By Lang Gibson, Director of Research, Structured Credit Products, Banc of America Securities
The year 2001 will go down as the first real test period for structured credit products, particularly those backed by corporates. Although long-term interest rates ended the year essentially unchanged, credit risk was seriously tested in all markets - especially in the post-Sept.11 period. Nevertheless, by the end of the year, spread levels in the most liquid credit markets were significantly tighter than they were at the beginning of the year discounting an economic recovery for the nine-month old recession.
However, primary CDO spreads lagged the post-Sept. 11 spread widening relative to other fixed income markets and they still remain at their widest levels reached in 2001 - particularly in sub-triple-A tranches - presenting attractive opportunities to increase allocations to CDOs, in our opinion. Heavy visible CDO volume in 2001 of $174.5 billion drove much of the year-end new issue spread widening and should remain a determining factor for primary CDO spread performance in the year to come.
However, in the wake of record CDO downgrades and corporate credit defaults in 2001, we cannot ignore the impact on pricing in the new issue market. Fortunately, most CDOs issued in 2001 were invested in cheap collateral, providing a cushion of excess spread for the high defaults and downgrades expected to plague the markets even after the current recession has officially run its course. In the present environment - which has CDO managers increasing collateral diversity in order to combat credit event risk - the quality and track record of the manager is becoming increasingly important. Therefore, although static deals will remain a hallmark of the synthetic market, we expect to see substantial growth in managed synthetics in 2002.
usurped cash by 30%
Although we believe that real supply was well over $200 billion in 2001, we have tracked visible supply of $174.5 billion (Figure 1). Visible synthetic CDO issuance of $98.5 billion surpassed cash CDO issuance of $76 billion by 30%. This figure is most likely understated, as a few dealers still choose not to report their synthetic CDOs.
We estimate that this overage would actually be closer to 75% if we had the statistics on all CDO issuance. In real supply terms, therefore, we estimate that synthetic CDO issuance reached $133 billion - 75% more than cash CDOs. This $209 billion estimate for real supply compares favorably to $188 billion in 2000 and $163 billion in 1999 (see Figure 2). Please see the Appendix for a summary of all deals.
Arbitrage CDOs accounted for 73% of all cash CDOs in 2001. With a 19% market share, cash balance sheet CLOs accounted for the majority of the remainder. Finally, market value, TruPS and EM CDOs accounted for the balance (8%). In the cash arbitrage CDO category, HY and IG collateral accounted for 55% and 45% of market share, respectively, reflecting the increasing attraction of IG collateral over the course of 2001, as more investors sought to avoid the high level of defaults in the high yield universe. Total deal count for cash CDOs was 169 in 2001.
HY loans accounted for 17% of all cash CDOs versus 23% for HY bonds, primarily due to the latter's smaller size relative to previous years and leveraged loan sourcing difficulties during the ramp-up phase. If it hadn't been for the sourcing problem, we would have seen substantially more HY loan CDO volume, because loan deals have been increasingly outperforming bond deals. Furthermore, in the first quarter of 2002, we have already seen evidence of investors aggressively seeking HY loan deals for this very reason. In 2001, the increasing attraction of structured product as a collateral class for CDOs allowed multi-sector issuance to match HY bonds at 23%. Despite the absence of any significant excess spread in the cash market for much of the year, IG corporates accounted for 10% of cash CDO volume, adding 19 deals to the pot.
reached $98.5 billion
Balance sheet and portfolio default swaps accounted for 47% and 42%, respectively, of visible synthetic CDO volume, with the remaining 11% accounted for by independently managed synthetics. However, if we exclude the one mega balance sheet CLO of the year (the $11.25 billion Amstel CLO), portfolio default swaps would have accounted for the majority of 2001 synthetic volume by an eight percentage point margin.
Also, in 2001, traditional CDO managers increased their use of synthetic technology for leveraging their investments in HY loans and IG corporates. At least 17 independent managers repackaged $10.6 billion of collateral into synthetic CDOs. Of this total, approximately $6.3 billion was accounted for by IG corporates in six deals, which averaged over $1 billion each. Similar to cash CLOs, synthetic CLOs averaged slightly more than $400 million each. For 2002, we expect the largest spurt in synthetic growth to occur in the managed category as more CDO managers jump on the bandwagon in an attempt to achieve advantageous economics in the credit derivatives market.
dominated 2001 CDOs
Investment-grade debt represented the most popular CDO collateral in 2001. Pure IG collateral was represented in 48% (28% IG corporate plus 20% multi-sector) of all cash and synthetic CDOs (see Figure 4). Additionally, approximately half of the collateral in the IG/HY loan category - which represents 28% of all CDOs - was investment grade. Only 20% of collateral was represented by HY bonds and leveraged loans combined. It was the combination of balance sheet CLOs (cash & synthetic), portfolio default swaps and the popularity of multi-sector deals that drove the predominance of IG collateral in CDOs for 2001.
CDOs have come to represent an increasingly larger share of the underlying collateral markets. In no sector was this more apparent than in the leveraged loan sector. In 2001, newly issued HY loans CLOs - both cash and synthetic - represented exactly 50% of all institutional leveraged loan volume (see Figure 6 above). The share of CDOs in other sectors is growing as well. Newly issued HY bond CDOs accounted for 23% of all HY bond issuance in 2001. In IG sectors, CDOs account for a much lower proportion, primarily because only the highest yielding bonds in these sectors provide a sufficient funding gap to achieve the requisite economics in a CDO. Both newly issued multi-sector and IG corporates accounted for 9% each of their respective underlying collateral classes.
$220 BN expected in 2002
In 2002, we expect the following trends in CDO issuance:
Visible volume for all CDOs in excess of $220 billion - a minimum increase of 26% from 2001 levels, with synthetics enjoying most of the growth and more dealers beginning to report their synthetic deals.
In cash CDOs, a preference for HY loan collateral, which has significantly outperformed HY bonds in the past (on a risk-adjusted basis) and has substantially more subordination than IG collateral.
In IG CDOs, an increasing preference for highly rated structured product, which requires synthetic technology to achieve the requisite economics. Sourcing risk will become an increasingly larger challenge for cash multi-sector deals that purchase lower rated structured products, as we expect consumer credit to deteriorate in the face of a rising unemployment rate over 2002.
Increasing use of synthetic technology by traditional CDO managers with credit derivatives expertise across asset classes (e.g., substantial growth in independently managed synthetics).
Continued dominance by European banks in securitizing both non-corporate and higher yielding corporate assets in balance sheet CLOs with more of an emphasis on economic capital relief as opposed to regulatory.
Tranched portfolio default swap activity to increasingly include high yield credits and structured product as the reference portfolio.
Even more tiering between CDO managers than was seen in 2001, as CDO performance data becomes more "publicly" available.
Sticky New Issue Pricing
Until Sept. 11, the year 2001 was generally characterized by a slight spread tightening in new issue CDO tranches. After the attacks, there was a delayed period before primary market transparency re-appeared. Significant spread widening occurred only in the sub-triple-A tranches, particularly in the As and below (Figure 2).
Relative to the corporate bond market, CDOs experienced substantially less spread widening in response to Sept. 11. The net result was that CDO funding gaps initially widened substantially post-attacks, as they were driven by widening asset spreads and relatively tight funding costs, which particularly benefited from exceptionally tight triple-A spreads. Since 70-90% of the capital structure consists of triple-As, the weighted average funding cost for a CDO only mildly increased over the year.
Double-As widened 520 basis points, with the largest widening occurring in the multi-sector segment, for which double-As priced at L+95 by year-end versus L+80 to L+85 in other sectors. In the single-A category, IG corporates and multi-sectors widened approximately 30 basis points over the year to come into line with HY sectors at approximately L+150. Finally, in triple-Bs, every sector widened 60 basis points with the exception of HY loans, which went out by only 46 basis points. Multi-sector CDOs had the widest pricing in triple-Bs by year-end at L+310, while HY loans had the tightest at L+265. This 45 basis points spread differential reflects investors' concern with the relatively low subordination for structured products, which will likely be tested this year if consumer credit deteriorates as we are expecting.
The relative value advantage of new issue CDOs can best be exemplified by the pickup to same-rated corporates over the year (Figure 3). The Fed induced corporate spread tightening at the beginning of the year resulted in a relative value pickup in CDOs that endured until Sept.11. After Sept. 11, corporate spreads widened substantially, taking back all of the relative value advantage for CDOs - and more, particularly in high yield. Then, in the last few months of 2001, corporate spreads tightened substantially to end much lower than they were at the start of the year. Meanwhile, new issue CDO spreads remained at their wides by the end of 2001. Consequently, in comparison to same-rated corporates, CDOs now present attractive relative value advantages that have not been seen in a long while. For instance, by the end of the year, primary CLO spreads were 162 basis point wider in triple Bs and 108 basis points wider in single As. In double-Bs, the CDO pick-up was 373 basis points.
On account of the recent divergence between corporates and CDOs, we recommend an overweight position in the new issue CDO sector versus corporates. Nevertheless, it is important to bear in mind the specifics of any new issue CDO, such as the manager and initial portfolio yield. For instance, to the extent the CDO ramped up over a period when corporate spreads were tight, there is much less of a cushion to support defaults in the deal. Thus, we recommend that investors use tools such as our CDO ROE Barometers, as well as intense deal and manager scrutiny to estimate the true funding gap during a deal's ramp-up period.
The year 2001 proved to be a critical testing ground for CDOs. Despite corporate defaults and rating migrations approaching levels not seen in a decade, the new issue market continued to show substantial growth, particularly in synthetics. Admittedly, much of the slack given up by HY CDOs was replaced by CDOs backed by IG credits. As the cash and synthetic markets become increasingly integrated both on the CDO and single-name front, we see no end in sight for the growth of structured credit products.
Note: For more information and relevant figures, please refer to BofA's Jan. 16 report, "2001 CDO Review & 2002 Outlook."