Clive, RE: Personally I'm not very comfortable with.....
Nor am I. The reason being that the earnings estimates by V/L are just that; future earnings estimates and based likely on best case situations. The inputs are: future earnings estimates for 1700 stocks x price projection multiple = VLAP. Further compounding the problem is basing the AP on 1700 stocks. Some will meet or beat earnings, some will match, some will miss, some will drop out of the universe, some will go out of business or get merged away. The measure is too broad.
If one believes we are in a secular bull market, and the market crash is just a corrective to it, then the projection might be close to being correct. Believers of same should do a standard deviation study to see in past markets where in the bell curve the projections actually come out to. If one takes the opposing viewpoint of being in a secular bear market, then the VLAP is not addressing the likely outcomes except when the markets revert back to bull mode and P/E's have likely shrunk along with prices.
Currently, company analyst's estimates are being challenged by stock strategists as being too high for current market conditions, thus any future projected VLAP or P/E is going to be too high. Price would go lower to adjust to the lowered earnings. In each of the past 3 recessions, 1982, 1990, 2001, earnings have fallen on average 22% peak to trough. That's quite a leeway and from the figures I've seen, we're not there yet. In deep recessions, P/E's can contract to single digits as in 1974 and 1980.
Now, look at current values; on 10/24 the S&P was at 876.77 with a 12 month trailing P/E of 17.07. Last week the S&P closed at 968.75 with a P/E of 18.86. The snapback rally we've had has exceeded the 60yr historical P/E of 17.82. So, currently stocks are basically fairly valued. Now, going forward does one believe the company analysts or the strategists?
Because P/E ratios have been subject to variants in analyzing them, GAAP earnings, Operating earnings, as reported earnings, or a blend of past and future earnings (as V/L uses) statistically it's a nightmare getting an objective view of what the earnings actually are, what is fair value, and what is the best ball park projection to use.
Be that as it may, then the ladder approach is likely a better solution vs. broad relative valuation levels. Now from what I see, you're ladder protocol is based it on past and present valuations, dividend yields, and simple mathematical price projections. I would suggest a third method based upon the actual market pricing values. Those values are unique to each tradeable, thus using a broad valuation measure would not be necessary.
Your ladder appears to be "stepped" to fixed percentages matched to various pricing levels. There is another way to put the pricing levels into a more real world context mathematically, so the broad ranges on the ladder can be narrowed down and the cash reserve values could be matched to the pricing levels using an alternate to the fixed percentages.
An individual ladder could be created for each index, ETF, stock, or bond to reflect the estimated price projections both up and down. Moreover, when price projections or market conditions warrant, a the ladder could be resized to current conditions.
The bottom line is that like AIM, whatever way the market goes, the ladder will follow it up or down with the cash reserve and it's done mechanically not by subjective estimates that may or may not be near the mark.
Best regards, Tim
Market Data 
Markets 
AI
