A.I.M. Users Bulletin Board

[ls7550]

Hi Vitaali

Running out of cash before the bottom isn't critical. More important is that you haven't fed a asset with your reserves that continues on down towards total failure. i.e. diversify and hold assets that are less inclined to total failure.

Also there are various forms of 'averaging'. Typically the 'average' is a great outcome as that avoids the worst case outcomes.

UK data from 1896 for 30 year after inflation total returns (UK over that period has somewhat reflected a global index as a number of the largest stocks are just listed out of London, predominately have global business activities). The US over that period was a right tail (great outcome), others saw much poorer outcomes. Global exposure will better ensure 'average' outcome, rather than poor outcome, but that average could be lower than the US average if the US continues to be a right tail good outcome (with the way China is going, who knows, Japan for instance during the 1970's/1980's rose from nowhere to knock the US back from 75% down to 25%. Japan's 1990's however saw their rise drop rapidly away again when the US again restored its predominance). That aside, here are some figures for UK since 1896 total 30 year after inflation, costs and taxes outcomes with dividends being reinvested

	worst	median 
 50/50	1.088	2.11 
 67/33	1.057	2.44 
 75/25	1.026	2.62 
 80/20	1.002	2.61 
 100/0	0.872	2.65

In the median (average) case you pretty much ended up with 2.11 to 2.65 times the inflation adjusted start date value after 30 years, a little under 3% annualised real gain after costs/taxes, whether you followed a 50/50, 75/25 or 100/0 stock/bond asset allocation. In the worst cases you barely broke even.

Contrast that with the median (average) gross total returns

50/50	3.04 
 67/33	3.99 
 75/25	4.30 
 80/20	4.45 
 100/0	5.31

Using median better ensures a more likely actual average, if you use arithmetic average the tendency is to see higher values due to rare extremely good cases, that in practice you're unlikely to actually encounter.

From the first of the above two tables, net of taxation/costs worst cases for accumulation (dividends/interest reinvested) weren't that dissimilar whether you used 50/50 stock/bonds or 100/0. The same relatively little differences are a common pattern. Weighting 50/50, 67/33, 75/25, 80/20 or even 100% stock didn't make much difference. Tax and cost efficiencies however did make a big difference.

Cost averaging is one approach to better ensure you avoid the worst case outcome, but in reality it doesn't make that much of a difference. What does make more of a difference is what Lichello indicated to be seed averaging. Which is somewhat like cost averaging but where you rebalance between holdings, looking to more equal weight individual assets. If asset A is up 25% and asset B is down 25% then reduce A to add to B.

There's a yet better form of averaging however that does also make a big difference - time averaging. Accumulating stocks in younger years by buying more stock each month/year/whatever is a form of time averaging, as is drawing down the portfolio in retirement years (taking $x of income from the portfolio each month/year). You can however scale that. If you start with a allocation of 50/50 and over 30 years increase that to 100/0 by uplifting the stock allocation by 1.67%/year, so 50/50 in year 1, 51.67/48.33 in year 2 ...etc. all the way up to 100/0, then that broadly is no different to having held 75/25 constantly. Yes in practice direct comparisons will see differences in outcomes (rewards), however its like more was being averaged into stocks over time (time diversification), that is more inclined to avoid the worst case outcome. If you lumped in and held continually 75/25 weightings from day 1 in total disregard of costs/taxes (high cost tax inefficient holdings), then as per the above tables after 30 years of accumulation you might just have the same amount in inflation adjusted terms as at the start (1.026 times 'more' - just 2.6% more inflation adjusted value than at the start). In contrast tax/cost efficient and avoiding the worst case through cost/seed/time averaging ... and you're more inclined to see 4.3 times more inflation adjusted value at the end of 30 years compared to the start date value (5% annualised real gain after discounting inflation). A negative factor however is that in better ensuring against the worst case outcome, averaging also ensures you wont hit the best case outcome either (difficult to see however from present day relatively low interest rates/inflation the next 30 years being a great/above-average outcome).

Cost averaging, holding some stock, some bond and rebalancing either back to target fixed weightings or using AIM (that might run out of cash), is only one part of the whole, and a relatively small part of 'averaging'. Don't be too concerned with that, there are more important factors to focus upon. If your AIM runs out of cash during a share price decline sequence, before the bottom, whilst yes it would be nice if it hadn't, you still achieved a degree of cost averaging, albeit not the optimal case. Similarly in not having deployed all cash in other cases equates to too much cash drag, cash that is never deployed. We can't predict how much cash is appropriate/optimal in advance, only in hindsight.

In drawdown, imagine you come into a lump sum of $1M. You might plan to draw that down over 30 years, but want to leave a inheritance or have funds available in the case of longevity. Say you're content with a 2.5% SWR (25K/year disposable income that rises with inflation each year). From the above, tables, if we opt for the 75/25 figure then that in the cost/tax efficient case yielded a 4.3 factor gain over 30 years in the median case. If we're cost/seed/time averaging we're more incline to achieve that average (no guarantees that the future will reflect the past however). So in order to 'insure' our $1M value we drop 1 / 4.3 = 23% of the $1M into a growth bucket, that on average grew to $1M value over 30 years. The 77% rest, we can invest in inflation pacing safe bonds and draw-down/spend those over 30 years (2.5% safe withdrawal rate (SWR)). And in the average case, once all of those had been spent, the growth bucket would be valued at $1M in inflation adjusted terms.

For that growth bucket, rather than 75/25 constant target stock/bond weightings, we might start with 50/50 and progressively increase that towards 100/0. That still averages 75/25 overall but that in effect added more into stock exposure over time (time diversification). That has lower early year risk (when its 50/50), higher late year risk (100/0), but that is a more appropriate weighting for younger heirs. We've in effect lowered our risk, and shifted it towards heirs who typically will have enough human capital to tolerate those levels of risk. One factor to be mindful of with such time averaging is that it does produce a bowl U shape type value progression. Portfolio value will curve downwards in earlier years, recoup (curve back up again) in later years. Without awareness of that you might be inclined to 'give-up' and rotate into something else upon seeing apparently poorer interim outcomes compared to others. Compare that approach to another who lumped all into 75/25 constant weighted in a cost/tax inefficient manner and at the worst possible time, who even though they cost-averaged (help stock and bonds and rebalanced between the two) saw a outcome of just having paced inflation over the 30 years (likely would have done just as well as having left the money in interest bearing cash deposit accounts). Or worse ... if they were spending money (funding retirement) and possibly have seen the money totally exhausted/spent after maybe 20 years.

Sorry for the intended brief posting becoming a lecture. Hope its of use.

Clive.