The "trend following" Permanent Portfolio signals

[ochotona]

This is the Permanent Portfolio forum, after all. Here is a signals service for running a tactical PP.

The technique is 12-month Absolute Momentum. If the 12-month return of the asset is greater than the return of T-bills, then hold it. If less than T-bills, then go to T-bills  or cash. You only trade once a month, on the same day of the month, as close as you can manage.

SIGNAL AS OF 1/15/2016

US Stocks: 0%
Long Treasuries: 25% 
Gold: 0%
T-bills / Cash: 75%

If you are playing the "invest in the top two assets out of four game", then:

Long Treasuries: 50% 
T-bills / Cash: 50%

We're not trading Long Treasuries, just the Stocks and Gold. Long Treasuries and Cash is about equal to a 10 year Treasury in volatility. I want to trade a minimum amount where it does the most good, and not trade everything.

If this thread goes dark, wake me up, and I'll report the latest. It's pretty easy, anyone can look this up for themselves at any time.

Zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz......

[ochotona]

Do you include interest and dividends in the calculation?

Yes, it's just the 12 month total return from Morningstar.com

I'm going to turn off trend-following on bonds. I think when you consider bonds and cash together as a unified barbell, who cares? Together, they are WAY less volatile and loss-inducing than stocks and gold.

This thread is not trying to invalid Harry Browne's basic finding that the low-correlations between the four asset classes compensate for each other. That's a historical truth. Nevertheless, beloved Mathjak107 did make a good point, that the un-correlated volatilities in the portfolio tend to just fight and thrash with each other, and the weak underlying positive return, if any, has a tough time to manifest.

I want to see what happens if we just suppress some of the individual asset volatility when it does the most damage (in stock bear markets, or gold post-1980, post-2011), using a simple, slow-trading, 100% rule-based approach, and see if anything good comes out of it.

When Faber trend-followed the PP using 10 month moving averages, he improved the CAGR to about 10% from 1972-2010, and it looks less volatile than the original PP (he calls it PERM, unfortunately). He writes, "....the timing helps due to the highly volatile gold and US equity components."

Over on www.portfoliovisualizer.com, I can do the kind of test I want using Dual Momentum on VTI and IAU, Cash and bonds we're not trendfollowing.

From 2006-2015,

                          CAGR  Sharpe  MaxDD
Trend-following  10.8%    0.64      26%

Buy & hold        8.25%    0.60      26%

I think most PP holders would not mind an extra 2% of CAGR over the past decade.

From a starting $10,000 investment,

Trend-following  $ 27,887

Buy & hold        $ 22,094

DELTA = 26%

[Dieter]

Wh did you pick trend following over moving averages? (Did link to a moving aberages study even.)

Why moving to cash? I played around a bit with moving averages at portfolio visualizer, and switching to int term treasures instead of cash helped.

What with ITT yielding what cash used to....

[ochotona]

Chose 1-year return vs moving average after studying Antonacci's work against Faber's work. Both work, results about the same. Antonacci trades much less, making the portfolio lazier (= better).

[ochotona]

Meb Faber's ETF "VAMO" is supposed to have active hedging risk-reduction. It's not buying / selling, but using futures, but I'm not convinced it's working properly. It's too new. But it would be cool if it did work, the only other active management would be you running to the UPS Store carrying many small and strangely dense little boxes to and fro...

[ochotona]

Portfolio 1: US Equities
Portfolio 2: Gold
Portfolio 3: 50% cash + 50% long US Treasuries

The equities and gold are where the risk lie in the PP.

[ochotona]

In another thread, it has been suggested that a catastrophic crack-up of long Treasuries could occur if interest rates go negative, then start to come back from negative towards positive. So I'll publish 1-year lookback TMOM signals for TLO as well as SCHB, and IAU. That will be Monday, Feb. 15.

[dualstow]

In another thread, it has been suggested that a catastrophic crack-up of long Treasuries could occur if interest rates go negative, then start to come back from negative towards positive.

I think that was suggested by you, wasn't it?
The main thing suggested by Craig was that it probably wouldn't be worth it to him to hold long treasuries below 1% yield.

[ochotona]

The trend-following PP signals are issued today for February 15, 2015, President's Day, a NYSE Holiday

Stocks: STAY IN CASH 
Gold:    HOLD               
Long Treasuries:  HOLD
Cash: HOLD (always)

This uses a 1-year total return of the asset criteria. It must be above the 90-day T-Bills interest rate, the risk free rate.

Optionally, as discussed earlier, you could just buy and hold cash and long Treasuries, and never go into and out of the Treasuries, because the cash - Treasuries barbell isn't that volatile compared to gold or stocks.

Thank you.

[Cortopassi]

You say stocks:  go to cash, but were they not already in cash on Jan 15?  Your first post shows them at 0%?

If that isn't the case, then this would bear out my biggest concern, you will get whipsawed out here at a short term bottom, market has some potential to rise next month, and buy back in higher.

[ochotona]

You say stocks:  go to cash, but were they not already in cash on Jan 15?  Your first post shows them at 0%?

If that isn't the case, then this would bear out my biggest concern, you will get whipsawed out here at a short term bottom, market has some potential to rise next month, and buy back in higher.

I clarified the prior post. Yes, you will get whipsawed from time to time. If that bothers you, don't do trend-following. I got whipsawed in Nov-Dec 2015. It wasn't fun, but at least I had a specific methodology to follow, to guide me past my emotions.

The Antonacci method makes a round-trip trade every 1.3 years, on average.

[ochotona]

The trend-following PP signals are issued based on the close on March 14, 2016:

Stocks: STAY IN CASH 
Gold:    HOLD               
Long Treasuries:  HOLD
Cash: HOLD (always)

This uses a 1-year total return of the asset criteria. It must be above the 90-day T-Bills interest rate, the risk free rate (0.34%).

Optionally, as discussed earlier, you could just buy and hold cash and long Treasuries, and never go into and out of the Treasuries, because the cash - Treasuries barbell isn't that volatile compared to gold or stocks.


Thank you.

[MachineGhost]

Why are you using the middle of the month?

[ochotona]

I'm going to get back onto the end-of-month schedule.

ALL FOUR PP ASSETS ARE "ON" AS OF 4/1/2016

[var]

are u using real money to do this?

[ochotona]

are u using real money to do this?

OH YES.... indeed... 60% of my entire portfolio is trend-following. This is a serious commitment.

[Cortopassi]

Ocho, the 10 month Ivy signals got back in stocks on April 1.  From the chart, looks like late 2015-2016 so far has been possibly the most whipsawed period of the past 20 years?  It's worked well in the past, but more than anything I am personalyl just using this as confirmation bias that I am doing the right thing by not doing this and just using PP bands.

[MachineGhost]

Ocho, the 10 month Ivy signals got back in stocks on April 1.  From the chart, looks like late 2015-2016 so far has been possibly the most whipsawed period of the past 20 years?  It's worked well in the past, but more than anything I am personalyl just using this as confirmation bias that I am doing the right thing by not doing this and just using PP bands.

No, 1994 had more whipsaws.  4 in a row if I recall correctly.

What trend following does is basically shift risk; in return for giving up half of the drawdown, you take on whipsaw and tracking risk.  No free lunch it seems.  The trick is that the when-invested gains overcome the cumulative whipsawed losses as well as the missed opportunity losses from not being fully invested at a bottom.  It does do that which is the only reason why it works.

I like downside risk management so long as I don't personally have to be responsible for it.  It's emotionally tough to be flittering in and out of the markets with huge chunks of your capital.  You have both fear of missing out as well as fear of being whipsawed.

Besides, the PP hedging only works by relying on the assets to be fully invested at all times.  Using trend following on the assets changes the correlated risk parity such that the optimal weights may be entirely different than a fully invested PP.