All that Glitters 9/7/12- Reducing Exposure

I am reducing my exposure to gold miner shares. This is after the 24% gain in the gold miner index (ticker GDX) since July 24 and my gold miner timing model going negative. The 6 month is now above 2 and the 1 year model is at 0.4, neither model is at "buy" levels. You will find the models updated for the latest data below. The latest 6-month model is now 2.12 and the 1-year model is 0.4.

I am most concerned with the that the 6- month model is now above 2. Historically, the odds appear against you when the indicator is at or above this level. The following is the historical returns on the Phily Gold/Silver Miner Index (ticker XAU) for various time period when this occurs.
 

	1w	3M	6M	1y	2y	total
Avg	-1.51%	-1.86%	0.57%	-2.38%	15.48%	16.31%
Median	-0.87%	-2.17%	0.80%	-3.45%	15.18%	-4.05%
Max	10.22%	23.58%	28.47%	44.44%	65.55%	137.59%
Min	-14.26%	-23.71%	-26.82%	-64.56%	-19.15%	-29.84%
St Dev	5.0%	11.1%	14.4%	25.8%	20.4%	45.6%
Sharpe	-0.30	-0.17	0.04	-0.09	0.76	0.36
#>0	32	30	37	33	44	32
%>0	45.7%	42.9%	52.9%	47.1%	77.2%	45.7%

This compares to a buy and hold results of...

	1w	3M	6M	1y	2y	total
Avg	0.28%	3.07%	6.41%	13.98%	30.62%	57.36%
Median	0.27%	2.14%	4.74%	13.33%	29.82%	30.56%
Max	25.37%	89.59%	88.09%	151.78%	203.74%	270.67%
Min	-30.86%	-62.22%	-62.18%	-64.56%	-50.89%	-31.67%
St Dev	5.4%	15.1%	21.6%	27.5%	32.5%	75.1%
Sharpe	0.05	0.20	0.30	0.51	0.94	0.76
#>0	351	366	389	421	447	464
%>0	53.3%	56.4%	61.2%	69.0%	80.1%	70.5%

The timing mining is in a range when the performance of gold miner shares have performed significantly worse relative to the buy-and-hold scenario. 

This is even considering the potential for QE3 or QE to infinity. If you follow the money supply, a major component in the timing model, you would know that these figures are reported with a lag. Due to this lag, I have to estimate money supply figures for at least two weeks. I do this by using seasonal factors, the size of the Federal Reserve's balance sheet, and other time series relationships. Even assuming an increase of nearly $300 billion in money supply, or more than 3 standard event from current levels, the timing indicators would only be about flat on the 1-year model and a 1.4 on the 6-month model. A hold at best. To put the relationship into a deeper perspective, during the last two QE events, money supply growth accelerated but still took 6 or so months to reflect the increased size of Fed's balance sheet. In addition, gold miner shares increased during/following QE1 but declined in QE2/Operation twist. So, the idea that QE will automatically lead to higher gold miner stock prices is not necessarily set in stone.

This is not to say I am completely out of precious metal miners or gold. I am selling out the principal balances in my investments, but maintain positions representing past gains.