Gilgongo wrote:I don't see much about SoR on these boards, and I get the impression elsewhere that it's more spoken about by financial advisors keen to present themselves as experts than something we should worry (too much) about.
Is that an accurate view, or is SoR something that people pay attention to (at least initially in retirement)? If so, I assume you have done something differently in the last 12 months than you did in the preceding years?
SoR risk is predominately a risk for those just entering into retirement/drawdown so a relatively specific risk. For accumulators a 'bad' sequence of returns, price declines, isn't a risk and is more inclined to benefit the investor as new money buys more shares than might otherwise have been the case. For those established into drawdown/retirement a bad sequence of returns might only reduce/give back some of gains already achieved, +50% up and then seeing a -33% decline isn't so bad/critical as seeing a -33% decline immediately after starting retirement/drawdown.
A bucket approach combined with averaging can substantially reduce SoR risk.
Consider a bad first decade where portfolio total accumulation rewards yield -5% annualised real (after inflation) decline. Assuming a 4% SWR is also being drawn then we can model that with
V = ( V - 0.04 ) * 0.95 ... repeated 10 times and were V initially = 1.0
1
0.912
0.828
0.749
0.674
0.602
0.534
0.469
0.408
0.349
0.294
... i.e. ends up with less than 30% of the inflation start date amount remaining at the end of 10 years of repeated -5% real annualised declines compounded with 4% SWR.
Compared to if no withdrawals, just the -5% annualised
1
0.950
0.903
0.857
0.815
0.774
0.735
0.698
0.663
0.630
0.599
that ends up with around 60% of the inflation adjusted start date value.
If instead we dropped 40% initially into bonds to draw that down over 10 years, and had the remainder 60% initial amount enduring that latter case outcome, then after 10 years with all of bonds spent we're left with 0.6 x 0.6 = 0.36, 36% of the inflation adjusted start date amount, which is better than the first case above that ended with 29.4% remaining.
Starting with 60/40 stock/bonds, spending the bonds such that it transitions to 100/0 stock/bonds, averages 80/20 stock/bonds. And is a form of time-averaging more into stocks that tends to do OK across a period when stocks declined. Somewhat like being both in drawdown and still accumulating at the same time that reduces SoR risk.
Combined with averaging the start date, not migrating all-in into the retirement asset allocation at day one, but instead averaging in half at day 1 and half after a year, helps to avoid having lumped all-in at the worst possible time. To facilitate that you might have a separate cash pot to cover the 6 months of spending whilst 'averaging into' the full retirement asset allocation.
The combination of both buckets and averaging has the effect of shifting the likes of Trinity Study data risk more towards the left, de-risks.
The 4% SWR rate becomes more like the 3% SWR rate type outcome. Shifts more into a 100% success rate type probability. At the expense of potentially/likely yielding slightly less overall upside rewards. Subjectively that 'cost' might be nothing, or could be 1% annualised i.e. comparing the difference between
100% stock and 80/20 average stock/cash and it depends upon which time period you might compare as to whether there was near no difference or a larger difference.
Negative real yields do have influence upon the 'bond' holdings. 40% in cash deposits to cover 10 years of 4% withdrawals/drawdown might see that not last the full 10 years due to cash deposit interest lagging inflation. But that can swing the other way around, former negative real yields transition to being positive real yields that balance out/wash. Whilst the next five years for instance might see cash interest lagging inflation by say 2%/year, if years 6 to 10 see cash earning +2% real then overall that all balanced out.