Sizing Up the Underwriters: A Five-Year Perspective on Underwriter Performance
Click to Print This Page
Portfolio managers and other investors shopping in the high-yield new-issue market have long been in need of a practical tool to facilitate the process of evaluating and selecting securities. When assessing the attraction of new offerings, portfolio managers will consider such factors as credit quality, ratings, covenants, call features, and spread. For borderline cases, however, managers should take note of the underwriter’s record in structuring and pricing deals that hold up well in the aftermarket.
The performance-based league tables published here are a tool for just such an eventuality. They rank underwriters by aftermarket performance of their deals, rather than dollar amount underwritten, with full credit given to the lead underwriter of each high-yield new issue. Each issue priced as part of a multitranche offering is treated as a separate observation.
The tables can also be used to assess the possible risk of excessive concentration in deals underwritten by a particular firm. The concentration cutoff for a firm with a good underwriting record might be higher than for one with a poor underwriting record. While investors should not attribute undue importance to minor differences in underwriters’ rankings, the average change in spread reported for the top- and bottom-ranked underwriters is statistically significant at a confidence level of no less than 86% and as high as 98%.
Drawing on new-issue data from 20 quarters (Q1 2003 through Q4 2007), this five-year perspective examines the long-term consistency of underwriters’ performance records and identifies perennial champions and laggards. Readers should note that the full five-year study period comprises a modest bull market for fixed income. In this environment, lower-rated credits such as Triple-C issues tend to benefit disproportionately from tailwinds, whereas issues at the upper end of the quality spectrum tend to underperform the overall high-yield index, as measured by total-return performance.
For example, in 2003, when the Merrill Lynch Master II Index posted its second-highest return since its inception in 1986 (28.15%), the Double-B subindex returned 19.45%, compared to 26.01% for Single-Bs and 60.99% for Triple-Cs. While Triple-Cs constituted less than 20% of total issues in 2003, their stellar performance would have biased this study to benefit underwriters whose portfolios of deals were overconcentrated in lower-tier issues. Conversely, the opposite would have been true in 2005 and 2007, when the High-Yield Master II Index posted below-average returns of 2.74% and 2.19%, respectively, and Double-Bs outperformed Triple-Cs by a significant margin. An overview of annual returns of the different rating-category subindexes is presented in Exhibit 1.
Exhibit 2 ranks underwriting firms by average tightening (-) or widening versus the Lehman Brothers US Corporate High-Yield Index in the first week following issuance. This measure reflects underwriters’ effectiveness in building good books for their deals.
If an investment bank regularly places bonds in weak hands, selling by “flippers” will cause the firm’s offerings to widen after they are freed to trade. Deals that tighten after issuance enhance the performance of portfolio managers relative to their benchmarks. Ranked underwriters led a minimum of 50 deals over the 2003–2007 time period.
NEW ISSUE PRICE PERFORMANCE
On average, leader Jefferies’ deals tightened by 18.7 basis points versus the Lehman Brothers US Corporate High-Yield Index. That compares with an average tightening of 10.4 basis points for all deals, 16.7 basis points for runner-up Merrill Lynch, and 16.5 basis points for third-place finisher Goldman Sachs.
Jefferies also posted above-average consistency of performance, as indicated by both its standard deviation of 28.7 basis points and semideviation of only 8.8 basis points, which came in below the group averages of 29.2 and 13.9 basis points, respectively. (Semideviation measures the standard deviation of yield widening versus the Lehman index, thereby analyzing downside risk only.)
A MAP OF THIS STUDY
The Lehman Brothers US Corporate High-Yield Index covers the universe of fixed-rate, noninvestment-grade debt with at least one year to final maturity regardless of call features and at least $150 million par amount outstanding. PIK (pay-in-kind) bonds, Eurobonds, and debt issues from countries designated as emerging markets are excluded. However, SEC-registered Canadian and global bonds of issuers in nonemerging markets are included, as are original issue zeroes, step-up coupon structures, and 144-As. With approximately 1,500 constituent issues, the Lehman Brothers US Corporate High-Yield Index serves as a proxy of the overall high-yield universe.
The “New Issue Price Performance” section of this study rates underwriters by the aftermarket performance of their 2003–2007 deals, as measured by average tightening versus the Lehman Brothers US Corporate High-Yield Index one week and four weeks after issuance. In the one-week category, Jefferies leads its peers, followed by Merrill Lynch in second place and Goldman Sachs in the third-place slot. In the four-week category, Goldman Sachs tops the list ahead of Credit Suisse and Merrill Lynch.
The “Long-Term Trends” section tracks patterns apparent in the trailing-twelve-months rankings and highlights individual underwriters whose deals showed improved or worsened quality or that performed consistently throughout the entire five-year period. Bank of America and Merrill Lynch were among the most improved, while Goldman Sachs and Morgan Stanley showed the greatest consistency. Performance of Deutsche Bank’s and Lehman Brothers’s deals worsened.
”Adjusted Performance Rankings and Consistency of Aftermarket Performance” calculates rankings based on average tightening in the aftermarket, taking into account the number of underwriters that qualified to enter our rankings in each quarter and trailing-twelve-months period. This modification propels Goldman Sachs and Merrill Lynch to the top of both lists.
Also considered is consistency of aftermarket performance, as measured by both standard deviation and semideviation of the average tightening or widening of individual new issues. That perspective propels UBS Investment Bank and Bear Stearns to the top quartile for both the quarterly and trailing-twelve-months rankings. “Discussion and Implications for Investors” summarizes the study’s findings and suggests some ways in which portfolio managers may apply these results to their scrutiny of new-issue securities.
Exhibit 3 focuses on price performance in the first four weeks after issuance. During this longer interval, credit concerns may surface that are not visible at the time of underwriting. Goldman Sachs won a decisive victory over runner-up Credit Suisse and third-place finisher Merrill Lynch to take the four-week lead, as the three firms posted average tightening of 16.6, 14.6, and 14.2 basis points, respectively.
As in the one-week performance category, Goldman Sachs claimed the worst-in-group consistency, corresponding with a standard deviation of 58.6 basis points and a semideviation of 59.2 basis points, compared to group averages of 45.7 and 28.9 basis points, respectively. This highlights great variability in the aftermarket performance of Goldman’s deals, ranging from a tightening of 164.4 basis points on one deal to a worst-in-group widening of 226.4 basis points on another.
Credit Suisse also posted below-average consistency, as measured by standard deviation, but its above-average semideviation suggests that average downside risk for deals underwritten by Credit Suisse was limited. Fourth-place Jefferies pulled off the difficult feat of combining above-average narrowing with below-average volatility, as measured by both standard deviation and semideviation, whereas Merrill Lynch and Deutsche Bank achieved above-average narrowing only with average or slightly above-average downside volatility.
Combining the information from the two tables, Credit Suisse, Deutsche Bank, Goldman Sachs, Jefferies, and Merrill Lynch experienced above-average tightening of their deals in both the one-week and four-week performance categories, while Bank of America, Citigroup, JPMorgan, Lehman Brothers, Morgan Stanley, and UBS Investment Bank saw below-average tightening over both time frames.
Credit Suisse and Jefferies combined above-average tightening with below-average downside risk. At the other end of the spectrum, Morgan Stanley earned the dubious distinction of combining below-average tightening with above-average downside risk.
In making their investment decisions, portfolio managers should keep in mind that certain underwriters focus on specific parts of the quality spectrum. For example, for the 2003–2007 study period, both Credit Suisse’s and Jefferies’ deal books were underconcentrated in high- and medium-tier issues but overconcentrated in low-tier issues. The two investment banks thereby benefited from the comparatively strong aftermarket performance of the lowest rating tier. (High-tier issues are defined as rated BB+ or higher on a senior-equivalent basis. Medium-tier issues are rated BB, BB-, or B+. Low-tier issues are rated B or below. Unlike nominal ratings, which reflect both probability of default and expected recovery in the event of a default, senior-equivalent ratings only measure default probability.)
Conversely, Morgan Stanley’s deals were overconcentrated in high-tier issues, underconcentrated in medium-tier issues, and showed approximately average representation of low-tier issues, causing the company to experience headwinds as a result of below-average aftermarket performance of the highest rating tier. Notably, Goldman Sachs managed to achieve above-average tightening despite a distribution of deals that roughly paralleled Morgan Stanley’s.
Illustrating the extent of skewing in the most recent years of our study period, fully 85.7% of Jefferies’ and 72.4% of Credit Suisse’s 2007 deals were rated B and below on a senior-equivalent basis, compared to a group average of 56.0%. For Morgan Stanley, 31.6% of 2006 deals were rated BB+ or above on a senior-equivalent basis, compared to a group average of 11.0%.
On average, deals of all underwriters tightened versus the Lehman Brothers US Corporate High-Yield Index for both the one-week and four-week performance categories, resulting in enhanced performance of portfolio managers who participated in a representative sample of these primary-market deals. An interpretation of the high–low ranges of individual underwriters in conjunction with their median performance figures reveals skewing in the distribution of deal performance toward yield tightening for all underwriters in the one-week performance category, and for all underwriters except Citigroup and UBS Investment Bank in the four-week performance category. Not unexpectedly, greater high–low ranges and standard deviations occur in the four-week performance category than the one-week performance category, indicating greater variability in deal performance over the longer time span.
Exhibit 4 shows a traditional league table for the years 2003 through 2007, with underwriters ranked by number of lead- managed deals. Credit Suisse held the top spot for deal volume with 300, ahead of runner-up Citigroup with 279 and third-place finisher JPMorgan with 259. Morgan Stanley led the group in average deal size with $459.5 million.
This section seeks out discernible trends among underwriter rankings over the five-year period. Quarter-to-quarter rankings show high volatility and can be based on as few as three deals per quarter for which complete pricing information is available. Therefore, the analysis only considers trailing-twelve-months statistics based on a minimum of 10 priced deals.
No trend is strictly monotonic, but certain underwriters have consistently out- or underperformed most of their peers. Additionally, the performance of certain underwriters has clearly improved or worsened over time. Interestingly, for the underwriters whose performances exhibit the most dramatic trends, 2005 or 2006 appears to have been the turnaround year.
Exhibits 5 and 6 depict the rankings of Bank of America and Merrill Lynch, the most improved underwriters. Rankings are based on the performance of their deals four weeks after issuance. (In general, one-week and four-week performance charts of individual underwriters display similar patterns, with few exceptions. The charts used here are those that most clearly support the author’s arguments.)
Among the worst performers in its peer group during 2003 and 2004, Bank of America moved to occupy a solid midfield position beginning in the second quarter of 2005. Merrill Lynch, whose rankings were volatile between the fourth quarter of 2003 and the second quarter of 2005, had moved all the way to the top of the rankings by the fourth quarter of 2006, and achieved the impressive feat of maintaining the top spot for five consecutive quarters.
In contrast, Deutsche Bank and Lehman Brothers exhibited declining rankings within their peer group. Their declines began in the second quarter of 2006, after they had achieved top-half rankings throughout most of 2003–2005 (see Exhibits 7 and 8).
Goldman Sachs achieved consistently strong rankings over the entire five-year period, with few outlying quarters, whereas Citigroup’s and Morgan Stanley’s rankings consistently trailed those of their peers. Goldman Sachs’s and Morgan Stanley’s rankings versus those of their peer group are shown in Exhibits 9 and 10.
ADJUSTED PERFORMANCE RANKINGS AND CONSISTENCY OF AFTERMARKET PERFORMANCE
Calculations of adjusted performance rankings normalize for the size of the peer group in any given quarter or rolling twelve-month period. A minimum of three deals with complete pricing information is required in order to include an underwriter in the quarterly rankings. Therefore, the size of the peer group will fluctuate based on the total number of issues coming to market in any given quarter.
For example, in the third quarter of 2007, reduced LBO (leveraged buyout) activity and market uncertainty associated with the global credit crunch resulted in issuance of only 19 new deals, with Bank of America and Jefferies the only two ranked underwriters. In contrast, in the first quarter of 2007, prolific new-issuance activity saw 137 offerings coming to market, divided between 14 different lead underwriters.
Because it is more difficult to place second in a peer group of fourteen than in a peer group of two, normalization for the size of the peer group is achieved by dividing the ranking of each underwriter in a given quarter or rolling twelve-month period by the size of the total peer group. This yields a number greater than zero and smaller than or equal to one, with numbers closer to zero indicating higher rankings and numbers closer to one indicating lower rankings. Averaging these numbers over the 20 individual quarters and rolling twelve-month periods, for both the one-week-post-issuance and four-weeks-post-issuance categories, produces the figures displayed in Exhibits 11 and 12.
In general, for each individual underwriter the averaged numbers for both the one-week-post-issuance and four-weeks-post-issuance categories lie reasonably close together. This suggests that underwriters are approximately equally effective or ineffective at building good books for their deals and rooting out potential hidden credit concerns prior to underwriting an issue.
For both the quarterly rankings (Exhibit 11) and rolling twelve-month rankings (Exhibit 12), Goldman Sachs occupied the top slot and Merrill Lynch was the runner-up. Deutsche Bank finished third in the quarterly rankings, while Jefferies took third place in the rolling twelve-month rankings.
On a quarterly basis, the top three underwriters’ deals performed better in the aftermarket than the bottom three underwriters’ deals, achieving 97% confidence. On a twelve-month basis, the confidence level was 98%.
An investigation of the consistency of aftermarket performance for the most part upholds the traditional risk–return relationship. That is, underwriters whose new issues exhibit the strongest average aftermarket tightening tend to trail their peers in consistency, as measured by both standard deviation (Exhibit 13) and semideviation (Exhibit 14). For investors, this means that the probability of getting stuck with a badly underperforming new issue is higher in the case of underwriters that top the performance rankings than in the case of underwriters that occupy the midfield or even trail their peers.
Exhibits 13 and 14 show consistency rankings based on a trailing-four-quarters analysis, in which the large underlying sample sizes diminish distortion by outliers and quarter-to-quarter volatility. UBS Investment Bank and Bear Stearns achieved high consistency ratings while trailing most of their peers in average aftermarket tightening. Conversely, Merrill Lynch and Goldman Sachs, which exhibited the highest aftermarket tightening, performed poorly in the consistency category.
Citigroup alone defied the risk–return tradeoff by combining average to below-average returns with above-average risk. Citigroup came in seventh in the adjusted quarterly performance rankings but occupied the bottom quartile in adjusted performance rankings on a rolling twelve-month basis as well as in the rolling twelve-month consistency rankings.
DISCUSSION AND IMPLICATIONS FOR INVESTORS
This analysis of the performance of 2,245 new issues in the aftermarket assigns each deal to a single lead underwriter. Assessing average tightening or widening versus the Lehman Brothers US Corporate High-Yield Index both one week and four weeks following issuance reveals that underwriter rankings differ little between the one-week and four-week categories. New issues tightened by an average of 10.4 basis points over the first week and by an average of 10.5 basis points over the first four weeks, with both distributions skewing toward yield tightening. Above-average aftermarket performance tended to be associated with below-average consistency, and vice versa.
The volatility of underwriter rankings for the 20 rolling twelve-month periods in our study displays no set patterns. Certain underwriters performed consistently at the top or bottom of their peer group (representing perennial good or bad bets). Others improved or worsened over the five-year span of our study. Still others exhibit a great deal of volatility in their rankings.
The absence of patterns in the volatility of rankings may arise from industry specializations or from shifts in preference among underwriters with respect to deal quality. For example, an underwriter that increased its concentration of lower-quality deals over the 2003–2007 period could advance in the performance ratings relative to a peer that made no changes to the quality of its mix of underwritings.
Conversely, a firm specializing in industry groups that went out of favor during our study period might have registered a drop in the rankings due to reduced aftermarket demand. (An investigation of the extent to which sector distribution may drive long-term results could prove interesting.)
As investors migrate toward lower-risk propositions in the new-issue market, the advantage is likely to shift away from underwriters that specialize in low-quality deals (and have benefited from the low-default environment of the past few years, combined with high market demand for new credit issues) to underwriters that specialize in medium- to high-quality deals.
Looking ahead to the end of 2008 and into 2009, the worldwide credit crunch, with its tighter lending standards and reduced LBO activity, will take a toll on the new-issuance market. Corporate borrowers with weaker balance sheets will find it more difficult to float new issues under less-than-punishing terms. Over time, this should result in an improvement in new-issuance quality from its 2007 low of only 24% of new issues rated BB- and higher by senior-equivalent rating. (In fact, the percentage of new issues rated BB- and higher by senior-equivalent rating rose to 40.7% during the first half of 2008 as weaker borrowers got shut out of the market amid heightened investor cautiousness.)
In addition, new-issuance activity and new-issue dollar volume will remain low in the near future in both the United States and Europe. Heightened volatility, depressed and uncertain pricing, reduced liquidity, and investors’ greater sensitivity to risk will all put a premium on insightful credit analysis.
The new-issue market remains an important vehicle for the allocation of capital to the high-yield asset class, as liquidity in the secondary market tends to be limited, particularly for small issues in out-of-favor industries. However, most portfolio managers will consider holding periods beyond four weeks when finalizing their investment decisions. The overall state of the market, credit-quality ratings, and recovery expectations in case of default will be among the major determinants of total-return expectations.
As different underwriters cater to different niches of the new-issue market (characterized, for example, by issue size and quality), portfolio managers may find their new-issue requirements served by only a subset of underwriters. While these managers will generally try to avoid excessive concentration in deals underwritten by a single firm, their investment mandates will probably steer them toward a particular subset of underwriters, or their evaluation of market conditions will cause them to favor one group over another.
–Karen Sterling , PhD, CFA, is vice president and senior analyst at Fridson Investment Advisors, LLC.
Acknowledgments: Four individuals were instrumental in bringing this work to fruition: my colleague, Marty Fridson, of Fridson Investment Advisors; my former colleague, Greg Braylovskij, of Strategic Value Partners, who compiled the early-years data included in this study; my husband, Adam Sterling, who asked the initial question that led to the publication of this paper; and Alvin Kressler, executive director of the New York Society of Security Analysts (publisher of the Investment Professional).
Note: Between the writing of this paper and its publication (in the Fall 2008 issue of the Investment Professional), the institutional landscape of high-yield issuance experienced seismic change. As a result, a portion of the analysis performed herein deals with companies that are no longer extant. However, the research has been published as originally planned in the hope that the resources presented here can be useful not only as forward-looking tools for analyzing league tables when making portfolio selections, but also as retrospective tools that can further an understanding of how we arrived where we stand today. KS.