Asset Allocation Critique

[WilliamG]

I am turning 59 1/2 next year and will begin regular withdrawals from my IRA(s). Have been looking at asset allocation for quite a while and have come up with what I believe will work for us and would appreciate any feedback/critique. Overall portfolio is 56/44 equity vs non-equity and all funds are Vanguard.

8% Total International Stock Index
8% REIT Index
8% Small Cap(Small Cap Index&Small Cap Val Index)
16% Value Index (large cap value)
16% Total Market Index

16% GNMA
16% Short Term Corporate
8% US Treasury Direct Holdings (I Bonds, TNotes&Bills)
4% Cash

Thanks in advance for your help!....William G

 

[Cut-Throat]

It looks like a decent enough mix to me.

You did not mention what percentage of withdrawal that you planned for the first 5-10 years of this portfolio.

If it is around 4%, I would think the cash is way too lean. You could be forced to sell stocks/bonds in a down market or when you don't want to. If it's less than 1%, the cash amount is probably fine.

 

[WilliamG]

Thanks for the reply Cut-Throat. Will be withdrawing in the 4% range, perhaps a little less. Am sensitive to not selling equities when the market is down, but would hope our US Treasuries and even Short Term Corp would fill the need for any liquidations during a down market. I Bonds are particularly useful here, as we have them reaching 5 years beginning next year which means they can be cashed at any time without even losing 3 mos. interest. See them almost as a cash substitute.... William G

 

[Ted]

WilliamG,

While your allocation to stocks soeems reasonable, I'd say that there is too much in "cash equivalents," which include not only the Vanguard Short Term Corporate Fund, but also whatever bonds are about to mature.

I would make three suggestions:

1. Drop the stock allocation to about 50%, and drop the "cash equivalent" allocation enough to put about 15% of you assets in one or more high yield bond funds. I'm not sure whether Vanguard's is open to new investors, but TIAA-CREF has one that is equivalent, and T.R. Price also has a good one. Fidelity's High Income Fund and Capital and Income Fund offer higher returns, but with greater volatility. (In terms of volatility, high yield bonds act about 2/3 like bonds and 1/3 like stocks.)

2. Put any long term bonds (greater than 5 years) into TIPs. Now that you are retired, the tax deferral of the gains on I Bonds probably does not compensate for the lower interest that they pay relative to TIPs.

3. Check the average maturity of your GNMA bonds. If it is more than about 5 years, I would sell those and buy TIPs to protect against loss of value if long term interest rates rise, which I think is very likely. When they do, you might consider putting some money back into GNMA or other longer term bonds.

 

[ats5g]

my thoughts

I hope no one is suggesting that William either sell or convert his I bonds into something else. Since the I bonds are reaching their 5 year anniversary next year, it is likely that William's I bonds are earning a rather high real rate (e.g 3.4%). That's much better than he can get on any maturity of TIPS now. You might hold onto those babies for as long as possible. Though I bonds are as good as cash for liquidation purposes, I would hold onto those I bonds with higher real rates for a long time, and possibly consider them as part of a TIPS allocation.

From http://www.publicdebt.treas.gov/sav/sbirate2.htm, the real rates on previously issued I bonds.

NOV 1, 2003 1.10%
MAY 1, 2003 1.10%
NOV 1, 2002 1.60%
MAY 1, 2002 2.00%
NOV 1, 2001 2.00%
MAY 1, 2001 3.00%
NOV 1, 2000 3.40%
MAY 1, 2000 3.60%
NOV 1, 1999 3.40%
MAY 1, 1999 3.30%
NOV 1, 1998 3.30%
SEP 1, 1998 3.40%

Though, I wouldn't buy anymore I bonds now.

I agree with Ted in adding TIPS, but I have some reservations on adding that much HY bonds. It's good he pointed out that HY bonds are part equity and part bonds. I think the higher junk credit quality of Vanguard and TC's HY funds makes them around 2/3 bonds, 1/3 stocks, but I think the Fidelity's funds would be more like 1/2 bonds, 1/2 equity (or even less bonds). Although, when things go bad for equities, and correlations increase, the HY funds may become more equity like just when equities are doing terrible (at the wrong time). So, if you are going to add HY bonds, lowering the equity allocation (as Ted suggested), would be the way to do it. Personally, I'd rather keep the equity risk in equities and keep my bonds high quality and inflation protected.

Another consideration may be where you plan on holding the HY bonds. If a taxable account, I would think that keeping the equity risk in equities would be more tax-efficient, as those gains would likely by taxed at lower capital gains rates, rather than ordinary income rates in HY bonds.

Given the recent decline in yields of HY bonds (b/c of the recent runnup), I would think that HY bonds are less attractive than they were, say, a year ago. So, I might limit my allocation to 10% HY. Perhaps I'm just a scared-ee cat. I have also read that HY bonds have high correlations to small cap stocks, in Bernstein's article on them:

http://www.efficientfrontier.com/ef/401/junk.htm

I suppose one could hold b/w 10-15% cash/ST bonds until he/she has reached SS eligibility age to hedge one's bets in case of another major market meltdown b/w then and now.

On the small cap index & small cap value index, there is likely some serious overlap b/w the two. Also, I don't think Vanguard's Small Cap Value index is as value titled as a fund like DFA's small value, and likely includes a lot of small cap blend as well as small cap value. So, if you just want small cap value and blend (and not small cap growth), you could just do one small cap fund, small cap value.

- Alec

 

[WilliamG]

Thanks Ted and Alec for taking the time to reply.

Yes, Alec, I was fortunate enuf to get a wad of I Bonds at 3.4 and 3.6 rates. These comprise 7% of the 8% I noted as allocated to Treasuries. The other 1% is a 5 year note I also bought in the same time period that will be maturing in 2005. It will mature when we will probably be needing to buy a new car. I try to incorporate a 3-5 year cash flow into my allocation planning as well as general diversification goals.

Both of you seem to think my "cash equivalent" or short bond allocation is too high. I have been influenced by Frank Armstrong and others who say longer maturities do not adequately compensate for higher risk. Also, a major purpose of this allocation is to have for liquidation in down market. Frank says to have enuf for 5-7 years, so that is how I came up with the allocation targets for cash, Treasury Direct and Short Term Corporate. Guess you guys don't agree with this approach? If not, weren't the years 2000 thru 2002 a challenge for raising spending money from a retiree's portfolio?

Thanks again for your feedback... William

 

[ats5g]

William,

I'm not disagreeing, but I don't think I'd want to use GNMA's. I agree with F. Armstrong that one would certainly NOT want to withdraw from equities when they're down (reverse DCAing). The longer maturities of nominal bonds, longer than 7-10 years for example, do not really compensate one for the increase inflation and interest rate risk, in my opinion. Not to mention that the longer the maturities and the lower the credit quality, the higher the likelihood of higher correlations to your equities. TIPS (on the other hand) are very good at hedging inflation risks for longer term bonds, and should have much lower correlations with equities than nominal long-term bonds. You cannot buy the longer maturities of TIPS (greater than 10 years I believe) from the Treasury any more, but you can buy the longer maturities from the secondary market.

Here's what I was thinking. It's definitely a good idea to take your cash flow situation (including upcoming purchases) into account. Since you have a relatively large purchase (a car) it should make sense to have more in cash then someone who doesn't have a large purchase coming up. You could do something like:

1) have one years withdrawal in a MM (perhaps something like ING Direct paying 2%) - 4% & 1% for car
2) have years 2-5 withdrawals in a ST bond fund like ST corporate or ST bond index - 16%
3) rest of bonds in high paying I bonds and TIPS - 7% I bonds & 16% TIPS

So, your bond allocation could look something like:

4% - MM
1% - Treasury note for car
16% -ST bonds (either low cost fund or ladder of Treasuries/CD's)
7% - I bonds
16% - TIPS (maybe laddered in 5-10 [or whatever] year maturities)

The incredibly good thing about the I bonds is that if equities do go on a multi year, over 4 years, bear market, and if you run through your MM & ST bonds, you still have the I bonds which can be cashed in whenever, yet still earn those high real returns. And obviously, you'd want to keep the TIPS for last to let them mature. Or, if you have another bond maturing, you can always defer sale of the I bonds for when you have a gap in income.

Jaye Jarrett (http://jjarrett.home.texas.net/) did a study on optimal fixed income allocations in retirement. I found it interesting:

http://jjarrett.home.texas.net/resFixedIncomePortion.pdf

If I read it correctly, Intermediate Term Gov't bonds (with maturities of around 5 years) were better than Tbills or longer maturity nominal bonds (of 20 years). Unfortunately, TIPS and HY bonds haven't been around long enough for much studying.

I think the thing about this method is that if equities go into a multi year down period early in one's retirement, one may have to reduce his/her withdrawal rate slightly for a year or two (or three) until the equities take off again. Being diversified the way your equities are should alleviate this somewhat, but not totally. Anytime equities go on massive tears, it would probably be good to rebalance out of them and save the excess returns for the down years (so you won't have to reduce your withdrawal all that much).

- Alec

 

[WilliamG]

Alec,
Thanks for the links! I have quickly perused the document and plan to return to it and the site later.

I have been thinking that I would add TIPS to the Treasury allocation in the future. I want to hold on to those high fixed return I Bonds as long as possible as you recommended. Have been using taxable funds to live on since retiring at 55, so getting low in $ to buy additional I Bonds even if I wanted to at current low rate. Thought would use Vanguard's TIPS fund, but could also set up a brokerage account in my IRA for purchasing bonds directly.

Main change I see in your recommended fixed income allocation is ditching the GNMA. If I read correctly in my quick perusal, the Jarett study didn't evaluate them due to lack of data. I currently "take the dividend" from GNMA in my IRA to my IRA money market from which I will be making my 59 1/2 withdrawals. Influenced to GNMA for high quality based securities and relatively high dividend (note, not taking a high risk of having to sell at a low NAV). Also have seen positive info on GNMA in Scott Burns analysis of govt bonds, and most recently Bill Gross' bond allocation advice: split equally between GNMA, Corporate and TIPS.

Since Short Term Corporate, US Treasury and Cash cover my safety stash for not liquidating equities in a downward spiral, Frank Armstrong would probably recommend putting the GNMA 16% in equities but I'm too conservative for that! Based on this further clarification of how I am viewing my fixed income allocation, do you still feel negative on GNMA? Would you increase equity? I guess there is also the alternative of using high yield without decreasing my current equity allocation.

Best regards, William

 

[adviceseeker]

I have read about TIPS but have not invested in them to date. It seems many people look at them as a crucial part of their portfolio, but I wonder if you or others could give me your thoughts on whether this is a good time to purchase the new 10-year TIP offerings (January) - 5-Year TIPS are no longer available. I would probably purchase them from TreasuryDirect to avoid any commissions/fees; I also understand the cost to sell them before maturity is about $34, which may be competitive. My question relates to the need to sell the TIPS before the 10-year maturity, and what the liklihood would be that this would result in a loss of principal. In other words, what is the real-world experience in selling before maturity.

The other issue is the currently low interest rates. Would it be better to spread out the purchases, or if interest rates rise could you sell the current TIP and re-invest at the higher rate? This is probably where the principal value is impacted on sales before maturity, but I am interested in hearing from others with practical experience.

 

[jdvalle]

Be sure that you fully understand how individual TIPs work before you add them to your taxable account. You'll receive an annual tax bill for the inflation adjustment of principle even though you won't receive the benefit until the maturity date. This is in addition to the taxes paid on the bi-yearly coupon distributions.

Under certain inflation scenarios, the taxes on your TIPs could exceed the dividend payouts. That's why TIPs are normally recommended for tax-deferred accounts.

If you want a more diverse mix of TIPs, consider the low cost Vanguard and Tia-Creff TIPs funds. As you point out, though, you'll be paying the fund expenses; the ER and the internal transaction costs. Treasury Direct is cost free.

 

[ats5g]

William,

I wouldn't use GNMA's (or other callable mortgage backed bonds) because of "negative convexity". If you don't understand what that is and can't explain it, find out before using GNMA bonds.

I found this in "The Handbook of Fixed Income Securities", by Frank Fabozzi on page 584 (I think):

"In recent years, duration has become more important for the evaluation of pass-throughs [like GNMA - mortgage pass through securities]... Duration is the weighted average time to receipt of the present value of both principal and interest cash flows. Duration is appealing because it can be used to measure the price sensitivity of a bond. That is, 'modified' duration expresses the amount the price (present value) will change given a small change in the yield used to discount the cash flows. Thus, duration has important applications as a measure of interest-rate risk. As with yield and average life calculations, both Macauley and modified duration are highly sensitive to the prepayment assumptions used to project cash flows. As a result, duration can significantly misestimate the actual price change of pass-throughs when interest rates decline. More importantly, a pass-through's duration changes as the expected prepayment rates used to calculate it change in response to changes in the general level of interest rates... A pass-through's price declines more quickly for a small change in interest rates as the general level of interest rates rise. Similarly, pass-through prices increase more slowly for successive declines in the general level of interest rates... This characteristic of pass-through's price behavior is generally referred to as 'negative convexity'."

Pages 796-797 also has some good graphs which may help to explain negative convexity.
____________________________________________
I also found this at http://www.creditunions.com/investments/articles/Misc/042803-cfo.asp

"Convexity:

Understanding the concept of positive and negative convexity is more important than understanding its definition. Basically, convexity is a measure of duration. It measures how sensitive a bond's price is to [interest] rate movements in relation to the bond's price in an opposite [interest] rate movement.

Positive convexity occurs when the price of a bond appreciates more given a downward move in its yield than it appreciates when there is an upward move in its yield... It says you will experience more price appreciation in a declining rate environment in relation to its price depreciation in a rising rate environment.

Negative Convexity occurs when the price of a bond depreciates more given an upward move in its yield in relation to its appreciation in a declining rate environment. Negative Convexity is what makes callable bonds [like GNMA's] a less attractive investment vehicle for some portfolio managers: in a declining interest rate environment, callable bonds will tend to trade close to par as the call date approaches, because of the issuer's [or borrowers] built-in option to call the bonds. This is a classic case of negative convexity; the price depreciation due to an increase in rates is unlimited, but the price appreciation due to a decline in rates is limited because issuers will often call these bonds if they can re-issue them at lower rates."
________________________________________

So, some people have described mortgage backed bonds like GNMA's as a "lose/lose scenario". When interest rates decline, the capital gain on the mortgage bonds is not as great [if any gain] as in non-callable bonds because people will refinance or prepay their mortgages, calling the mortgage bonds [from you] and forcing you to reinvest at lower interest rates. But, when interest rates rise, people do not refinance or prepay their mortgages as much, leaving the mortgage holders [you] with the lower paying mortgage bonds for longer, lengthening their duration. The duration of GNMA's decrease when interest rates decline (when you don't want it to) and increase when interest rates rise (when you don't want it to).

You certainly do get a higher yield for investing in GNMA's, but at a higher risk than bonds of the same maturity. No free lunch here. E.F. Moody has written a rather scathing critique of GNMA's here:

http://www.efmoody.com/investments/gnma.html

If you've already decided that you can only stand around 55% equity, then I wouldn't increase your equity allocation. I like the addition of TIPS because (in my opinion) after 5 years or so you have to start worrying about inflation. Individual TIPS are the surest way to hedge this risk. Even ST bonds can be slighly positively correlated to inflation. And what does terrible during times of higher inflation? Equities and longer term nomial bonds (which GNMA's can turn into as rates rise – increasing their correlation with equities!!).

I think we can also say that individual TIPS have less reinvestment risk than nominal bonds of similar maturities. This is because of the inflation adjustments to the principle of the TIPS bond itself, which you cannot get [for cetain] until the TIPS bond matures. This is also why (as jdvalle said) individual TIPS are probably best held in tax deferred accounts.

That help (or just make things more confusing)?

- Alec

 

[Bob_Smith]

 That's why TIPs are normally recommended for tax-deferred accounts.

You're right about that, jdvalle. I'm in the process of shuffling things around to accommodate a large portion of individual TIPS. What I found is that 403(b) plans don't allow TIPS. So I can't open a 403(b) brokerage account at Vanguard. Since I'm still employed, I can't transfer $ from the 403(b) to an IRA (and I would lose any future protection from creditors if I did). That leaves my IRAs and SIMPLE plans. Then I learn that Vanguard doesn't allow SIMPLE brokerage accounts either. So I must shift $ from the SIMPLE plans to our traditional IRA plans. Nothing is ever simple.

 

[adviceseeker]

While it seems TIPS are a good idea in a tax deferred account, if I am retiring in 2004, and my tax deferred accounts are maxed out (401K does not offer TIPS and don't qualify for IRA contributions), I am looking at TIPS as an alternative to taxable CD's. In that event, are the 10-year TIPS a good alternative now. If interest rates increase over the next several years, what will this do to the principal amount assuming an early sale on the secondary market?

 

[Ted]

Here are a few more considerations about TIPs:

It is true that both the interest and annual increase in principal of TIPs are taxable. As with other bonds, that makes it desirable to hold them in a retirement account, BUT this is only important while a person is still working and has a large amount of earned income that -- in effect -- causes their investment income to be taxed in their highest bracket. Once a person is retired, their income is almost entirely from investments (and social security). Not only is the tax rate much lower, but there isn't usually much advantage to deferring taxes to the future, when they will have to be paid anyway, presumably at the same rate (or possibly even higher!).

In fact, an early retiree presumably is not drawing social security benefits, and so may be in a lower tax bracket now than when the s.s. benefits start.

If long term interest rates in general rise -- which is very likely -- the driving force will probably be an increase in inflation. This will cause the market value of most bonds to drop. By law, it will cause the principal value of TIPs to increase. It is possible that longer term TIPs could still temporarily lose some market value (causing their yield to rise) but it is extremely unlikely that they would lose as much as other bonds of similar maturity. I certainly wouldn't worry about much loss on 10 year TIPs.

Jdvalle pointed out the that it is theoretically possible for the taxes on TIPs (in a particular year) to exceed the interest payment. Well, yes, but for that to happen there would need to be a hugh increase in inflation that would produce huge losses in conventional bonds (and probably stocks too). TIPs are very liquid, and it would always be convenient and not too costly to liquidate some to supplement the income from the interest.

Ats5g said that a disadvantage of TIPs and high yield bonds is that they haven't been in existence long enough to have established a track record. That's somewhat true, but a person should remember that past performance, while relevant, is not the only consideration in investing for the future. To its credit, FIRECalc contains a provision for "modelling" the hypothetical past performance of TIPs. This model contains the simplifying assumption that the annual return of TIPs would have been the inflation rate plus the interest rate that is specified by the user. It does not account for fluctuations in the market value of the TIPs -- but it is reasonable to believe that this would have been fairly minor. Playing around with FIRECalc, I find that TIPs with a base yield of 2.5% are indicated to support a higher rate of withdrawal (in conjunction with stocks) than the other fixed income options.

With regard to high yield bonds, I have done some research of my own on their performance relative to stocks and bonds over the past 12 years or so and will give some further comments on that after I refine it more. A general comment, though, is that a lot of people are as ignorant about high yield bonds as they are about TIPs. All they know is that they are "risky" and so they write them off. Well, it's true that they are risky, but as an asset class are not nearly as risky as stocks. So it is completely irrational for people who readily invest large amounts of their money in stocks to reject high yield bonds out-of-hand as being "too risky." As with other financial assets, they have their place in a diversified portfolio.

 

[unclemick]

Ted

We're using Vanguard High Yield Corporate as a small portion of our income stream - with no intention of selling or rebalancing. Look forward to your research as emotional backup when interest rate fluctuation moves the NAV in a negative manner.

 

[LA_Newsboy]

Howdy all. Can i get some advice on the idea of holding Fidelity's Inflation-Protected Bond Fund in an IRA?

I have an IRA in Mutual Beacon that grown too large for my comfort and I like the idea of adding a inflation-protected bond component to my portfolio.

Any and all thoughts are welcome. Thanks.

 

[wzd]

I can't comment on the bond funds, since I have not looked at that particular fund. However, be sure you understand the difference between individual bonds and bond funds before buying one. They are quite different. In fact it would be interesting if anyone has studies on the impact of each when studying asset allocation. I have seen a number of Equity/Bond studies, and haven't looked into bond vs. bond fund assumptions in them, but am starting to. I currently have a mostly equity with a bond ladder to get me thru my SEPP program.

Anyone have pointers to studies that compare bond and bond fund asset allocation performance?

Wayne

 

[eytonxav]

http://Re: Asset Allocation Critique

Hi Alec, To your point regarding convexity for GNMAs, what do you think about Scott Burn's view on this issue, see: http://www.dallasnews.com/business/scottburns/columns/2003/stories/070303dnbusburns.160ca.html

I am wondering if the recent restatement of earnings news is a concern for anyone holding a GNMA bond??

Thx,

Doug

 

[Bob_Smith]

However, be sure you understand the difference between individual bonds and bond funds before buying one.   They are quite different.  In fact it would be interesting if anyone has studies on the impact of each when studying asset allocation.

Wayne, I don't know of any studies, but I prefer individual TIPS. My thinking is that with funds, I'm continually exposed to interest rate/inflation risk. So if coupon payments go to almost zero, that's where I'll go too if I'm in a fund. I'm totally exposed. But with individual bonds I can lock in around 2.5% for twenty-five years if I hold to maturity. If rates go up dramatically, I lose, but since I'm planning to hold until maturity, I won't realize any losses at all. My purpose is not to make money, but to lock in a real rate of return to pay the bills, so I think I'm far better off with individual bonds than with a fund. Thoughts anyone? Am I missing something in my thinking?

 

[John Galt]

Hi Bob_Smith!

I understand what you are saying re. bonds, but
2.5% for a 25 year bond You must be kidding.

John Galt

 

[wzd]

Bob - I certainly agree. Note that my post mentioned a bond ladder - which to me implies individual bonds. Funds do behave quite differently, and I think the SWR studies are for individual bonds - but am not sure.

Wayne

 

[Bob_Smith]

Hi Bob_Smith!
I understand what you are saying re. bonds, but 2.5% for a 25 year bond  You must be kidding.

Hi John,

I can see what you're saying, but consider this: bonds have only produced a real return of about 2.2% on average between 1926 and 2001. Post WW2 (1946-2001) they've returned only 1.3%. If you examine the sub-periods within those long stretches you'll see that the last 20 years or so have produced incredible real returns, while from 1946-1981 (35 years) bonds actually produced a NEGATIVE real return. I'm concerned that the future for bonds could look more like 1946-1981 than 1982 to the present. The problem is, I have no way of knowing. TIPS lock in 2.5% plus whatever inflation happens to do. There is no upper limit. There is a floor, however. If I buy TIPS at 2.5%, I can lock in a 2.5% REAL (not total) return, which is better than bonds have done on average since 1926. Plus I am not exposed to market risk if I hold to maturity. If I buy regular bonds at 6%, I'm exposed to whatever the markets dish out. All inflation needs to do is exceed about 3.5% and I'd be losing ground with regular bonds paying 6%, relative to TIPS at 2.5%. And that isn't all that far fetched. Plus, I think returns have a way of regressing to the mean, so a 2.5% real return locked in for 25 years isn't all that bad. With TIPS I may kick myself if regular bonds end up producing a higher real return over the next 25 years, instead of the 2.5% I'm shooting for, but my retirement is relatively secure nevertheless. With regular bonds I could end up at McDonald's. So I could probably do better, as you point out, but I wouldn't know for sure until it's too late to fix it.

 

[Bob_Smith]

Bob - I certainly agree.  Note that my post mentioned a bond ladder - which to me implies individual bonds.  Funds do behave quite differently, and I think the SWR studies are for individual bonds - but am not sure.
Wayne

I agree Wayne. It has been my impression that SWR studies are based on individual bonds, although I haven't really examined the Shiller data. Since he goes back to 1871, there were no funds for much of that time. But I'm not sure how a bond fund would have done compared to individual bonds, and I would like to see studies comparing funds to individual bonds. I wonder if that's even doable. I have used bond funds during the accumulation phase, but as I approach retirement I'm biased against them. It just seems that you control your destiny with individual TIPS vs. getting whatever the market produces in funds.

 

[haha]

It has been my impression that SWR studies are based on individual bonds, although I haven't really examined the Shiller data.

I thought that Prof. Shiller concerned himself with equity market behavior, not bonds. And to the best of my (limited) knowledge, he has never examined SWR.

Have I missed something?

Mikey

 

[Bob_Smith]

I thought that Prof. Shiller concerned himself with equity market behavior, not bonds. And to the best of my (limited) knowledge, he has never examined SWR.

I think the Shiller data includes the S&P Composite Price Index, the PPI, and commercial paper (not bonds) and those figures were used by others to examine SWR. I thought bonds were included too, but it appears that data came from other sources.