Archive for the ‘Bonds’ Category

Retirement Balanced Portfolio: Asset Super Class Performance From 1928

Monday, June 29th, 2015

Allocation all boils down to Owning, Loaning and Reserving.  Everything else is a variation on implementation of these three super classes.

Retirement portfolios need some Owning (for capital and income growth), some Loaning (for steady income and portfolio volatility dampening), and some Reserving (for dependable cash availability for withdrawals, and less reason to panic when the market is panicking).

The term “asset class” gets tossed around a lot and sometimes asset types that are merely subcategories of true asset classes are mislabeled as “asset classes”.  While the use of the term is certainly arguable, the three most basic asset classes are: (1) cash reserves, (2) loans made to others, and (3) ownership of something like stocks, real estate or commodities.  We think of them as “super classes”, because just about everything else falls under one of these as a subcategory (even derivatives).

LOR_ThreePrimarySuperClasses

Three specific examples of Loan, Own and Reserve are the 10-Year Treasury Bonds (Loan), the S&P 500 (Own) and 90-day Treasury Bills (Reserve).

They, like the super classes to which they belong, are substantially minimally or even negatively correlated over reasonable periods of time.

While stocks and bonds have periods of high correlation, such as in panics, where there may be a rush to cash for safety and liquidity; for the most part they react differently to different circumstances, making them good pairs, along with cash reserves in a portfolio where risk (volatility) control is an important goal.

For most asset categories really long-term data is hard to find, but for 10-year US bonds, US large-caps stocks (S&P 500 and precursors), and 90-day T-Bills; we have data back to 1928 as proxies for the Loan, Own and Reserve super classes.

Let’s see how those super classes behaved in total return, price return and income return in an annually rebalanced tax-deferred or non-taxable account from 1928 through 2014.

The portfolio we use here is 60% US large-cap stocks, 31% 10-year Treasury Bonds, and 9% of 90-day T-Bills. US large-cap stocks are the S&P500 and its precursors.

That is a specific modification of a generalized 60/40 portfolio to make room for a three bucket (9%) operating reserve for retirement accounts.  The reserve is important to create fairly high certainty of capital available for withdrawal for up to 3 years, which is helpful to avoid panic in a steep down stock or bond market.

There are very few periods where a balanced portfolio is down for more than 3 years, so the 9% 3-bucket reserve can help a retiree glide over a bear market without selling assets at fire sale prices to fund living expenses.  Selling assets at depressed prices in retirement is a good way to outlive your assets — and that is not a good thing.

In practice, the 60% in stocks would likely include different market-cap sizes, and some international in most cases; and the 31% in bonds would likely include different durations and perhaps qualities, and maybe some international; and the 9% in reserves (90-day T-Bills as proxy here) would probably consist of 3% in a money market fund (duration about 90 days),  3% in an ultra-short bond fund (duration about 1 year) and a short-term bond fund (duration about 2.5 to 3 years).

Figure 1 is a chart showing the annual total return of the 60/31/9 portfolio from 1928 (87 years), along with the three-year moving average of total return:

Figure 1: Total Return – 60 stock / 31 bonds / 9 reserve

60-31-9 total return

There were some substantial dips on an annual basis, but the 3-year average was almost always positive, and where negative was mostly only mildly so.  Except for the Great Depression where the 60/31/9 declined over 14%, the worst decline was less than 4.5%

There were only 7 years of the 3-year average that were negative, and only two periods had back-to-back negative years: 1931-1932 and 1941-1942.  Two of the 7 down years had declines of less than 0.5%.

Specifically for individual years, there were 19 out of 87 years with a negative return (21.84%) of the time.

The periods of back-to-back negative periods were:

  • 4 years from 1929-1932
  • 2 years from 1940-1941
  • 2 years from 1973-194
  • 2 years from 2001-2002

The traditional 60/40 fund uses the S&P 500 index and the Aggregate Bond index.  That fund had total return less than negative 22% in 2008 (ref: Vanguard Balanced Fund VBIAX), whereas this 60/31/9 portfolio illustration uses Treasury debt for both the 31 and the 9.  Since 2008 was a year of panic, Treasury debt outperformed Aggregate Bonds.  This example portfolio declined 15.56% in 2008.

Both VBIAX and this example portfolio declined more than 22% and 15% from their intra-year 2007 peak to their 2008 intra-year bottom, but this memo is only about calendar year performance.

Figure 2 is a chart showing the price return separate from the income return for the 61/31/9 allocation.

Figure 2: Price Return and Income Return – 60 stock / 31 bonds / 9 reserve

61-31-9 price and income return separately

As you might expect the income return is far less volatile and more reliable than the price return.  That is important to retirees who are drawing on their portfolios.

The brown line is the price return and the green line is the income return.  The dashed black line is the 4% return level that relates to the 4% rule of thumb for retirement withdrawals.

While total return had 19 negative years of 87, price return had 29  (33.33% of the time).  It was the income return that saved the portfolio from the 10 extra negative years.

The specific back-to-back negative price return periods were:

  • 4 years from 1929-1932
  • 3 years from 1939-1941
  • 3 years from 1946-1948
  • 2 years from 1973-1974
  • 2 years from 1977-1978
  • 3 years from 2000-2002

With regard to the 9% 3-bucket reserve, individual bonds are probably not a practical alternative for most retirees for a variety of reasons. Here are three low-cost Vanguard funds that could probably fulfill the objectives and uses of the 9% reserve.

9 pct reserve bucket

 

 

Survey of Technical Condition and Portfolio Attributes of Key Bond Fund Types

Sunday, April 12th, 2015

There may seem to be a bewildering array of bond fund types. This survey of key bond types may help you chose which if any are appropriate for your particular circumstances and portfolio.

We looked at 24 types of bond funds in terms of price chart slope:

US Aggregate Bonds (BND)
US ST Bonds (BSV)
US IT Bonds (BIV)
US LT Bonds (BLV)
US ST Govt Bonds (VGSH)
US IT Govt Bonds (VGIT)
US LT Govt Bonds (VGLT)
US Extended Duration Govt Bnds (EDV)
US ST Corp Bonds (VCSH)
US IT Corp Bonds (VCIT)
US LT Corp Bonds (VCLT)
US HY Corp Bonds (HYG)
ST Natl Tax Exempt Bonds (VWSTX)
IT Natl Tax Exempt Bonds (VWIUX)
LT Natl Tax Exempt Bonds (VWLTX)
HY Natl Tax Exempt Bonds (VWAHX)
Total Foreign Bonds (hedged) (BNDX)
Global Ex US HY Bonds (Euro, Sterling, Loonie) (HYXU)
Emerg Mkt Dollar Bonds (VWOB)
Emerg Mkt HY Dollar Bonds (EMHY)
US Inflation Protected Treasuries (VTIP)
US Mortgage Backed Bonds (VMBS)
US Floating Rate Bank Loans (BKLN)
US Preferrred Stocks (PFF)

We included a preferred stocks ETF in the list of “bond” funds. because preferred stocks can be arguably said to be closer to bonds than to common equity in their characteristics; and in this period of very low interest rates preferred stocks have been adopted by many investors as bond substitutes in the search for yield.

Figure 1: Technical Condition of Key Bond Fund Types
(Click to enlarge)

The most basic form of chart analysis is a binary determination — is the  slope of the price pattern is UP or NOT

Figure 1 below includes 6 binary Yes/No (1/0) measurements of the slope of the price chart as of Friday April 10, 2015 (from most important on top of list and least important on bottom of list):

  • 12 mo direction (is the 12 month average now higher than 10 days ago? Y/N)
  • 3 month av / 12 mo av (is the 3 month average above the 12 mo average? Y/N)
  • 3 mo direction (is the 3 month average now higher than 10 days ago? Y/N)
  • 1 mo / 3 mo (is the 1 month average now higher than 10 days ago? Y/N)
  • 1 mo direction (is the 1 month average above the 3 month average? Y/N)
  • Price / 1 mo (is the last price above the 1 month average? Y/N)

1

 

We have highlighted those with all 6 binary answers as Yes (1) with a bold border.  Of the 24 (as of Friday April 10, 2015), 13 of the 24 has a “perfect” 6 binary YES (1) slope ratings:

BND, BSV, BIV, VGSH, VGIT, VCSH, VCIT, VWIUX, VWLTX, VWAHX, BNDX, VMBS, PFF

Figure 1 also presents 4 other chart price pattern attributes:  12 mo, 3 mo and 1 mo Z-scores (number of standard deviations the price is from the corresponding moving average price); and the 14-day money flow index (a volume weighted version of the relative strength index — itself a measure of the tendency of the price to position itself within its average daily range).

For Z-scores look out for +/- 2 or more, as being possibly too far from its average to be likely able to maintain that distance.  For Money Flow (MFI14) values over 70 are seen as overbought, and under 30 as oversold).

This binary data was produced using custom formulas in Equis Metastock with data feed from Reuters.  Note that Reuters adds suffixes to many securities to identify where they are traded.

Figure 2: Bond Fund Type Rolling Returns, Trailing Yield and Expense Ratio
(Click to enlarge)

Note that some of the fund types have a 12 month trailing yield that is greater than the 12 month total return (returns are as of last month-end(.  That means the price declined (See high yield corporate bonds, short-term municipal bonds, emerging market Dollar denominated bonds, emerging market high yield bonds, inflation protected Treasury bonds, and floating rate bank loans).

No adjustment has been made to express municipal bond yields on an after-tax basis.  A 1-to-1 comparison of presented yield between taxable and tax-exempt bonds should not be made — rather an estimate of your pre-tax equivalent yield is up to you to estimate.

2

Figure 3: Bond Fund Type Spread to Aggregate Bonds, Duration and Credit Quality
(Click to enlarge)

Take note of the different credit qualities and average durations, not just the yields on the different fund types.

3

Also keep in mind that duration for a fund (note is called “average” duration) has potentially quite different implication in a period of changing interest rates when compared to a single bond with a the same duration. A bond fund of any particular average duration may not respond to interest rate changes in the same way that a single bond would do.  Similarly two bonds funds of the same average duration may respond differently as well

Consider these two hypothetical bond funds, each containing 10 individual bonds, and each bond fund having an average duration of 5.5 years.  The first fund has 10 bonds each with a duration of 5.5 years.  The second fund as 10 bond each with a different duration (1,2,3,4,5,6,7,8,9, and 10 years), also with an average duration of 5.5 years.

When interest rates change, and the shape of the yield curve changes, these two funds will no react to the same degree.

This is not even to consider the mix of credit qualities that might react differently due to perceived changes in credit risk that may be happening at the same time that interest rates are changing.  And, to the extent that the fund uses derivatives to emulate bonds, a liquidity crisis could create an entirely different ball game for those funds versus funds that only hold real bonds.

Figure 4 Bond Fund Type Spread of Credit Qualities Within Each Fund
(Click to enlarge)

When thinking about your risk level, it may be helpful to look not only at average credit quality, but also the spread of quality levels that result in the average quality level, as shown in Figure 4.

4a

Figure 5: Bond Fund Type Composition by Type of Bonds

(Click to enlarge)

When thinking about your risk level, it may be helpful to look not only at average credit quality (and the spread of qualities), but also at the type of bond: government issued, corporate issued, securitized, municipal and synthetic (derivative).

5

The data in Figures 2-5 is from Morningstar Advisor Workstation.

Price Charts for Bond Fund Types
(click images to enlarge)

Here are the charts for each of the bond funds in the table above, listed in the same order as in the tables.  These charts were produced mid-day April 15, whereas the tables were produced as of end-of-day Friday April 10.

You can judge for yourself how useful the binary evaluation of slope and moving average stacking tells the story, and how well it assets you in keeping your eyes honest as you look at charts, and how using a simple set of binary decisions enables you to compare many charts, each of which may have more information that you remember.

Each chart shows the 1 year percentage total return (as calculated by StockCharts.com) in black, the 200-day exponential moving average in gold, the 50-day exponential moving average in dashed blue, and he ratio of the funds performance to the performance of BND (aggregate US bonds) in green in the lower panel.

bonds1

bonds2bonds3bonds4