Test of Boucher Strategies: Using Treasury Yields to Trade the S&P 500

14Nov08

This is a test of two of Mark Boucher’s strategies from his book “The Hedge Fund Edge”. I have to rat myself out here – I usually can’t sit still long enough to read whole books (hence the reason I read short and concise blogs) – so I’m basing this test on a description from a great blog, Trader’s Narrative.

Both strategies are based on the idea that stocks and bonds compete for investors, and when bond yields are attractive enough, investors will begin to move away from stocks and towards bonds (a bad omen for stocks).

STRATEGY #1: 1-YEAR CHANGE IN 10-YEAR TREASURY YIELDS

Strategy in green and S&P 500 buy & hold in blue from 1963:

[logarithmic-scale, return on cash = half of nearest 13-week UST]

Rules: Go long the S&P 500 at today’s close when 10-year US Treasury yields have increased less than 9% over the last 12 months. Close the position and move to money market when yields have increased by more than 9%.

Note that when I say 9%, I mean as a percentage of the previous rate. So a move from 5.0% to 5.5% would be an increase of 10% (not 0.5%). Also note that Boucher originally used 30-year treasuries, but more data is available for the 10-year flavor and the test results are basically the same.

Since 1963, this strategy averaged 7.6 trades per year and was long 70% of the time. In the 38 years prior to 2000, this strategy consistently matched or outperformed the market, and sidestepped all major drawdowns. That’s the good news. The bad news is that since 2000, the strategy has done poorly. Treasuries yields have mostly fallen over that time, but unfortunately, so have stocks.

Stats for the number-lovers are included below. I’ll share my thoughts about this strategy after I discuss strategy #2.

 

STRATEGY #2: 10-YEAR vs 3-MONTH TREASURY YIELDS

Strategy in green and S&P 500 buy & hold in blue from 1962:

[logarithmic-scale, return on cash = half of nearest 13-week UST]

Rules: Go long the S&P 500 at today’s close when the ratio of 10-year to 3-month US Treasury yields is greater than 1.15. Close the position and move to money market when the ratio is less than 1.15.

Since 1962, this strategy averaged 7.1 trades per year and was long 66% of the time. Results were similar to the previous strategy, although this version was long during the 1987 crash (probably more coincidence than a statement about the strategy) and underperformed the market a good deal during the bull run of the late 90’s.

And for the number-lovers:

 

MY THOUGHTS

The results show that these particular approaches to measuring the attractiveness of bonds versus stocks worked well in the distant past, but not so well in the last decade or so.

I am a pretty active trader, usually holding positions between 1 and 5 days. I’m very wary of systems like this one that take such long-term positions for two reasons: (a) there are so few historical trades to study that the risk of curve-fitting is increased and (b) because there are so few trades, we usually have to test a very long period of time, and as we’ve shown multiple times on this blog, the markets are constantly evolving. I think the failure of this strategy to outperform in the last decade can be attributed to these points.

In my next post, I’m going to demonstrate a system that takes a much shorter-term, and I think much more effective, approach to trading the bond vs stock relationship. More to follow.

Happy Trading,
ms

 

To stay up to date with what’s happening at the MarketSci Blog, we recommend subscribing to our RSS Feed or Email Feed.

Filed under: Trading Strategies, Treasuries & Interest Rates   |  10 Comments