A more volatile PP, part two

[Pointedstick]

I've posted on this subject before but now, armed with more data, figured I'd give it another go.

Something has always nagged at me about the PP. It has one growth asset (total stock index), two flight-to-safety assets (gold and 30 year treasury bonds), and one stability asset (cash). Unfortunately, stocks aren't as volatile as bonds and especially gold, so when money that flocked to the gold and bonds leaves and goes to stocks, one less-volatile asset has to cover the losses from two more-volatile assets. I think stocks don't have the punch to hold the portfolio in this case. As we've seen recently, they only provide enough to really break even, if that. And when the stock market is blowing down the barn doors, we aren't able to take as much advantage of it.

I've always wondered whether a broad stock market index just didn't have enough volatility to it. Armed with Sophie's excellent spreadsheet, I decided to see what happened if we replaced the total stock index component with a small-cap value index.

We already know that the PP does fairly well withdrawing 5% from cash:




So what if we take this PP variant with SCV stocks instead of a broad stock index and withdraw that same 5%?



Yes, that's 22 million dollars after 42 years of withdrawals. I guess 5% wasn't so risky after all! How about the completely ludicrous withdrawal rate of 7% per year, or $70,000/year in this case?



Keep in mine these are not just growth charts, they're growth charts including yearly inflation-adjusting withdrawals!

What do you think of this?

[AdamA]

What do you think of this?

Very interesting.

Which spreadsheet of Sophie's are you referring to?

I'd like to look at it. 

How did you get your stock growth rates for each scenario?

[Pointedstick]

Which spreadsheet of Sophie's are you referring to?

I'd like to look at it. 

How did you get your stock growth rates for each scenario?

https://docs.google.com/file/d/0B5v98TX ... sp=sharing

I got the SCV performance figure from the Simba/Jono Bogleheads backtesting spreadsheet. I believe that's where she got the total stock index numbers from, too.

[AdamA]

Which spreadsheet of Sophie's are you referring to?

I'd like to look at it. 

How did you get your stock growth rates for each scenario?

https://docs.google.com/file/d/0B5v98TX ... sp=sharing

I got the SCV performance figure from the Simba/Jono Bogleheads backtesting spreadsheet. I believe that's where she got the total stock index numbers from, too.

Got it. 

What is the average return on the SCV stocks over that time period?

[Pointedstick]

Which spreadsheet of Sophie's are you referring to?

I'd like to look at it. 

How did you get your stock growth rates for each scenario?

https://docs.google.com/file/d/0B5v98TX ... sp=sharing

I got the SCV performance figure from the Simba/Jono Bogleheads backtesting spreadsheet. I believe that's where she got the total stock index numbers from, too.

Got it. 

What is the average return on the SCV stocks over that time period?

Average return from 1972 - 2012 was 15.06%, as opposed to 11.57% for TSM.

[Tyler]

One word of caution -- remember that timing matters.  The early 70s was the best possible time to start a PP, with gold going absolutely bananas coming off the gold standard.  Baselining your SWR right before a run-up (in any portfolio) can make retirement results look quote rosy compared to, say, retiring in 2007 right before a heavy-stock portfolio tanked in 2008 (but keeping that inflation-adjusted initial withdrawal rate).

That doesn't mean SCV won't have a (potentially significant) positive effect.  It's more a statement about basing final results on a single start point.  Think of it as "timing bias".

FWIW, my own calculations (trying to account for this) have shown that the PP is still an extremely good portfolio for retirement.  I'd just be a little more cautious about 5%+ WRs, as the start point can start to affect the results more and more the higher you go.

EDIT: For reference, I came to the conclusion that at 4% and below (with a standard PP) the start point for retirement was essentially irrelevant.

[sophie]

PointedStick, that is fascinating.  I know this has come up before and I think very much worth considering.  You will all recall that Harry Browne originally suggested picking a basket of 20 "highly volatile" stocks, but then changed to recommending index funds for simplicity - not necessarily because they had appropriate volatility for the portfolio.  But, volatility is exactly what makes the PP work.  That paper that (I think) MachineGhost posted some time ago on the benefits of rebalancing for volatility harvesting was brilliant:  among other things there was a nice demo showing that if you invest in two assets with zero gain, you will end up with a gain if you rebalance periodically to take advantage of non-correlated volatility.

So the more volatile the assets, the more gains you'll get from rebalancing.  This runs counter to the standard index-investing philosophy, which assumes that gains come from stocks (more risk = more return), and that bonds are there only to reduce risk.

Couple of housekeeping observations.  One thing not included in the spreadsheet is expenses. SCV funds tend to have higher expenses than, say, an S&P 500 index.  It's worth modifying the spreadsheet to add these in, to get a more realistic backtest.  Also, there's no guarantee that SCV will continue to provide higher returns as it did in the past, because "the secret" is now out and markets are too highly efficient for that, but I think you're right to expect it will continue to be more volatile than the large caps.

Tyler had modified the spreadsheet to automate rebalancing, as I recall.  Did you do the same, and can one of you repost the spreadsheet with the index from Simba's spreadsheet (which I assume is S&P 500) and the small cap value #s?  Where did you find those?

[Pointedstick]

Couldn't one get that desired effect by switching out the 30 years bonds for 10 year bonds and lowering the % of gold relative to the other assets?

Perhaps something like:
40% TSM
40% 10 year Treasuries
10% Gold
10% Cash

That would reduce the cash percentage, which it turns out is very important to preserving the portfolio value during withdrawals, since it allows you to avoid selling the other assets until a rebalance band is hit normally. Here's what 5% withdrawal rate looks like for that portfolio:

[sophie]

Couldn't one get that desired effect by switching out the 30 years bonds for 10 year bonds and lowering the % of gold relative to the other assets?

Perhaps something like:
40% TSM
40% 10 year Treasuries
10% Gold
10% Cash

I expect so, but why dial down the bond volatility?  If anything, it would be great to dial it up e.g. by using zeros.

Dialing back gold is a reasonable way to approach the problem, and could be especially useful for people with limited space for gold in their tightly controlled retirement plans and taxable savings.  Something like 30% stocks, 30% bonds, 20% gold, 20% cash.  Rebalancing gets more complicated though.  To go with 40% rebalancing, you'd have to use 18%/42% bands for stocks and bonds, and 12%/28% for gold and cash.

[Pointedstick]

That doesn't mean SCV won't have a (potentially significant) positive effect.  It's more a statement about basing final results on a single start point.  Think of it as "timing bias".

FWIW, my own calculations (trying to account for this) have shown that the PP is still an extremely good portfolio for retirement.  I'd just be a little more cautious about 5%+ WRs, as the start point can start to affect the results more and more the higher you go.

Absolutely. I wish FireCalc allowed such precise parameters. Nonetheless, I think it stands up theoretically because increasing volatility is what we want and we can still withdraw from cash.

Couple of housekeeping observations.  One thing not included in the spreadsheet is expenses. SCV funds tend to have higher expenses than, say, an S&P 500 index.  It's worth modifying the spreadsheet to add these in, to get a more realistic backtest.  Also, there's no guarantee that SCV will continue to provide higher returns as it did in the past, because "the secret" is now out and markets are too highly efficient for that, but I think you're right to expect it will continue to be more volatile than the large caps.

Tyler had modified the spreadsheet to automate rebalancing, as I recall.  Did you do the same, and can one of you repost the spreadsheet with the index from Simba's spreadsheet (which I assume is S&P 500) and the small cap value #s?  Where did you find those?

All true. And you're right, the SCV effect might not stand up. The thing that I think helps is that we're not actually out for better performance, just more volatility. I didn't want to use a levered index so SCV seemed like the next best option.

I didn't use Tyler's chart with the rebalancing modifications. Is that posted somewhere? It should probably be put on the sticky thread. We should get expenses added, too.



TSM numbers 1972 - 2012:

Code: Select all

16.9
-18.1
-27.2
38.7
26.7
-4.2
7.5
23
32.7
-3.7
20.8
22
4.5
32.2
16.1
1.7
18
28.9
-6
34.7
9.8
10.62
-0.17
35.79
20.96
30.99
23.26
23.81
-10.57
-10.97
-20.96
31.35
12.52
5.98
15.51
5.49
-37.04
28.70
17.09
0.96
16.25

SCV numbers 1972 - 2012:

Code: Select all

7
-26
-18.2
54.5
53.6
21.8
21.8
35.4
25.4
14.9
28.5
38.6
2.3
31
7.4
-7.1
29.5
12.4
-21.8
41.7
29.1
23.8
-1.5
25.8
21.4
31.8
-6.5
3.35
21.88
13.7
-14.2
37.19
23.55
6.07
19.24
-7.07
-32.05
30.34
24.82
-4.16
18.56

[iwealth]

There's no safety in a 5% WR.

Just an example:

Starting in 1972 with $1 million and a 5% WR, I see it grows to about $5.3 million in 25 years out to 1997...and I'm withdrawing about $195,000 from the account that year or 3.6%.

Starting in 1987 with $1 million and a 5% WR, it only grows to $1.7 million in 25 years...and I'm withdrawing $100,000 from the account that year or about 6%. Barring some anomalously positive performance over the following few years, it won't last. (EDIT: And that is using total stock market - if you use SCV during that time period, you only have $1.45 million left after 25 years)

Even with a 50/50 portfolio, there are start years where an 8-9% WR would have worked. But, cherry-picking such years isn't of much use.

[Ad Orientem]

First I think Pointed has a good point (no pun). With half the permanent portfolio tied up in long and short term bonds and another quarter in gold, a non productive asset whose primary purpose is as a currency backstop, there is something of an anti-prosperity bias in the PP. But that's what is so attractive about it. This is where I put the money I don't want to lose.

Harry never intended that everyone would use the PP for all of their investments or as the principal means of making money. So rather than trying to fix something that works very well for what it's supposed to do (keep you from losing money) I have been suggesting that one way to compensate for the anti-prosperity bias is to just create a VP. Put 80-90% (or whatever your comfort level dictates) into an HB-PP and the remaining 10-20% in a broad based stock index fund for long term capital appreciation. Just understand that your VP is going to swing pretty wildly at times.

[BP]

PS: I brought up one of your old threads back in May 2013: http://gyroscopicinvesting.com/forum/pe ... /#msg66006

I ran the following:

1.  VOLATILE 4x25 PP at etfreplay.com from Jan 29, 2008 (farthest allowed back because of one of the ETFs used) EDV, VBR, IAU, and SHY - CAGR 7.8%, Max. Drawdown 13.55%  (SPY - 5.6% Max DD - 51.49%).

2.  Regular 4x25 PP at etfreplay.com from Jan 29, 2008, using VTI, IAU, TLT, and SHY - CAGR 3.1% Max Drawdown 3.85 %

Note how the volatile 4x25PP beat SPY quite handily with much less max DD.

The last I heard was that MG was going to do some testing and get back to the forum with the results.

[craigr]

Ad's point is one I've made for years as well. If someone wants to take on more chance of growth, then just hold a smaller "core" for the Permanent Portfolio and add in more stocks.

And by stocks I think a broadly based stock fund works pretty well. You could just buy Vanguard Total All World Index (VT) and call it a day with wide domestic/international diversification.

Something like:

20% Cash
20% Bonds
20% Gold
40% Stocks (Vanguard Total)

Unfortunately this is not a science. Backtesting can only do so much…Also consider the volatility carefully. Everyone thinks they can handle portfolio swings by taking concentrated risks, but most people can't. It calls for careful introspection and you won't know it until you are down some money as to how much it hurts you personally.

[Kriegsspiel]

Ad's point is one I've made for years as well. If someone wants to take on more chance of growth, then just hold a smaller "core" for the Permanent Portfolio and add in more stocks.

And by stocks I think a broadly based stock fund works pretty well. You could just buy Vanguard Total All World Index (VT) and call it a day with wide domestic/international diversification.

Something like:

20% Cash
20% Bonds
20% Gold
40% Stocks (Vanguard Total)

Unfortunately this is not a science. Backtesting can only do so much…Also consider the volatility carefully. Everyone thinks they can handle portfolio swings by taking concentrated risks, but most people can't. It calls for careful introspection and you won't know it until you are down some money as to how much it hurts you personally.

This is what I do. I basically have a PP, then just *more* TSM and TISM. I'm also going to be adding real estate to my VP.

[AdamA]

Average return from 1972 - 2012 was 15.06%, as opposed to 11.57% for TSM.

It seems like a good idea.

The hard thing for me would be that I might second guess myself if it underperformed from time to time.  You have to be a true believer in the small cap and value premiums in order to stick to this.

It's worth noting that over the past 10 years, (since Aug 4, 2003), small cap value has returned 125% vs. the S&P's 85%.

A 50/50 mix in the same time period returned around 100% (that's without any rebalancing). 

I think if someone were uncomfortable using only SCV stocks, they might mix SCV and the S&P according to their own comfort.

How does an SCV index fund do as far as taxes go? 

[JMoyle]

Has anyone looked at Sound Mind Investing's Dynamic Asset Allocation stratedgy? They have also automated the strategy into a Mutual Fund symbol SMIDX. Looks like a very promising VP!

[Ad Orientem]

And the fund carries an expense ratio of almost 2% !

Special seating for fund managers with ER's of 2%...

[brownehead]

Pointedstick try with emerging markets. They are even more volatile and in my tests they worked quite well for the PP.

[D1984]

Pointedstick try with emerging markets. They are even more volatile and in my tests they worked quite well for the PP.

If you are using the numbers from the Simba speadsheet for emerging markets historical return data then please be aware that anything it has for EM from before 1987 or '88 (whenever the MSCI EM index was incepted) is basically junk. They just took their data from IFA who in turn created their data from a mishmash of different kinds of international value and international small value that was NOT actually EM at all and then substituted it for EM since the MSCI EM index didn't exist before the late 80's.

The actual EM returns data for much of this period is available as the S&P IFC EM index (it started on 1-1-1976). Just a quick glance at it and you begin to see why and how backtesting using historical EM returns based on the Simba data is a seriously foolish idea. The real EM returns from circa 1980 to the mid 80s were a disaster as the 1979-80 oil shock knocked many of the non-oil producing emerging Asian economies for a loop and then the early 80s Latin American debt crises did the same to most of the Central and South American emerging markets. In 1981-82 alone EM stocks lost more than 25% each year (-27.55% in 1981 and -31.86% in 1982); in fact, every year from 1981-84 was a losing one for EM stocks on average! Despite this the Simba spreadsheet shows more than double the actual return EM stocks earned for 1980, a positive number for 1981, a 1.5% loss for 1982, a 32% gain (WTF?) in 1983, a nearly 10% gain in 1984, and approximately double the actual IFC EM returns (which were positive) for '85 and '86 as well.

Like I said, complete and utter garbage. Using the pre-MSCI EM Simba data is a perfectly serviceable way to GIGO a backtest. I have the monthly and annual IFC EM returns from 1976-87 if you want them posted here for backtesting purposes.

[koekebakker]

Thanks for clarifying D1984.

Another thing to consider is that you want the stock part of your portfolio to move in the same direction as the total stock market. SCV and EM correlation to TSM might be too low for the PP.

[BP]

Interesting article on SCV from Rick Ferri:

http://www.rickferri.com/blog/strategy/ ... ri+Blog%29

[brownehead]

Thank you for the clarification D1984, I didn't know that and will apreciate the annual figures. However, I think all we agree that EM are more volatile than TSM and I guess that even with the right data the CAGR will be better too.

Koekebakker, you are right about SCV and EM not being perfectly correlated to TSM, but what we really want is correlation with prosperity (TSM is just a proxy for it) and I think both SCV and EM allways have their glory days in every prosperity cycle, when money looks for new and risker assets.

[JMoyle]

Brownhead, Basics of the DAA:

Acknowledged as mimicking principles of Harry's PP, the DAA holds the top 3 performing of six ETFs (designed to react to economic conditions) on a monthly basis. The ETFs are stock index, LT bonds, ST bonds, REIT, Gold, Int. stock. At the end of each month, if you are holding the top 3 performing funds for the month...you do nothing. If one or all of the top 3 drop out of the top 3 performing funds, you sell the fund, or funds, and replace a fund, or funds, with what now comprises the top 3 performing funds. If one of the top 3 funds falls significantly during the month you are holding the fund then you are only in the fund for that month. On the flip side, if you are not holding a rising fund, you are only behind on its rise for a month. SMIDX is a mutual fund that follows this strategy for the holder. Looks to me to be a great VP option.

[brownehead]

Brownhead, Basics of the DAA:

Acknowledged as mimicking principles of Harry's PP, the DAA holds the top 3 performing of six ETFs (designed to react to economic conditions) on a monthly basis. The ETFs are stock index, LT bonds, ST bonds, REIT, Gold, Int. stock. At the end of each month, if you are holding the top 3 performing funds for the month...you do nothing. If one or all of the top 3 drop out of the top 3 performing funds, you sell the fund, or funds, and replace a fund, or funds, with what now comprises the top 3 performing funds. If one of the top 3 funds falls significantly during the month you are holding the fund then you are only in the fund for that month. On the flip side, if you are not holding a rising fund, you are only behind on its rise for a month. SMIDX is a mutual fund that follows this strategy for the holder. Looks to me to be a great VP option.

I don't see how that strategy mimicks the PP principles, but anyways I would't pay 2% for it.