Regulation after the Financial Crisis. Impact on Corporate Bond Market Liquidity

Table of Content

List of Abbreviations

List of Figures

1. Introduction

2. Liquidity provision in equity and corporate bond markets
2.1. Equity markets
2.2. Corporate bond markets

3. Post-crisis regulatory provisions
3.1. Basel 2.5
3.2. Basel III
3.3. The Volcker Rule

4. First-order impact on dealer liquidity provision
4.1. The business model of market makers
4.2. Channels of regulatory impact
4.3. Foundations of contemporary literature
4.4. Evidence of the regulatory impact on traditional dealers

5. Changes in market structure and behavior
5.1. Second-order effect (1): Shift from principal trading to matchmaking
5.2. Second-order effect (2): Migration of market making to non-bank dealers

6. Evidence of aggregate corporate bond market liquidity
6.1. Price-based liquidity measures
6.2. Trading-based liquidity measures
6.3. Liquidity under stress

7. Conclusion

References

Appendix

List of Abbreviations

Abbildung in dieser Leseprobe nicht enthalten

List of Figures

Abbildung in dieser Leseprobe nicht enthalten

1. Introduction

After the financial crisis, regulators intervened to enhance the resilience of the banking system. Their provisions range from capital and liquidity standards to the prohibition of single activities considered too risky. However, concerns arise that post-crisis regulations harm liquidity by imposing constraints on its providers. When liquidity is low, investors that want to trade large volumes must wait for counterparties or accept to trade below market prices. Therefore, in certain financial markets like that for corporate bonds, intermediaries emerged to facilitate market functioning. They enable investors to trade immediately, reconciling imbalances in supply and demand (CGFS, 2014).

Illiquidity is costly for the economy as investors require compensation for holding riskier bonds. Amihud and Mendelson (2012) provide cross-sectional and time-series evidence of the resulting illiquidity discount. Hence, if regulations reduced liquidity, they would cause a depreciation of prices. Also, lower liquidity implies a higher cost of debt and transaction costs, as well as a less efficient resource allocation. The regulatory impact on liquidity is, therefore, highly important for policymakers and investors.

The paper aims to answer the question of whether post-crisis regulatory interventions caused a decline in liquidity. To serve this purpose, it investigates how individual provisions affect the market making business and how the corporate bond market changed in response to regulations. The paper approaches the issue by structuring theoretical and empirical evidence of corporate bond liquidity. It develops regulations' impact levels from particular to aggregate, facilitating a perspicacious analysis. Important to note, the study attempts to assess neither welfare effects nor the desirability of regulations.

Furthermore, the main focus is on the U.S. corporate bond market. It is the largest market of corporate debt and provides comprehensive data from bond transactions. Nevertheless, it will become evident that findings also apply to other jurisdictions.

The current literature shows that the Basel Accords and the Volcker Rule are the most relevant post-crisis regulations for corporate bond markets. They entail provisions that increase the inventory and financing costs of bank-affiliated dealers. Market making, however, is a business that bases on low margins and high volumes. Its profitability is, therefore, vulnerable to changes in marginal costs. Besides, regulations constrain dealers' ability to manage risks and impose limits on their inventories. Bank-affiliated dealers respond to the regulations by reducing capital committed to market making. They are less willing to absorb large order imbalances and more reluctant to hold risky bonds.

In addition to the withdrawal of bank dealers, regulations induce major changes in the structure and behavior of the corporate bond market. Non-bank dealers enter the market and compete with traditional dealers for the provision of immediacy. Before the implementation of regulations, bank dealers marginalized less efficient non-banks. Because regulations increase their market making costs, however, they surrender parts of their business. Instead, bank-affiliated dealers shift to matchmaking, which is another mode of trade facilitation. They match customer orders in place of assuming ownership and related risks. Consequently, investors must wait for suitable counterparties to execute trades. They cannot trade immediately but pay lower transaction costs (Schultz, 2017). The electronification of corporate bond markets coincides with the impact of regulations. New trading venues emerge and lower the entry barriers for non-banks. Also, new technologies increased the efficiency of matchmaking.

Many empirical studies indicate that corporate bond market liquidity is higher than or indifferent from pre-crisis levels. Particularly, liquidity proxies that measure transaction costs like the bid-ask spread suggest that liquidity improved. However, other research findings object to this notion. They find that the cost of immediacy increased, and the turnover declined (Anderson and Stulz, 2017; Choi and Huh, 2019). Reconciling this seemingly contradictory evidence reveals shortcomings in conventional liquidity measures. These do not capture dimensions of liquidity that are highly relevant to investors, like costs associated with delayed execution or forgone trades.

The evidence of the regulatory impact does not suggest that it impaired liquidity in all respects. Instead, regulations induced a structural change to the corporate bond market that has different implications for investors. Those that heavily depend on immediacy services and trade large volumes are likely affected most.

The paper continues as follows. Section 2 and Section 3 lay the foundations for the study. They compare liquidity provision in equity and corporate bond markets and introduce the most relevant regulatory provisions. Section 4 explores how regulations directly impact traditional dealers (“first-order”). Section 5 expands the analysis to behavioral responses to regulations that alter the corporate bond market structure (“second-order”). Section 6 then structures evidence of liquidity measures, identifies contradictions, and presents solutions suggested in the literature. Section 7 concludes.

2. Liquidity provision in equity and corporate bond markets

Equity and corporate bond markets differ substantially in terms of asset and market structure. While corporations tend to have a few classes of shares, they issue bonds varying in maturity, seniority, and collateral (Adrian et al., 2017). In 2018, the average weekly turnover of corporate bonds was only 1.7%, compared to 3.6% on the part of NYSE-listed stocks (SIFMA, 2019)¹. This fragmentation and infrequent trading results in corporate bonds being traded in over-the-counter (OTC) markets. Meanwhile, equity investors can trade on exchanges without the need for intermediaries.

2.1. Equity markets

Most exchanges are either pure electronic limit order markets or structured as hybrids. Hybrids also allow for ex-post liquidity provision by designated specialists (Parlour & Seppi, 2008). Thereby, limit order markets are order-driven. Any investors can provide executable quotes by posting non-marketable limit orders (CGFS, 2014). That is, in equity markets, investors can provide liquidity multilaterally in limit order books.

Examining a contemporary trading venue helps illustrate this mechanism. Xetra is an electronic open limit order book operated by Deutsche Börse AG². Occupying more than 90% of overall share trading, it is the leading German exchange (Deutsche Börse AG, 2020b). At Xetra, investors can electronically place limit and market orders. Limit orders are pre-commitments that specify the amount an investor is willing to trade below or above a price limit. Market orders, in contrast, demand immediate execution at best obtainable market prices. Xetra's market model covers “the trading forms auction and continuous trading for on-exchange trading” (Deutsche Börse AG, 2019, p. 16). That means Xetra conducts an opening auction to initiate the trading day, an intraday auction, and a closing auction that terminates active trading. Between the auctions, continuous trading takes place. Thereby, algorithms match orders without human agency³.

Xetra promptly checks arriving orders against the existing orders. If no execution is possible, they add to the limit order book (Deutsche Börse AG, 2019, p. 22). Whether orders provide liquidity depends on the limit relative to the current bid and ask in the book. Thereby, market or limit orders that obtain immediacy presumably pursue active trading motives and, hence, consume liquidity. In other words, non-marketable limit orders that add to the limit order book provide liquidity (Parlour & Seppi, 2008).

The NYSE is the leading stock exchange in the U.S. It operates a hybrid model that allocates each stock to a sole specialist. In active markets, particularly on the part of retail­sized orders, transactions take place without intermediation. Then, designated specialists merely act as brokers and maintain the limit order book. However, if trading is infrequent or risks to cause price fluctuations, specialists are obliged to provide liquidity. That is, they deal on own account and earn the bid-ask spread (Bodie, Kane, and Markus, 2011, pp. 67-68). In recent years, transactions on the NYSE have further shifted away from intermediation. As Jovanovic and Menkveld (2016) describe, trading of NYSE-listed stocks migrated from floor-based to limit order markets, and high-frequency traders contribute liquidity replacing human specialists.

2.2. Corporate bond markets

As aforementioned, the heterogeneity of corporate bonds results in a fragmented market. Another market characteristic is high institutional ownership. Insurance companies or pension funds, for instance, pursue buy-and-hold strategies (Friewald and Nagler, 2016). Hence, inactive trading and market fragmentation complicate the search for suitable counterparties. Institutions also transact large volumes, likely having adverse price impact in an open limit order book setting. Therefore, investors in corporate bonds require active intermediation by market makers in OTC markets (Bessembinder, Spatt & Venkataraman, 2020). That is, while liquidity provision in equity markets migrates towards multilateralism, it remains bilateral in fixed income.

Juxtaposed to order-driven equity markets, corporate bond trading is quote-driven. Thereby, only market makers provide quotes either continuously or upon customers' request for quotation (RFQ). The contact between institutional clients and market makers can, for instance, occur by phone⁴. Their relationships tend to endure and involve negotiations (Bessembinder, Spatt & Venkataraman, 2020). Most market makers are bank-affiliated securities brokers and dealers (dealers). These serve a dual function. Confronted with customer orders, they can opt to act as dealers and trade on own account (principal trading). Otherwise, as a broker, they match orders without assuming ownership (agency trading). Besides, non-bank liquidity providers like hedge funds or asset management firms emerged post-crisis (Choi and Huh, 2019).

Principal trades are transactions in which dealers consent to commit capital. They mirror existing order imbalances as counterparties and enable immediate trade execution. Consequently, positions appear on balance sheets until dealers absorb reverse trades. Friewald and Nagler (2016) find that dealers need five to six weeks to liquidate half of an individual bond inventory. It illustrates the fundamental risk involved in principal trading. For assuming this risk, dealers earn the bid-ask spread. The further customer liquidity demand deviates inventory from target levels, and the lower the dealers' risk limit, the wider is the bid-ask spread that dealers require to undertake a principal trade⁵.

In comparison, agency trades refer to transactions that do not affect dealers' balance sheets. Instead of trading on own account, they act on behalf of their customers. While principal trades reconcile orders over time, agency trades merely match current supply and demand. Hence, trade execution requires an available counterparty, which can be a customer or another dealer. Investors, therefore, have to pause and bear transaction fees as well as costs associated with the execution delay (Saar et al., 2019). It elucidates that agency trading does not provide immediacy in the narrower sense. Notably, the waiting time of liquidity-demanding investors that approach dealers first is unobservable.

Prearranged (or riskless principal) trades are economically similar to agency trades. These affect the balance sheet for a short time, not putting it at risk. For instance, Bessembinder et al. (2018) designate trades that dealers offset within 60 seconds as prearranged. Given the low aggregate liquidity of corporate bonds, the matched counterparty was de facto known before the first trade. Henceforth, matchmaking serves as a hypernym for both agency and riskless principal trades.

In recent years, electronification has altered liquidity provision in corporate bond markets. Dealers enhanced their business with trading algorithms, and all-to-all platforms emerged to pool liquidity from market participants. These order-driven platforms enable buy-side investors to provide liquidity. They thus reduce the need for market making analogously to equity markets. However, they remain less prominent. High transparency and price changes against volume make institutions dependent on active intermediation (BIS, 2016; CGFS, 2016). Today, almost 20% of U.S. corporate bonds, primarily retail­sized, trade on the leading all-to-all platform MarketAxess (Saar et al., 2019).

3. Post-crisis regulatory provisions

The financial crisis unveiled significant shortcomings in the financial system. Severe solvency and liquidity concerns of sizable banks pointed to the relevance of systemic risk. This risk refers to knock-on effects created by interconnected markets that render the whole economy unstable. Regulatory authorities, however, noted that banks accounted for neither systemic risk nor failure costs in their decision-making (Haselmann et al., 2019). Hence, the Basel Committee on Banking Supervision (BCBS)⁶ revised existing regulations to enhance the safety and soundness of the banking system.

This section provides a basic understanding of post-crisis regulatory provisions. It focuses on the most relevant measures for market making in corporate bonds. These are the revisions to Basel II, called Basel 2.5, Basel III, and the Volcker Rule.

3.1. Basel 2.5

Basel 2.5, formally revisions to the Basel II market risk framework, pertains to a set of measures that increased capital requirements for market risk and countered regulatory arbitrage. For European banks, compliance was mandatory from January 2012. In the U.S., it became effective one year later. There, authorities adopted the revision after a period of review and amendments on June 7, 2012 (Getter, 2015).

Market risk refers to losses from banks' trading operations due to falling market prices. Previous regulations underestimated this risk. During the crisis, banks' trading books incurred losses well beyond pre-crisis capital requirements⁷ (BCBS, 2011a). Therefore, Basel 2.5 requires banks to determine minimum capital based on stressed Value at Risk (sVaR, see Appendix B, Figure 2). The measure accounts for adverse movements of market variables during one year of turmoil. All else being equal, sVaR already increased the minimum capital as pre-crisis estimates of VaR had based on data from a low- volatility period. However, Basel 2.5 also introduced a new capital formula, at least doubling capital requirements for market risk (Hull, 2018, pp. 378-379).

Regulatory arbitrage denotes attempts to minimize capital requirements by exploiting regulatory loopholes. A precedent was the shifting of assets from trading to banking books. It was possible to reduce regulatory capital as requirements were higher for traded assets⁸. Therefore, Basel 2.5 introduced the incremental risk charge (IRC). It stipulates capital as the maximum obtained from trading or banking book calculations (Hull, 2018, pp. 379-380). The ICR, thus, raised the capital needed for banks' trading activities.

3.2. Basel III

After the crisis, regulators turned their attention from individual banks to the resilience of the whole financial system. The BCBS, therefore, instigated a set of macroprudential banking regulations called Basel III (see BCBS, 2011b). Starting in 2014, they gradually phased-in the framework to allow for a consistent and coordinated adoption by banks. The phase-in agreement scheduled full implementation of all measures for 2019⁹. In the U.S., regulators announced the adoption of most Basel III recommendations on July 9, 2013. Thereby, national authorities have the discretion to adjust provisions to the regional context. U.S. authorities, for instance, removed references to external credit ratings. They presumed that flawed ratings contributed to the financial crisis (Getter, 2015).

Capital requirements: During the crisis, the banking system was fragile due to insufficient amounts of high-quality assets. Basel III, thus, augmented both the quality and quantity of required capital. Qualitatively, regulators tightened capital definitions. By dividing Tier 1 capital (T1) into Common Equity T1 (CET1) and Additional T1, they stressed the importance of common equity. Whereas hitherto, banks could deploy any form of equity to satisfy T1 requirements; they henceforward need common equity or retained earnings. Quantitatively, regulators enhanced the required CET1 from 4 to 4.5% and the total T1 from 4 to 6% of risk-weighted assets (RWA)¹⁰.

Furthermore, Basel III introduced the leverage ratio (LR) as a backstop for risk-based capital requirements¹¹. It addressed the procyclicality of minimum capital that proved to be a source of systemic risk. During the crisis, rising volatility elevated risk-based capital requirements. It forced banks to deleverage, amplifying the systemic shock. Basel III mitigates this risk by demanding banks to hold enough equity in good times. Specifically, it stipulates that banks must maintain a CET1 ratio of 3% relative to an exposure measure. Relevant to market makers, securities financing adds to this measure (BCBS, 2014). Section 4 outlines the function of securities financing and its implications for dealers.

Basel III also introduced two macroprudential capital buffers to increase banks' resilience under stress. The capital conservation buffer requires banks to withhold additional CET1 equal to 2.5% of RWA. Analogously, national authorities can stipulate a countercyclical buffer that increases the CET1 ratio by 0 to 2.5%¹². If common equity falls short of its required level (7 to 9.5%, see Appendix B, Figure 3), regulators stipulate minimum retained earnings that restrict distributions (Hull, 2018, pp. 383-384).

Liquidity requirements: Banks' short-term financing of long-term investments makes them vulnerable to sudden liquidity dry-outs. As regulators noticed, this maturity mismatch fueled the propagation of liquidity shocks. Basel III, therefore, implemented two liquidity standards. The liquidity coverage ratio (LCR) requires banks to hold liquid assets. Precisely, they must hold enough high-quality liquid assets (HQLA) to cover net cash outflows during 30 days of stress. Vice versa, the net stable funding ratio (NSFR) demands banks to use less liquid funding. Financing of assets that mature in less than a year must at least be commensurate with the assets' maturity (BCBS, 2011b).

3.3. The Volcker Rule

The Volcker Rule is a U.S. federal regulation that refers to section 619 of the Dodd-Frank Act. It prohibits institutions that capitalize on public sector backstops from engaging in proprietary trading and owning hedge or private equity funds. The rule proposed to eliminate the moral hazard created by government guarantees. Deposit insurance (FDIC), for instance, reduces depositors' risk sensitivity and might, thus, incline banks towards excessive risk-taking¹³ (Bessembinder, Spatt & Venkataraman, 2020). Compliance with the Volcker Rule was obligatory by June 2015, more than a year after its implementation.

Proprietary trading constitutes trading on own account to benefit from asset price movements. That is, it pursues investment motives. Proprietary trading resembles market making to the extent that banks commit capital. As shown in the second section, however, market making serves a vital function in fixed income. Therefore, the Volcker Rule comprises an exemption that permits market making provided it meets reasonably expected near term customer demand (Allahrakha et al., 2019).

Hence, regulators, as well as financial institutions, must tell bona fide market making from proprietary trading. For instance, the Financial Stability Oversight Council (2011, pp. 27-31) defined distinctive features. Among others, continuous bid and ask shall signal that liquidity provision is the sole trading purpose. To allow for consistent identification of proprietary trading, the Volcker Rule requires banks to report a multitude of daily measures. These include inventory turnover and aging, the volatility of daily trading profits, as well as internal risk and position limits. Supervisors then interpret the reported metrics. They, for example, consider slower inventory turnover or higher profit volatility as indicative of proprietary trading (Schultz, 2017).

On January 1, 2020, an amendment to the Volcker Rule responding to industry complaints became effective. It modified reporting requirements and aligned their extent with the bank's trading activities. It also reversed a presumption that, hitherto, assumed assets remaining on the balance sheet for fewer than 60 days originated from proprietary trading. The deadline for compliance with the amended Volcker Rule is January 1, 2021 (Federal Register, 2019). Hence, the unamended Volcker Rule is pertinent for this paper.

Although the Volcker Rule is specific U.S. regulation, Rischen and Theissen (2019) emphasize its applicability to European institutions. These likely comply with U.S. regulations due to U.S. market prominence and market interconnectedness. Furthermore, the Liikanen report recommended the E.U. to assume regulations similar to the Volcker Rule. Despite its eventual withdrawal, it likely led to anticipation on the part of banks.

[...]

¹ The turnover (ratio) is the volume traded relative to the total volume. These are the NYSE market capitalization ($20,679.5 bn), and the amount outstanding of U.S. corporate bonds ($9,200.7 bn)

² Deutsche Börse AG also operates the Frankfurt Stock Exchange as an on-floor trading venue with specialists running a closed limit order book (Appendix A.2 contrasts open and closed limit order books).

³ Appendix A.1 details the trading process at Xetra and rationales for regular auctions. For less liquid stocks, there are designated sponsors that allow for seamless trading by adding quotes to the limit order book.

⁴ Emerging single- or multi-dealer platforms make quotes of one or many dealers electronically available. Retail investors usually reach market maker through retail brokers.

⁵ Section 4 casts light on dealers' business model and risks involved in market making. It lays the foundations for understanding channels of regulatory impact.

⁶ The BCBS is a group of central banks and banking supervisory authorities from all major economies. The Bank of International Settlements (BIS) in Basel hosts the BCBS, yet they are distinct entities.

⁷ The trading book accounts for assets that banks (intend to) trade actively. The banking book, in contrast, consists of positions that banks hold until maturity.

⁸ The rationale for the unequal treatment rests on different exposures. While market risks are relevant for the trading book, risks of banking book positions arise from possible defaults of counterparties (credit risk).

⁹ The phase-in start was initially planned for 2013 but postponed to 2014 due to political delays. Furthermore, many countries did not obtain full implementation in 2019 (Financial Stability Board, 2019)

¹⁰ See BCBS (2011b, pp. 12-19) for an exhaustive account of capital definitions and requirements.

¹¹ It also safeguards against model and measurement errors by reducing the discretion in computing capital. Besides, U.S. authorities introduced a supplementary leverage ratio (SLR) that added a surcharge of 2% to the LR for the most significant institutions (Adrian et al., 2017).

¹² Confronted with the Covid-19 pandemic, the BaFin reduced the buffer in Germany to 0% (BaFin, 2020).

¹³ Further regulatory advantages with similar effects are implicit “too big to fail” subsidies or short-term liquidity provision by the Fed (called discount window).