Have Liquidity and Trading Activity in the Canadian Provincial Bond Market Deteriorated?

Introduction

Since the 2007–09 global financial crisis, the total value of Canadian provincial bonds outstanding has more than doubled (Chart 1). At the end of the second quarter of 2018, provincial bonds made up 24 per cent of the total outstanding bonds denominated in Canadian dollars. There are now more provincial bonds outstanding than Government of Canada bonds (GoCs).

After GoCs and other federal government-guaranteed bonds, provincial bonds are considered some of the safest and most liquid bonds in Canada. Unlike GoCs, which are issued by auction on a predetermined schedule, provincial bonds are issued through an underwriting syndicate without a pre-announced schedule.

However, concerns about the secondary market for provincial bonds were raised in the 2016 Canadian Fixed-Income Forum (CFIF) Survey on Market Liquidity, Transparency and Market Access (Bank of Canada 2016). In the survey, most respondents felt that provincial bond liquidity had declined between 2014 and 2016. They generally attributed this perceived deterioration to recent regulatory or technological changes that reduced dealers’ ability or willingness to act as market makers.

A resilient provincial bond market

In this note, we find evidence that the provincial bond market has been resilient since 2010. On the one hand, we find that a proxy for the bid-ask spread has worsened. On the other hand, all other measures of liquidity and trading activity are stable or have improved. For example, our price-impact proxy and the turnover ratio have remained stable since 2010. These results are consistent across bonds from different provinces, older and newer bonds, and bonds of different sizes.

Alberta provides an interesting case study of the resilience of provincial bond markets in recent years. After the fall in oil prices in 2014–15, trading activity for Alberta bonds improved despite large fiscal deficits, widening credit spreads and downgrades by credit rating agencies. The positive effect of increased bond issuance on trading activity is another sign that provincial bond markets are resilient.

Chart 1: The total value of outstanding provincial bonds has surpassed that of Government of Canada and corporate bonds

Data and methodology

Measures of liquidity and trading activity are calculated using data from Thomson Reuters (TR) and the Canadian Depository for Securities (CDS). The TR data are used for bond characteristics, such as issue date, maturity date, coupon rate and amount outstanding. CDS data include transaction information: the date, price and quantity of each provincial bond trade. The sample period is from January 2010 to August 2018.

The liquidity measures used are proxies for the price impact (Amihud 2002) and the bid-ask spread (Roll 1984). These are the same proxies used in Gungor and Yang (2017). More details on the calculation of the proxies are provided in the Appendix. Fontaine et al. (2017) find that these proxies are reliable measures of liquidity, even for infrequently traded bonds. For both proxies, a higher value indicates lower liquidity.

We also use the measures of trading activity in Gungor and Yang (2017): the monthly number of trades, the average trade size and the turnover ratio (see Appendix). We supplement those with a measure of zero-trade days—the share of trading days in a month where no transaction is recorded for a given bond. A higher share indicates that a bond is less frequently traded. This measure is taken from Dick-Nielsen, Feldhütter and Lando (2012) and Chen, Lesmond and Wei (2007).

Liquidity and trading activity are stable for the market as a whole

Chart 2 shows the daily average price-impact and bid-ask proxies for all outstanding provincial bonds since 2010. On average, the price-impact proxy has remained stable. In contrast, the level of the bid-ask proxy has edged higher since around 2013, pointing to modest deterioration in liquidity.

These two contrasting results hold when looking at bonds issued by different provinces. However, both proxies show resilience during the taper tantrum (2013) and after the oil price fall in 2014–15. In each case, both proxies deteriorated markedly, but the shocks were transitory and both proxies returned rapidly to previous levels.

Chart 2: Aggregate liquidity in the provincial bond market remains stable

Investigating market participants’ concerns

Our proxies may not reveal other dimensions of liquidity. In recent years, market participants have expressed concern about their ability to trade rapidly and in large quantities. Participants may have increasingly broken down their large trades into smaller ones if bulk transactions have become costlier or more difficult to conduct. Participants may also trade less than before, as a lower average liquidity pushes them to extend bond holding periods. Finally, adverse liquidity conditions may increasingly lead market participants to withhold trades for days (or longer), because they may find it advantageous to wait for market conditions to improve. However, our data provide no evidence that any of these changes have occurred in the provincial bond market.

First, the data provide little indication that trades are being broken down into a larger number of smaller trades. Chart 3 shows that, while the number of trades has increased, the average trade size has remained stable. Second, we find that market participants as a group do not trade less: the turnover ratio (Chart 4) is mostly unchanged throughout the period. Together, these charts show that the number of trades is growing in proportion to the outstanding value of provincial bonds. Finally, it is unlikely that market participants now wait for days or longer before executing trades. Chart 4 shows that the number of zero-trade days has declined since 2010.

Chart 3: Number of trades has increased while trade size is stable

Chart 4: Turnover is stable while zero-trade days have declined

Trading activity depends on bond characteristics

Another concern of market participants is that a decline in trading activity has affected a subset of bonds, notably those that are older or smaller. We find instead that trading activity has been generally stable across bonds of different ages and sizes.

Older bonds and smaller bonds are no less likely to trade in 2018 than they were in 2010. However, they are traded less often than newer and larger bonds. Chart 5a shows the number of trades for the oldest one-third of bonds outstanding and the newest one-third. Since 2010, the number of trades has increased for both groups of bonds. Chart 5a also shows the one-third of bonds with the largest issue size and the one-third with the smallest issue size. Again, we find that the number of trades has risen for both groups.

We also find that turnover for older and smaller provincial bonds has been stable since 2010. Chart 5b reports the turnover ratio for both the oldest one-third and newest one-third of bonds, as well as for the one-third of bonds with the largest issue size and the one-third with the smallest issue size. For most groups, turnover has generally remained stable since 2010, although it has declined slightly for the largest bonds.

Chart 5a: Newer and larger bonds are traded more frequently

Chart 5b: Turnover ratios are higher for newer and larger bonds

Activity in Alberta bonds has surged

Alberta bonds provide an interesting case study for the effect of issue size on trading activity. After the fall in oil prices in 2014–15, Alberta—a major oil producer—turned to bond markets to finance its budget deficit. The value of outstanding Alberta bonds rose from less than $19 billion in December 2013 to more than $60 billion in April 2018 (Statistics Canada 2018). Before 2015, Alberta had issued only a single Canadian-dollar bond with an issue size above $2 billion; since then it has issued seven, including a $6.2 billion bond in 2017. Alberta bonds were eventually downgraded by credit rating agencies and their yield increased relative to that of other provinces.

Yet measures of trading activity show marked improvements following the issuance of several large bonds, which could be interpreted as a sign of resilience for provincial bond markets. Alberta bonds went from being seldom traded on the secondary market to being traded more than 1,000 times a month (Chart 6a). The average trade size and turnover ratio increased, while the share of zero-trade days declined. In each case, the levels of these measures are close to those of Ontario and Quebec, Canada’s two largest bond issuers (Chart 6b, Chart 6c and Chart 6d). This evolution was not mirrored by British Columbia bonds, for example. Indeed, little has changed since 2010 for the trading activity and liquidity of bonds issued by other provinces. These results indicate that, in the case of Alberta, the positive effect of issue size on liquidity and trading activity has dominated the negative effect of lower credit quality.

Chart 6a: Trading activity for Alberta bonds increased after the fall in oil prices

Chart 6b: The size of Alberta bond trades has increased

Chart 6c: The turnover ratio of Alberta bonds has increased

Chart 6d: The share of zero-trade days has decreased for Alberta bonds

Conclusion

Our empirical results help track the evolution of the provincial bond market. We find that a proxy for the bid-ask spread has increased modestly since 2010. However, we also find that a proxy for price impact as well as measures of trade size, the number of trades and turnover have been stable or have improved since 2010. Trading activity changes with the age, size and issuers of provincial bonds, yet measures for every subcategory have improved or been stable since 2010. One exception stands out: the trading activity of Alberta bonds has grown considerably in recent years as a result of the large increase in issuance after the oil price shock. We interpret the evidence as a sign of resilience in provincial bond markets.

Appendix

Data

We restrict our analysis to Canadian-dollar-denominated bonds directly issued by provincial governments that have traded at least once in the secondary market. Hence, we exclude bonds issued by provincial agencies, or by other entities but guaranteed by provincial governments. To exclude retail trades, we include only those trades with a value of at least $100,000. These trades account for 27 per cent of the total number of trades but less than 0.2 per cent of trading volume.

Amihud’s illiquidity ratio

We calculate the ratio as follows: $$Amihud_{i,d}=Average\left(\frac{|r_{i,t}|}{DVOL_{i,t}}\right),$$ where \(r_{i,t}\) and \(DVOL_{i,t}\) are the return and dollar value of trading volume for bond \(i\) at time \(t\), respectively. The transaction-level illiquidity ratio is averaged over a day d to find the daily illiquidity ratio. The higher the illiquidity ratio of a bond, the less liquid the asset is.

Roll’s effective spread

We calculate the effective spread as follows: $$Roll_{i,t}=2\sqrt{-Cov(r_{i,t},r_{i,t-1})},\,\,if\,Cov(r_{i,t},r_{i,t-1})<0,$$ where \(r_{i,t}\) is the return of bond \(i\) at time \(t\) computed from consecutive trades.

Trading activity measures

The number of trades is the number of individual trades conducted in a month. Average trade size is the average dollar amount for individual trades over a month. For a given bond, the monthly turnover ratio is the ratio of the total volume of trades conducted during that month (number of trades multiplied by the value of trades) scaled by the bond’s total outstanding value. A higher number of trades, greater trade size and a higher turnover ratio indicates increased trading activity. $$Turnover_{i,t}=\frac{Trading\,volume_{i,t}}{Outstanding_{i,t}},$$ where \(Trading\,volume_{i,t}\) is the product of the number of trades and average trade size, and \(Outstanding_{i,t}\) is the average outstanding amount for month \(t\).

References

1. Amihud, Y. 2002. “Illiquidity and Stock Returns: Cross-Section and Time-Series Effects.” Journal of Financial Markets 5 (1): 31–56.
2. Bank of Canada. 2016. “Canadian Fixed Income Forum: CFIF Survey Results on Liquidity, Transparency and Market Access in Canadian Fixed Income Markets.”
3. Chen, L., D. Lesmond and J. Wei. 2007. “Corporate Yield Spreads and Bond Liquidity.” Journal of Finance 62 (1): 119–149.
4. Dick-Nielsen, J., P. Feldhütter and D. Lando. 2012. “Corporate Bond Liquidity Before and After the Onset of the Subprime Crisis.” Journal of Financial Economics 103 (3): 471–492.
5. Fontaine, J. S., J. Gao, J. Sandhu and K. Wu. 2017. “Do Liquidity Proxies Measure Liquidity in Canadian Bond Markets?” Bank of Canada Staff Analytical Note No. 2017-23.
6. Gungor, S. and J. Yang. 2017. “Has Liquidity in Canadian Government Bond Markets Deteriorated?” Bank of Canada Staff Analytical Note No. 2017-10.
7. Roll, R. 1984. “A Simple Implicit Measure of the Effective Bid-Ask Spread in an Efficient Market.” Journal of Finance 39 (4): 1127–1139.
8. Statistics Canada. 2018. “Bonds outstanding, shown at par value, by currency of payments and issuers, Bank of Canada,” Table: 10-10-0130-01. Available at https://www150.statcan.gc.ca/t1/tbl1/en/cv.action?pid=1010013001.