November 29, 2016
The last few years have proven to be momentous on the US Investment Grade new issue calendar, and 2016 is positioning likely to be the biggest yet. The latest data through September 30 indicates that this market is on track for $1.7T in new issue volume, which would make it the 7th consecutive year of growth and double that of 2010’s $850B new issue volume.
Needless to say, this cumulative $8.5T new issue volume has pushed the investment grade market to its limits in several ways:
- Total amount outstanding has recently hit $8.4T, an all-time high.
- Number of unique CUSIPs traded every month is more than 15,600, an all-time high.
- Market turnover has fallen to 66%, a clear but modest uptick from its all-time low.
Identifying New Issues
Measuring new issuance – primary market activity – can be a challenge. There is no centralized listing venue or other systematic means to track which bonds are available for trading. The most common methods are labor-intensive, and involve tracking SEC filings and speaking directly with underwriters and capital markets desks.
However, Finra’s TRACE provides unfettered insight into secondary market activity. This permits the market to observe the first trade report of a new issue, and use that as its issue date. For the vast majority of securities, this approach is accurate and appropriate. On the margin, though, it is not perfect. For example, it will miss those new issues which do not trade in the secondary market (meaning they never appear on TRACE); or it may assign an incorrect issue date (usually by no more than a few days) to those new issues which have TRACE-reported grey market activity.
New Issues on TRACE
The new issuance calendar is often held accountable for “crowding out” secondary market trading in other (more seasoned) issues. After all, new issues are often heavily marketed, priced at a discount, and experience significant turnover. These help attract a lot of attention from investors and dealers, leaving less mindshare for everything else.
Of course, we also know that, despite seeing relative heavy activity in their first few days of trading, new issue activity has a short half-life and the bonds often quickly settle into a much lower level of trading.
How do these forces balance themselves today and historically? We can look at the “new issue market share” (or “TRACE share” for short) by examining TRACE activity in new issues vs. total TRACE activity.
We can do this by looking at the TRACE share using a definition applied to three different time periods: (a) volume in the first 5 trading days of a new issue; (b) volume in the first 20 trading days of a new issue, or roughly 4 weeks; and (c) volume in the first 90 calendar days of a new issue, or roughly 3 months.
This data will be aggregated monthly. Meaning, we aggregate the daily volume meeting the above definition in a given month, and divide by the total monthly TRACE volume.
Let’s walk through a simplified example using the first 5 trading day method. On the 1st trading day of the month, we’ll look at the total volume in all bonds that are within 5 trading days of their first appearance on TRACE. Some of those bonds may have been issued that same day, some the prior day, and so on. On the 2nd trading day of the month, we perform the same calculation using bonds that are within 5 trading days of their first appearance on TRACE. This set of bonds is a slightly different set than the 1st day of the month. We take the sum of each day’s new issue volume for each trading day of the month, and divided by the total TRACE volume for the month.
Analogous mechanics apply to the 20 trading day and 90 calendar day methods.
Results for First 5 Trading Days
If we look at the new issue share using the first 5 trading day method, we can see that 5-10% of monthly TRACE volume can generally be attributed to new issues with an overall monthly average of 7.2%. The low month of 2.9% occurred in Apr 2010, and the high month of 12.3% occurred in Sep 2013. While 2012 had the highest 12-month average at 8.5%, 2016 is shaping up to be the second most active at 7.9% so far. The busiest months of the year tend to be March, May, September and November.
Of course, this variation is a function of the size of the new issue calendar in a particular month. Given that a very small number of bonds on any given day meet the new issue definition, though, their average turnover is clearly very high.
Results for First 20 Trading Days
If we look at the new issue share using the first 20 trading day method, we can see that 10-15% of monthly TRACE volume can generally be attributed to new issues with an overall monthly average of 12.6%. That’s nearly double the 5 day method but it would include 4 times the number of bonds.
Not surprisingly, the pattern of activity mirrors that found in the 5 day method. The low month of 6.2% occurred in Apr 2010, and the high month of 19.3% occurred in Sep 2013. While 2012 had the highest 12-month average at 14.1%, 2016 is shaping up to be the second most active at 14.0% so far. The busiest months of the year tend to be March, May, September and November.
Again, we suspect that this variation is a function of the size of the new issue calendar in a particular month.
Results for First 90 Calendar Days
If we look at the new issue share using the first 90 calendar day method, we can see a range of 18-24% of monthly TRACE volume that can generally be attributed to new issues with an overall monthly average of 20.7%. Of course, this includes roughly 3 times the bonds of the 20 trading day method, and roughly 1/4 of all bonds issued in a rolling 12 month period. Since there are 300-400 new issues (individual CUSIPs) per month, as much as a 25% of TRACE volume can flow to those 1000 CUSIPs, leaving 75% of TRACE for all other outstanding issues.