone concern capital protects depositors and senior creditors in the event of insolvency.

Graph depicts an expected and unexpected losses. — **Figure 4.2** Expected and unexpected losses.

Regulation has increased the amount of CET1 capital that a bank must hold to 4.5% in relation to the 8% minimum capital ratio. Additionally, certain additional CET1 requirements are being introduced that will require a bank to improve its capital ratios. These are:

Capital conservation buffer. This is intended to ensure that banks build up capital buffers outside periods of stress which can be drawn down as losses are incurred. The buffer is set at 2.5% above the regulatory minimum capital requirement. Banks can still operate as normal when below this requirement but are subject to constraints over distributions such as dividend payments.
Countercyclical capital buffer. This aims to ensure that banking sector capital requirements take account of the economic climate, and the primary objective of the buffer is to restrict the activities of banks during periods of excess credit growth. It is intended to be set between 0% and 2.5%, and the value will change dynamically depending on the macro-economic environment, with quantities such as the deviation in the ratio of lending to gross domestic product (GDP) from the long-term trend being considered as metrics for setting the ratio.
G-SIB surcharge. This is a surcharge for globally-systemically-important banks (G-SIBs). It is implemented based on a BCBS methodology that defines a bank as potentially being within one of five G-SIB buckets, with surcharge increasing from 1% to 2.5%.

The above requirements are summarised in Figure 4.3, which shows the 2019 values in the case where there is a phase-in. Note that the requirements can sometimes be region specific (i.e. regulators in one country may impose tighter standards than those in another).

Schematic illustration of an overview of capital ratios required since 2019. — **Figure 4.3** Overview of capital ratios required since 2019.

There are also two final components which should be considered:

Pillar 2 requirements. Whilst Pillar 1 of the Basel Accord deals with minimum capital ratios, Pillar 2 forms a supervisory review process, with one of its aims being to address firm-specific risks that are not adequately covered by Pillar 1. National regulators can, therefore, require a bank to hold additional capital under Pillar 2.
Buffers. Banks will inevitably hold more capital than the regulatory minimum to prevent having to raise more capital under tight timescales. This buffer should be considered an additional cost.

4.2.3 Risk Type

From the point of view of any valuation adjustment, it will only be relevant to consider capital requirements that apply directly to a transaction or business. Any capital requirements that cannot be allocated in this way are just general costs, like electricity or office equipment, and can only be considered implicitly.

Broadly speaking, banks must hold regulatory capital against the following risks:

market;
credit; and
operational.

Accordingly, there are separate requirements within each of these areas to determine RWAs. The RWAs are a measure of the total assets, weighted for their riskiness. Whilst market and credit risk methodologies are transaction based, those for operational risk are based on much more general quantities such as net revenue. It is therefore not possible to incorporate operational risk capital costs directly into transaction-specific valuation adjustments. In contrast, it is important to understand the different capital charges for credit and market risk and how these relate back to the cost of individual transactions.

In terms of different requirements, the following are important considerations under credit and market risk capital:

Market risk.
- Market risk capital charge. This is the capital charge for market risk across all trading desks in a bank (except the desk responsible for CVA). It has been subject to a significant rethink in recent years under the Fundamental Review of the Trading Book (FRTB). The FRTB requirements are described in BCBS (2019b).
- The CVA capital charge. This is a market risk capital charge specific to CVA trading, which is partially aligned with the FRTB rules above.
Credit risk.
- Counterparty credit risk (CCR) capital charge. This is the specific credit risk capital charge that is relevant for OTC derivatives and other transactions subject to CCR.
- Leverage ratio. This is a general metric that aims to limit the leverage of a bank and contains the exposure (in the sense of credit risk) as a key component.

Note that capital charges are treated as mutually exclusive and additive, whereas risks may offset one another. For example, from a counterparty risk perspective, it is not easy to completely disentangle the credit risk (CCR capital) and market risk (CVA capital). Furthermore, the methodologies for market risk and CVA are separate and additive, whereas a bank may see an offset between the market risk of CVA and other market risks.

There are also two other components that are more indirect but may influence bank capital requirements. These are:

Bank stress tests. Banks are subject to a variety of stress tests which assess their capability to withstand unfavourable economic conditions. Banks are typically required to maintain a certain CET1 capital ratio from these stress tests.
Capital floors. Regulation imposes floors whereby banks may be forced to use standardised methodologies rather than internal models.

Most capital charges are continuous in nature – i.e. the larger the position, the bigger the capital required. Note that some of the components above are not capital charges per se but rather conditions that must be met. This applies to the leverage ratio, bank stress tests, and capital floors. This makes pricing these components more difficult, as will be discussed in more detail in Section 5.4.3.

The above components will be reviewed below. Chapter 13 will give a more detailed account of the regulatory methodologies used to define capital requirements.

4.2.4 Market Risk Capital

Banks take a significant amount of market risk in the form of interest rate, inflation, foreign exchange (FX), commodity, equity, and credit spread risks when trading derivatives and other transactions with clients. Whilst most trading desks aim to hedge most of their market risk, this is not always possible and certain residual risks remain, often in the form of more subtle components such as basis risk.

Regulation has therefore required that banks hold capital against market risk. One of the benefits of the introduction of value-at-risk (VAR) models in the mid-1990s (Section 2.6.1) was the ability to quantify residual market risk across a portfolio that may be reasonably well balanced. VAR, or a similar metric, also allows the aggregation of different forms of market risk (e.g. interest rate and FX) in a consistent framework.

More recently, the FRTB (BCBS 2013e) is trying to define a more consistent set of market risk regulatory capital rules, balancing risk sensitivity and complexity and promoting consistency. One aspect of the FRTB is the use of expected shortfall (ES) as a replacement for VAR (Section 2.6.1) as the risk measure for market risk capital based on internal models. This is due to the view that ES is better at capturing tail risk: VAR measures only a single quantile and does not consider the loss beyond this level, whereas ES averages all losses above the confidence level. Since the ES is naturally a more conservative measure, it is proposed to replace 99% VAR with 97.5% ES (for a normal distribution the two are almost identical). The standard 10-day time horizon for VAR is also split into several categories based on the liquidity of the underlying portfolio. These changes may create larger capital requirements, especially for seemingly-well-hedged positions.

4.2.5 CVA Capital

In BCBS (2009) it is stated that, in the preceding crisis period, only one-third of counterparty risk-related losses were due to defaults, with the remaining two-thirds being driven by CVA volatility. At the time, only default losses (credit risk)⁵ and not MTM (market risk) were considered. Basel, therefore, introduced the CVA capital charge to capitalise the market risk of CVA.

The CVA capital charge, in general, has been quite controversial and has attracted criticism. Some of the criticism concerns the significant increase in capital requirements, whilst some is more subtle and relates to the precise methodology. For example, the European Banking Authority (EBA 2015b) reports that the CVA capital charge as defining the potential increase in CVA is ‘very conservative’ as it implies that the CVA of most banks surveyed could increase between 10 and 20 times over 10 days.

Basel III defines global standards on capital requirements that are generally being implemented consistently, albeit with small differences and timescales depending on the region. However, one extremely significant divergence has occurred in Europe via the Capital Requirements Directive IV (CRD IV), which is the mechanism through which the EU implements Basel III standards on capital. In the EU, the implementation of Basel standards for CVA capital (via CRD IV) exempted banks from holding CVA capital towards sovereign counterparties.⁶ This was followed in 2013 by similar exemptions covering non-financial counterparties (below the clearing threshold) and pension funds (temporarily, in line with a clearing exemption). These exemptions were themselves controversial since they meant that an EU bank would have to hold significantly less capital trading with certain counterparties than a non-EU bank would.⁷ A study by the EBA reported that without the exemptions the total CVA capital of 18 banks surveyed would more than double from €12.4bn to €31.1bn, with sovereign exemptions being the most material (EBA 2015b).

There has been clear resistance by some European countries to the CVA capital exemptions. In theory, national regulators can impose additional capital charges via Pillar 2 charges (see Section 4.2.2).⁸ In 2015, the EBA opined that ‘EU exemptions on the application of CVA charges should be reconsidered or removed, since they leave potential risks uncaptured’.⁹ However, it was suggested that this be done as part of the broader implementation of the FRTB (see Section 4.2.4).

The aforementioned FRTB process resulted in a consultative document in 2015 (BCBS 2015a) proposing changes to the CVA capital charge. Whilst these changes do make the CVA treatment generally more in line with market risk under the FRTB, the CVA capital regulation is somewhat simpler to reflect the complex nature of CVA. These changes are scheduled for implementation in 2022 (BCBS 2017).

The regulatory uncertainty over the implementation of the new CVA capital charge and the lack of clarity in the removal of the European exemption is a good example of the difficulty of pricing capital into transactions (discussed later in Section 19.2).

4.2.6 CCR Capital

CCR capital is a specific case of credit risk capital in general. Credit risk capital is defined as a product of three components, all defined by regulatory rules:

Probability of default (PD). The default probability of the counterparty in question.
Loss given default (LGD). The loss given default (or equivalently 100% minus the recovery rate).
Exposure at default (EAD). The exposure to the counterparty.

The above components are defined by various methodologies discussed in Chapter 13. For basic credit products, such as loans, the definition of EAD is trivial since it just represents the amount lent to an obligor. For products subject to counterparty risk, defining EAD is much more complex because the value of a portfolio can be both positive and negative and extremely volatile. Hence, there are specific methodologies for EAD that are relevant only for the assessment of CCR capital for derivatives and other transactions.

4.2.7 Leverage Ratio

An often-cited cause for the severity of the GFC was excessive on- and off-balance-sheet leverage of banks even when seemingly-strong risk-based capital ratios were in place. This has led to the introduction of the leverage ratio (LR) – sometimes referred to as the supplementary leverage ratio (SLR) – described in BCBS (2014a). The introduction of a simple, transparent, non-risk-based LR is designed to act as a credible supplementary measure to the risk-based capital requirements.

Although increasing capital requirements should reduce leverage, capital methodologies can be risk- and asset-type specific and are sometimes based on internal models which have been observed to vary significantly from bank to bank. Certain products and trading strategies can effectively create extremely high leverage – collateralised OTC derivatives and repos being two examples. The LR is seen as complementing regulatory capital requirements as a fairly simple approach to limiting the overall leverage within a bank. The LR can, therefore, be seen as a backstop to capital requirements, which mitigates the inherent uncertainties in these approaches. From the point of view of capital requirements, a bank must have enough capital to meet traditional capital ratios and comply with the leverage ratio.

The LR requires that the ratio of a bank's Tier 1 capital to their exposure must be greater than a certain amount, thereby restricting leverage (which is effectively the reciprocal of this amount):

(4.1)

The exposure in Equation 4.1 aims to capture the total exposure, on and off balance sheet, of the bank and includes products such as derivative exposures, securities finance transactions (e.g. repos), and other off-balance-sheet exposures. The methodology for defining the exposure for derivatives typically follows methodologies for defining minimum regulatory capital with certain restrictions (see further discussion in Section 13.4.6). It is relatively punitive for transactions such as OTC derivatives and repos and is relatively conservative in its treatment of collateral.

Note that the LR measure, by design, does not account in any way for credit quality because it only looks at exposure and not default probability (or LGD). Hence, higher-credit-quality entities will have the same impact as the same exposure to weaker credits. Since capital ratios are credit-quality sensitive, this means that transactions and businesses with good underlying credit risk are more likely to be LR-consuming (this means that the implicit capital requirement under the LR formula is higher than the usual capital requirement). LR-consuming business may eventually cause the bank to be LR constrained (close to breaching the condition, unless exposure is reduced or capital increased). The treatment of collateral also means that bilateral-collateralised and centrally-cleared transactions are more costly from an LR point of view.

A bank that is (or is close to) being LR constrained may consider this to be more relevant than its actual capital charges. However, managing business with respect to the LR is difficult since it represents a binary condition on capital rather than a continuous requirement. Furthermore, a bank must only meet the leverage requirement overall, not for a particular activity. For example, if a business grows, then their capital requirements will probably increase more or less in line with this growth and will not be related to capital requirements in the rest of the bank. This business may also catalyse its bank breaching the LR condition above, but this will happen at a discrete point and will depend on other business in the bank. It is, therefore, more difficult to envisage how the LR can be incorporated into business decisions and priced via a valuation adjustment measure such as capital value adjustment (KVA) (Section 19.2.6).

Note also that in the US, separate from Basel III, an ‘enhanced supplementary leverage ratio’ has been proposed for the largest banks, with a more conservative 5% minimum.

4.2.8 Capital Floors

Related to the LR is the use of capital floors based on the results of standardised approaches (BCBS 2014g). This requirement has been driven by the observation that internal models can produce large variations in capital charges across banks for the same exposure and very small capital charges in certain situations. Such results suggest significant model risk and perhaps overly-optimistic assumptions within a bank's internal models. The use of a floor based on standardised capital methodologies is intended to mitigate against excessively-low internal model numbers and limit variation between banks based on their internal models. Note that banks with internal model approval for a component (e.g. CCR capital) will also need to calculate capital under the standardised regulatory methodology.

A capital floor works by limiting a bank's aggregate calculation of RWAs (measured with methodologies including internal models) to be no lower than a certain percentage of the RWAs that would result if the bank had used only standardised methodologies. The value of the floor in this context is difficult: too high and banks will have no incentive to use internal models, but too low and it will be unlikely to have any impact.¹⁰ The current intended final value of the floor (following a phase-in where the floor is increased gradually) is 72.5%.

A capital floor creates a similar problem as the LR for managing businesses and pricing transactions, due to the fact that it is binary and an aggregate bank measure. For example, a business may have an internal model capital requirement that is less than the standardised methodology floor, but unless this is the case at the bank level, the business's capital requirements will still be driven by its internal model.

4.2.9 Large Exposure Framework

The large exposure framework (BCBS 2014e) has been developed as a way to limit the maximum loss a bank could face in the event of a sudden counterparty failure, to prevent such an event also endangering the solvency of the bank. This is seen as complementing risk-based capital standards, which are not primarily focused around the potential default of a single counterparty. Indeed, the regulatory standards for credit risk capital under internal models (Section 13.2.2) are based on the notional of an infinitely large portfolio.

The large exposure framework is focused on the potential default of a single counterparty or group of connected counterparties. Connected counterparties – for example, where one counterparty has ownership or control over another – must be treated as a single counterparty. Large sectorial or geographical concentrations, whilst recognised as being problematic, are not considered. The only counterparties exempted from the framework are sovereigns and their central banks and public sector entities that are treated as sovereigns under capital requirements. Similarly, exposures guaranteed by such sovereign counterparties are excluded. CCPs are currently excluded, pending further observations.

A large exposure is defined as an exposure to a counterparty, or group of connected counterparties, that is larger than 10% of the bank's eligible capital base. Eligible capital corresponds to CET1 capital (Section 4.2.2). Such exposures must not be more than 25% of a bank's eligible capital base (this figure is lowered to 15% for a G-SIB's exposure to another G-SIB).

For counterparty-risk-related instruments, the exposure is defined using the Standardised Approach for Counterparty Credit Risk (SA-CCR) (Section 13.4.3). Certain risk mitigation techniques (e.g. collateral) can be accounted for in reducing the exposure.

4.2.10 Bank Stress Tests

The GFC left many banks and financial institutions severely undercapitalised. The aim of stress tests is to prevent this by analysing a bank's capability to withstand unfavourable economic circumstances. Such circumstances are defined by a series of hypothetical scenarios, generally intended to be unlikely but plausible, which may include aspects such as a fall in GDP, equity and house prices, together with an increase in unemployment and interest rates and the weakening of a local currency. Examples of regulatory stress tests are the Comprehensive Capital Analysis and Review (CCAR) and the Dodd–Frank Act Stress Testing (DFAST) in the US, and the EBA and Bank of England stress tests in Europe.

One of the key outputs of stress tests is a bank's CET1 capital ratio, which must remain above a certain defined level. This is, therefore, another binary, aggregative bank measure that must be met and may – implicitly – generate a capital requirement.

4.2.11 Prudent Valuation

Specific to the EU, Prudent Valuation (EBA 2015a) requires that financial instruments recorded at fair value from an accounting perspective (Section 5.3.3) are assessed based on their prudent (conservative) value. The prudent value aims to capture the 90% confidence level worst value. The difference between the prudent and fair values – known as additional valuation adjustment (AVA) – must be deducted from CET1 capital and can, therefore, be seen as an additional capital charge.

Amongst the components that must be assessed with AVA are:

unearned credit spreads;
investing and funding costs; and
model risk.

The first component above clearly relates to CVA consideration via the credit spread inputs. Additionally, the second component above could be seen as requiring some consideration in relation to funding value adjustment (FVA) and potentially also margin value adjustment (MVA), although the reporting of these components within fair value is not yet completely standard. Finally, the last term in the list potentially relates to all valuation adjustments (to the extent that they are part of fair value) since they are all typically model driven.

4.3 LIQUIDITY

4.3.1 Overview

Banks tend to engage in activities such as funding long-dated illiquid assets (e.g. long-term loans) with short-term liquid liabilities (e.g. deposits). Such activities entail maturity transformations (borrowing money on shorter time frames than lending money out) and liquidity transformations (funding relatively illiquid assets with more liquid liabilities). Some of the high-profile failures in the GFC were directly related to these maturity and/or liquidity transformations (e.g. financing long-term lending with short-term borrowing).

One of the experiences of the GFC was that many banks, irrespective of the size and quality of their capital bases, had problems arising from not managing liquidity risk appropriately. It also became clear how quickly the liquidity of assets can decline, and how sharply the funding costs of banks can increase.

The inappropriate approach to liquidity risk by banks was broadly due to relying on short-term transactions to fund longer-term assets. This resulted in a need to refinance short-term borrowing regularly, which created liquidity risk due to the possibility that this would become more expensive or even not possible at all. Strong capital bases would not mitigate such problems, and regulators therefore saw the need to implement specific international liquidity standards for the first time, formulated in a similar way to global capital standards.

The BCBS has developed two complementary minimum standards for funding liquidity:

The liquidity coverage ratio (LCR). This aims to ensure that a bank has the resources to survive a hypothetical adverse liquidity scenario of 30 days (BCBS 2013a).
The net stable funding ratio (NSFR). This aims to create longer-term funding stability by ensuring that assets are financed with sources of funding that are relatively stable compared to the profiles (liquidity and maturity) of the assets. The NSFR has a time horizon of one year (BCBS 2014b).

Neither the LCR nor the NSFR aims to prevent a bank from offering various forms of maturity or liquidity transformation, but rather aims to ensure that such transformations are not extreme and that the bank has a structure of assets and liabilities that is sustainable in the long term.

Note that the LCR and NSFR have the same features as some of the capital-based regulations, such as the LR, capital floors, and stress tests, in that they are binary regulations that apply at the bank level. They also create the need to assess two separate funding strategies: the traditional one that the bank believes is appropriate, and the one that maintains compliance with the LCR and NSFR. For certain assets, the latter strategy will be different from the former. In particular, ‘self-funded’ products such as collateralised derivatives and repos may be thought of as requiring virtually no funding but will require implicit LCR and NSFR funding via a required stable funding component.

4.3.2 High-quality Liquid Assets

At the core of the LCR and NSFR regulation is the concept of unencumbered high-quality liquid assets (HQLAs). This generally refers to assets that can be readily converted into cash (with minimal credit and liquidity risk) to meet liquidity needs in a timely fashion.

HQLAs can consist of the following types of assets (see BCBS 2013a):

Level 1. These are assets of excellent liquidity and credit quality, including cash, central bank reserves, and certain securities guaranteed by sovereign, supranational, or similar entities (subject to certain conditions). These assets can comprise an unlimited share of the pool and are not subject to a haircut.
Level 2A. These are less-liquid assets, such as certain securities guaranteed by sovereign, supranational, or similar entities, corporate and covered bonds (subject to certain conditions), and commercial paper. These cannot comprise more than 40% of the total liquidity pool and have a 15% haircut applied.
Level 2B. The most illiquid assets, where allowed,¹¹ can only comprise 15% of the liquidity pool and are subject to a 25% or 50% haircut. This includes certain corporate bonds, residential mortgage-backed securities (RMBS), covered bonds, and equities.

4.3.3 Liquidity Coverage Ratio

The LCR aims to ensure that banks have sufficient HQLAs to survive a liquidity stress event, which is associated with aspects such as increased market volatility, reduction of current funding (e.g. partial run-off of deposits), inability to access secured and unsecured funding markets, and potential outflows arising from a downgrading of the bank's external credit rating. This requirement can be expressed as:

(4.2)

One particularly important aspect of the LCR that is particularly relevant for xVA is the liquidity needs related to downgrade triggers embedded in transactions. A bank must include in the total net liquidity outflows 100% of any additional collateral that would have to be posted, or other contractual outflows (e.g. termination), in the event of a three-notch downgrade of the bank's external credit rating.¹² This applies to derivatives transactions where there is contractual collateral posting, termination, or replacement in the event of the bank's rating deteriorating to a specified level, and can be related to both FVA (in the case of variation margin requirements being triggered) or MVA (independent amount or initial margin).

Under the LCR, banks must also consider increased liquidity needs related to changes in the valuation of derivatives or other transactions and the change in the valuation of the collateral in relation to such transactions. This can have an impact on FVA (Section 18.3.6) and MVA (Section 20.3.3).

Returning to the point about two funding strategies, suppose an A-rated bank must post a certain amount of collateral in the event that its credit rating is downgraded by three notches. The traditional funding strategy would probably consider that it would not be necessary to fund this payment prior to a downgrade (since the relatively significant downgrade to BBB is perceived to be extremely unlikely). On the other hand, the LCR implies that this contingent payment should be funded in its entirety via the requirement to hold 100% of HQLAs. However, suppose the bank has a relatively healthy LCR well in excess of 100%. Would it be relevant to impose a funding cost on originating this type of business even though it is already pre-funded via the bank's healthy LCR? We will discuss this question in more detail in Section 5.4.3.

4.3.4 Net Stable Funding Ratio

Whilst the LCR is relatively short term, the purpose of the NSFR is to ensure that banks hold a minimum amount of stable funding over one year to prevent excessive maturity and liquidity transformation. The NSFR is calculated by dividing a bank's available stable funding (ASF) by its required stable funding (RSF) (Equation 4.3), and uses defined weights to proxy the funding requirements of assets (receivables) and funding benefits of liabilities (payables). As with the LCR, this ratio must always be greater than 100%:

(4.3)

The NSFR requires a minimum number of stable sources of funding at a bank relative to the liquidity profiles of the assets, as well as the potential for contingent liquidity needs arising from off-balance-sheet commitments, over a one-year horizon. The NSFR aims to limit over-reliance on short-term wholesale funding during times of buoyant market liquidity and encourage better assessment of liquidity risk across all on- and off-balance-sheet items.

Most transactions for a bank are either assets or liabilities, and so the NSFR analysis is relatively straightforward. For example, if a bank funds a certain amount of residential mortgage lending (RSF of 65%) with the same amount in ‘stable deposits’ (ASF 95%) then its NSFR (assuming no other assets or liabilities) will be a healthy 95/65 = 146%. On the other hand, funding these same assets with funding of less than one year (typically ASF 50%) would lead to a non-compliant NSFR of 50/65 = 77%. Clearly the ASF and RSF factors and the timescale of one year are subjective, but beyond this the methodology is relatively simple.

Transactions such as derivatives are more problematic from an NSFR perspective because they can be both assets (when they have a positive value) and liabilities (when they have a negative value). They can also move from one to another as market variables such as interest rates move. With respect to collateralised transactions, the margin/collateral can either be an asset (received) or liability (posted). Overall, due to having a combination of collateralised and uncollateralised transactions, a bank may have an overall derivatives portfolio that is positive (‘asset heavy’) or negative (‘liability heavy’) even if the market risk is well hedged overall.

The NSFR, therefore, recognises that some transactions can be both assets and liabilities and that collateral may be paid in either direction. This means that gross derivatives assets are cancelled by the same amount of gross derivatives liabilities, leaving no net RSF requirement. Additionally, cash collateral received cancels with collateral posted, leaving no required funding.

However, the NSFR also takes into account (implicitly) that the balance between derivatives assets and liabilities is likely to be imperfect and also dependent on market conditions, and may therefore be transient (i.e. a bank may need to fund derivatives assets over a short timescale). There are therefore certain features specific to derivatives which receive an asymmetrical treatment:

Whilst net derivatives assets have a 100% RSF weight and require funding, net derivatives liabilities have a 0% ASF weight. This means that a bank with a derivatives book that provides funding overall via a net surplus of liabilities over assets does not provide stable funding (i.e. generate ASF), according to the NSFR. Put another way, the NSFR assumes that collateral generated from derivatives cannot fund other parts of a bank's balance sheet. This is an important factor in FVA quantification (Chapter 18).
Twenty percent of derivatives payables have a 100% RSF factor. This means that even funding benefits have an associated funding cost under the NSFR. The intention of this is to reflect the large variability in derivatives exposures and the fact that a funding benefit can quite quickly become a cost when negative values become positive. However, this component has been contentious, and the BCBS has communicated that regulators may, at their discretion, lower the value of this factor to no less than 5%.¹³
Collateral (variation margin) received can only reduce the funding cost of a derivative if it is cash and it complies with certain conditions.¹⁴ This means that collateralised trades, which may be considered to have very small funding requirements, may still have a large contribution to the NSFR if, for example, bonds are permissible as collateral.¹⁵

The above constraints on the way in which a bank funds its balance sheet would most obviously be a consideration for FVA quantification (Section 18.3.4).

Beyond FVA, there is the consideration of the impact the NSFR has on initial margin and MVA (Chapter 20). Received initial margin has a 0% ASF factor, which is not surprising since it is usually segregated and therefore has no funding benefit. Posted initial margin has a relatively high RSF factor of 85% (for example, as mentioned above, residential mortgages have an RSF of 65%). Therefore, if a bank is posting initial margin using securities with an ASF of less than 85%, then the NSFR introduces an additional funding cost.

Recall again that the NSFR is binary and imposed at the bank level. A given trade does not therefore, per se, have an NSFR component. Hence, it is not easy for a bank to have a robust policy on pricing such requirements and it is unclear whether a bank that is comfortably in compliance with the NSFR would feel the need to charge for RSF on new transactions.

4.4 CLEARING AND MARGINING

4.4.1 Central Clearing

From 2007 onwards the GFC triggered grave concerns regarding counterparty risk, catalysed by events such as Lehman Brothers, the failure of monoline insurers, and the bankruptcy of Icelandic banks. Counterparty risk in OTC derivatives, especially credit derivatives, was identified as a major risk to the financial system. There were also related operational and legal issues linked to aspects such as collateral management and close-out processes which result directly from counterparty risk mitigation. A CCP offers a potential solution to these problems as it guarantees performance, potentially reducing counterparty risk, and provides a centralised entity where aspects such as collateral management and default management are handled.

One of the largest perceived problems with bilateral OTC derivatives markets is the close-out process in the event of a major default, which can take many years and be subject to major legal proceedings (for example, see Figure 2.11). By contrast, CCPs can improve this process by establishing and enforcing the close-out rules, ensuring continuity and thereby reducing systemic risk. The default management of OTC derivatives by CCPs was viewed as being highly superior to bilateral markets in the aftermath of the Lehman bankruptcy. Although bilateral markets have made progress in certain aspects (see, for example, the adoption of the ISDA close-out protocol discussed in Section 6.3.5), they still cannot claim to be as coordinated as CCPs in this regard.

In contrast to OTC derivatives, the derivatives market that was cleared via CCPs was much more stable during the GFC. CCPs such as LCH.Clearnet coped well with the Lehman bankruptcy (Section 2.5.3) when many other elements of the OTC derivative market were operating poorly. One of the reasons for this is that, unlike most market participants, CCPs had actually envisaged and prepared for such a situation. Hence, whilst CCPs still experienced problems (such as identifying the positions of Lehman's clients), they were able, with help from their members, to transfer or close out a large volume of Lehman derivatives positions without major issues. Indeed, within a week of Lehman's bankruptcy, most of their outstanding OTC-cleared positions had been hedged, and within another week most of their client accounts had been transferred. Centrally-cleared OTC derivatives were seemingly much safer than their bilateral equivalents, although this may be questioned, especially given the recent Nasdaq default (Section 10.1.3).

In the aftermath of the GFC, policymakers (not surprisingly) embarked on regulatory changes that seemed largely aimed at moving risk away from global banks and the dangerous bilateral OTC derivatives market. This seemed to be driven generally by the view that the size, opacity, and interconnectedness of the market were too significant. One aspect of these policy changes was greater bank capital requirements for OTC derivatives. Another aspect was in relation to mandatory central clearing for certain products, with CCPs seemingly emerging as a panacea for financial markets' stability. For example:

How do we establish good regulatory structure without destroying the incentive to innovate, without destroying the marketplace? We agree that we need to improve our regulations and to ensure that markets, firms, and financial products are subject to proper regulation and oversight. For example, credit default swaps – financial products that ensure against potential losses – should be processed through centralized clearinghouses. (George Bush, 15 November 2008)

As a part of financial reform, important legislative changes with respect to the OTC derivatives market were introduced. In September 2009, G20 leaders agreed that all standardised OTC derivatives would, in the future, need to be cleared through CCPs. This was done with the belief that a CCP can reduce systemic risk, operational risks, market manipulation, and fraud, and contribute to overall market stability. It is interesting to note that the original push towards central clearing seemed to be much lighter. The G20 meeting in 2008 defined a regulatory goal as:

Strengthening the resilience and transparency of credit derivatives markets and reducing their systemic risks, including by improving the infrastructure of over-the-counter markets. (G20 declaration, Washington, November 2008)

Less than a year later, the clearing mandate was clear, and the focus on credit derivatives had expanded greatly to cover potentially all OTC derivatives:

All standardized OTC derivative contracts should be traded on exchanges or electronic trading platforms, where appropriate, and cleared through central counterparties by end-2012 at the latest. OTC derivative contracts should be reported to trade repositories. Non-centrally cleared contracts should be subject to higher capital requirements. We ask the FSB [Financial Stability Board] and its relevant members to assess regularly implementation and whether it is sufficient to improve transparency in the derivatives markets, mitigate systemic risk, and protect against market abuse. (G20 declaration, Pittsburgh, September 2009)

In 2010, both Europe (via the European Commission's formal legislative proposal for regulation on OTC derivatives, CCPs and trade repositories) and the US (via the Dodd–Frank Wall Street Reform and Consumer Protection Act) put forward proposals that would commit all standardised OTC derivatives to be cleared through CCPs by the end of 2012. As a result, policymakers seemed to focus on CCPs as something close to a panacea for counterparty risk, especially with respect to the more dangerous products such as credit default swaps.

An important distinction with clearing is whether a counterparty is clearing directly (i.e. is a member of a CCP) or indirectly (i.e. is not a CCP member). In the latter case, a counterparty will have to clear through a clearing member, and the interactions in such a tri-party set-up are clearly important to understand. In general, large banks will be clearing members, and smaller participants will be non-clearing members and clear through a large bank. There will be some intermediate cases, such as smaller banks and financial institutions, where it will be necessary to weigh up the pros and cons of clearing directly and indirectly for a given asset class or product. This is discussed in more detail in Section 8.2.5.

Clearing is a process that occurs after the execution of a trade in which a CCP may step in between counterparties to guarantee performance. The main function of a CCP is, therefore, to interpose itself directly or indirectly between counterparties to assume their rights and obligations by acting as a buyer to every seller and vice versa. This means that the original counterparty to a trade no longer represents a direct risk, as the CCP to all intents and purposes becomes the new counterparty. CCPs essentially reallocate default losses via a variety of methods, including netting, margining, and loss mutualisation. Obviously, the intention is that the overall process will reduce counterparty and systemic risks.

CCPs are not a new idea and have been a part of the derivatives landscape for well over a century in connection with exchanges. An exchange is an organised market where buyers and sellers can interact in order to trade. Central clearing developed to control the counterparty risk in exchange-traded products and limit the risk that the insolvency of a member of the exchange may have. CCPs for exchange-traded derivatives is arguably a good example of market forces privately managing financial risk effectively. The two clearing structures, bilateral and central, share many common elements such as netting and margining, but also have fundamental differences. The fact that neither has become dominant suggests that they may each have their own strengths and weaknesses that are more or less emphasised in different markets.

CCPs provide a number of benefits. One is that they allow netting of all trades executed through them. In a bilateral market, an institution being long a contract with counterparty A and short the same contract with counterparty B has counterparty risk. However, if both contracts are centrally cleared, then the netted position has no risk. CCPs also manage margin requirements from their members to reduce the risk associated with the movement in the value of their underlying portfolios. CCPs also allow loss mutualisation; one counterparty's losses are dispersed throughout the market rather than being transmitted directly to a small number of counterparties with potential adverse consequences. Moreover, CCPs can facilitate orderly close-out by auctioning off the defaulter's contractual obligations with netting, thus reducing the total positions that need to be replaced, which reduces price impact. A well-managed centralised auction mechanism can be liquid and result in smaller price disruptions than uncoordinated replacement of bilateral positions during periods of pronounced uncertainty. CCPs can also facilitate the orderly transfer of client positions from financially-troubled intermediaries. The margins and other financial resources they hold protect against losses arising from this auction process.

The general role and mechanics of central clearing are:

A CCP sets certain standards for its clearing members.
The CCP takes responsibility for closing out all the positions of a defaulting clearing member.
To support the above, the CCP maintains financial resources to cover losses in the event of a clearing member default:
- variation margin to closely track market movements;
- initial margin to cover liquidation or close-out costs above the variation margin; and
- a default fund to mutualise losses in the event of a severe default.
The CCP also has a documented plan for the very extreme situation where all their financial resources (initial margin and the default fund) are depleted.¹⁶ For example:
- additional calls to the default fund;
- variation margin gains haircutting; and
- selective tear-up of positions.

It is important to note that many end users of OTC derivatives (e.g. pension funds) will access CCPs through a clearing member and will not become members themselves. This will be due to the membership, operational, and liquidity requirements related to being a clearing member. Some end users will also be exempt from the clearing obligation and will, therefore, be able to choose whether to clear via a CCP or not.

4.4.2 Bilateral Margin Requirements

Imposing a clearing mandate on all standardised OTC derivatives is clearly intended to encourage the central clearing of as many transactions as possible. However, for a particular product to be ‘clearable’ there must be enough incentive for CCPs and clearing participants to develop a clearing capability for that particular product. Obviously, if banks feel that clearing will be more expensive, then they will try and avoid this – for example, by transacting products that are not yet clearable. There are certain potential loopholes here: a clearable product can easily be made non-clearable by changing the contractual terms (for example, the maturity or reference rate). The introduction of higher capital charges (Section 4.2) for non-cleared OTC derivatives was one indirect way to promote clearing. However, a more direct way is to mandate margining/collateral practices in bilateral markets that are close to clearing. This is the aim of the bilateral margin requirements (BMR), often known as the uncleared margin rules (UMR). The rules are formulated by the Basel Committee and International Organization of Securities Commissions (BCBS-IOSCO 2012, 2013b, 2015, 2019a).

The mandatory margin regime takes into account two types of collateral:

Variation margin. This covers the variability of the portfolio of transactions in question and is already relatively common in bilateral derivative markets.
Initial margin. This represents extra collateral to cover costs in the event of a default and is historically uncommon (as ‘independent amount’) in bilateral derivative markets.

The rationale for the above rules (in particular initial margin) seems to be to reduce systemic risk (due to the fact that a large fraction of OTC derivatives will not be centrally cleared) and promote central clearing. It could be argued that high capital charges alone would achieve this objective. However, there is a key difference between initial margin and capital charges; for example, BCBS-IOSCO (2013b) notes that:

Initial margin follows a ‘defaulter pays’ approach rather than a ‘survivor pays’ one, and therefore does not consume the financial resources of surviving entities.
Initial margin is more ‘targeted’ to a specific portfolio and is furthermore adjusted over time to reflect changes in the risk of that portfolio in market conditions.

Initial margins, in terms of amount and liquidity, will be broadly in line with those required by CCPs and therefore they can be seen to promote central clearing and narrow the gap between the treatment of clearable and non-clearable OTC derivatives.

The specific details of the BMR are discussed in more detail in Section 7.4, but the key requirements are outlined below:

Variation margin should be taken to reflect current exposure with no threshold (a threshold as described in Section 7.3.4 represents an amount below which margin is not exchanged).
Initial margin should be taken against the potential increase in exposure in the future and should be calculated to cover a confidence level of (at least) 99%.
A time horizon for close-out of 10 days should be assumed for initial margin calculation purposes.
A key issue with initial margin requirements is segregation. Without any specific segregation (such as custodial or segregation agreement), a party having posted initial margin would essentially only be treated as an unsecured creditor in the event of the default of the margin receiver. Since non-segregated initial margin leads to counterparty risk (in case the receiver defaults), then initial margin segregation is required.
Margin must be liquid, not just in normal market conditions but also in volatile periods, to the extent that it can be sold rapidly and predictably even in a time of financial stress.
Appropriate haircuts must be taken against the value of cash or securities posted as margin.

4.4.3 Exemptions

Unlike capital and liquidity requirements, which generally apply to banks, clearing and margining regulation applies to potentially all users of OTC derivatives. However, certain relatively unsophisticated end users, who use derivatives for hedging purposes, cannot realistically be expected to clear or conform to the UMR and are therefore given exemptions. Furthermore, some products are exempt from these mandates. To determine whether a given transaction must be cleared or subject to margin requirements, the following questions must be asked:

Are both parties subject to mandatory clearing or margin requirements (‘covered entities’)?
Is the product itself subject to mandatory clearing or UMR?

If the answer to both the above questions is yes, then the transaction is subject to mandatory clearing or bilateral margining requirements. A transaction can, therefore, be exempt, either due to the nature of either one of the entities trading or due to the product type itself.

In general, the following entities are exempt from the mandates:

sovereigns;
central banks;
multilateral development banks;
the Bank for International Settlements; and
other non-financial institutions that are not systemically important.

The final exemption above clearly has to be carefully defined and generally reflects end users undertaking hedging activities. However, the precise definitions are slightly different depending on the region in question:

Clearing exemptions under US regulation (Dodd–Frank). Exemptions for non-financial entities entering into transactions in order to hedge their commercial risk.
Clearing exemptions under EU regulation (EMIR). Non-financial counterparties (known as NFC-) are exempt. A counterparty is designated NFC+ when it crosses a threshold in terms of OTC derivative notional. These thresholds are set by asset class and are reviewed periodically. Entities designated as NFC+ are obliged to clear and comply with the UMRs, but NFC- entities are exempt. Pension funds are also currently given a temporary exemption, which is due to expire in 2020. This is intended to avoid dramatic shifts that could force major changes in asset allocation and gives CCPs more time to develop models for the main pension scheme products, namely interest rate and particularly inflation derivatives. Note that pension funds are not exempt from the UMR.
Bilateral margin exemptions (global guidelines via the BCBS-IOSCO). Covered entities are financial firms and systemically important non-financial entities. There is a notional threshold below which entities are exempt and a further counterparty-level threshold below which initial margins need not be posted (Section 7.4).

A significant transaction-related exemption is that of certain FX products. Market participants lobbied hard to argue that FX represents a special case since contracts are generally short-dated, and existing market practice such as CLS (continuous linked settlement) reduces risk significantly already (see discussion in Section 3.1.2 for a description of settlement risk and CLS functionality). Under EMIR, physically-settled FX forwards, FX swaps, and the FX ‘leg’ of cross-currency swaps are subject only to variation margin requirements (not initial margin). Similar exemptions exist in the US. Despite the clearing exemption, some non-deliverable forwards (cash-settled FX forwards) have been cleared on a voluntary basis.

Whilst physically-settled FX products create problems due to gross settlement and settlement risk (Section 3.1.2), there has been debate over the exemptions. For example, Duffie (2011) argues that whilst many contracts are short-dated, some FX transactions have significant counterparty risk due to the exchange of notional together with aspects such as the relatively high FX volatilities, fat tails, and sovereign risk. There are also potential regulatory arbitrages from categorising products differently.¹⁷

Another important point is that, generally, the clearing and the UMR apply only to trades done after the regulation requiring these changes was implemented and therefore only apply to new and not legacy transactions. Market participants can decide to voluntarily adopt clearing (this is referred to as ‘backloading’) or change their margining/collateral agreements to follow the UMR.

One potential issue, however, arises with respect to ‘frontloading’ of CCP trades. This requires derivatives counterparties to retrospectively clear any transaction executed between the authorisation of a CCP that clears the product and the start of a formal mandate. Once a CCP has been authorised to do business in a given region, regulators will decide if product offerings from that CCP will lead to mandatory clearing of these products. However, trades executed in the time period between CCP authorisation and this decision would need to be cleared if their remaining maturity is above a threshold. This means that a bilateral transaction may later become subject to mandatory clearing, which introduces confusion as the costs of bilateral and cleared transactions are different. This will give market participants a difficult decision as, when pricing and executing the trades, they would be uncertain as to whether or not they would be hit by a clearing mandate in the future.¹⁸

4.4.4 CCP Capital Requirements

Given the clearing mandate, there is also a need to define capital requirements for centrally-cleared transactions. Like the bilateral market, this would have the potential to distinguish:

the credit quality (e.g. default probability and LGD) of the CCP;
the exposure of the portfolio; and
the margining terms and their impact on reducing exposure.

Prior to the clearing mandate, exposure to a CCP via contracts such as derivatives (and assets held by the CCP as a custodian) were given a zero exposure under the so-called Basel II framework (2004) as long as the CCP exposures were fully margined on a daily basis.

Under current capital rules, CCP exposures are subject to capital charges in order to reflect the fact that CCPs are not risk free, and also to potentially differentiate between the quality of different CCPs. Under Basel III, a low-risk weight of 2% is applied to qualifying CCPs (QCCPs). QCCPs are recognised by local regulators as conforming to global principles and accordingly should be safer than non-qualifying CCPs (to which banks are exposed to higher capital charges). Whilst this element is not risk sensitive (i.e. the 2% applies for all QCCPs), there is another capital charge (related to the so-called default fund discussed in Section 8.2.4) which is risk sensitive. This capital charge will, therefore, be larger for a CCP that is deemed more risky.

CCP capital rules are dependent on whether clearing is done directly or indirectly. In the latter case, the 2% risk weight may be higher, but there will be no default fund-related capital charge since indirect clearers (i.e. those clearing through clearing members) do not contribute to the CCP default fund.

CCP capital charges and calculation methodologies will be discussed in more detail in Section 13.6.