The liquidity system as I last defined it on 10/31/16 provided a buy signal today for IWM. I had revised my financial indicator such that the new index was:
= Price of XLF x .0386 * price of XLRE. (1)
XLRE is an equity REIT ETF. I was calculating this index each day to determine a high, low, and close value for the index. Usually XLRE was not hitting its high at the same time as XLF. So the highs and lows for the index were theoretic, coupling the intra-day highs of each stock no matter when they occurred during the day. If this index hit a 20-day high during the day, the strategy was to buy IWM if not already in it.
I had also tried using ICF and VNQ, two other equity REITs, to calculate the index, but I have now determined that is redundant, and so I will just use XLRE.
Also I have recently realized that I made a mistake in the formulation of my indicator with equation 1, because it weights the REIT sector and the financial sector (excl. REITs) equally. I wanted the sector weights to be like they were before real estate was removed from the financial sector. This equation, I think, accomplishes that.
Index2 = price of XLF + (.139 x price of XLRE).(2)
Regrettably index1 has performed better since October 1 2016 because index 2 had me selling IWM on October 26 as a result of a 55-day low, and then index 1 promptly had me buying again on October 27. Then the Russell 2000 quickly hit a new 55-day low, and the system sold it again, and as a result missed the Trump rally, because the system also has a rule of no repurchase of IWM for two months after the last purchase of IWM. Index1 did not have this problem, because after the sell on October 26, it didn't signal a buy until 11/9, which was more than 2 months since the previous purchase of IWM. There was no sell after 11/9, until 3/22/17.
I have no idea how an index based on an equal weighting of financial and real estate stocks will do in the future, or did in the past. I can't back-test it because the two sectors were measured as one, and there are no historical indexes that separate them. So I will monitor, and report, the buys and sells based on both indexes from now on, and we see how they do.
I will dump the buying of PTY. I was only including PTY for years where Shiller's PE indicated that stocks were over-priced. So now the systems will buy 100% IWM when the financial index indicates to do so. IWM will be sold when either the index drops by 2ATR (average true range) from the 20-day high which signaled a buy, or hits a 55-day low, or the Russell 2000 drops by 2ATR from the point at which it was bought, or the Russell 2000 hits a 55-day low, which ever occurs first.
So system 1 will be based on Index1, which signaled a buy today, and system 2 will be based on Index2. System 2 is still in FLOT (see March 22 blog). System 1 sells 1877 shares of FLOT at $50.85. There was also a $.055 dividend from FLOT, and so the proceeds are $95,548. This bought 694 shares of IWM at a price of $137.53.
Traders or timers must adapt. I've learned this before, but it is very frustrating. I'm also thinking about how to modify the rule of not buying IWM again, until 2 months after the last time it was purchased. The rule worked well in back-tests, but the Trump rally punished it.
Disclosure: I am/we are long FLOT.