The current B grade bond spread to US Treasuries is 260 basis points for 10 year bonds, and the average default rate since 1970 has been about 6%, and recovery rates about 50%. So on average, a B grade bond portfolio, with no turnover, should expect to return slightly less than Treasuries (260-300= - 40).
The historical record of high yield bond portfolios is consistent with this observation. Bloomberg’s B rated bond series starts only in 1992 when B spreads were a little above what they are today, around 300 basis points (thus making it a fair comparison start date). Merrill Lynch’s HY Index and Government Bond Index from that date show similar increases, a little over 8% annualized. The annualized standard deviation is also similar (about 7% annualized), as for HY bonds the extra volatility of the spread risk is mitigated by the fact spread movements are inversely correlated with interest rates (spreads go up when the economy is weak and rates go down).
Thus a simple forward looking calculation (260 bp spread vs 300 bp expected loss) or historical performance suggest you do not get rewarded for taking on extra credit risk.
Given the extra risk of HY bonds, why invest in them now? Most people I know say they know that on average HY bonds are a bad trade, but they will get out before the next recession, which currently appears at least a few years away. That kind of reasoning invokes the greater fool theory (I'll sell to some fool before he notices the impending recession). But I imagine that bond yields will spike not in the next recession (which is well off), but at the faintest hint of trouble as holders simultaneously try to cash in on the greater fool theory. In other words, the next HY crack-up will not occur during the next recession or the next financial crisis, but rather the next symptom of an impending symptom of a recession, which will happen much sooner.
The stated yield may be high, but the net returns are not.
The historical record of high yield bond portfolios is consistent with this observation. Bloomberg’s B rated bond series starts only in 1992 when B spreads were a little above what they are today, around 300 basis points (thus making it a fair comparison start date). Merrill Lynch’s HY Index and Government Bond Index from that date show similar increases, a little over 8% annualized. The annualized standard deviation is also similar (about 7% annualized), as for HY bonds the extra volatility of the spread risk is mitigated by the fact spread movements are inversely correlated with interest rates (spreads go up when the economy is weak and rates go down).
Thus a simple forward looking calculation (260 bp spread vs 300 bp expected loss) or historical performance suggest you do not get rewarded for taking on extra credit risk.
Given the extra risk of HY bonds, why invest in them now? Most people I know say they know that on average HY bonds are a bad trade, but they will get out before the next recession, which currently appears at least a few years away. That kind of reasoning invokes the greater fool theory (I'll sell to some fool before he notices the impending recession). But I imagine that bond yields will spike not in the next recession (which is well off), but at the faintest hint of trouble as holders simultaneously try to cash in on the greater fool theory. In other words, the next HY crack-up will not occur during the next recession or the next financial crisis, but rather the next symptom of an impending symptom of a recession, which will happen much sooner.
The stated yield may be high, but the net returns are not.
HedgeFundGuy - am 2005-01-19 21:42 - Rubrik: Finance
stxx meinte am 20. Jan, 00:11:
Where is the mistake in my calculation?
Today's 10y Treasury rate is at 4.18%. Adding a 260 BP spread yields an expected return of 6.78%. Assuming a total default in 6% of all cases the expected return on a HY investment reduces to 6.78%*0.94 = 6.373% which is higher than the expected return of a 10y Note.In your graph the HY bonds were performing better until 2001 when the global economy went into a downturn. The underperformance stopped in mid 2002 and since then the single Bs again outperformed the AAA. Hence, most of the time holding risky bonds yielded better returns than holding riskless bonds.
Are durations of the Gov't bond index and the HY bond index the same? I remember that those HY indexes have significantly smaller durations (up to two year depending on the index) than a JPM EMU benchmark. Such a difference would not adequately reflect the remuneration on taking credit risk.
HedgeFundGuy antwortete am 20. Jan, 00:44:
oops
A 6.78% yield portfolio, with a 6% default rate, and a 50% recovery rate, returns 3.78% (excluding convexity issues). It's 0.94*6.78+.06*-50, not 0.94*6.78.
stxx antwortete am 20. Jan, 01:46:
Ok, I made a pretty bad mistake calculating either 6.78% or 0% performance (assuming a recovery that equals current market price) ;D. Nevertheless, what are the durations of the indexes?
HedgeFundGuy antwortete am 20. Jan, 02:36:
The Gov't index is a blend of Treasuries >5yrs, so I figure it's a similar duration to the Merrill HY Index. as for
Hence, most of the time holding risky bonds yielded better returns than holding riskless bonds.
On average, over the entire 12 years, they were the same. I don't think it's reasonable to 't exclude the bad years because they bring the average down.
stxx antwortete am 20. Jan, 03:47:
First,
... sorry about my unclear formulations. I did not want to emphasize the point that risky bonds are better investments. I just wanted to say that most of the time investors are rewarded for taking risk. But, as it seems in your example, this reward is not of a permanent nature which I absolutely don't understand.I did a quick simulation on the parameters default prob (varying 30% around its mean) and recovery value (15% off by its mean) by introducing iid random variables (... and having no idea about their true distributions). After 1,000 runs I got an expected return of 3.34% with 59% of the returns within a range of 2.8% to 4.2%. Hence, the expected return should be less.
At my internship I had to play around with a few benchmarks. When stretching the duration for half a year to 4.5 years then the performance over the analysed time span (1 year) increased by 40 BP. If this is not the case here then I resign and accept your argument to be valid :D
Please note that I have a bad day doing calculations today.
HedgeFundGuy antwortete am 20. Jan, 15:01:
good idea to put the recovery rate assumption in the text!This sample period included a big decline in rates, so anything that increases duration generates better returns.
In any case, HY seems a poor investment.
finance bob meinte am 20. Jan, 22:57:
About the risk-return trade-off
I´ve got a question about the statement about the risk-return relationship mentioned earlier (...that taking additional risk doesn´t provide any premium).Why should a "risky" investment in the fixed-income market provide a risk premium if the market values the securities efficiently?
In your calculations you have started with the actual YTM of the bonds, then you have taken out the risk factors (6% default and 50% recovery rate) and the result was a yield very close to that of a riskless treasury bond (3,78% to 4,18%).
Since I´m a fan of efficient markets I´d say that your calculation is rather another proof of the Efficient Markets Hypothesis than a puzzle.
However I´m worried about some mistakes in my conclusions too so please don´t hesitate to correct me!
HedgeFundGuy antwortete am 20. Jan, 23:05:
Well, you could be right, but only if credit risk is considered nonsystematic, or diversifiable. Most researchers think it is systematic and nondiversifiable.
stxx antwortete am 21. Jan, 01:28:
Another thing that is puzzling me: "If credit risk is not rewarded by the financial markets, why do investors not simply engage in an asset swap transaction and go short the B and long the AAA cash flow? With a large enough portfolio (to realize the 6% default prob.) I would have a free lunch (most likely)?"
finance bob antwortete am 21. Jan, 15:05:
Yeah that´s right. So if hedge funds (or other traders) have the possibility to exploit this free lunch then in the long-run bond prices should adjust so that the extra yield of a risky bond (=spread) exactly offsets the risk of default.(?)
HedgeFundGuy antwortete am 21. Jan, 16:26:
positive return, poor sharpe
I agree that the 3% return seems a reasonable estimate, but if you look at the annualized volatility of going long the AAA and short the B Index, you have an annualized stdev of about 8%. That means a sharpe of merely 3%/8%, or .38, which is not going to excite many investors.
