Portfolio Design

Advanced concepts in stock investment portfolio design.  Fundamentals, technical analysis and many other related topics are discussed.

Sector Rotation And Market Timing Combined Part 4

In Sector Rotation And Market Timing Combined Part 3 - A New Market Timer, a new market timer signal based on the "Market Wheel", was developed. In Part 4, this indicator is added to the suite of indicators that provide a bull/bear consensus, deciding whether to be invested in the market with one of nine sector ETFs, or out of the market holding 10Y Treasury Notes (ETF Symbol: IEF). 

Stock market analysis fueled by Portfolio123.


For this design, the ranking system from Part 3 has modified, by adding 5 market timing indicators described in past posts (one of them is proprietary). These are shown below.

If you have your own version of market timing signals then you can use those instead. Note that I have lowered the voting threshold from 50 to 40. With all of the market timers included, there is little sensitivity to threshold value but 40 gives slightly better results.

 

Stock market analysis provided by Portfolio123.

Sector rotation ranking system - TRUE portion of the conditional node

The FALSE part of the conditional node is modified and now consists of three nodes as shown below.

Sector rotation ranking system - FALSE portion of the conditional node

The formula for the Bull/Bear node is:

Eval(NodeRank("MT Indicators") > 40, $XLY | $XLK | $XLI | $XLB | $XLE | $XLP | $XLV | $XLU | $XLF, $IEF)

The formula for Income Yield is:

Eval($XLY, NodeRank("SP1"), Eval($XLK, NodeRank("SP2"), Eval($XLI, NodeRank("SP3"), Eval($XLB, NodeRank("SP4"), Eval($XLE, NodeRank("SP5"), Eval($XLP, NodeRank("SP6"), Eval($XLV, NodeRank("SP7"), Eval($XLU, NodeRank("SP8"), Eval($XLF, NodeRank("SP9"), 0)))))))))

The formula for Momentum is:

 EMA(50)/EMA(200)


Pay particular attention to the node weightings.  The weighting for the Bull/Bear node must be larger than for the Income Yield node, and the Income Yield node weighting must be larger than the momentum node. Otherwise, the results won't be what you want.

This ranking system is essentially momentum-based with falling valuation sectors removed from contention. If anything, this exercise demonstrates that there really is something to be said for combining value and momentum.

The custom universe is shown below.

Sector rotation custom ETF universe

The simulation backtest results are shown below.

Sector rotation simulation backtest results

The risk measurement statistics show an annualized return greater than 19%, maximum drawdown less than 24%, and an alpha exceeding 16%.

Sector rotation simulation risk measurement figures

The one negative for this system is the low average return per trade of 2.13% before fees and slippage.

Sector rotation simulation trade statistics

This is where skill for improving profit per trade comes into play, but I'll leave that to the reader's ingenuity. I would like to note that the technique described here is limited only by one's imagination. One value factor was used (IACTTM/MktCap) with the addition of momentum.  By using this strategy in conjunction with other fundamentals and/or technical indicators, who know what you can come up with!