Where do the losses in HY show up? [High Yield bonds]

[Kevin M]

I wish I could find that source of historical VG SEC yields that Kevin referred to.

You can find SEC yield (and price) history for any Vanguard fund on the Vanguard website. Go to the Price & Performance tab, click on Search for more historical price information below the several-day price history at the right, then enter start date and end date. You can only pull about one year at a time, but I wrote a Google Sheets script that can cycle through multiple pages/years and retrieve the results into a spreadsheet. That's what i use to get the yields to create the charts.

Kevin

[restingonmylaurels]

Good question. Analyzing the 15-year's of returns (2003-2017 inclusive) broken out by capital return and income return provided by Vanguard here, https://investor.vanguard.com/mutual-fu ... ve-returns, the cumulative 15-year capital return is 1.2% (0.1% annualized), and 15-year cumulative income return is 183.7% (7.2% annualized). So not negative, but slightly positive capital return.

One thought is that perhaps the return was lower than expected based on the initial yield. SEC yield on 12/31/2002 was 8.37%. So 15-year annualized return was a bit more than 1 percentage point lower than the initial yield, which is within the typical range of initial yield prediction error. However, most bond funds probably had 15-year returns that were higher than the initial yield--at least that's my guess based on looking at 10-year returns and initial SEC yields for several bond funds in previous analyses (maybe I'll check this).

Kevin

Kevin,
One of the implications of what you have shown in your posts on this thread (that the losses are not showing up in a lower NAV but are reducing total return vis a vis SEC yield) potentially solve one of the problems with holding this fund.

That is, the asymmetry of taxation that when the income is taxed at ordinary marginal rates but if the credit losses are hitting the NAV, then they would only be recovered through reductions in capital gains at the capital gains rates and only when the fund shares are sold.

If the credit losses are flowing through instead as reduced distributions, then this solves the taxation problem, as these would reduce the tax paid on ordinary income instead of capital gains and would also solve the timing problem, as the reduction would happen in the year of the losses instead of building up over time in the NAV.

Whether intended or it, would you say this is what is in reality happening?

I don't think this is what is happening, because interest rates were declining over this period. Lower interest rates cause bond funds to have capital gains; these gains go away if a bond is held to maturity, but are realized if the bond is sold before maturity. Defaults and downgrades cause bond funds to have capital losses. Vanguard's high-yield fund sells most bonds before maturity, so it gets capital gains from falling rates, and capital losses from downgraded bonds. It still gets almost all of the coupons as payments, as it usually sells bonds before they default.

David,

So in essence you are saying the situation has been something like the following (my weak attempt at fund accounting, using hypothetical numbers).

Code: Select all

NAV - starting    7.50
Interest earned   6.00
Dividends paid   (6.00)
Credit losses    (1.50)
Appreciation      1.50

NAV - ending      7.50

I wish I could find that source of historical VG SEC yields that Kevin referred to.

In lieu of that, I find the S&P 500® High Yield Corporate Bond Index had a YTM of 4.96% 5 years ago and a now has a YTM of 5.56%. Does that support the notion that interest rates have been falling sufficient to offset the credit losses?

OK, with the pointer from Kevin M, I can look at the SEC yield at two points in time and look for that impact of falling interest rates David G has pointed out.

If I that say 15 years, from the beginning of 2003 to the beginning of 2018, I find the NAV is essentially the same. But the SEC yield has fallen significantly on a percentage basis.

!/7/2003 $5.93 8.32%
1/2/2018 $5.92 5.00%

As the NAV is flat, I would then assume that all of the decrease in the SEC yield is due to lower interest payments (the falling interest rates David refers to).

David said that he thought that the credit losses may be being balanced against the sales of appreciated securities (due to the falling interest rates).

Using the Rick Ferri loss of 1.5% from credit events, that would seem to be a loss of 22.5% over 15 years. The SEC yield fell 40% in those 15 years (8.32% to 5.00%).

If those two are required to be essentially balanced for my fund accounting to work to keep the NAV constant 15 years later, then it would mean what? The credit losses were actually higher? Something else?

[grabiner]

OK, with the pointer from Kevin M, I can look at the SEC yield at two points in time and look for that impact of falling interest rates David G has pointed out.

If I that say 15 years, from the beginning of 2003 to the beginning of 2018, I find the NAV is essentially the same. But the SEC yield has fallen significantly on a percentage basis.

!/7/2003 $5.93 8.32%
1/2/2018 $5.92 5.00%

As the NAV is flat, I would then assume that all of the decrease in the SEC yield is due to lower interest payments (the falling interest rates David refers to).

David said that he thought that the credit losses may be being balanced against the sales of appreciated securities (due to the falling interest rates).

Using the Rick Ferri loss of 1.5% from credit events, that would seem to be a loss of 22.5% over 15 years. The SEC yield fell 40% in those 15 years (8.32% to 5.00%).

And for a fund with a 4.4-year duration, the decrease in SEC yield from 8.32% to 5.00% causes a capital gain of 4.4*3.32%=15%. (I don't know how consistent the duration has been over these 15 years.)

If those two are required to be essentially balanced for my fund accounting to work to keep the NAV constant 15 years later, then it would mean what? The credit losses were actually higher? Something else?

Given my computation, the credit losses should be an annual 1%, not 1.5%. Since this fund holds higher-quality junk (mostly BB and B) and you are looking at a 15-year period with only one recession, that seems reasonable.

[restingonmylaurels]

OK, with the pointer from Kevin M, I can look at the SEC yield at two points in time and look for that impact of falling interest rates David G has pointed out.

If I that say 15 years, from the beginning of 2003 to the beginning of 2018, I find the NAV is essentially the same. But the SEC yield has fallen significantly on a percentage basis.

!/7/2003 $5.93 8.32%
1/2/2018 $5.92 5.00%

As the NAV is flat, I would then assume that all of the decrease in the SEC yield is due to lower interest payments (the falling interest rates David refers to).

David said that he thought that the credit losses may be being balanced against the sales of appreciated securities (due to the falling interest rates).

Using the Rick Ferri loss of 1.5% from credit events, that would seem to be a loss of 22.5% over 15 years. The SEC yield fell 40% in those 15 years (8.32% to 5.00%).

And for a fund with a 4.4-year duration, the decrease in SEC yield from 8.32% to 5.00% causes a capital gain of 4.4*3.32%=15%. (I don't know how consistent the duration has been over these 15 years.)

If those two are required to be essentially balanced for my fund accounting to work to keep the NAV constant 15 years later, then it would mean what? The credit losses were actually higher? Something else?

Given my computation, the credit losses should be an annual 1%, not 1.5%. Since this fund holds higher-quality junk (mostly BB and B) and you are looking at a 15-year period with only one recession, that seems reasonable.

From memory, I believe the duration has been higher, we could say it was 5.0 years over this history.

In using your calculation methdology on the two shares classes over their entire price/yield history, with a constant NAV:
Inv: 9.63% (Jan. 3 1993) – 6.41% (Dec. 4 2018) = 3.22 * 5.0 = 16.1 / 26 years = 0.62% credit loss/year
Adm: 9.16 (Nov. 12 2001) - 6.41% (Dec. 4 2018) = 2.75 * 5.0 = 13.75 / 17 years = 0.81% credit loss/year

But the NAV has not stayed constant. For Inv shares, it has decreased from $7.43 to $5.59 = 24.7% or about 0.95% per year over 26 years. For Adm shares, it has decreased from $6.33 to $5.59 = 11.7% or about 0.69% per year over 17 years.

I would think, as NAV is the denominator in the SEC yield calculation, that its decrease would be inflating the SEC yield by masking some of the credit losses.

That would seem to indicate that the NAV decrease should be added back to the SEC yield decrease to arrive at the annual credit loss number. So the historical credit loss may be 0.62% + 0.95% = 1.57% per year for Inv and 0.81% + 0.69% = 1.50% per year for Adm.

These would then seem to line up with Kevin’s calculation of the difference between SEC yield and total return over time but is that the proper calculation to use when the NAV is changing or am I double counting?

[Electron]

Here is a chart showing the NAV and Dividend trends going back to 7-10-89. The long term decline in NAV has been attributed to multiple factors including rating downgrades, early calls, buying bonds at a premium, and bid-ask spread. The fund has paid out a few relatively small capital gains distributions. It's not clear if bond premium has been amortized over the life of the fund. The bid-ask spread can rise significantly in times of distress in the high yield market. Any market timing puts the fund at a disadvantage in terms of bond pricing. In regards to defaults, I'd expect that bonds would typically be sold after downgrades and before any actual default. The NAV trend seems to have stabilized in recent years.

The dividend touched a new low in January. Since then it has bottomed out and turned up modestly.





I recently compared the performance of Wellesley Income and Vanguard High Yield Corporate on Morningstar and was surprised with the results. Performance has been nearly identical over the last three years. However, Wellesley has generally outperformed over the last fifteen years with the exception of the ten year period.

Total Return Annual Percentage 12-07-18

Code: Select all

Fund     YTD     1 Yr   3 Yr    5 Yr    10 Yr   15 Yr

VWIAX 	-1.28 	-0.42   5.39    5.44     8.59   6.73    Wellesley Income
VWEAX 	-0.80   -0.46   5.22    4.17    10.49   6.10    High Yield Corporate

[Kevin M]

Here is a chart showing the NAV and Dividend trends going back to 7-10-89.<snip>

The NAV trend seems to have stabilized in recent years.

I showed an even longer NAV history in this post on the previous page:
viewtopic.php?f=10&t=265108#p4236519

As I said there, I don't really think you can say that there's any steady NAV trend. You can find very long periods where the NAV ends up where it started. Then there are the periods when it never recovers to its previous highs.

Kevin

[Valuethinker]

Here is a chart showing the NAV and Dividend trends going back to 7-10-89. The long term decline in NAV has been attributed to multiple factors including rating downgrades, early calls, buying bonds at a premium, and bid-ask spread. The fund has paid out a few relatively small capital gains distributions. It's not clear if bond premium has been amortized over the life of the fund. The bid-ask spread can rise significantly in times of distress in the high yield market. Any market timing puts the fund at a disadvantage in terms of bond pricing. In regards to defaults, I'd expect that bonds would typically be sold after downgrades and before any actual default. The NAV trend seems to have stabilized in recent years.

The dividend touched a new low in January. Since then it has bottomed out and turned up modestly.





I recently compared the performance of Wellesley Income and Vanguard High Yield Corporate on Morningstar and was surprised with the results. Performance has been nearly identical over the last three years. However, Wellesley has generally outperformed over the last fifteen years with the exception of the ten year period.

Total Return Annual Percentage 12-07-18

Code: Select all

Fund     YTD     1 Yr   3 Yr    5 Yr    10 Yr   15 Yr

VWIAX 	-1.28 	-0.42   5.39    5.44     8.59   6.73    Wellesley Income
VWEAX 	-0.80   -0.46   5.22    4.17    10.49   6.10    High Yield Corporate

Thank you.

One would think that chart would put people off. However I think the recency effect dominates.

Even if we assume no repeat of the 1989-90 situation, arrest of Michael Milliken and bankruptcy of Drexel Burnham Lambert, the volatility of these things is still impressive.

Right now the market is offering very little risk premium for the additional risk taken on in HY bonds. It feels like a "heads the borrower wins, tails the investor loses" kind of situation.

Your Mileage May Vary.

[fennewaldaj]

Even if we assume no repeat of the 1989-90 situation, arrest of Michael Milliken and bankruptcy of Drexel Burnham Lambert, the volatility of these things is still impressive.

I was somewhat surprised by this but the 1989-1990 situation looks like a non event when plotted in portfolio visualizer

https://www.portfoliovisualizer.com/bac ... sisResults

granted a similar event now would likely result in a negative 3 year return since starting yields are like 7%

[Valuethinker]

Even if we assume no repeat of the 1989-90 situation, arrest of Michael Milliken and bankruptcy of Drexel Burnham Lambert, the volatility of these things is still impressive.

I was somewhat surprised by this but the 1989-1990 situation looks like a non event when plotted in portfolio visualizer

https://www.portfoliovisualizer.com/bac ... sisResults

granted a similar event now would likely result in a negative 3 year return since starting yields are like 7%

I'd have to check but I think it would look worse on a very short time horizon (ie the bonds went down *a lot* but then recovered when defaults were not as bad as expected) -- real volatility. I agree that it may simply be the coupons (12-15% in those days?) compensated.

I think the main problem with HY is the confusion between Original Issue (that Milliken "invented") and Fallen Angels-- the latter due show an anomaly that is not explained just by equity risk but results from a regulatory anomaly in the market - that some investors cannot hold sub investment grade.

[Electron]

Thanks Kevin and Valuethinker for your comments. Here is a very interesting history on the high yield market.

https://www.investopedia.com/articles/i ... tdowns.asp

"The junk bond market grew exponentially during the 1980s from a mere $10 billion in 1979 to a whopping $189 billion by 1989, an increase of more than 34% each year. Throughout this decade, junk bond yields averaged around 14.5% with default rates just a little over two at 2.2%, resulting in annual total returns for the market somewhere around 13.7%."

"In a change that took perhaps as little as 24 hours, new junk bonds basically disappeared from the market with no rebound for about a year. This resulted in investors losing a net 4.4% on the high-yield market in 1990 – the first time the market had returned negative results in more than a decade."

This information adds some background to Kevin's earlier comment that most of the NAV degradation took place from 1978 through 1990. Michael Milken had a large part in creating the junk bond market where new bonds were issued with a junk rating for the first time starting in the early 1980s. As Valuethinker mentioned, Drexel Burnham Lambert declared bankruptcy in 1990.

It would be very interesting to see the early portfolio holdings in Vanguard High Yield which started operation at the end of 1978. One might guess that the portfolio included fallen angels as well as newly issued junk bonds. As I recall, a high profile default would have major impact on the entire high yield market. Another unknown is whether the fund distributed any capital gains in the early years.

There might be an interesting comparison that could reveal more about Vanguard High Yield. We need to find a Treasury or Investment Grade bond fund that declared little in the way of capital gains distributions. The Net Asset Values could be compared which might reveal more about Vanguard High Yield in the 1980s. The fund would undoubtedly have been impacted by any periods of large redemptions or market timing.

[fennewaldaj]

Even if we assume no repeat of the 1989-90 situation, arrest of Michael Milliken and bankruptcy of Drexel Burnham Lambert, the volatility of these things is still impressive.

I was somewhat surprised by this but the 1989-1990 situation looks like a non event when plotted in portfolio visualizer

https://www.portfoliovisualizer.com/bac ... sisResults

granted a similar event now would likely result in a negative 3 year return since starting yields are like 7%

I'd have to check but I think it would look worse on a very short time horizon (ie the bonds went down *a lot* but then recovered when defaults were not as bad as expected) -- real volatility. I agree that it may simply be the coupons (12-15% in those days?) compensated.

I think the main problem with HY is the confusion between Original Issue (that Milliken "invented") and Fallen Angels-- the latter due show an anomaly that is not explained just by equity risk but results from a regulatory anomaly in the market - that some investors cannot hold sub investment grade.

It looks like the max drawdown of vanguards fund over that period was 12.5% per portfolio visualizer. This occurred over a 3 month period. That is monthly data i think so it could have been worse looking at daily data.

[grabiner]

Even if we assume no repeat of the 1989-90 situation, arrest of Michael Milliken and bankruptcy of Drexel Burnham Lambert, the volatility of these things is still impressive.

I was somewhat surprised by this but the 1989-1990 situation looks like a non event when plotted in portfolio visualizer

https://www.portfoliovisualizer.com/bac ... sisResults

granted a similar event now would likely result in a negative 3 year return since starting yields are like 7%

I'd have to check but I think it would look worse on a very short time horizon (ie the bonds went down *a lot* but then recovered when defaults were not as bad as expected) -- real volatility. I agree that it may simply be the coupons (12-15% in those days?) compensated.

I think the main problem with HY is the confusion between Original Issue (that Milliken "invented") and Fallen Angels-- the latter due show an anomaly that is not explained just by equity risk but results from a regulatory anomaly in the market - that some investors cannot hold sub investment grade.

It looks like the max drawdown of vanguards fund over that period was 12.5% per portfolio visualizer. This occurred over a 3 month period. That is monthly data i think so it could have been worse looking at daily data.

Does your chart go back more than ten years? I checked with Morningstar, which shows a 25% drawdown from 8/31/2008 to 11/30/2008. (The worst drawdown over a non-month-ending period was 27% from 9/9/08 to 12/6/08.)

Edit: Sorry, I missed the reference to the correct period in the "over that period". Your data is correct.

[Kevin M]

Even if we assume no repeat of the 1989-90 situation, arrest of Michael Milliken and bankruptcy of Drexel Burnham Lambert, the volatility of these things is still impressive.

I was somewhat surprised by this but the 1989-1990 situation looks like a non event when plotted in portfolio visualizer

https://www.portfoliovisualizer.com/bac ... sisResults

granted a similar event now would likely result in a negative 3 year return since starting yields are like 7%

I'd have to check but I think it would look worse on a very short time horizon (ie the bonds went down *a lot* but then recovered when defaults were not as bad as expected) -- real volatility. I agree that it may simply be the coupons (12-15% in those days?) compensated.

I think the main problem with HY is the confusion between Original Issue (that Milliken "invented") and Fallen Angels-- the latter due show an anomaly that is not explained just by equity risk but results from a regulatory anomaly in the market - that some investors cannot hold sub investment grade.

It looks like the max drawdown of vanguards fund over that period was 12.5% per portfolio visualizer. This occurred over a 3 month period. That is monthly data i think so it could have been worse looking at daily data.

Does your chart go back more than ten years? I checked with Morningstar, which shows a 25% drawdown from 8/31/2008 to 11/30/2008. (The worst drawdown over a non-month-ending period was 27% from 9/9/08 to 12/6/08.)

The discussion was specifically about the 1989-1990 period, and the max drawdown (monthly data) was indeed -12.5% from Aug 1990 through Oct 1990, with recovery by Apr 1991. This was the second-worst drawdown, with the worst being -28.90% from Jun 2007 through Nov 2008, with recovery by Sep 2009.

The PV link provided above does not work. Here is a working link.

Kevin

[Electron]

Here is a very interesting article about Michael Milken and junk bonds published in November 1990.

Note the comment about the quality of junk bonds issued in the early 1980s being superior to those issued later.

https://www.nytimes.com/1990/11/22/busi ... -junk.html

Vanguard High Yield was launched slightly ahead of the period associated with Michael Milken and Drexel Burnham Lambert. One might speculate that Vanguard initially invested in a mixture of fallen angels, lower investment grade credits, and perhaps some higher quality credits for liquidity. It's unfortunate that early annual reports are not available as it would be very interesting to see if Vanguard purchased any of the Drexel sponsored junk bonds. They could have purchased bonds as new issues or in the secondary market after having done a detailed analysis in either case. The article I referenced earlier mentioned that the default rate in the 1980s was only slightly higher than 2%.

"A study by Barrie A. Wigmore, a limited partner of Goldman, Sachs & Company, found the quality of junk bonds issued in the early 1980's was vastly superior to those issued later."

"Not surprisingly, it is the later bonds that have most often defaulted, and those who bought such bonds in recent years generally have suffered significant losses in the collapse of the junk bond market, while bonds issued earlier have generally provided profits for buyers."

"When Mr. Milken first entered the securities business, in the 1970's, junk bonds were relegated to an obscure part of the financial community. Known as "fallen angels," most of these securities were bonds issued by companies that once had good credit standing, but whose credit had deteriorated. Many investors automatically sold bonds that ran into such problems, driving them down to prices that Mr. Milken rightly argued more than offset the increased risk. He developed a following by buying and trading such bonds."

"The business grew into underwriting and selling junk bonds for companies that did not qualify as investment-grade issuers. Buyers of the bonds generally prospered, more and more money became available for such investments, and credit standards started to deteriorate."