1. Withdraw 4% of the portfolio value on January 1 of the year of retirement.
2. In subsequent years, at the beginning of the year, prior to rebalancing the portfolio if necessary, withdraw the previous year's withdrawal amount incremented by the previous year's inflation rate.
The portfolios considered are:
1. SP500 with dividends (VFINX).
2. Mid cap value (FDVLX)
3. 30% VFINX, 30% FDVLX, 40% Aggregate Bond Fund (VBMFX)
4. 30% VFINX, 30% FDVLX, 20% VBMFX, 20% Long Term Treasury Fund (VUSTX)
5. 60% FDVLX 40% VUSTX
The first figure depicts the compound annual growth rate (Net CAGR) of the portfolio net of withdrawals computed for the period starting in the year of retirement thru 2014. Note that except for the VFINX, Net CAGR is always positive. Thus, although the VFINX portfolio would have sustained the withdrawal schedule, for retirement beginning in some of the years around 2000, the portfolio value at the end of 2014 would have been lower than its starting value. For every other portfolio, if the retirement commenced in any year starting in 1987-2014, at the end of 2014 the portfolio value would be higher than the starting value.
In the following figure are shown the maximum excess (beginning of the year) drawdowns for various portfolios. This quantity is defined to be the difference between the initial starting value of the portfolio and its lowest value at the time of withdrawal in any of the subsequent years, less 4%, as at best 4% will be the drawdown of the portfolio at the beginning of retirement. Clearly the excess drawdowns of the all equity portfolios will be unacceptable for most retirees. However, for the 60/40 portfolios, the excess drawdowns appear to be a good measure of what can be expected from well constructed portfolios. (For example, the portfolio of JNJ, MCD, CL, KO and PG with the same withdrawal schedule would have had the maximum excess drawdown of about 24%, similar to the 20% for the 60/40 portfolios.). (click to enlarge).
I
]]>1. Withdraw 4% of the portfolio value on January 1 of the year of retirement.
2. In subsequent years, at the beginning of the year, prior to rebalancing the portfolio if necessary, withdraw the previous year's withdrawal amount incremented by the previous year's inflation rate.
The portfolios considered are:
1. SP500 with dividends (VFINX).
2. Mid cap value (FDVLX)
3. 30% VFINX, 30% FDVLX, 40% Aggregate Bond Fund (VBMFX)
4. 30% VFINX, 30% FDVLX, 20% VBMFX, 20% Long Term Treasury Fund (VUSTX)
5. 60% FDVLX 40% VUSTX
The first figure depicts the compound annual growth rate (Net CAGR) of the portfolio net of withdrawals computed for the period starting in the year of retirement thru 2014. Note that except for the VFINX, Net CAGR is always positive. Thus, although the VFINX portfolio would have sustained the withdrawal schedule, for retirement beginning in some of the years around 2000, the portfolio value at the end of 2014 would have been lower than its starting value. For every other portfolio, if the retirement commenced in any year starting in 1987-2014, at the end of 2014 the portfolio value would be higher than the starting value.
In the following figure are shown the maximum excess (beginning of the year) drawdowns for various portfolios. This quantity is defined to be the difference between the initial starting value of the portfolio and its lowest value at the time of withdrawal in any of the subsequent years, less 4%, as at best 4% will be the drawdown of the portfolio at the beginning of retirement. Clearly the excess drawdowns of the all equity portfolios will be unacceptable for most retirees. However, for the 60/40 portfolios, the excess drawdowns appear to be a good measure of what can be expected from well constructed portfolios. (For example, the portfolio of JNJ, MCD, CL, KO and PG with the same withdrawal schedule would have had the maximum excess drawdown of about 24%, similar to the 20% for the 60/40 portfolios.). (click to enlarge).
I
]]>Obviously the future performance of the portfolio relative to SPY may be entirely different, but the results suggest that this basket may prove to be a good long term investment, a conclusion that is also supported by the various performance metrics shown below.
CAGR (%) | MaxDD (%) | Sharpe | Sortino | Volatility (%) | |
---|---|---|---|---|---|
Eq. Wt. | 14.3 | 33.2 | 0.91 | 1.69 | 12.6 |
Risk Parity | 16.2 | 34.8 | 1.04 | 1.85 | 12.7 |
VFINX | 9.9 | 50.9 | 0.54 | 0.88 | 14.5 |
BRK-A | 15.7 | 44.5 | 0.68 | 1.36 | 20.6 |
This is the allocation for the annually rebalanced portfolio according to a modified risk parity method for the current year 2015 (YTD return : 4.4%, 2014 return : 27.7%)
FBIOX 13.3%
FRESX 24.4%
TLT 43.8%
FDFAX 16.9%
FSCSX 1.6%
]]>Obviously the future performance of the portfolio relative to SPY may be entirely different, but the results suggest that this basket may prove to be a good long term investment, a conclusion that is also supported by the various performance metrics shown below.
CAGR (%) | MaxDD (%) | Sharpe | Sortino | Volatility (%) | |
---|---|---|---|---|---|
Eq. Wt. | 14.3 | 33.2 | 0.91 | 1.69 | 12.6 |
Risk Parity | 16.2 | 34.8 | 1.04 | 1.85 | 12.7 |
VFINX | 9.9 | 50.9 | 0.54 | 0.88 | 14.5 |
BRK-A | 15.7 | 44.5 | 0.68 | 1.36 | 20.6 |
This is the allocation for the annually rebalanced portfolio according to a modified risk parity method for the current year 2015 (YTD return : 4.4%, 2014 return : 27.7%)
FBIOX 13.3%
FRESX 24.4%
TLT 43.8%
FDFAX 16.9%
FSCSX 1.6%
]]>A recent article on SA by a fundamental analyst par excellence provides a good example (http://seekingalpha.com/article/2470375-yield-on-cost-a-vitally-important-consideration-for-retired-investors ). The case study described in this article considers a portfolio that was initiated with $3M at the beginning of 2006. Even after annual withdrawals that increased every year as the dividends accrued from the twenty stocks of the portfolio, it ends up with $5.6M at the end of 2013, thus weathering the intervening shocks in the market with admirable and rarely matched success.
I have compared the the performance of this all stock portfolio with that of various portfolios based the Naïve Graham allocation strategy (http://seekingalpha.com/instablog/709762-varan/2990923-naive-graham-passive-investing-according-to-the-master ) . The following are the relevant details for each portfolio:
Year | Amount withdrawn |
2006 | $88,845 |
2007 | $105,108 |
2008 | $116,076 |
2009 | $124,513 |
2010 | $134,035 |
2011 | $147,111 |
2012 | $161,281 |
2013 | $174,271 |
It might be noted that the timing of the withdrawals may affect the results, but the withdrawal at the beginning of the year probably biases the results against the allocation portfolios.
The amount left in the portfolios at the end of 2013 is given in the following table:
Portfolio | Value at EOY 2013 |
VTI/TLT | $6.5M |
IJJ/TLT | $5.8M |
IJS/TLT | $5.9M |
iShares Value | $5.7M |
iShares Gowth | $6.5M |
Fidelity Value | $6.3M |
Fidelity Growth | $6.7M |
All Stock | $5.5M |
For the iShares Growth portfolio, the following ETFs were used in addition to TLT: IVW, IJK and IJT. For the Fidelity Growth portfolio, I used the following funds: FBGRX, FMCSX and FCPGX. For the other portfolios the same funds were used as in the original post on Naïve Graham referenced above.
]]>A recent article on SA by a fundamental analyst par excellence provides a good example (http://seekingalpha.com/article/2470375-yield-on-cost-a-vitally-important-consideration-for-retired-investors ). The case study described in this article considers a portfolio that was initiated with $3M at the beginning of 2006. Even after annual withdrawals that increased every year as the dividends accrued from the twenty stocks of the portfolio, it ends up with $5.6M at the end of 2013, thus weathering the intervening shocks in the market with admirable and rarely matched success.
I have compared the the performance of this all stock portfolio with that of various portfolios based the Naïve Graham allocation strategy (http://seekingalpha.com/instablog/709762-varan/2990923-naive-graham-passive-investing-according-to-the-master ) . The following are the relevant details for each portfolio:
Year | Amount withdrawn |
2006 | $88,845 |
2007 | $105,108 |
2008 | $116,076 |
2009 | $124,513 |
2010 | $134,035 |
2011 | $147,111 |
2012 | $161,281 |
2013 | $174,271 |
It might be noted that the timing of the withdrawals may affect the results, but the withdrawal at the beginning of the year probably biases the results against the allocation portfolios.
The amount left in the portfolios at the end of 2013 is given in the following table:
Portfolio | Value at EOY 2013 |
VTI/TLT | $6.5M |
IJJ/TLT | $5.8M |
IJS/TLT | $5.9M |
iShares Value | $5.7M |
iShares Gowth | $6.5M |
Fidelity Value | $6.3M |
Fidelity Growth | $6.7M |
All Stock | $5.5M |
For the iShares Growth portfolio, the following ETFs were used in addition to TLT: IVW, IJK and IJT. For the Fidelity Growth portfolio, I used the following funds: FBGRX, FMCSX and FCPGX. For the other portfolios the same funds were used as in the original post on Naïve Graham referenced above.
]]>RS-GMR-ETF: IJJ, IEV, ILF, EPP, EEM, TLT
RS-GMR-LETF: MVV, IEV, ILF, EPP, EEM, TLT
RS-GMR-MF: FDVLX, FIEUX, FEMKX, FLATX, FPBFX, VUSTX
Period | CAGR | Sharpe (Sortino) | Max. Drawdown | Min. Annual Return | |
RS-GMR-ETF | 2003-2014 | 28.6% | 1.3 (2.8) | 17.2% | 6.5% |
RS-GMR-LETF | 2007-2014 | 31.5% | 1.12 (2.11) | 22.4% | 4.1% |
RS-GMR-MF | 1991-2014 | 20.7% | 0.97 (1.93) | 24.6% | -24.6% |
YTD Returns
RS-GMR-ETF 13.1%
RS-GMR-LETF 4.1%
RS-GMR-MF 14.1%
For August 2014, both of the ETF strategies are going to be invested in EEM.
The following figures display some results for the RS-GMR-ETF and the RS-GMR-MF baskets.
Disclosure: The author is long EEM.
Additional disclosure: This is not investment advice in any form.
]]>RS-GMR-ETF: IJJ, IEV, ILF, EPP, EEM, TLT
RS-GMR-LETF: MVV, IEV, ILF, EPP, EEM, TLT
RS-GMR-MF: FDVLX, FIEUX, FEMKX, FLATX, FPBFX, VUSTX
Period | CAGR | Sharpe (Sortino) | Max. Drawdown | Min. Annual Return | |
RS-GMR-ETF | 2003-2014 | 28.6% | 1.3 (2.8) | 17.2% | 6.5% |
RS-GMR-LETF | 2007-2014 | 31.5% | 1.12 (2.11) | 22.4% | 4.1% |
RS-GMR-MF | 1991-2014 | 20.7% | 0.97 (1.93) | 24.6% | -24.6% |
YTD Returns
RS-GMR-ETF 13.1%
RS-GMR-LETF 4.1%
RS-GMR-MF 14.1%
For August 2014, both of the ETF strategies are going to be invested in EEM.
The following figures display some results for the RS-GMR-ETF and the RS-GMR-MF baskets.
Disclosure: The author is long EEM.
Additional disclosure: This is not investment advice in any form.
]]>The Naïve Graham strategy described in an earlier post (http://seekingalpha.com/instablog/709762-varan/2990923-naive-graham-passive-investing-according-to-the-master ) appears to yield satisfactory results when applied to a basket of 2X leveraged funds consisting of MVV, SSO, DDM and UBT. The results shown here were obtained by using the six fund strategy described in the earlier post with the following approach:
For the period 2007-to date, the method yield the following results:
CAGR 26.5%
Number of Years of Losses 0
Minimum Annual Return 11% (2008)
Maximum Drawdown 26.7%
Sharpe Ratio 1.14
Sortino Ratio 2.19
The equity curve, the Manhattan Allocation diagram and the raw allocation diagram are shown in the following figures, with the label $UBT representing the simulated version of UBT. (A comparison of the results obtained from the simulated version with the results that used the actual time history of UBT during the period 2011-to date did not show any significant differences.)
Disclosure: The author has no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.
Additional disclosure: This is not investment advice.
]]>The Naïve Graham strategy described in an earlier post (http://seekingalpha.com/instablog/709762-varan/2990923-naive-graham-passive-investing-according-to-the-master ) appears to yield satisfactory results when applied to a basket of 2X leveraged funds consisting of MVV, SSO, DDM and UBT. The results shown here were obtained by using the six fund strategy described in the earlier post with the following approach:
For the period 2007-to date, the method yield the following results:
CAGR 26.5%
Number of Years of Losses 0
Minimum Annual Return 11% (2008)
Maximum Drawdown 26.7%
Sharpe Ratio 1.14
Sortino Ratio 2.19
The equity curve, the Manhattan Allocation diagram and the raw allocation diagram are shown in the following figures, with the label $UBT representing the simulated version of UBT. (A comparison of the results obtained from the simulated version with the results that used the actual time history of UBT during the period 2011-to date did not show any significant differences.)
Disclosure: The author has no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.
Additional disclosure: This is not investment advice.
]]>The Naïve Graham strategy, which is the simplest one that can be derived from Graham's idea - one that the he probably would not approve, though it difficult to say so definitively as the investment vehicles such as ETFs did not exist at that time - leads to results that are noteworthy. Prior to describing the details of the strategy, we summarize the results for a number of cases. For the purposes of comparison with much simpler buy and hold alternatives, also shown are the results for a small cap value fund, DFSVX, and a balanced fund PRWCX, both of which have historically yielded superior performance compared to their respective peers.
Funds | CAGR (%) | Sharpe Ratio (Sortino) | Max. Drawdown (%) | Min. Annual Return (%) |
2003-2013 Two Fund Baskets | ||||
VTI TLT (Market) | 12.7 | 1.1 (2.0) | 14.3 | 3.0 |
IJJ TLT (Mid Cap Value) | 12.8 | 1.0 (1.9) | 14.3 | 4.6 |
IJS TLT (Small Cap Value) | 13.1 | 1.0 (1.9) | 15.4 | 2.9 |
DFSVX | 13.7 | 0.6 (1.1) | 61.2 | -36.8 |
PRWCX | 10.4 | 0.8 (1.2) | 36.6 | -27.2 |
2007-2013 Six Fund Baskets (component assets shown at the end of the post) | ||||
Guggenheim Value | 12.9 | 0.9 (1.4) | 24.6 | -7.15 |
Guggenheim Growth | 17.1 | 1.3 (2.4) | 15.8 | -0.9 |
Vanguard Value | 13.35 | 1.1 (1.95) | 13.6 | 4.8 |
Vanguard Growth | 14.8 | 1.2 (2.0) | 20.1 | -6.6 |
DFSVX | 6.8 | 0.4(0.6) | 61.2 | -36.8 |
PRWCX | 7.7 | 0..6(0.8) | 36.6 | -27.2 |
Inasmuch as the risk-based performance metrics are substantially better than those of the alternatives considered here, the strategy has much to recommend itself.
The strategy entails rebalancing a basket of stock and bond ETFs at the beginning of every quarter on the basis of their relative returns during the immediately preceding quarter. The weights of the various ETFS are determined such that the total weights of each of the two classes of ETFs (stocks and bonds) are between 25% and 75% as advised by Graham.
The Method for Two Fund Baskets
This method applies to a basket of two funds, one a stock fund, and the other a bond fund.
The Method for Six Fund Baskets
This method applies to a basket of six funds, half of them being stock funds, and the rest bond funds.
Rank | Weight (%) |
1 | 35 |
2 | 25 |
3 | 15 |
4 | 12 |
5 | 8 |
6 | 5 |
The precise values of the weights are not very important: the only requirement is that the total for the top three ranked funds be 75% and it be 25% for the bottom three funds (even the totals of 75% and 25% weights may be replaced by slightly different weights for stocks and bonds). If the rank-based weights are in decreasing order, the allocation is determined purely by relative strength: the funds with higher returns in a quarter are assumed to likely perform better than the other lower ranked funds in the next quarter.
The following equity ETFs were used in the various baskets:
The ETF TLT and the mutual funds FLBIX and VUSTX, all based on long term treasuries, were used for the bond portion of the six fund baskets. In actual implementation, once the weights at the beginning of a quarter are determined, FLBIX and VUSTX may be replaced by TLT. (Alternatively one may just use three copies of TLT in the computations, but for the sake of simplicity we have used three distinct funds here.)
The main advantage of this strategy is that the allocation is determined without any complex computations, in sharp contrast to the other methods such as risk parity or the maximum diversified portfolio algorithm, and yet the returns are quite satisfactory. For the baskets considered here, the volatility of the returns and the maximum drawdown are also much lower than those of the stock funds alone. Just as the asset allocation methods that minimize volatility end up yielding portfolios which have superior returns over periods that span multiple market cycles, it appears that the portfolios whose allocation is based on returns alone may have low volatility.
As an example, the equity growth curve and the allocation diagram for the Guggenheim Growth basket are shown in the following figures.
Disclosure: The author is long IJS.
Additional disclosure: This is not investment advice.
]]>The Naïve Graham strategy, which is the simplest one that can be derived from Graham's idea - one that the he probably would not approve, though it difficult to say so definitively as the investment vehicles such as ETFs did not exist at that time - leads to results that are noteworthy. Prior to describing the details of the strategy, we summarize the results for a number of cases. For the purposes of comparison with much simpler buy and hold alternatives, also shown are the results for a small cap value fund, DFSVX, and a balanced fund PRWCX, both of which have historically yielded superior performance compared to their respective peers.
Funds | CAGR (%) | Sharpe Ratio (Sortino) | Max. Drawdown (%) | Min. Annual Return (%) |
2003-2013 Two Fund Baskets | ||||
VTI TLT (Market) | 12.7 | 1.1 (2.0) | 14.3 | 3.0 |
IJJ TLT (Mid Cap Value) | 12.8 | 1.0 (1.9) | 14.3 | 4.6 |
IJS TLT (Small Cap Value) | 13.1 | 1.0 (1.9) | 15.4 | 2.9 |
DFSVX | 13.7 | 0.6 (1.1) | 61.2 | -36.8 |
PRWCX | 10.4 | 0.8 (1.2) | 36.6 | -27.2 |
2007-2013 Six Fund Baskets (component assets shown at the end of the post) | ||||
Guggenheim Value | 12.9 | 0.9 (1.4) | 24.6 | -7.15 |
Guggenheim Growth | 17.1 | 1.3 (2.4) | 15.8 | -0.9 |
Vanguard Value | 13.35 | 1.1 (1.95) | 13.6 | 4.8 |
Vanguard Growth | 14.8 | 1.2 (2.0) | 20.1 | -6.6 |
DFSVX | 6.8 | 0.4(0.6) | 61.2 | -36.8 |
PRWCX | 7.7 | 0..6(0.8) | 36.6 | -27.2 |
Inasmuch as the risk-based performance metrics are substantially better than those of the alternatives considered here, the strategy has much to recommend itself.
The strategy entails rebalancing a basket of stock and bond ETFs at the beginning of every quarter on the basis of their relative returns during the immediately preceding quarter. The weights of the various ETFS are determined such that the total weights of each of the two classes of ETFs (stocks and bonds) are between 25% and 75% as advised by Graham.
The Method for Two Fund Baskets
This method applies to a basket of two funds, one a stock fund, and the other a bond fund.
The Method for Six Fund Baskets
This method applies to a basket of six funds, half of them being stock funds, and the rest bond funds.
Rank | Weight (%) |
1 | 35 |
2 | 25 |
3 | 15 |
4 | 12 |
5 | 8 |
6 | 5 |
The precise values of the weights are not very important: the only requirement is that the total for the top three ranked funds be 75% and it be 25% for the bottom three funds (even the totals of 75% and 25% weights may be replaced by slightly different weights for stocks and bonds). If the rank-based weights are in decreasing order, the allocation is determined purely by relative strength: the funds with higher returns in a quarter are assumed to likely perform better than the other lower ranked funds in the next quarter.
The following equity ETFs were used in the various baskets:
The ETF TLT and the mutual funds FLBIX and VUSTX, all based on long term treasuries, were used for the bond portion of the six fund baskets. In actual implementation, once the weights at the beginning of a quarter are determined, FLBIX and VUSTX may be replaced by TLT. (Alternatively one may just use three copies of TLT in the computations, but for the sake of simplicity we have used three distinct funds here.)
The main advantage of this strategy is that the allocation is determined without any complex computations, in sharp contrast to the other methods such as risk parity or the maximum diversified portfolio algorithm, and yet the returns are quite satisfactory. For the baskets considered here, the volatility of the returns and the maximum drawdown are also much lower than those of the stock funds alone. Just as the asset allocation methods that minimize volatility end up yielding portfolios which have superior returns over periods that span multiple market cycles, it appears that the portfolios whose allocation is based on returns alone may have low volatility.
As an example, the equity growth curve and the allocation diagram for the Guggenheim Growth basket are shown in the following figures.
Disclosure: The author is long IJS.
Additional disclosure: This is not investment advice.
]]>