Customer Liquidity Provision in Canadian Bond Markets

Introduction

This note assesses the provision of bond market liquidity by institutional customers (i.e., pension funds, hedge funds, mutual funds and insurance companies) in Canada. Customer liquidity provision occurs when a dealer, after filling an order from a customer, quickly makes an offsetting trade with an institutional investor.

Customer liquidity on a large scale could have positive and negative effects on bond markets. It can diversify the supply of liquidity beyond a small group of dealers and brokers, thus making bond markets more competitive and robust. However, customer liquidity may be more sensitive to a deterioration in market conditions and thus a less reliable source of liquidity. For example, sudden redemptions at a mutual fund can force the fund to switch from supplying liquidity to demanding it (Arora 2018). In contrast, dealers have a broad set of funding sources and may be more able to provide liquidity for their clients through a range of market conditions.

For our analysis, we use new data from Canada’s Market Trade Reporting System (MTRS) and find there is customer liquidity provision in 9 per cent of corporate bond trading in Canada. The share is lower for provincial bonds: 4 to 6 per cent, depending on the province. Customer liquidity is concentrated in relatively illiquid bonds, occurs on days with high trading volume and is more frequent when liquidity is scarce. One interpretation of these results is that customer liquidity supplements dealer liquidity when liquidity demand is high.

What is customer liquidity provision?

Customer liquidity provision is a type of agency intermediation (Hyun, Johal, and Garriott 2017) in which a dealer manages its inventory risk by matching some of its client orders to another counterparty using an offsetting trade. Customer liquidity provision occurs when the offsetting counterparty is an institutional investor and not another dealer.

Customer liquidity may be passive or active. If passive, a customer is approached by a dealer looking to offset some other customer’s trade. Dealers often offer price concessions to motivate the customer to transact quickly and supply liquidity. If active, a customer quotes a schedule of prices and quantities to the dealer (or even the market) and trades selectively to take advantage of price dislocations.

Customer liquidity is rare but occasionally supplements dealer liquidity

We use data from the MTRS from June 2016 to August 2017. MTRS contains a record of trades at broker-dealers registered with the Investment Industry Regulatory Organization of Canada, the self-regulatory organization for the Canadian securities market. The data contain the date, time, price, quantity, security and counterparty type for each trade.

Following Choi and Huh (2017), we first separate dealers’ trades into customer liquidity trades and dealer liquidity trades. We identify customer liquidity trades as those that offset customer trades intermediated by a single dealer within 15 minutes.1 This is a conservative metric; customer liquidity provision may take place over a longer time frame and may involve more than one dealer, but including other transactions would be more likely to mix customer liquidity with dealer liquidity.

Chart 1 shows the volume shares of customer liquidity provision for bonds issued by corporates and provinces; the share is equal to the volume of customer liquidity trades divided by the total trading volume. Customer-supplied liquidity is uncommon in Canada, averaging between 4 and 9 per cent, depending on the type of bond, and is significantly less common than in the United States (Choi and Huh 2017). Customer liquidity provision is more frequent for less liquid securities such as corporate debt and the debt of provinces with low quantities of debt outstanding (i.e., provinces other than Ontario and Quebec). This is consistent with dealers seeking client liquidity to reduce inventory risk.

Chart 1: Customer liquidity provision is rare across major Canadian bond classes

Sources: MTRS 2.0, Triton; estimate for US corporates is from Choi and Huh (2017) Last observation: 2017/08/21

Next, we ask whether average spreads differ for trades that we identify as drawing on customer liquidity. The spread is:

$$Spread=2\,Sign ×\frac{(Traded\,Price-Reference\,Price)}{(Reference\,Price)}$$

where Sign equals 1 for a dealer buy and -1 for a sell. The traded price excludes trades of less than $1 million to eliminate retail trades. The reference price is the average price, weighted by volume, of all interdealer trades greater than $100,000 for the bond on the same day. Spreads for customer-supplied liquidity include both legs of matched customer trades; we do this so we can directly compare total liquidity costs of dealer- and customer-supplied trades.

Chart 2 shows the average spreads for trades where liquidity was provided by customers and by dealers. Trades drawing on customer liquidity have higher average spreads than trades that draw solely on dealer liquidity. This contrasts with the US results seen in Choi and Huh (2017), in which trades using customer liquidity have lower average spreads. Our result is surprising because, as Choi and Huh point out, the effective spread for trades with some customer-liquidity supply should be biased downward by the prices of the offsetting trades. Yet, despite the potential bias, we still measure a larger spread for these trades.

Chart 2: Trades that draw on customer liquidity have higher average spreads

Sources: MTRS 2.0, TritonLast observation: 2017/08/21

Our explanation for this difference is that, in Canada, institutional investors are a discretionary source of liquidity for dealers, who use their customers selectively because customers demand a price concession to supply liquidity. If our explanation is correct, we should observe that the use of customer liquidity eats into dealer revenues and that dealers use customers mainly on days when liquidity is scarce or when volume is high (i.e., days when they are more likely to run into risk constraints).

We check the prediction that dealers make less revenue when they draw on customer liquidity. As in Choi and Huh (2017), we compute a metric of the round-trip revenue that dealers earn on their intraday positions. We collect all positions that are offset in a single day, and we compare revenue on trades offset using customers (in 15 minutes, as before) with the same revenue from the dealer trades. Revenues on intraday inventories are computed on a first-in-first-out basis.

Chart 3 shows that dealers make less revenues on their intraday trading positions when they use customer liquidity, which is consistent with our hypothesis. They appear to be paying to use this supplementary source of liquidity.

Chart 3: Using customer liquidity reduces dealer revenues

Sources: MTRS 2.0, TritonLast observation: 2017/08/21

Next, we confirm that this customer liquidity is more prevalent when liquidity is scarce. Chart 4 shows average spreads for dealer-liquidity trades on two types of days: those with low shares of customer-supplied liquidity and those with high shares. We do this to see if a higher cost of liquidity in general can explain the increased use of customer liquidity.

Chart 4: Customer liquidity is more prevalent when liquidity is expensive

Sources: MTRS 2.0, TritonLast observation: 2017/08/21

Indeed, Chart 4 shows that days with higher shares of customer liquidity provision have higher spreads, indicating that some part of the higher spreads we observed for customer liquidity in Chart 2 are the result of trades taking place when liquidity is scarce.

Last, we check whether dealers use expensive customer liquidity to manage trading exposures on days with high volumes. The intraday exposures of dealers may exhibit greater risks on these days, so dealers might use customers as discretionary sources of liquidity. We compute the average trading volume for days with high shares of customer liquidity against days with low shares to see if customer liquidity is used more on high-volume days.

Chart 5 confirms that trading volumes are significantly larger on days with higher customer liquidity. Together, Chart 4 and Chart 5 show that customer liquidity is more prevalent when liquidity is scarce or volumes are high. In our interpretation, customer liquidity is used to accommodate trading as a secondary source of liquidity on more active trading days.

Chart 5: Customer liquidity is more prevalent when trading volumes are higher

Sources: MTRS 2.0, TritonLast observation: 2017/08/21

While uncommon now, customer liquidity may eventually increase

We find the share of customer liquidity is small in Canada. However, data are only available from June 2016 on. Survey evidence suggests that the share of customer liquidity may be rising. Responses to the Survey on Market Liquidity, Transparency, and Market Access (Canadian Fixed-Income Forum 2016) indicate that a sizable minority of customers are supplying liquidity to markets: 30 per cent of buy-side respondents said they have increased their short-term trading to take advantage of short-term price dislocations. An equal number said they use their portfolio to provide liquidity.

In addition, forthcoming regulations may lead to a greater share of client liquidity in Canada. The implementation of the Net Stable Funding Ratio and the Fundamental Review of the Trading Book will likely increase the cost to bank-owned dealers of carrying large inventories of bonds. After other capital and liquidity standards were implemented (e.g., Liquidity Coverage Ratio, Leverage Ratio), banks generally reduced their capital commitment to market-making in the United States (Adrian, Boyarchenko and Shachar 2017; Bessembinder et al. 2017; Schultz 2017; Bao, O’Hara and Zhou 2016), while the volume share of non-bank dealers grew in the same period (Bao, O’Hara and Zhou 2016). Finally, further proliferation of electronic bond trading platforms, together with a broader array of trading protocols, could make it possible for customers to supply liquidity at lower costs.

  1. 1. We exclude trades where the security has been issued in the previous 30 days, as well as trades where the security’s term-to-maturity is less than one year. We focus only on fixed-rate, vanilla bonds to make it easier to compare the spreads and revenues of corporate bonds (which are more likely to have idiosyncratic features that may influence pricing and liquidity) and of provincial bonds.[]

References

  1. Adrian, T., N. Boyarchenko and O. Shachar. 2017. “Dealer Balance Sheets and Bond Liquidity Provision.” Federal Reserve Bank of New York Staff Reports No. 803. Available at https://www.newyorkfed.org/medialibrary/media/research/staff_reports/sr803.pdf?la=en
  2. Arora, R. 2018. “Risk of Redemption Runs in Corporate Bond Mutual Funds.” Bank of Canada Staff Analytical Note (forthcoming).
  3. Bao, J., M. O’Hara and A. Zhou. 2016. “The Volcker Rule and Market-Making in Times of Stress.” Federal Reserve Board Finance and Economics Discussion Series. Available at https://www.federalreserve.gov/econresdata/feds/2016/files/2016102pap.pdf
  4. Bessembinder, H., S. Jacobsen, W. Maxwell and K. Venkataraman. 2017 "Capital Commitment and Illiquidity in Corporate Bonds.” Journal of Finance (forthcoming). Available at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2752610
  5. Canadian Fixed-Income Forum. 2016 CFIF Survey Results on Liquidity, Transparency and Market Access in Canadian Fixed Income Markets. Available at https://www.bankofcanada.ca/wp-content/uploads/2016/10/cfif-survey-overview-031016.pdf
  6. Choi, J. and Y. Huh. 2017 “Customer Liquidity Provision: Implications for Corporate Bond Transaction Costs.” Available at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2848344
  7. Hyun, D., J. Johal and C. Garriott. 2017. “Do Canadian Broker-Dealers Act as Agents or Principals in Bond Trading?” Bank of Canada Staff Analytical Note 2017-11. Available at https://www.bankofcanada.ca/wp-content/uploads/2017/09/san2017-11.pdf
  8. Schultz, P. 2017. “Inventory Management by Corporate Bond Dealers.” Available at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2966919

Acknowledgements

We would like to thank Jason Allen, Jean-Sébastien Fontaine and Darcey McVanel for many helpful comments and suggestions. We would also like to thank Joshua Fernandes for excellent research assistance.

Avis d’exonération de responsabilité

Les notes analytiques du personnel de la Banque du Canada sont de brefs articles qui portent sur des sujets liés à la situation économique et financière du moment. Rédigées en toute indépendance du Conseil de direction, elles peuvent étayer ou remettre en question les orientations et idées établies. Les opinions exprimées dans le présent document sont celles des auteurs uniquement. Par conséquent, elles ne traduisent pas forcément le point de vue officiel de la Banque du Canada et n’engagent aucunement cette dernière.