Bill Sharpe's preferred portfolio

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
User avatar
watchnerd
Posts: 5981
Joined: Sat Mar 03, 2007 11:18 am
Location: Seattle, WA, USA

Re: Bill Sharpe's preferred portfolio

Post by watchnerd » Wed Jan 15, 2020 12:47 am

pascalwager wrote:
Tue Jan 14, 2020 9:39 pm


For June 30, 2015
Right, which means it's not really correct 5 years later given the market port shifts over time.

You don't need a multicolor graph to say:

'Pick your TIPS allocation at whatever level you like. Put the rest of your funds in the market portfolio'.

Which is what he has said, verbally in YouTube interviews and all that graph is really showing given the market portfolio is fluid over time.
70% Global Market Weight Equities | 15% Long Treasuries 15% short TIPS & cash || RSU + ESPP

pascalwager
Posts: 1803
Joined: Mon Oct 31, 2011 8:36 pm

Re: Bill Sharpe's preferred portfolio

Post by pascalwager » Wed Jan 15, 2020 3:05 am

watchnerd wrote:
Wed Jan 15, 2020 12:47 am
pascalwager wrote:
Tue Jan 14, 2020 9:39 pm


For June 30, 2015
Right, which means it's not really correct 5 years later given the market port shifts over time.

The market portfolio changes every day.

You don't need a multicolor graph to say:

'Pick your TIPS allocation at whatever level you like. Put the rest of your funds in the market portfolio'.2

It's saying versus showing.

Which is what he has said, verbally in YouTube interviews and all that graph is really showing given the market portfolio is fluid over time.
Sharpe is attributing risk in a simple way. TIPS are riskless because they don't have default or inflation risk. The market portfolio is risky because it contains stocks and corporate bonds and non-inflation protected sovereign bonds. Bernstein and Edesess and others do likewise.

The only reason I referred to the graph is because you didn't believe that Sharpe intended for the TIPS to be allocated based on percentage (of TIPS and market portfolio) and regularly rebalanced.
Retired, pension, no SS | Bond funds: TIPS, TBM | Global stocks: total market, large value, small company, emerging market funds

User avatar
watchnerd
Posts: 5981
Joined: Sat Mar 03, 2007 11:18 am
Location: Seattle, WA, USA

Re: Bill Sharpe's preferred portfolio

Post by watchnerd » Wed Jan 15, 2020 3:30 am

pascalwager wrote:
Wed Jan 15, 2020 3:05 am

Sharpe is attributing risk in a simple way. TIPS are riskless because they don't have default or inflation risk. The market portfolio is risky because it contains stocks and corporate bonds and non-inflation protected sovereign bonds. Bernstein and Edesess and others do likewise.

The only reason I referred to the graph is because you didn't believe that Sharpe intended for the TIPS to be allocated based on percentage (of TIPS and market portfolio) and regularly rebalanced.
Fair enough.

But I still find it all pretty underwhelming, given his track record, basically:

1. Hold market weight in stocks.

A fair number of BHers seem to do it independent of Sharpe.

2. Hold market weight in bonds, including international.

Jury is out on if this is really optimal, but Vanguard seems to be moving in that direction.

3. Combination of #1 + #2 = ~60/40 port

That's comforting, but given all the other research that takes the ubiquitous 60/40 as the baseline for 'balanced', it's just icing on the cake for that AA.

4. Have a riskless LMP portfolio in TIPS

Also not conceptually new. Whether it counted as 'bucket investing' our not depends on if one is allowed to balance out of riskless into risky.


It's all pretty common stuff that has been discussed by others, elsewhere.

I don't mean to sound mean, but I expected something more from a guy that has a Nobel prize.
70% Global Market Weight Equities | 15% Long Treasuries 15% short TIPS & cash || RSU + ESPP

DB2
Posts: 900
Joined: Thu Jan 17, 2019 10:07 pm

Re: Bill Sharpe's preferred portfolio

Post by DB2 » Wed Jan 15, 2020 9:42 am

watchnerd wrote:
Wed Jan 15, 2020 3:30 am

I don't mean to sound mean, but I expected something more from a guy that has a Nobel prize.
What else do you expect? There's not much more to it all when it comes down to it (unless one buys into the smart alpha, momentum, etc.) It's like someone saying, "Jack Bogle was brilliant. He started Vanguard and created the index in a time where it was inconceivable given the industry mindset. You mean he only recommends two or at the most three stinking index funds to hold forever while adjusting an AA? That's it?! I expected more from such a smart guy, sheesh."

User avatar
watchnerd
Posts: 5981
Joined: Sat Mar 03, 2007 11:18 am
Location: Seattle, WA, USA

Re: Bill Sharpe's preferred portfolio

Post by watchnerd » Wed Jan 15, 2020 10:53 am

DB2 wrote:
Wed Jan 15, 2020 9:42 am


What else do you expect? There's not much more to it all when it comes down to it (unless one buys into the smart alpha, momentum, etc.) It's like someone saying, "Jack Bogle was brilliant. He started Vanguard and created the index in a time where it was inconceivable given the industry mindset. You mean he only recommends two or at the most three stinking index funds to hold forever while adjusting an AA? That's it?! I expected more from such a smart guy, sheesh."
Say: "Nothing more needs to be said"

Reading all the RISMAT stuff is just a re-hash of known ideas.

There is value in collecting it all, but a task more suitable to a writer of popular books for the layman.
70% Global Market Weight Equities | 15% Long Treasuries 15% short TIPS & cash || RSU + ESPP

palanzo
Posts: 809
Joined: Thu Oct 10, 2019 4:28 pm

Re: Bill Sharpe's preferred portfolio

Post by palanzo » Wed Jan 15, 2020 1:20 pm

watchnerd wrote:
Wed Jan 15, 2020 3:30 am
pascalwager wrote:
Wed Jan 15, 2020 3:05 am

Sharpe is attributing risk in a simple way. TIPS are riskless because they don't have default or inflation risk. The market portfolio is risky because it contains stocks and corporate bonds and non-inflation protected sovereign bonds. Bernstein and Edesess and others do likewise.

The only reason I referred to the graph is because you didn't believe that Sharpe intended for the TIPS to be allocated based on percentage (of TIPS and market portfolio) and regularly rebalanced.
Fair enough.

But I still find it all pretty underwhelming, given his track record, basically:

1. Hold market weight in stocks.

A fair number of BHers seem to do it independent of Sharpe.

2. Hold market weight in bonds, including international.

Jury is out on if this is really optimal, but Vanguard seems to be moving in that direction.

3. Combination of #1 + #2 = ~60/40 port

That's comforting, but given all the other research that takes the ubiquitous 60/40 as the baseline for 'balanced', it's just icing on the cake for that AA.

4. Have a riskless LMP portfolio in TIPS

Also not conceptually new. Whether it counted as 'bucket investing' our not depends on if one is allowed to balance out of riskless into risky.


It's all pretty common stuff that has been discussed by others, elsewhere.

I don't mean to sound mean, but I expected something more from a guy that has a Nobel prize.
Nice summary.

Why would one need to balance out of riskless LMP into risky? With 40 bonds those will provide sufficient funds to rebalance to 60/40 during a downturn.

User avatar
watchnerd
Posts: 5981
Joined: Sat Mar 03, 2007 11:18 am
Location: Seattle, WA, USA

Re: Bill Sharpe's preferred portfolio

Post by watchnerd » Wed Jan 15, 2020 1:29 pm

palanzo wrote:
Wed Jan 15, 2020 1:20 pm


Nice summary.

Why would one need to balance out of riskless LMP into risky? With 40 bonds those will provide sufficient funds to rebalance to 60/40 during a downturn.
Because otherwise you're using a bucket approach, which has a higher SWR failure rate according to the Estrada paper.

That being said, there comes a point where if the COL:portfolio size multiple is high enough (say, 40x), it probably doesn't matter.
70% Global Market Weight Equities | 15% Long Treasuries 15% short TIPS & cash || RSU + ESPP

palanzo
Posts: 809
Joined: Thu Oct 10, 2019 4:28 pm

Re: Bill Sharpe's preferred portfolio

Post by palanzo » Wed Jan 15, 2020 2:13 pm

watchnerd wrote:
Wed Jan 15, 2020 1:29 pm
palanzo wrote:
Wed Jan 15, 2020 1:20 pm


Nice summary.

Why would one need to balance out of riskless LMP into risky? With 40 bonds those will provide sufficient funds to rebalance to 60/40 during a downturn.
Because otherwise you're using a bucket approach, which has a higher SWR failure rate according to the Estrada paper.

That being said, there comes a point where if the COL:portfolio size multiple is high enough (say, 40x), it probably doesn't matter.
My reading of the Estrada paper is that when he models the bucket approach he uses two buckets. Bucket one is bills and bucket 2 is stocks. Of course this underperforms static allocations.

What he does not do is model bucket one as cash equivalents and bucket two as 60/40 as we are discussing above. Perhaps I have misread his paper but as I mentioned in a previous post on Estrada I don't agree with his approach and assumptions.

User avatar
watchnerd
Posts: 5981
Joined: Sat Mar 03, 2007 11:18 am
Location: Seattle, WA, USA

Re: Bill Sharpe's preferred portfolio

Post by watchnerd » Wed Jan 15, 2020 2:46 pm

palanzo wrote:
Wed Jan 15, 2020 2:13 pm


My reading of the Estrada paper is that when he models the bucket approach he uses two buckets. Bucket one is bills and bucket 2 is stocks. Of course this underperforms static allocations.

What he does not do is model bucket one as cash equivalents and bucket two as 60/40 as we are discussing above. Perhaps I have misread his paper but as I mentioned in a previous post on Estrada I don't agree with his approach and assumptions.
The devils are definitely in the details.

Here's the hole I find in Sharpe's concept -- without knowing the relative size of the risky portfolio vs riskless, you don't know if the risky portfolio will generate enough returns to refresh the riskless side.

And by defining 'risky' as stocks + bonds, you're putting a drag on the risk portfolio.

Example:

1,000,000 to allocate.

1. 50% in riskless = $500k in TIPS
2. 50% in risky = $500k in 60/40

Weighted, that gives me:

TIPS = 50% of port
Risky Bonds = 20% of port
Stocks = 30% of port

Total: 30% stocks / 70% bonds

You can call it whatever you want, do mental accounting to call one side 'riskless' and the other side 'risky', but that's what you have.

Whether a 30/70 portfolio has good enough returns to support your retirement needs will vary immensely, but...

1. Calling it what it truly is (30/70) is going to make planning a lot easier when it comes using retirement planning calculators

2. Figuring out withdrawal patterns is a lot easier if you do it in aggregate for the whole port, per the usual methods. And if you're going to use something like variable withdrawals on the whole port, why bother with the mental risky vs riskless accounting?

3. True risk management is a lot more accurate if you view it aggregate. Using the above allocations:

Risk Decomposition:

Riskless:
31.32%: TIPS
Risky:
6.49%: Global Bonds (USD Hedged)
27.04%: US Stock Market
35.16%: Global ex-US Stock Market

https://www.portfoliovisualizer.com/bac ... tion4_1=15


That reality is obscured by the 'risky' vs 'riskless' mental accounting.


4. If the argument in favor of holding the global market weights in everything is that it avoids having to choose how to tilt, the question of what % of total assets should be allocated to TIPS removes that. Doing so *does* require an active decision, a tilt, and thus undermines the whole premise to begin with.

In the above example, a 30/70 port is *not* the global market weight port, no matter how much you might take comfort in pretending that part of it is.
70% Global Market Weight Equities | 15% Long Treasuries 15% short TIPS & cash || RSU + ESPP

palanzo
Posts: 809
Joined: Thu Oct 10, 2019 4:28 pm

Re: Bill Sharpe's preferred portfolio

Post by palanzo » Wed Jan 15, 2020 3:34 pm

watchnerd wrote:
Wed Jan 15, 2020 2:46 pm
palanzo wrote:
Wed Jan 15, 2020 2:13 pm


My reading of the Estrada paper is that when he models the bucket approach he uses two buckets. Bucket one is bills and bucket 2 is stocks. Of course this underperforms static allocations.

What he does not do is model bucket one as cash equivalents and bucket two as 60/40 as we are discussing above. Perhaps I have misread his paper but as I mentioned in a previous post on Estrada I don't agree with his approach and assumptions.
The devils are definitely in the details.

Here's the hole I find in Sharpe's concept -- without knowing the relative size of the risky portfolio vs riskless, you don't know if the risky portfolio will generate enough returns to refresh the riskless side.

And by defining 'risky' as stocks + bonds, you're putting a drag on the risk portfolio.

Example:

1,000,000 to allocate.

1. 50% in riskless = $500k in TIPS
2. 50% in risky = $500k in 60/40

Weighted, that gives me:

TIPS = 50% of port
Risky Bonds = 20% of port
Stocks = 30% of port

Total: 30% stocks / 70% bonds

You can call it whatever you want, do mental accounting to call one side 'riskless' and the other side 'risky', but that's what you have.

Whether a 30/70 portfolio has good enough returns to support your retirement needs will vary immensely, but...

1. Calling it what it truly is (30/70) is going to make planning a lot easier when it comes using retirement planning calculators

2. Figuring out withdrawal patterns is a lot easier if you do it in aggregate for the whole port, per the usual methods. And if you're going to use something like variable withdrawals on the whole port, why bother with the mental risky vs riskless accounting?

3. True risk management is a lot more accurate if you view it aggregate. Using the above allocations:

Risk Decomposition:

Riskless:
31.32%: TIPS
Risky:
6.49%: Global Bonds (USD Hedged)
27.04%: US Stock Market
35.16%: Global ex-US Stock Market

https://www.portfoliovisualizer.com/bac ... tion4_1=15


That reality is obscured by the 'risky' vs 'riskless' mental accounting.


4. If the argument in favor of holding the global market weights in everything is that it avoids having to choose how to tilt, the question of what % of total assets should be allocated to TIPS removes that. Doing so *does* require an active decision, a tilt, and thus undermines the whole premise to begin with.

In the above example, a 30/70 port is *not* the global market weight port, no matter how much you might take comfort in pretending that part of it is.
I agree with all your points. The TIPS represent about 12 years of withdrawals at 4%. Personally I would not keep so much in the LMP although others might. If one models an overall 60/40 allocation with a TIPS LMP then the risk profile is not as bad as you have shown.

Estrada was talking about Bucket one representing 1-5 years of withdrawals. As you say, the devil is definitely in the details.

Personally I am staying with a market weight 60/40 portfolio.

User avatar
watchnerd
Posts: 5981
Joined: Sat Mar 03, 2007 11:18 am
Location: Seattle, WA, USA

Re: Bill Sharpe's preferred portfolio

Post by watchnerd » Wed Jan 15, 2020 3:43 pm

palanzo wrote:
Wed Jan 15, 2020 3:34 pm


I agree with all your points. The TIPS represent about 12 years of withdrawals at 4%. Personally I would not keep so much in the LMP although others might. If one models an overall 60/40 allocation with a TIPS LMP then the risk profile is not as bad as you have shown.

Estrada was talking about Bucket one representing 1-5 years of withdrawals. As you say, the devil is definitely in the details.

Personally I am staying with a market weight 60/40 portfolio.

I'm 70/30 now. My equities are market weight. I'll probably move to 60/40 in about 2 years.

Personally, I'm just going to count my LMP as part of the 40% bonds.

We already do this, as we have a short TIPS / cash allocation as part of our bond allocation.
70% Global Market Weight Equities | 15% Long Treasuries 15% short TIPS & cash || RSU + ESPP

palanzo
Posts: 809
Joined: Thu Oct 10, 2019 4:28 pm

Re: Bill Sharpe's preferred portfolio

Post by palanzo » Wed Jan 15, 2020 4:01 pm

watchnerd wrote:
Wed Jan 15, 2020 3:43 pm
palanzo wrote:
Wed Jan 15, 2020 3:34 pm


I agree with all your points. The TIPS represent about 12 years of withdrawals at 4%. Personally I would not keep so much in the LMP although others might. If one models an overall 60/40 allocation with a TIPS LMP then the risk profile is not as bad as you have shown.

Estrada was talking about Bucket one representing 1-5 years of withdrawals. As you say, the devil is definitely in the details.

Personally I am staying with a market weight 60/40 portfolio.

I'm 70/30 now. My equities are market weight. I'll probably move to 60/40 in about 2 years.

Personally, I'm just going to count my LMP as part of the 40% bonds.

We already do this, as we have a short TIPS / cash allocation as part of our bond allocation.
Likewise. I count my LMP as part of the 40% bonds. Do you hold the short TIPS in taxable?

pascalwager
Posts: 1803
Joined: Mon Oct 31, 2011 8:36 pm

Re: Bill Sharpe's preferred portfolio

Post by pascalwager » Wed Jan 15, 2020 4:06 pm

I don't find Sharpe referring to the TIPS portfolio as an LMP. It just seems to be an overall risk reduction component that includes inflation risk reduction. Also, Sharpe is only targeting "many retirees", not younger investors.

And anyone is free to put all of their retiree savings into the high-performing T. Rowe Price New Horizons Fund (small/mid growth) which one former BH member actually did back in 2012 (he didn't like cash or bonds because of inflation risk and low-yields).
Retired, pension, no SS | Bond funds: TIPS, TBM | Global stocks: total market, large value, small company, emerging market funds

User avatar
watchnerd
Posts: 5981
Joined: Sat Mar 03, 2007 11:18 am
Location: Seattle, WA, USA

Re: Bill Sharpe's preferred portfolio

Post by watchnerd » Wed Jan 15, 2020 4:17 pm

pascalwager wrote:
Wed Jan 15, 2020 4:06 pm
I don't find Sharpe referring to the TIPS portfolio as an LMP. It just seems to be an overall risk reduction component that includes inflation risk reduction.
I think it's completely open to interpretation what the TIPS part is.

He basically punts on the question and said 'TBD between you and your advisor'.

If the advisor likes LMP, it will probably be LMP oriented. If not, it would be smaller, just a 'buffer'.
70% Global Market Weight Equities | 15% Long Treasuries 15% short TIPS & cash || RSU + ESPP

User avatar
CyberBob
Posts: 3315
Joined: Tue Feb 20, 2007 2:53 pm

Re: Bill Sharpe's preferred portfolio

Post by CyberBob » Thu Feb 27, 2020 11:41 am

31 January 2020 FTSE numbers (in millions of USD):

31,312,957 - U.S. stocks - FTSE Global All-Cap index, U.S. breakdown
24,996,509 - ex U.S. stocks - FTSE Global All-Cap index minus U.S.
23,147,130 - U.S. bonds - FTSE US Broad Investment-Grade index
19,961,800 - ex U.S. bonds - FTSE World Broad Investment-Grade index minus U.S.

So that equates to allocation percentages of:

31.50% US stocks (VTSAX,FSKAX,VTI,ITOT,SCHB,SPTM)
25.14% ex US stocks (VTIAX,FTIHX,VXUS,IXUS,ACWX,CWI)
23.28% US bonds (VBTLX,FXNAX,BND,AGG,SCHZ,SPAB)
20.08% ex US bonds (VTABX,FBIIX,BNDX,IAGG)

Or, alternatively:

56.64% world stocks (VT,VTWAX)
43.36% world bonds (BNDW)

User avatar
CyberBob
Posts: 3315
Joined: Tue Feb 20, 2007 2:53 pm

Re: Bill Sharpe's preferred portfolio

Post by CyberBob » Fri Mar 27, 2020 2:08 pm

29 February 2020 FTSE numbers (in millions of USD):

28,695,054 - U.S. stocks - FTSE Global All-Cap index, U.S. breakdown
22,955,022 - ex U.S. stocks - FTSE Global All-Cap index minus U.S.
23,470,130 - U.S. bonds - FTSE US Broad Investment-Grade index
20,046,610 - ex U.S. bonds - FTSE World Broad Investment-Grade index minus U.S.

So that equates to allocation percentages of:

30.15% US stocks (VTSAX,FSKAX,VTI,ITOT,SCHB,SPTM)
24.12% ex US stocks (VTIAX,FTIHX,VXUS,IXUS,ACWX,CWI)
24.66% US bonds (VBTLX,FXNAX,BND,AGG,SCHZ,SPAB)
21.06% ex US bonds (VTABX,FBIIX,BNDX,IAGG)

Or:

54.27% world stocks (VT,VTWAX)
45.73% world bonds (BNDW)
Last edited by CyberBob on Thu Apr 30, 2020 1:21 pm, edited 1 time in total.

User avatar
asset_chaos
Posts: 1815
Joined: Tue Feb 27, 2007 6:13 pm
Location: Melbourne

Re: Bill Sharpe's preferred portfolio

Post by asset_chaos » Fri Mar 27, 2020 3:53 pm

CyberBob wrote:
Fri Mar 27, 2020 2:08 pm
The allocations over time seem pretty consistent. You can see just a little dip in the December 2018 and February 2020 stock percentages. It will be interesting to see what the March Corona Virus-affected numbers will look like.
I set up a M* portfolio of the global markets portfolio to use as a benchmark. It currently says around 53:47 stocks:bonds.
Regards, | | Guy

User avatar
CyberBob
Posts: 3315
Joined: Tue Feb 20, 2007 2:53 pm

Re: Bill Sharpe's preferred portfolio

Post by CyberBob » Mon Mar 30, 2020 1:55 pm

Two interesting links:

William Sharpe blog, where he talks about the World Bond-Stock portfolio and the corresponding Vanguard funds used to build it.
https://retirementincomeanalysis.blogspot.com/?m=1

FTSE Russell calculator for Sharpe's Adaptive Asset Allocation.
https://research.ftserussell.com/Analyt ... Home/Index

User avatar
CyberBob
Posts: 3315
Joined: Tue Feb 20, 2007 2:53 pm

Re: Bill Sharpe's preferred portfolio

Post by CyberBob » Thu Apr 30, 2020 1:41 pm

31 March 2020 FTSE numbers (in millions of USD):

24,645,470 - U.S. stocks - FTSE Global All-Cap index, U.S. breakdown
19,522,099 - ex U.S. stocks - FTSE Global All-Cap index minus U.S.
23,296,690 - U.S. bonds - FTSE US Broad Investment-Grade index
19,344,680 - ex U.S. bonds - FTSE World Broad Investment-Grade index minus U.S.

So that equates to allocation percentages of:

28.39% US stocks (VTSAX,FSKAX,VTI,ITOT,SCHB,SPTM)
22.49% ex US stocks (VTIAX,FTIHX,VXUS,IXUS,ACWX,CWI)
26.84% US bonds (VBTLX,FXNAX,BND,AGG,SCHZ,SPAB)
22.28% ex US bonds (VTABX,FBIIX,BNDX,IAGG)

Or:

50.88% world stocks (VT,VTWAX)
49.12% world bonds (BNDW)

CFM300
Posts: 1953
Joined: Sat Oct 27, 2007 5:13 am

Re: Bill Sharpe's preferred portfolio

Post by CFM300 » Sat May 02, 2020 1:42 pm

I just wanted to thank those who continue to update this thread. Very interesting and useful. :beer

User avatar
CyberBob
Posts: 3315
Joined: Tue Feb 20, 2007 2:53 pm

Re: Bill Sharpe's preferred portfolio

Post by CyberBob » Sat Jun 13, 2020 9:28 am

31 May 2020 FTSE numbers (in millions of USD):

29,290,091 - U.S. stocks - FTSE Global All-Cap index, U.S. breakdown
21,821,490 - ex U.S. stocks - FTSE Global All-Cap index minus U.S.
23,080,940 - U.S. bonds - FTSE US Broad Investment-Grade index
20,431,910 - ex U.S. bonds - FTSE World Broad Investment-Grade index minus U.S.

So that equates to allocation percentages of:

30.95% US stocks (VTSAX,FSKAX,VTI,ITOT,SCHB,SPTM)
23.06% ex US stocks (VTIAX,FTIHX,VXUS,IXUS,ACWX,CWI)
24.39% US bonds (VBTLX,FXNAX,BND,AGG,SCHZ,SPAB)
21.59% ex US bonds (VTABX,FBIIX,BNDX,IAGG)

Or:

54.02% world stocks (VT,VTWAX)
45.98% world bonds (BNDW)

pascalwager
Posts: 1803
Joined: Mon Oct 31, 2011 8:36 pm

Re: Bill Sharpe's preferred portfolio

Post by pascalwager » Tue Jun 16, 2020 6:32 pm

Here's a link to an article about RISMAT by Joe Tomlinson (financial planner). He refers to it as a financial economics textbook and lists some asset allocation practices which might appear suspect to someone inclined toward total market investing. The linked article also provides a couple of interesting Sharpe interview links.

https://www.advisorperspectives.com/art ... t-planning

And thanks to CyberBob for continuing the WBS portfolio AA updates.

User avatar
Maestro G
Posts: 114
Joined: Fri Aug 03, 2007 7:55 pm
Location: San Francisco

Re: Bill Sharpe's preferred portfolio

Post by Maestro G » Thu Jun 18, 2020 2:49 am

CyberBob wrote:
Sat Jun 13, 2020 9:28 am
31 May 2020 FTSE numbers (in millions of USD):

29,290,091 - U.S. stocks - FTSE Global All-Cap index, U.S. breakdown
21,821,490 - ex U.S. stocks - FTSE Global All-Cap index minus U.S.
23,080,940 - U.S. bonds - FTSE US Broad Investment-Grade index
20,431,910 - ex U.S. bonds - FTSE World Broad Investment-Grade index minus U.S.

So that equates to allocation percentages of:

30.95% US stocks (VTSAX,FSKAX,VTI,ITOT,SCHB,SPTM)
23.06% ex US stocks (VTIAX,FTIHX,VXUS,IXUS,ACWX,CWI)
24.39% US bonds (VBTLX,FXNAX,BND,AGG,SCHZ,SPAB)
21.59% ex US bonds (VTABX,FBIIX,BNDX,IAGG)

Or:

54.02% world stocks (VT,VTWAX)
45.98% world bonds (BNDW)
Hi CyberBob,

Perhaps I am confused, but it appears to me from the FTSE AA Calculator (bottom text) https://research.ftserussell.com/Analyt ... Calculator, that for the end of May the MCAP percentages were exactly the opposite:

54% World Bond
46% World Stock

What am I missing here?

Thanks!

Maestro G
Everything should be made as simple as possible, but no simpler. Most daily market noise is "a tale told by an idiot, full of sound and fury, signifying nothing.”

User avatar
asset_chaos
Posts: 1815
Joined: Tue Feb 27, 2007 6:13 pm
Location: Melbourne

Re: Bill Sharpe's preferred portfolio

Post by asset_chaos » Fri Jun 19, 2020 12:03 am

I believe cyberbob uses the FTSE World BIG bond index and the FTSE US BIG bond index monthly fact sheets for the market value of US and ex-US bonds. The links go to the most recent month's sheet, which are both dated 31 May 2020. Similarly for stocks is the FTSE Global All Cap Stock index fact sheet, which is dated 29 May 2020. The numbers cyberbob quotes are the correct numbers for end of May according to those sheets. And a quick math check on percentages confirms those are right for the dollar figures too. But I do see what you mean: that FTSE calculator page certainly does say the global stock:bond ratio at end of May is 46:54.

I guess the question is what's that FTSE calculator page really showing to be so markedly different from other FTSE data? As the BIG in the bond indices stands for Broad Investment Grade, could the calculator be including the non-investment grade bond market, and is it large enough to make that much difference? All the homepage for the calculator says is
Adaptive Asset Allocation Policy weights are calculated by FTSE Analytics based on data sourced from FTSE Russell and Thomson Reuters Datastream for bonds. Bond data generally reflects Datastream All Lives Total Market government bond data combined with market-leading corporate bond data.
which may imply they're using different bond data from their own BIG bond index data. From a brief description of the Thomson Reuters Datastream database, for bonds they say one can
Access a wide coverage of government and corporate bonds, warrants, convertibles, including global international and domestic issues in an increasing range of major currencies. In total, over 500,000 active bonds and convertibles and 1.4 million active warrants across 72 markets are available.
As the FTSE World BIG fact sheet says it covers around 13,000 investment grade only bond issues, and since I don't think the World BIG index includes warrents and convertibles, it may be the FTSE calculator uses a much more expansive definition of bonds to include convertibles, warrents, and non-investment grade bonds. Perhaps that is where the difference is.
Regards, | | Guy

User avatar
Maestro G
Posts: 114
Joined: Fri Aug 03, 2007 7:55 pm
Location: San Francisco

Re: Bill Sharpe's preferred portfolio

Post by Maestro G » Sat Jun 20, 2020 2:22 am

asset_chaos wrote:
Fri Jun 19, 2020 12:03 am
I believe cyberbob uses the FTSE World BIG bond index and the FTSE US BIG bond index monthly fact sheets for the market value of US and ex-US bonds. The links go to the most recent month's sheet, which are both dated 31 May 2020. Similarly for stocks is the FTSE Global All Cap Stock index fact sheet, which is dated 29 May 2020. The numbers cyberbob quotes are the correct numbers for end of May according to those sheets. And a quick math check on percentages confirms those are right for the dollar figures too. But I do see what you mean: that FTSE calculator page certainly does say the global stock:bond ratio at end of May is 46:54.

I guess the question is what's that FTSE calculator page really showing to be so markedly different from other FTSE data? As the BIG in the bond indices stands for Broad Investment Grade, could the calculator be including the non-investment grade bond market, and is it large enough to make that much difference? All the homepage for the calculator says is
Adaptive Asset Allocation Policy weights are calculated by FTSE Analytics based on data sourced from FTSE Russell and Thomson Reuters Datastream for bonds. Bond data generally reflects Datastream All Lives Total Market government bond data combined with market-leading corporate bond data.
which may imply they're using different bond data from their own BIG bond index data. From a brief description of the Thomson Reuters Datastream database, for bonds they say one can
Access a wide coverage of government and corporate bonds, warrants, convertibles, including global international and domestic issues in an increasing range of major currencies. In total, over 500,000 active bonds and convertibles and 1.4 million active warrants across 72 markets are available.
As the FTSE World BIG fact sheet says it covers around 13,000 investment grade only bond issues, and since I don't think the World BIG index includes warrents and convertibles, it may be the FTSE calculator uses a much more expansive definition of bonds to include convertibles, warrents, and non-investment grade bonds. Perhaps that is where the difference is.
Perhaps, but it’s curious to me that the percentages are exactly reversed for the asset classes :?: Could be just a coincidence, but odd.

Additionally, it is those exact percentages and this calculator that William Sharp references in his blog.

Thanks for your response Asset Chaos.

Maestro G
Everything should be made as simple as possible, but no simpler. Most daily market noise is "a tale told by an idiot, full of sound and fury, signifying nothing.”

User avatar
CyberBob
Posts: 3315
Joined: Tue Feb 20, 2007 2:53 pm

Re: Bill Sharpe's preferred portfolio

Post by CyberBob » Sat Jun 20, 2020 9:47 am

Maestro G wrote:
Thu Jun 18, 2020 2:49 am
CyberBob wrote:
Sat Jun 13, 2020 9:28 am
31 May 2020 FTSE numbers (in millions of USD):

29,290,091 - U.S. stocks - FTSE Global All-Cap index, U.S. breakdown
21,821,490 - ex U.S. stocks - FTSE Global All-Cap index minus U.S.
23,080,940 - U.S. bonds - FTSE US Broad Investment-Grade index
20,431,910 - ex U.S. bonds - FTSE World Broad Investment-Grade index minus U.S.

So that equates to allocation percentages of:

30.95% US stocks (VTSAX,FSKAX,VTI,ITOT,SCHB,SPTM)
23.06% ex US stocks (VTIAX,FTIHX,VXUS,IXUS,ACWX,CWI)
24.39% US bonds (VBTLX,FXNAX,BND,AGG,SCHZ,SPAB)
21.59% ex US bonds (VTABX,FBIIX,BNDX,IAGG)

Or:

54.02% world stocks (VT,VTWAX)
45.98% world bonds (BNDW)
Hi CyberBob,

Perhaps I am confused, but it appears to me from the FTSE AA Calculator (bottom text) https://research.ftserussell.com/Analyt ... Calculator, that for the end of May the MCAP percentages were exactly the opposite:

54% World Bond
46% World Stock

What am I missing here?

Thanks!

Maestro G
I think asset_chaos' post is correct, and that the most likely cause of the disparity is different indexing methodology.

It seems from Sharpe's original video interview that the four indexes he originally mentioned four years ago aren't necessarily perfect, but that they were simply the best freely-available information at the time. He even mentions that some aspects of the bond indexes are uncertain and that it's not obvious whether they are even hedged or not.

It looks like Sharpe is advocating the FTSE AAAP calculator numbers now as more exacting. In this blog post it looks like he finalizes his methodology, simplifying the funds needed down to two; Total World Stock (VT) and Total World Bond (BNDW), getting the most accurate available world bond/stock breakdown using the FTSE AAAP numbers, and even showing a very handy and simple spreadsheet example of how to adjust for any delay between the AAAP numbers and current valuations.

So it looks like I may no longer have to post numbers, as anyone wanting to follow Sharpe's World Bond/Stock can find all the needed information in this blog post.

User avatar
Maestro G
Posts: 114
Joined: Fri Aug 03, 2007 7:55 pm
Location: San Francisco

Re: Bill Sharpe's preferred portfolio

Post by Maestro G » Sat Jun 20, 2020 11:12 am

CyberBob wrote:
Sat Jun 20, 2020 9:47 am
Maestro G wrote:
Thu Jun 18, 2020 2:49 am
CyberBob wrote:
Sat Jun 13, 2020 9:28 am
31 May 2020 FTSE numbers (in millions of USD):

29,290,091 - U.S. stocks - FTSE Global All-Cap index, U.S. breakdown
21,821,490 - ex U.S. stocks - FTSE Global All-Cap index minus U.S.
23,080,940 - U.S. bonds - FTSE US Broad Investment-Grade index
20,431,910 - ex U.S. bonds - FTSE World Broad Investment-Grade index minus U.S.

So that equates to allocation percentages of:

30.95% US stocks (VTSAX,FSKAX,VTI,ITOT,SCHB,SPTM)
23.06% ex US stocks (VTIAX,FTIHX,VXUS,IXUS,ACWX,CWI)
24.39% US bonds (VBTLX,FXNAX,BND,AGG,SCHZ,SPAB)
21.59% ex US bonds (VTABX,FBIIX,BNDX,IAGG)

Or:

54.02% world stocks (VT,VTWAX)
45.98% world bonds (BNDW)
Hi CyberBob,

Perhaps I am confused, but it appears to me from the FTSE AA Calculator (bottom text) https://research.ftserussell.com/Analyt ... Calculator, that for the end of May the MCAP percentages were exactly the opposite:

54% World Bond
46% World Stock

What am I missing here?

Thanks!

Maestro G
I think asset_chaos' post is correct, and that the most likely cause of the disparity is different indexing methodology.

It seems from Sharpe's original video interview that the four indexes he originally mentioned four years ago aren't necessarily perfect, but that they were simply the best freely-available information at the time. He even mentions that some aspects of the bond indexes are uncertain and that it's not obvious whether they are even hedged or not.

It looks like Sharpe is advocating the FTSE AAAP calculator numbers now as more exacting. In this blog post it looks like he finalizes his methodology, simplifying the funds needed down to two; Total World Stock (VT) and Total World Bond (BNDW), getting the most accurate available world bond/stock breakdown using the FTSE AAAP numbers, and even showing a very handy and simple spreadsheet example of how to adjust for any delay between the AAAP numbers and current valuations.

So it looks like I may no longer have to post numbers, as anyone wanting to follow Sharpe's World Bond/Stock can find all the needed information in this blog post.
Hi Cyber Bob,

Yes, exactly, that’s the post I was referring to and thanks again for confirming. It is interesting to me that there was that wide a swing or divergence (8%) in a fairly limited amount of time, the recent crash and bounce notwithstanding. In any case, thanks again for providing the numbers and calculations to date. It does seem now that those will be much more easily referenced going forward.

BTW, I have always loved your avatar! I assume that cute cat is yours?

Best,

Maestro G
Everything should be made as simple as possible, but no simpler. Most daily market noise is "a tale told by an idiot, full of sound and fury, signifying nothing.”

User avatar
CyberBob
Posts: 3315
Joined: Tue Feb 20, 2007 2:53 pm

Re: Bill Sharpe's preferred portfolio

Post by CyberBob » Sat Jun 20, 2020 12:25 pm

Maestro G wrote:
Sat Jun 20, 2020 11:12 am
BTW, I have always loved your avatar! I assume that cute cat is yours?

Best,

Maestro G
Her name is Gracie. She’s gray (hence the name) and white with a perfect milk mustache. She was born in a barn on a farm in 2007 and rescued/catnapped at four weeks old by a friend of my mother when the mean farmer said he had enough barn cats and threatened to drown her in a bucket. She’s been my constant companion ever since. She’s also a great dog, as unlike most cats she always comes running when I call her name. That close-up photo is how I see her most days as she likes to sit between me and my computer keyboard and help me type. :D

User avatar
Maestro G
Posts: 114
Joined: Fri Aug 03, 2007 7:55 pm
Location: San Francisco

Re: Bill Sharpe's preferred portfolio

Post by Maestro G » Sat Jun 20, 2020 12:52 pm

CyberBob wrote:
Sat Jun 20, 2020 12:25 pm
Maestro G wrote:
Sat Jun 20, 2020 11:12 am
BTW, I have always loved your avatar! I assume that cute cat is yours?

Best,

Maestro G
Her name is Gracie. She’s gray (hence the name) and white with a perfect milk mustache. She was born in a barn on a farm in 2007 and rescued/catnapped at four weeks old by a friend of my mother when the mean farmer said he had enough barn cats and threatened to drown her in a bucket. She’s been my constant companion ever since. She’s also a great dog, as unlike most cats she always comes running when I call her name. That close-up photo is how I see her most days as she likes to sit between me and my computer keyboard and help me type. :D
A wonderful companion and a nice end to the story: how someone could kill or even think about killing an innocent animal like that is beyond me. Also, another argument against the myth that most cats are indifferent and standoffish to their “owners.” When a cat behaves that way, it usually says a lot more about the owner than the cat.

Hope you and Gracie stay well! :beer

Maestro G
Everything should be made as simple as possible, but no simpler. Most daily market noise is "a tale told by an idiot, full of sound and fury, signifying nothing.”

User avatar
Taylor Larimore
Advisory Board
Posts: 29542
Joined: Tue Feb 27, 2007 8:09 pm
Location: Miami FL

Re: Bill Sharpe's preferred portfolio

Post by Taylor Larimore » Sat Jun 20, 2020 4:28 pm

CyberBob wrote:
Sat Jun 20, 2020 12:25 pm
Maestro G wrote:
Sat Jun 20, 2020 11:12 am
BTW, I have always loved your avatar! I assume that cute cat is yours?

Best,

Maestro G
Her name is Gracie. She’s gray (hence the name) and white with a perfect milk mustache. She was born in a barn on a farm in 2007 and rescued/catnapped at four weeks old by a friend of my mother when the mean farmer said he had enough barn cats and threatened to drown her in a bucket. She’s been my constant companion ever since. She’s also a great dog, as unlike most cats she always comes running when I call her name. That close-up photo is how I see her most days as she likes to sit between me and my computer keyboard and help me type. :D
CyberBob:

Lovely story!

Thank you and best wishes.
Taylor
Jack Bogle's Words of Wisdom: "The Vanguard mutual structure, where we operate the funds for the benefit of shareholders and at cost, is one of the great ideas of all times."
"Simplicity is the master key to financial success." -- Jack Bogle

pascalwager
Posts: 1803
Joined: Mon Oct 31, 2011 8:36 pm

Re: Bill Sharpe's preferred portfolio

Post by pascalwager » Tue Jun 23, 2020 6:31 am

CyberBob wrote:
Sat Jun 20, 2020 12:25 pm
Maestro G wrote:
Sat Jun 20, 2020 11:12 am
BTW, I have always loved your avatar! I assume that cute cat is yours?

Best,

Maestro G
Her name is Gracie. She’s gray (hence the name) and white with a perfect milk mustache. She was born in a barn on a farm in 2007 and rescued/catnapped at four weeks old by a friend of my mother when the mean farmer said he had enough barn cats and threatened to drown her in a bucket. She’s been my constant companion ever since. She’s also a great dog, as unlike most cats she always comes running when I call her name. That close-up photo is how I see her most days as she likes to sit between me and my computer keyboard and help me type. :D
Wonderful story. I've always been "curious as a cat" about the gentle and pleasing avatar. My own cat, Pennypacker, likes to perch behind my head when I'm seated on the swivel chair at the computer (and reading Bill Sharpe's blog). She was also rescued as a kitten--from the underfloor insulation in a crawlspace under a house.

I notice that the AAAP calculator consistently provides a higher bond/stock ratio than produced by Sharpe's earlier methodology.
Retired, pension, no SS | Bond funds: TIPS, TBM | Global stocks: total market, large value, small company, emerging market funds

djm2001
Posts: 3
Joined: Fri May 29, 2020 6:23 am

Re: Bill Sharpe's preferred portfolio

Post by djm2001 » Wed Jun 24, 2020 11:58 am

I question Sharpe's latest suggestion to use the FTSE AAAP weights rather than the original method of using the market caps of the indexes that the funds actually track.

Using market-cap of X but then buying Y means that
  • we're assuming that X and Y have similar fundamentals.  In this case, we're substituting investment grade bonds for non-investment grade bonds.  Seems questionable.
  • we introduce tracking error, lose the self-balancing property, and increase the need to rebalance.
I plan to stick with using the index market caps method, and hope that the indexes will improve their coverage over time (rather than trying to access the missing coverage via potentially invalid substitutes).

P.S., Of course, I'm assuming here that FTSE's bond indexes mirror Bloomberg Barclay's corresponding indexes more closely than the AAAP data. This seems to be the case base on asset_chaos's post.

djm2001
Posts: 3
Joined: Fri May 29, 2020 6:23 am

Re: Bill Sharpe's preferred portfolio

Post by djm2001 » Fri Jun 26, 2020 11:13 pm

I made the following using Google Sheets to track the current values of Sharpe's World Bond-Stock Fund proxies:
The backing spreadsheet [edit: fixed link] projects the factsheet marketcaps forward to current time using current price and dividend history. This requires occasional manual update when new factsheets are released, and when dividends occur. I plan to keep this up-to-date since I plan to follow this portfolio myself.

The backing sheet also tracks some bond assets that are not covered by VBTLX -- namely, TIPS and munis. (The 2-Fund and 4-Fund allocations linked above ignore those.)

Note that, for VTSAX, I use the CRSP US Total Market Index (and project it forward) since that is what VTSAX tracks. The CRSP index covers 3.4K+ companies, while the FTSE US All Cap Index covers only 1.8K companies. I found that it gives a slightly higher market cap than using FTSE US All Cap Index, as expected.
Last edited by djm2001 on Sun Jul 05, 2020 7:44 am, edited 2 times in total.

User avatar
spdoublebass
Posts: 668
Joined: Thu Apr 27, 2017 10:04 pm
Location: NY

Re: Bill Sharpe's preferred portfolio

Post by spdoublebass » Sat Jun 27, 2020 12:10 am

djm2001 wrote:
Fri Jun 26, 2020 11:13 pm
I made the following using Google Sheets to track the current values of Sharpe's World Bond-Stock Fund proxies:
The backing spreadsheet projects the factsheet marketcaps forward to current time using current price and dividend history. This requires occasional manual update when new factsheets are released, and when dividends occur. I plan to keep this up-to-date since I plan to follow this portfolio myself.

The backing sheet also tracks some bond assets that are not covered by VBTLX -- namely, TIPS and munis. (The 2-Fund and 4-Fund allocations linked above ignore those.)

Note that, for VTSAX, I use the CRSP US Total Market Index (and project it forward) since that is what VTSAX tracks. The CRSP index covers 3.4K+ companies, while the FTSE US All Cap Index covers only 1.8K companies. I found that it gives a slightly higher market cap than using FTSE US All Cap Index, as expected.

What about EM bonds? Do you just take what is inside of BNDX? Or do you look at the different Market Values inside of the yield book indices?
I'm trying to think, but nothing happens

djm2001
Posts: 3
Joined: Fri May 29, 2020 6:23 am

Re: Bill Sharpe's preferred portfolio

Post by djm2001 » Sat Jun 27, 2020 5:15 am

TL;DR: Yes, I just take what's in BND + BNDX for EM bonds and high-yield corporate bonds.

Long version:

To start with, BND and BNDX's indexes lack some coverage for:
  • high-yield corporate bonds ("junk bonds"), indexed by S&P US High-Yield Corporate Bond Index; market cap of $1.54T as of May 29.
  • emerging market sovereign bonds.  These are split by issuing currency into:
    • USD-issued EM sovereign bonds, indexed by FTSE's EMUSDGBI; market cap of $0.7T as of Mar 31.
    • local-currency EM sovereign bonds, indexed by FTSE's EMGBI - EMUSDGBI.; market cap of $2T as of Mar 31.
[edit: To put the above market caps in perspective, note that the four-fund proxy's total index market cap is $96T currently; and is $98T if you also include TIPS and muni bonds.]

I originally found indexes and funds to track the above, but eventually dropped them after finding significant overlap with BND / BNDX. I apply a "no overlap" criterion to the set of funds used to "cover" the market.  I'd rather leave the gaps unfilled than try to fill them with potentially invalid substitutes (i.e., funds with different fundamentals) and/or introduce significant double counting of assets.

According to Vanguard's "Emerging Market Bonds -- Beyond the headlines" whitepaper, the overlaps as of 2012 were as follows:
  • 22% of USD-issued EM bonds were already in BND.  This intersection accounted for 2% of BND.
  • 28% of local-currency EM bonds were already in BNDX.  This intersection accounted for 6% of BNDX.
The landscape has probably changed since 2012, but the point remains.  The reason for the overlap comes down to BND/BNDX broad bond indexes' criteria for a bond to be considered "investment grade", and the fact that some EM bonds meet that bar.

It should be pointed out that EM sovereign bond indices have their own criteria that eliminate smaller bonds from the tail (e.g., minimum market size, minimum issue size).  At some point, one has to give up on the tail.

I have not tried to figure out the overlap between BND and high-yield corporate bonds, but I assume some overlap exists from high-yield corporate bonds that meet the AGG's bar for "investment grade". [edit: I am probably wrong about this. Barclays' US Agg and High-Yield bond indexes are separated by bond ratings, so probably don't overlap.]

On that note, I wish there was a way to easily find the overlap between two given bond ETFs.  I know of this tool, but it doesn't work for bond funds.

Post Reply