Author Topic: Portfolio Charts - 2 Fund Portfolio with a 6% SWR  (Read 840 times)

index

  • Bristles
  • ***
  • Posts: 341
Portfolio Charts - 2 Fund Portfolio with a 6% SWR
« on: August 01, 2020, 11:09:58 AM »
I have been playing around with some risk parity portfolios similar to the all weather portfolio and golden butterfly portfolio.

I like the new EFT's RPAR and NTSX.

NTSX is a 90/60 fund from Wisdom Tree with a 0.2% ER:

It can be simulated with:
90% S&P 500
60% Intermediate Term Treasury (7yr AVG)
-50% Cash

RPAR is a new risk parity fund with a 0.5% ER:

It can be simulated with:
25% Global Equities and Emerging Markets
15% Commodity Producers (either use commodities or add this to Global Equities above)
18% Gold
42% Long Term Treasury (20+yr AVG)
21% 15+ YR TIPs

The leverage comes from buying treasury futures.

I simulated a 55% RPAR/45% NTSX portfolio at portfolio charts using:

40.5% TSM USA
8.75% TSM Ex US
5% EM
34.6% LT Bonds (could not add TIPs)
27% IT Bonds
8.25% Commodity
10% Gold

34% Leverage

Deepest drawdown - 15%
Recovery - 3.5 yrs
SWR - 6%

Andy R

  • Bristles
  • ***
  • Posts: 319
Re: Portfolio Charts - 2 Fund Portfolio with a 6% SWR
« Reply #1 on: August 01, 2020, 09:23:43 PM »
We currently have historically high equity valuations and historically low interest rates returning virtually nothing from bonds of which more than half your portfolio is made up of. Might be worth taking that into consideration.

secondcor521

  • Magnum Stache
  • ******
  • Posts: 2954
  • Age: 51
  • Location: Boise, Idaho
  • Big cattle, no hat.
    • Age of Eon - Overwatch player videos
Re: Portfolio Charts - 2 Fund Portfolio with a 6% SWR
« Reply #2 on: August 02, 2020, 01:12:24 AM »
I would read up on the perils of backtesting and overfitting.

mrmoonymartian

  • Bristles
  • ***
  • Posts: 284
  • Age: 38
  • Location: Brisbane
Re: Portfolio Charts - 2 Fund Portfolio with a 6% SWR
« Reply #3 on: August 02, 2020, 01:32:10 AM »
Shouldn't you put -34% BIL if you're levering? Then the result becomes nothing to write home about:

-21% drawdown
recovery 10 years
SWR30 5.1%

2Birds1Stone

  • Walrus Stache
  • *******
  • Posts: 6032
  • Location: Earth
  • K Thnx Bye
Re: Portfolio Charts - 2 Fund Portfolio with a 6% SWR
« Reply #4 on: August 02, 2020, 03:57:58 AM »
I would read up on the perils of backtesting and overfitting.

Dis, so much.

MustacheAndaHalf

  • Magnum Stache
  • ******
  • Posts: 2834
Re: Portfolio Charts - 2 Fund Portfolio with a 6% SWR
« Reply #5 on: August 02, 2020, 05:09:09 AM »
@index - The length of retirement matters, but you haven't mentioned it.

I think you're also ignoring the bond environment.  Long-term bonds look good because for decades they've had falling yields, which gives bonds a capital gain.  So if you back test using data from even the past 30-40 years, you will only see an environment in which bond yields have been falling.

But right now 30-year treasury bonds yield 1.2% - there isn't room to fall further.  To get the past 30-40 years of bond performance, you'd need to see treasury yields go negative, and continue dropping.  I could be wrong, but that strikes me as very unlikely.  So anyone relying on the past 30-40 years of historical bond performance is very likely to be wrong about future bond performance.

And most of your portfolio is bonds, so that may be a significant risk to your retirement.

DalioGold10

  • 5 O'Clock Shadow
  • *
  • Posts: 43
Re: Portfolio Charts - 2 Fund Portfolio with a 6% SWR
« Reply #6 on: August 02, 2020, 01:00:04 PM »
@index - The length of retirement matters, but you haven't mentioned it.

I think you're also ignoring the bond environment.  Long-term bonds look good because for decades they've had falling yields, which gives bonds a capital gain.  So if you back test using data from even the past 30-40 years, you will only see an environment in which bond yields have been falling.

But right now 30-year treasury bonds yield 1.2% - there isn't room to fall further.  To get the past 30-40 years of bond performance, you'd need to see treasury yields go negative, and continue dropping.  I could be wrong, but that strikes me as very unlikely.  So anyone relying on the past 30-40 years of historical bond performance is very likely to be wrong about future bond performance.

And most of your portfolio is bonds, so that may be a significant risk to your retirement.

Very true and very good points with respect to the bond allocation.
Over the last 30 to 40 years bonds have been in bull market. With 30 yrs US treasuries at 1.2%, there is huge downside for bonds in a portfolio.
Cash relative to bonds look a better asset class in this environment, although "cash is trash over the medium and long term" :) For a retiree, cash is a better option to cover living expenses for 2-4 years depending on risk tolerance and also to reduce SWR.

Personally, I have reduced bonds allocation to a minimum, actually divested all my long term bonds few months ago.

vand

  • Pencil Stache
  • ****
  • Posts: 969
  • Location: UK
Re: Portfolio Charts - 2 Fund Portfolio with a 6% SWR
« Reply #7 on: August 03, 2020, 01:17:46 AM »
I'm not going to give much attention to any portfolio with a leveraged position. Historically the best performing portfolio is a 300% leveraged stock-only portfolio.. even if you blow up once or twice, eventually you'll get lucky and do great. Except nobody does that with good reason.

That said, I have been wondering if anyone has ever done a regression fit comparing Sharpe Ratio vs SWR.  Or even better, two separate regressions, one comparing Sharpe Ratio vs SWR and another comparing CAGR vs SWR. We know that growth and volatility are both important factors in supporting higher SWRs, so this sort of analysis is really useful in my mind to see exactly how much each factor matters.

index

  • Bristles
  • ***
  • Posts: 341
Re: Portfolio Charts - 2 Fund Portfolio with a 6% SWR
« Reply #8 on: August 03, 2020, 08:53:24 AM »
Shouldn't you put -34% BIL if you're levering? Then the result becomes nothing to write home about:

-21% drawdown
recovery 10 years
SWR30 5.1%

Portfolio Charts doesn't use the daily carry rate and BIL is the best we can use. So I agree with including the -34% BIL. The other issue with portfolio charts is there is no provision for TIPs. Remember the 34.6% LT Treasury portion of the portfolio is actually 11% LT TIPs and 23.6% LT Treasuries. The coupon portion of a 30 year TIP is closer to the 10-yr rate. So the TIP price is going to respond more like an IT bond to changes in interest rate, but bond will math the inflation rate similar to BIL.

How do you model that? 1:1 BIL : IT bond? 

That gives you a SWR of 5.6% but the 10 yr drawdown still stands. 7.5% vs 6.8% on the pinwheel portfolio.

@index - The length of retirement matters, but you haven't mentioned it.

I think you're also ignoring the bond environment.  Long-term bonds look good because for decades they've had falling yields, which gives bonds a capital gain.  So if you back test using data from even the past 30-40 years, you will only see an environment in which bond yields have been falling.

But right now 30-year treasury bonds yield 1.2% - there isn't room to fall further.  To get the past 30-40 years of bond performance, you'd need to see treasury yields go negative, and continue dropping.  I could be wrong, but that strikes me as very unlikely.  So anyone relying on the past 30-40 years of historical bond performance is very likely to be wrong about future bond performance.

And most of your portfolio is bonds, so that may be a significant risk to your retirement.

So I agree with you to a certain extent except for the early 70's when bond yields were increasing which is when the portfolio had the worst returns. The portfolio is attempting to follow a risk parity chart similar to the pinwheel and golden butterfly portfolio:



No one really knows what bond return is going to look like in the future. Germany has negative rates. The portfolio is seeking to weight the bonds heavier because they are not sensitive enough to produce adequate diversification. If rates increase, it will probably be to control inflation. The EM, Ex-US, Gold, TIPs, and Commodities should balance that risk while the LT bonds are getting killed.   

I'm not going to give much attention to any portfolio with a leveraged position. Historically the best performing portfolio is a 300% leveraged stock-only portfolio.. even if you blow up once or twice, eventually you'll get lucky and do great. Except nobody does that with good reason.

That said, I have been wondering if anyone has ever done a regression fit comparing Sharpe Ratio vs SWR.  Or even better, two separate regressions, one comparing Sharpe Ratio vs SWR and another comparing CAGR vs SWR. We know that growth and volatility are both important factors in supporting higher SWRs, so this sort of analysis is really useful in my mind to see exactly how much each factor matters.

The leverage is coming from treasury futures which is the tamest form that can be used. The leverage is used to increase your exposure to the "safe" treasuries without killing your returns by over allocating to them. Read up on rolling treasury futures. It is pretty standard in the financial industry and nothing like the 2x and 3x ETFs on daily index futures you are referencing. The two ETFs I highlighted do this for you and at pretty low ERs - 0.36 combined which is a bp over a robo advisor charges for standard ETF rebalancing.

@index - The length of retirement matters, but you haven't mentioned it.

I think you're also ignoring the bond environment.  Long-term bonds look good because for decades they've had falling yields, which gives bonds a capital gain.  So if you back test using data from even the past 30-40 years, you will only see an environment in which bond yields have been falling.

But right now 30-year treasury bonds yield 1.2% - there isn't room to fall further.  To get the past 30-40 years of bond performance, you'd need to see treasury yields go negative, and continue dropping.  I could be wrong, but that strikes me as very unlikely.  So anyone relying on the past 30-40 years of historical bond performance is very likely to be wrong about future bond performance.

And most of your portfolio is bonds, so that may be a significant risk to your retirement.

Very true and very good points with respect to the bond allocation.
Over the last 30 to 40 years bonds have been in bull market. With 30 yrs US treasuries at 1.2%, there is huge downside for bonds in a portfolio.
Cash relative to bonds look a better asset class in this environment, although "cash is trash over the medium and long term" :) For a retiree, cash is a better option to cover living expenses for 2-4 years depending on risk tolerance and also to reduce SWR.

Personally, I have reduced bonds allocation to a minimum, actually divested all my long term bonds few months ago.


I agree with you that bonds don't look attractive right now, but this is essentially market timing. I mean the market sitting at an all time high when we have record unemployment and negative GDP doesn't look very attractive either... As we learned from Europe, there are no rules against negative yields.

Edited image size!
« Last Edit: August 03, 2020, 08:57:00 AM by index »

EvenSteven

  • Bristles
  • ***
  • Posts: 399
  • Location: St. Louis
Re: Portfolio Charts - 2 Fund Portfolio with a 6% SWR
« Reply #9 on: August 03, 2020, 09:00:14 AM »
Quote
I mean the market sitting at an all time high when we have record unemployment and negative GDP doesn't look very attractive either... As we learned from Europe, there are no rules against negative yields.

Economic numbers look bad, but not that bad!

index

  • Bristles
  • ***
  • Posts: 341
Re: Portfolio Charts - 2 Fund Portfolio with a 6% SWR
« Reply #10 on: August 03, 2020, 09:33:24 AM »
Sorry. A negative 32.9% growth rate versus a negative 9% in 2008.

ChpBstrd

  • Handlebar Stache
  • *****
  • Posts: 2185
Re: Portfolio Charts - 2 Fund Portfolio with a 6% SWR
« Reply #11 on: August 03, 2020, 12:09:22 PM »
I'm not going to give much attention to any portfolio with a leveraged position. Historically the best performing portfolio is a 300% leveraged stock-only portfolio.. even if you blow up once or twice, eventually you'll get lucky and do great. Except nobody does that with good reason.

That said, I have been wondering if anyone has ever done a regression fit comparing Sharpe Ratio vs SWR.  Or even better, two separate regressions, one comparing Sharpe Ratio vs SWR and another comparing CAGR vs SWR. We know that growth and volatility are both important factors in supporting higher SWRs, so this sort of analysis is really useful in my mind to see exactly how much each factor matters.

I think this is a great comment. I too would like to know the optimal Sharpe for FIRE because picking the exact right AA for any given investment era resembles fortune telling. The best AA for the 1980s would have killed you in the 1970s, and the best AA for the 1970s would have killed you in the 1990s. When I start spinning narratives about the next 5 years, Iím deep in speculation, and thatís different than risk management.

index

  • Bristles
  • ***
  • Posts: 341
Re: Portfolio Charts - 2 Fund Portfolio with a 6% SWR
« Reply #12 on: August 05, 2020, 06:09:47 PM »
So I've been back testing and this got a little more complicated but still a pretty easy allocation to maintain.

45% RPAR
23% VTI
10% VBR
10% NTSX
7% BAR
5% IEMG

The results for backtesting are:

32% Total Stock Market
10% Small Cap Value
9% Total Stock Market ex US
7.3% Emerging Market

18.9% 20+ Yr Treasury
15.5% 7 Yr Treasury and TIPs

15.1% Gold
6.8% Commodities

-14.5% BIL (Leverage from NTSX and RPAR)

Results in a portfolio that is:

58.3% Equity
34.4% Bonds
21.9% Real Assets

Portfolio Charts shows:

Average Return- 7.4%
Std. Deviation- 10.9%
Longest Drawdown - 3.5 years
SWR for 30 yrs - 6%
Perpetual withdrawal rate - 5%


This beats both the Pinwheel Portfolio with 2 less positions. The 33% VTI/VBR position can be split up and allocated to your desired factor. 







mrmoonymartian

  • Bristles
  • ***
  • Posts: 284
  • Age: 38
  • Location: Brisbane
Re: Portfolio Charts - 2 Fund Portfolio with a 6% SWR
« Reply #13 on: August 06, 2020, 02:17:36 AM »
FYI, 10% VBR would be 6% MCV + 4% SCV. Wouldn't make much difference compared to the 10% SCV you're using. But if you do want SCV only, you'd need VIOV or similar instead of VBR.

helloyou

  • Stubble
  • **
  • Posts: 141
Re: Portfolio Charts - 2 Fund Portfolio with a 6% SWR
« Reply #14 on: August 06, 2020, 02:45:40 AM »
Sorry I don't know all acronyms, but what does SWR means?

index

  • Bristles
  • ***
  • Posts: 341
Re: Portfolio Charts - 2 Fund Portfolio with a 6% SWR
« Reply #15 on: August 06, 2020, 09:30:35 AM »
FYI, 10% VBR would be 6% MCV + 4% SCV. Wouldn't make much difference compared to the 10% SCV you're using. But if you do want SCV only, you'd need VIOV or similar instead of VBR.

Thanks. You are right, I should have used VIOV.

What I like about this portfolio is outside of RPAR, NTSX, Gold, and emerging markets, you can tilt the remaining 33% however you like. Small cap value backtests well, but momentum, tech, low volatility, quality, value, market cap, etc are all on the table.

Sorry I don't know all acronyms, but what does SWR means?

SWR = Safe Withdrawal Rate

This recognizes the importance of standard deviation and sequence of returns in your portfolio. Say you are one of the people who like 100% VTI on this site, if you plan to retire 15 years from now, there is a chance you portfolio can have a 0% return due to the sequence of returns and standard deviation of a 100% equity portfolio. When you go to withdrawal for retirement, the safe withdrawal rate to preserve your funds during retirement will actually be lower at around 3.8% to account for the possibility of long periods of zero or negative growth. The above portfolio and other more diverse portfolios have a smaller standard deviation which reduces the risk of long periods of underperformance and allows a retiree to safely withdrawal funds at a higher rate and preserve their stache. This is helpful to accumulators too who want to more accurately be able to project their retirement date.