Nice summary!ochotona wrote: ↑Tue May 22, 2018 8:41 pmFor the benefit of new readers:
http://gemsignals.blogspot.com/2018/05/ ... ap-of.html
Tactical Asset Allocation + HBPP an intriguing combo
Moderator: Global Moderator
Re: A recap of the GEM methodology
GEM goes back to US Large Caps
I have an awful, busy schedule then vacation to somewhere without Internet, so I traded the GEM early, today, instead of June 1. I sold all of my international ETFs and Mutual funds (SWISX, SCHF, SCHE) and went 100% US Large Cap (SCHX). SCHX is now in the lead from the 1-year total return perspective. I'll be back to the regular schedule next month.
There is a small chance this trade was in vain, that in the next four days things will reverse and GEM will stay in International, but I really can't wait around until June 1.
There is a small chance this trade was in vain, that in the next four days things will reverse and GEM will stay in International, but I really can't wait around until June 1.
Re: Dual Momentum GEM + HBPP a great combo, easy to test
An S&P500 mutual fund (SWPPX) traded using:
CAGR SD Best Yr Worst Yr MaxDD Sharpe Sortino US Mkt Correl
FAGIX signals 11.13% 9.45% 32.27% -8.33% -16.73% 0.98 1.74 0.51
SWPPX signals 9.41% 10.37% 32.27% -7.51% -16.17% 0.75 1.23 0.60
Buy & Hold 5.96% 14.39% 32.27% -36.72% -50.72% 0.35 0.50 0.99
FAGIX got out earlier in 2000, got in earlier in 2003. Got out earlier in 2008, got in earlier in 2009. Has even done better 2015-May 2018, which has been a bad time for trendfollowing, during which it has lagged Buy & Hold.
FAGIX = Fidelity® Capital & Income Fund (junk bonds). FAGIX itself is not amenable to trading, due to frequent trading rules, but just using it as a signal is beneficial.
CAGR SD Best Yr Worst Yr MaxDD Sharpe Sortino US Mkt Correl
FAGIX signals 11.13% 9.45% 32.27% -8.33% -16.73% 0.98 1.74 0.51
SWPPX signals 9.41% 10.37% 32.27% -7.51% -16.17% 0.75 1.23 0.60
Buy & Hold 5.96% 14.39% 32.27% -36.72% -50.72% 0.35 0.50 0.99
FAGIX got out earlier in 2000, got in earlier in 2003. Got out earlier in 2008, got in earlier in 2009. Has even done better 2015-May 2018, which has been a bad time for trendfollowing, during which it has lagged Buy & Hold.
FAGIX = Fidelity® Capital & Income Fund (junk bonds). FAGIX itself is not amenable to trading, due to frequent trading rules, but just using it as a signal is beneficial.
Re: Dual Momentum GEM + HBPP a great combo, easy to test
Did a simple backtest. Probably not doing it right. Got a CAGR of over 8%. How do you set it up?
https://www.portfoliovisualizer.com/tes ... odWeight=0
Thanks!
https://www.portfoliovisualizer.com/tes ... odWeight=0
Thanks!
Re: Dual Momentum GEM + HBPP a great combo, easy to test
https://www.portfoliovisualizer.com/tes ... odWeight=0
What do you think of this one?
Then if you turn Single Absolute Momentum "off" you go back to regular GEM.
-
- Senior Member
- Posts: 109
- Joined: Sat Sep 24, 2016 5:39 pm
- Location: NJ and India
Re: Dual Momentum GEM + HBPP a great combo, easy to test
Are international stocks, VEU, still part of the strategy?
Re: Dual Momentum GEM + HBPP a great combo, easy to test
I traded on Friday, which was a bit early, but no, I'm not in VEU or CWI anymore. Only S&P 500, so VOO or SCHX for Schwab users.
Re: Dual Momentum GEM + HBPP a great combo, easy to test
Thank you! A great composition!
Where do you look up FAGIX signals?
Where do you look up FAGIX signals?
ochotona wrote: ↑Sun May 27, 2018 1:18 pm
https://www.portfoliovisualizer.com/tes ... odWeight=0
What do you think of this one?
Then if you turn Single Absolute Momentum "off" you go back to regular GEM.
Re: Dual Momentum GEM + HBPP a great combo, easy to test
FAGIX data is at Morningstar, Stockcharts, Fidelity of course. It's a Fidelity mutual fund.
Re: Dual Momentum GEM + HBPP a great combo, easy to test
Ochotona, Happyman,
OK, a little while back I said I was going to try to get some hypothetical returns for foreign equities so as to extend the backtest before 1971 which is as far back as the GEM backtest page goes; I needed the returns from 1963 or so (actually January 1962 since GEM requires a one-year lookback) as that is when rates started significantly rising and pretty much either stayed flat, rose slightly, or rose a lot every year until the end of 1969...I wanted to be able to do this backtest in a rising-rate environment since it involves switching into intermediate-term bonds when stocks haven't beat the three-month T-Bill.
Well, after some laborious manual chart scraping of Morningstar data for the only two foreign funds that go back that far (TEPLX - Templeton Growth which is a foreign large growth/large blend fund, and SCINX - which is now an index fund based on the valuation/earnings/dividend-based CROCI index but before 2013 it was DWS International and before the early 1990s it was Scudder International hence the ticker SCINX...before it switched to being an index fund it was an actively-manged foreign large blend fund) and based on the "foreign large blend" category, I have the monthly data for foreign stock returns; I am using an average of "foreign large blend" (which being from Morningstar is survivorship-bias free...back in 2007 or so they recomputed all of their mutual fund category return data so that it is survivorship-bias free), TEPLX, and SCINX.
I also copied down the S&P 500 TR monthly returns from a post on Seeking Alpha so I have those.
I can try and copy them from my Excel spreadsheet to post here but I'd rather just post the whole spreadsheet and let everyone download it. Is it possible to actually attach a small .XLS file on this forum or do I need to have it hosted somewhere else?
The foreign stock returns data mentioned above show that foreign stocks and US stocks both did somewhat badly in 1962 and then from 1963 to 1966 foreign stocks underperformed US stocks by quite a significant margin; in 167 foreign stocks underperformed US stocks somewhat; from 1968 to 1969 foreign stocks seriously outperformed US stocks (including actually having a positive year in 1969 when US stocks lost money), and for 1970 US stocks made about three percent and foreign stocks (represented by the MSCI EAFE TR in 1970 since it started on 12/31/1969) lost money.
Now I am only missing one more thing to complete the backtest: Monthly intermediate-term Treasury bond returns so as to simulate an intermediate-term bond fund. Morningstar has the Ibbotson Intermediate-Term Government Bond and Ibbotson Intermediate-Term Corporate bond monthly returns back to at least 1962 monthly and 1926 annually but subscribing to their Encorr or Principia or corporate memberships cost thousands of dollars which means I don't have access to them. I do have daily 5-year Treasury yields back to 1-1-1962 so if someone could create and post a spreadsheet that can compute monthly returns based on changes in interest rates over the month then that would give as a fairly accurate approximation of an intermediate-term government bond fund monthly return series. I am aware that I could just use Treasury bills as the "switch into when stocks haven't outperformed T-bills over the past twelve months" asset but that would be kind of cheating since this was a time of rising rates and thus intermediate bonds would have done slightly worse than T-bills and I want the backtest to fully reflect that (plus Mr. Antonacci's backtest on the GEM webpage used intermediate-term bonds and I want to replicate his methodology as closely as possible).
Can someone please help with the bond returns data? Thank you.
OK, a little while back I said I was going to try to get some hypothetical returns for foreign equities so as to extend the backtest before 1971 which is as far back as the GEM backtest page goes; I needed the returns from 1963 or so (actually January 1962 since GEM requires a one-year lookback) as that is when rates started significantly rising and pretty much either stayed flat, rose slightly, or rose a lot every year until the end of 1969...I wanted to be able to do this backtest in a rising-rate environment since it involves switching into intermediate-term bonds when stocks haven't beat the three-month T-Bill.
Well, after some laborious manual chart scraping of Morningstar data for the only two foreign funds that go back that far (TEPLX - Templeton Growth which is a foreign large growth/large blend fund, and SCINX - which is now an index fund based on the valuation/earnings/dividend-based CROCI index but before 2013 it was DWS International and before the early 1990s it was Scudder International hence the ticker SCINX...before it switched to being an index fund it was an actively-manged foreign large blend fund) and based on the "foreign large blend" category, I have the monthly data for foreign stock returns; I am using an average of "foreign large blend" (which being from Morningstar is survivorship-bias free...back in 2007 or so they recomputed all of their mutual fund category return data so that it is survivorship-bias free), TEPLX, and SCINX.
I also copied down the S&P 500 TR monthly returns from a post on Seeking Alpha so I have those.
I can try and copy them from my Excel spreadsheet to post here but I'd rather just post the whole spreadsheet and let everyone download it. Is it possible to actually attach a small .XLS file on this forum or do I need to have it hosted somewhere else?
The foreign stock returns data mentioned above show that foreign stocks and US stocks both did somewhat badly in 1962 and then from 1963 to 1966 foreign stocks underperformed US stocks by quite a significant margin; in 167 foreign stocks underperformed US stocks somewhat; from 1968 to 1969 foreign stocks seriously outperformed US stocks (including actually having a positive year in 1969 when US stocks lost money), and for 1970 US stocks made about three percent and foreign stocks (represented by the MSCI EAFE TR in 1970 since it started on 12/31/1969) lost money.
Now I am only missing one more thing to complete the backtest: Monthly intermediate-term Treasury bond returns so as to simulate an intermediate-term bond fund. Morningstar has the Ibbotson Intermediate-Term Government Bond and Ibbotson Intermediate-Term Corporate bond monthly returns back to at least 1962 monthly and 1926 annually but subscribing to their Encorr or Principia or corporate memberships cost thousands of dollars which means I don't have access to them. I do have daily 5-year Treasury yields back to 1-1-1962 so if someone could create and post a spreadsheet that can compute monthly returns based on changes in interest rates over the month then that would give as a fairly accurate approximation of an intermediate-term government bond fund monthly return series. I am aware that I could just use Treasury bills as the "switch into when stocks haven't outperformed T-bills over the past twelve months" asset but that would be kind of cheating since this was a time of rising rates and thus intermediate bonds would have done slightly worse than T-bills and I want the backtest to fully reflect that (plus Mr. Antonacci's backtest on the GEM webpage used intermediate-term bonds and I want to replicate his methodology as closely as possible).
Can someone please help with the bond returns data? Thank you.