Take the recent financial crisis as an example Larry. All that has happened is that prices have realigned to the gold value. So for many years (decades), priced in terms of gold, stocks have been in a bubble, but have since reverted back down.
Buying stocks at any point during that prolonged bubble amounted to no different to overpaying for stocks. Even cost averaging in (regular purchases) doesn't avoid that, it just costs averages in at an overall average over-priced level.
The general/potentially better approach is to cost average DOWN the average cost of stock. 50-50 in stocks and cash and rebalancing back to those weightings is one method to cost-average down the average cost of stock.
Buy and hold with s% stock, c% cash (or bonds) and yearly rebalancing is a form of AIM and in some cases you'll see little difference between that and AIM. Where AIM potentially benefits is that its more of a dynamic weighting rather than constant weighting and it adjust the 'constant' up and down as prices decline and rise. The counter to that however is that it can exhaust cash reserves, but if it does so near a bottom and reversal then that works out well. If the price subsequently to having exhausted cash continues lower then it can work out bad.
In a good case, where prices perhaps decline 50% ($100 down to $50 price) at which time the cash reserve is exhausted, and then the price rebounds 100% from ($50 to $100 price) instead of breaking even as does a buy-and-holder, in doubling up the exposure as the price declined then the position ends up in profit.
The Permanent Portfolio is another similar constant weighting approach, but it relies on maintaining an overall constant value of the total fund value. If for example I invest $50 in stock, and $50 in cash, a total of $100, and stocks decline 50% whilst cash remains level, then I now have $75 which I split equally (constant weighting) between stocks and cash = $35 invested in stocks, $35 invested in cash. In contrast the PP invests $25 in each of four assets, of which one is stocks. If stocks halve in price but the other three collectively gain $12.50 then I still have $100 and constant weight rebalance and add to stocks to make the amount in stocks $25 again.
In proportionate terms, the 50-50 approach after rebalancing had $35 compared to $50 at the very start i.e. 70% amount of the original amount of exposure. The PP has 100% of the original amount in stocks after rebalancing. AIM works in a somewhat similar manner but instead of relying upon the total fund value to remain constant, it instead scales up the stock value exposure.
Backtesters often fail to consider appropriate choices of start and end dates. A fairer approach is to measure between historic peaks (peak to peak), or historic troughs (trough to trough) as generally p-p and t-t will have a similar slope to each other and a similar slope to that of fair value to fair value. The long term average return from stocks reflects that same fair-value to fair-value (or p-p or t-t) slope. Many stock investors however never achieve that same slope as more often (including cost average in'ers) achieve more of a peak to trough or peak to fair value (lower) slope. Cost averaging down is more likely to achieve fair value to fair value type slope gains or possibly better against the amount of funds they on average have exposed to stocks. The best cases arise if stocks are held over a trough to peak period, in which cases the gains can be outstanding even over decades of time.
If a 50-50 weighted cost average down approach can work ok, another way to look at that is : if a blend of 50% small cap value and cash achieve a somewhat similar reward to 100% Total Stocks, then instead of investing 50% of the total in stocks why not use a 50-50 sub-split i.e. overall 25% SCV and 75% cash. Which is sort of the basis upon which Larry Swedroe's fat tails works.
As to OCroft's idea. As I understand it, you monitor the difference between stock value and portfolio control over time. i.e. excluding SAFE generally AIM is based on Portfolio Control minus Stock Value as the amount that needs to be traded. i.e. if PC is $10,000 and current SV is $9000 its indicating $1000 needs to be added. If PC is $10,000 and SV is $11,000 its indicating $1000 needs to be sold. What OCroft is suggesting is that you monitor that difference and only initiate an actual trade when the current trend of that difference changes to the opposite direction. If in April the PC - SV indicated +$1000, and then that increased to $1100 in May, but then declined back to $1000 in June, then that reversal in the increasing amount is the time to make the actual trade (in this case buy $1000). Again I've ignored SAFE in this example for simplicity.
Best. Clive.
News 
Market Data 
Discover 
AI
