A.I.M. Users Bulletin Board

[ls7550]

Hi jaiml

In that set of charts I'd set dividends to be added to cash to see differences to that of reinvesting dividends, the differences were small/negligible. So the cash value increased as dividends got added to cash. Basically I posted the wrong charts, but other than that cash rebuild had very similar appearances/outcomes.

My recent studies are indicating that AIM of real S&P price has become more positively biased and hence higher cash being indicated due to post 1985 US dividend tax changes that led to more of earnings being retained. Looking like Dow/Gold stochastic is the better choice since the 1980's. Basically the lower historic Dow/Gold ratio value is near zero whilst the upper is near 40, so a quick guide is ( current - 0 ) / ( 40 - 0 ) ... or simply current Dow/Gold / 40. With Dow at around 36,000, gold at 1800 = Dow/Gold of 20 and a stochastic of 20 / 40 = 0.5 (50% cash). If the Dow/Gold hits new highs, say 45 then the divisor is changed to that.

Increasing the trade signals seems to work well, SAFE=0%, min trade size $1, throws off more/faster buy trades and has you all-in quicker. In effect buying into each pull-back. If historically the worst case peak to trough that resulted in 4% SWR sustaining 30 years then if you can average in at a lower than that peak, say 5% less, then the conceptual SWR rises to 4 / 0.95 = 4.2%. More relevant however is that Dow/Gold can indicate when prices are relatively high and keep back more cash such that if a big drawdown does occur relatively soon after purchase the cash can significantly dilute down that hit, and early year dips (SORR) is more critical than big dips after a number of years when reasonable gains might already have been achieved (give back of other peoples money rather than a hit against ones own capital).

I'm seeing a wide range of how long before all-in occurred, in cases of where low stock/high cash was initially indicated and a sustained Bull period followed then cash could have lasted for many years. In other cases typically after declines after starting cash could be burnt through quickly. Generally I'm seeing AIM tending to do better than lumping in if bear/bunny periods followed, in some cases with considerable/large differences in outcome, or near comparing slightly lagging if a sustained Bull period followed initialisation. At least when initialised to Dow/Gold stochastic based initial stock % (cash %) weightings.

I've also been testing gold, but where the stochastic is reversed for that i.e. 1 - stochastic for the amount of cash% value (or read the stochastic value to = % amount of gold to buy).

When SAFE and MTS are pretty much discounted, it pretty much becomes a PC - SV (or SV - PC) buy (sell) type monthly review/measure/indicator (where actual sales are ignored and in both cases (sell or buy), PC is increased by half the indicated trade size amount. Quite a nice simple method, more so with the simpler Dow/Gold value divided by 40 type stochastic indicator calculation method.

Purely by eye back in the late 1960's the Dow was around 900 whilst gold was around 35 so a Dow/Gold ratio of around 26, which was around a new high in the Dow/Gold ratio so the stochastic was pretty much indicating 100% cash. AIM wont work with zero stock value (PC=0, SV=0 unchanging), so dropping say 10% into stock to get AIM up and running would have been reasonable. As per that above Dow chart link the Dow then dived up to the mid 1970's so you'd have averaged in all that cash during those years whilst having minimised drawdowns compared to all-stock.

1999 and Dow was very high, gold was very low, so a high Dow/Gold ratio. Again seeing the Dow/Gold ratio breaking through into new historic highs, so low initial %stock exposure (high %cash).

Early 1980 and Dow/Gold ratio was down near 1.0 levels. So a very low %cash would have been indicated, near/at all-in from the offset and that partook of the great 1980/1990 up run in stocks.

More recently and its pretty much saying neutral, 50/50, no clear/obvious over or under valuation. So half lumped in, the other half averaged in progressively at pullbacks.

Clive