The popular stigma attaching to junk bonds was also reinforced by a widely circulated academic study. In the Journal of Finance, Paul Asquith and David Mullins, both then at Harvard, contended that past research, by systematically underestimating the true default risk of high-yield bonds, had painted an unduly rosy picture of the market.

At the time, I was convinced that the market’s problems were temporary. In a 1990 article, I wrote: “The system needs to be “cleansed” of the excesses of the past few years. The next wave of junk bond issues – and there will almost certainly be one (although whether the issuers will be publicly or privately placed is not at all clear) – will reflect more conservative capital structures and financing strategies. Prices of leveraged transactions will come down and the proportion of equity underlying such levels will rise.)”

In the 10 years since, many of those predictions have become reality. The default rates noted by Asquith and Mullins did not presage a permanent increase. Instead, default rates averaged less than 2% per year from 1992-1998. Perhaps most important, during the 1990s – and, indeed, over the entire 25-year life of the modern high-yield market – investors have essentially gotten what they bargained for. They have earned a rate of return that, at roughly 200 basis points over the return on 10-year Treasuries, is commensurate with junk bonds’ intermediate level of risk – higher than that of investment grade bonds, but lower than that of common stocks.

Some of the same phenomena that we observed in 1989-1990 have again surfaced in 1999-2000. Default rates soared to 4.15% in 1999, with a record $23.5 billion in bonds falling into default. And last year, borrowers defaulted on another $30.2 billion (Figure 1), a second consecutive record year of defaults. With defaults rising, investors’ required yield spread over Treasuries rose to 944 basis points as of December 31, 2000. By one metric – the return on high-yield bonds relative to Treasury note returns – the year 2000 was the worst performing year in the history of the market. The benign credit cycle of 1993-1998, when the default rate was below 2.0% each year, has clearly given way to a more turbulent and stormy environment.

So, will the next 18 months turn out to be as difficult and tumultuous as 1990-1991? Will default rates rise to approximately 10% as they did in both 1990 and 1991? And, will the market almost cease to function? Or, will returns rebound to almost the unbelievable annual level of over 40%, as they did in 1991?

In the early 1990s, the high-yield bond market fell dramatically, and then rose even more dramatically. In both 1990 and 1991, default rates exceeded 10.0% of the market – much larger than the previous high of 5.8% in 1987. The total amount of debt defaulting in each of these two years was over $18 billion. The pundits who predicted the demise of the market looked like sages when total returns to high-yield investors turned out to be -8.5% in 1990 – only the second year since 1978 that total returns were negative. And since Treasuries earned a positive 6.9% return that year, the return spread of junk bonds was a shocking –15.4%. At the end of 1990, the average historical annual return (starting from 1978, when the data were first compiled) to high-yield investors fell to 9.96% per year, and the return spread was a mere 0.19% per year. Clearly, this was inadequate compensation for the added risk of high-yield bonds.

Then came the turning point in 1991. Despite a second consecutive year of a default rate over 10%, high-yield investors earned a total return of 43.2%, the highest ever recorded in the history of the market (Figure 2). Investors realized that the worst was over and that the excesses of the 1980s had been purged. What remained were, for the most part, viable companies whose bonds were not going to default. As the operating performance of these companies continued to improve, the prices of their bonds made a spectacular recovery. The relationship between default rates and total returns is shown in Figure 3.

There is a striking parallel between the increasing default rates in 1989-1990 and 1999-2000; note the dip in returns in 1990 and 2000 and finally the resurgence in 1991. The question remains whether there will be a comparable resurgence in 2001.

Despite the high returns in 1991, however, the high-yield market shrunk rapidly, from a high of $189 billion in 1989 to $163 billion in the middle of 1992. Between 1987 and 1989, an average of $30 billion in high-yield bonds were sold each year. But in 1990 and 1991, the volume of new issues was $1.4 billion, and $10 billion, respectively.

Since 1991, however, the growth in new issues has been nothing short of spectacular, with over $100 billion of new issues in each of the last three years of the decade, and $45 billion in 2000. Between 1997 and 1999, new issuance of high-yield bonds accounted for over half of the bonds issued by industrial companies. At the end of the decade, about $600 billion of high- yield bonds were outstanding, as compared to under $200 billion at the start of the decade. This $600 billion represents roughly a third of the entire corporate bond market in the U.S.

During the 1990s, the annual return spreads between junk bonds and Treasuries rose from near zero at the end of 1990 to almost 2% per year last year. As reported in Figure 2, total compound annual returns on high-yield bonds for the 23-year period from 1978 through 2000 averaged 1.9% per year over the returns of 10-year U.S. Treasuries. This means that a $1,000 investment in high-yield bonds in 1978 would have been worth over $10,400 at the end of 2000, as compared to just $7,000 for 10-year Treasuries. And if one subtracts the average annual losses from defaults of about 2.45% per year over the period 1978-2000 from the average promised yield spread (4.76%) over that same period, the result (2.31%) is quite close to the realized annual return spread. Thus, one can attempt to predict future relative returns in the high-yield market by comparing current yield spreads to actual losses from the primary risk component – defaults.

Deteriorating Credit Quality

But even as issuance of high-yield debt continued to soar in recent years, there were signs of trouble. The default rate in 1999 spiked sharply to 4.15% from 1.6% in 1998 – the first time the rate topped 4% since 1991. One of several apparent reasons for the increase in defaults in 1999 was the seeming deterioration in credit quality of newly issued bonds. Over one fourth of the 125 issues that defaulted in 1999 had been outstanding less than 12 months before they defaulted – and 55% had been outstanding less than 24 months. These percentages compare with just 4% and 20% from the period 1991-1998, and 7.7% and 24.3% for 1971-1999. In 2000, that proportion of defaulted debt that had defaulted within two years of issuance dropped to 38%. Still, as much as 69% of the non-performing high-yield bonds had been issued within the previous 36 months.

To better understand these mortality statistics, however, it is important to analyze the purpose of the financing. Whether companies are using junk bonds to fund LBOs (as they did in great numbers in the late 1980s, but not in the late 1990s), growth opportunities, or just to refinance debt (as they have done in most years), can tell us a good deal about whether these one- or two-year mortality results are truly symptomatic of a decline in credit quality or can be explained by other factors.

There was, in fact, a decline in credit quality between 1997 and 1999. As mentioned earlier, high-yield new issu-ance as a percentage of all corporate bond issuance increased dramatically over the same three-year period. And within the high-yield sector, the percentage of new issues rated B and CCC also increased. Indeed, in 1999, B rated bonds comprised 66% of high-yield issuance and 31% of all new corporate bond issuance! CCC-rated bonds were particularly evident in the 1998 cohort, with $9.3 billion representing 10% of all high yield issuance – a jump from previous years.

At the start of 2000, I said that investors would likely require additional promised yields to compensate them for the uncertainty about possible higher default rates in the next few years. The large spike in yield spreads in 2000 – 447 basis points – seemed to bear out my predictions. (Actually, the default risk spike was only 315 basis points, since Treasuries declined by 132 basis points in 2000.)

Other Reasons for the Increase in Defaults

In addition to the deterioration in credit quality and the earlier occurrence of defaults, four other factors contributed to the sizeable increase in 1999 and 2000: (1) the recent increase in new issuance; (2) the Russian default in 1998; (3) a number of “sick” industries despite the economy’s overall strength; and (4) banks’ reluctance to refinance or give additional waivers to the marginal firm.

Because of the huge new issuance during 1997-1999, some increase in default rates is expected as these new issues age. In the absence of any other developments, two simple principles known as “regression to the mean” and the mortality or “aging” effect would have led us to expect both the default amounts and the default rate in the last two years to increase vis-à-vis prior years.

But the surge in the default rate to over 4% in 1999, and over 5% in 2000, was caused by additional factors. One important consideration, though difficult to document with statistics, is the ability of distressed firms to refinance their indebtedness. Borrowers found it increasingly difficult to refinance in the aftermath of Russia’s 1998 default and the flight-to-quality that ensued. Without the Russian contagion, the default rate would surely have been lower.

In recent years, there were notable concentrations of defaults in a number of chronically or newly ailing industrial sectors. Such sectors as energy, retailing, communications, healthcare, leisure/entertainment, and shipping were hit hardest. In 2000, newly hard-hit sectors have been telecommunications, steel, and movie theatre complex companies, as well as some large asbestos-related companies.

The energy sector’s difficulties reached their peak fairly early in 1999, while retailing and textiles have long experienced chronic problems. Industries such as communications and healthcare became new “leaders” in defaults, reflecting the frenetic new issuance in the former and the overcapacity and governmental regulation of fees in the latter. In sum, despite a vigorous economy driven by technology and productivity growth, a number of sectors have been ailing, and going forward some will continue to flounder.

Finally, there is anecdotal evidence of an increasing trend of banks and other lenders who are no longer willing to waive violations of covenants after just a few prior violations. There appears to be pressure from the Federal Reserve Board in the last two years for banks to record higher loss reserves and actual charge offs when bank profits are at high levels. Coupled with some indications of a slowing of the United States economy and higher interest rates, on lower quality issues, these factors are acting to increase the likelihood of defaults on bank loans and publicly held bonds. The Fed, however, seems to have finally let up on this pressure by lowering interests rates by 50 basis point in January 2001 and again by 50 in mid-March.

The Difference Between Then and Now

Although storm clouds hang over today’s high-yield market, the current situation differs in a number of important respects from a decade ago. Viewed from a purely statistical standpoint, 10% default rates in the near future are certainly possible, but not likely. Statistical analysis would suggest there is something like a 2.5% probability of default rates returning to their 1990 and 1991 highs.

The market, however, is not anticipating such a dire scenario, since yield spreads were 9.44% as of December 31, 2000, as compared to over 10.5% at the end of 1990. It is also important to recognize that a high percentage of those distressed issues in 1990 were the result of LBOs and other highly leveraged transactions (HLTs). Although HLTs made a strong comeback in the ’90s, they are far more conservatively financed today than their ’80s counterparts. Defaults from highly leverage restructurings in 1999-2000 did not account for any material amount of defaults. And the outlook is for this source to continue to be less important. I examined the proportion of total new high- yield bonds issued for a number of stated reasons, including acquisitions, leverage restructurings (e.g., LBOs), capital expenditures and other general corporate investments, and the refinancing of existing debt, between 1986 and 1999. And while the latter category has been the most important use of new debt financing every year since 1986, the levels of refinancing in 1997-1999 are not exceptionally high – about 44%. That’s below the average over this 14-year period. One reason for this is that, although Treasury rates did fall in these years from 1996 levels, the yields on high-yield debt actually increased, making refinancing more expensive.

Overall, in recent years about 20% of high-yield bond new issuance was used for acquisitions and only 4-5% for leveraged restructurings. This compares to 10-15% for acquisitions and well over 30% for LBOs and recapitalizations in the years leading up to the market’s problems a decade ago. Since leveraged restructurings can lead to unsustainable levels of debt and possible financial distress, the new issue market was decidedly more risky in the earlier period.

One of the important similarities between 1990 and 2000, however, is the proportion of the market that is distressed. If we define distressed bonds as those with a yield-to-worst at least 10% (1000 basis points) above the risk free rate, 28% of high-yield bond issues were in this precarious position at the end of 1990, as compared to about 30% as of December 31, 2000.

In Figure 4, we see that in early 1990 the proportion of distressed and defaulted bonds was 41%, with 28% distressed (the total market includes defaulted bonds in this case). A bit more than one-third of the 28% actually defaulted in each of the years 1990 and 1991. At the end of 1999 the distressed proportion was 9% of the market, and it grew to 17% in mid- 2000, and to 31% by the end of 2000, admittedly a dramatic increase. A good deal of this increase, however, is due to the big decrease in treasury yields (Figure 2). If one-third of the distressed proportion again defaults in the next 12 months, we will have a 10% default rate. Incidentally, 10% of a market that is over $600 billion works out to a default total of over $60 billion for the next 12 months. I do not believe this will occur, however, even with the sudden crisis in California utilities and a renewed scare of asbestos-related bankruptcies.

I believe that the default rate will be in the 6.5-7.5% range over the next 12 months and will not reach the higher levels that Moody’s and some other analysts are forecasting. And, I am persuaded by the past and by the market’s dynamics, that returns will be substantial after the peak of defaults, and perhaps even before the peak – whenever it occurs. Indeed, in the first two months of 2001, returns have been over 8.0% on high-yield bonds.

A Word on the Economy

The relationship between overall economic activity and default rates has always been tricky. Clearly, depressed economic growth and declining corporate profits and cash flows are related, in a negative sense, to default rates. But, the lead-lag relationship is not very stable over time. Still, the economic recession at the start of the 1990s clearly was an additional factor that helped push default rates to double-digit levels. Despite a slowdown in economic growth, few economists are forecasting a recession in the next year or two. And, with the recent Fed interest rate cuts, we do not foresee the same economic pressures on default rates in 2000 and 2001 as occurred a decade ago. Admittedly, there is great uncertainty today.

Déjà vu All Over Again?

So is it déjà vu all over again? Yes and no.

Despite the apparent similarities, there are sufficient differences between 1990 and 2000. As a result, the current market downturn will be less severe and less dramatic than its 1989-1990 predecessor. The high-yield bond market will weather this downturn, just as it did in the early 1990s. The present deterioration in credit quality will run its course, as investors refuse to continue providing capital to undercapitalized entities. New issue activities will no doubt fall off, and defaults will probably continue at levels that, although unsettling, are not catastrophic. Just so, subsequent returns will be impressive, although not of 1991’s magnitude. Indeed, in the early weeks of 2001, new issues of high-yield bonds surged as interest rates were lowered. A key question will be if this increase can be sustained. No doubt, the market will be hit by periodic setbacks, and more bad news in the coming months.

But, as long as the vast majority of issuing entities in the high-yield market remain viable enterprises, the market for high-yield bonds will retain its position as an important major source of finance for companies worldwide and a legitimate and profitable asset class for investors.

Edward I. Altman is the Max L. Heine professor of finance at NYU Stern and a consultant to Salomon Smith Barney in the high-yield and distressed debt areas.