Monthly Archives: January 2012

A simple stock/bond timing model – part two

Last week I shared a timing model that switches back and forth between stocks and bonds. The strategy changes holdings based on what’s been performing better in the recent past.  Well, I have another variation to show below.   This time we’ll look at correlations between stocks (S&P 500) and the 10 year treasury rate.  I’ll be using the same rolling time period from the previous model (90 days).  (Note: The exact number of days didn’t make too much of an impact in the previous model.  I tested 60, 90, and 120 day parameters.  All had very similar results.)

Below is a graph of the correlation between stocks and the 10 year treasury rate since 1962.  It looks noisy but remember there’s a lot of years in there.

I wanted to see if this correlation had any potential usefulness.  People talk about bond investors being smarter than stock investors.  Perhaps if stocks and bonds are moving in the same direction, it’s safer to bet on bonds.   Well, this seems to have been a good bet for most of the time period I looked at (some exceptions).  Look at the graph below.  The simple model invests in stocks when the correlation is below 0% and in cash when the correlation is above 0%.  (Note: I only includes gains from stocks here to highlight the stock moves the strategy manages to catch. Actual performance will be better when adding in bonds instead of cash.)

The model manages to catch a good portion of the upside in stocks while avoiding some of the nasty draw-downs (e.g. 1987 crash, 2000′s tech bust, late 2000′s credit crisis and 2011 plunge).  This strategy is weak when stocks are falling and treasury rates are rising at the same time.  This is what led to the crap-tacular performance in the 70′s.  The model would have suffered through most of the ugly bear market of 73-74 resulting in poor performance for that decade.

I tested the strategy to time entries into VFINX and VUSTX for the same period as last week’s model.  The correlation strategy performed even better.

In conclusion, this strategy seems to work well most of the time.  Performance can seriously suffer though when stocks are falling and treasury rates are rising at the same time for extended periods.

A simple stock/bond timing model

I was reading this post from Dynamic Hedge and got inspired to create something similar.  I created a model that switches back and forth between stocks (SP 500) and bonds (long term govt bonds) based on what’s been outperforming in the recent past.  The model is pretty simple and I’m surprised by how well it seems to have worked.  Check out the equity curve below.

Not bad, eh?  I used VFINX and VUSTX as my stock and bond proxies.  I’d really like to test this back further.  Will be looking for govt bond data to push it back.  I’d also like to add in gold as a third asset and see what it does.

Here’s what my timing signal looks like.  Long stocks when above 0; long bonds when below 0.

I’m sure many people have developed similar strategies.  I hope to explore the idea more.  Will keep you updated on new developments.

-Mike

Allocation change

The Alpha strategy switched from cash to XIV at today’s close.  I didn’t think this would happen till tomorrow.  I missed the switch this afternoon and plan to buy tomorrow morning.

VIX futures contango continues to increase.  This gives XIV a nice tailwind going forward.  We’ll see where she goes from here.

Allocation change

The Alpha strategy went from XIV to cash at today’s close. The last trade generated a 31% profit.  February and March VIX futures just barely settled with the right spread to trigger the switch today.  I expect the strategy to hop back into XIV in the near future (possibly next week).

Below is the trade history since inception. 

Volatility model to signal a move to cash this week

The Alpha strategy will likely move from XIV to cash this Wednesday or Thursday.  There could easily be a move back to XIV after a week or two.

See the Alpha strategy page for a review of the rules.  This switch will be due to the average front month premium over the prior 20 days rising above -5%.  I’ve plotted out the spread and average spread below.

Allocation change

The TSP strategy changed allocations from the G fund to the C fund at today’s close.

TSP allocation change possible tomorrow

A close above 1,282 in the S&P 500 tomorrow will signal an allocation change for the TSP strategy.  The model would allocate everything to the C fund at tomorrow’s close.

Last year was a crappy one for this strategy with a small loss overall.  I’m hoping the next trade is a good one.  Truthfully, I’m not very optimistic about the stock market.  The same was true though in September of 2010. This is when the last all-in move to the C fund was triggered.  Since I can’t really trust my intuition, I plan to stick to the model and live through the drawdowns when they come (as they must).

2011 Strategy Performance

2011 has come to a close.  Time to see how my trio of strategies did for the year…

TSP

The TSP model lost about 2.3% in 2011.  It lagged this year thanks to the fast drop in the stock market in August.  It started out the year in a 60/40 G fund/ C fund allocation and switched to a 100% G fund allocation in early August.  It remained all in the G fund through the end of the year.

Core

The Core portfolio gained about 2.5% in 2011.  Volatility was low as expected.  The gains in treasuries and gold holdings more than made up for losses from stocks in the portfolio.

Alpha

The Alpha model gained about 97.7% in 2011.  It wasn’t officially created in its current form till Sept 30.  The chart below shows returns had the model been followed all year.  The beauty of this strategy is it can earn high returns regardless of whether the market is rising or falling.  It’s incredibly volatile though.

 

Wishing everyone the best of luck in 2012.

-Mike

Follow

Get every new post delivered to your Inbox.

Join 137 other followers