I fully agree with you that any approach has to be carefully and compared with the alternatives before drawing any conclusions. My, now slightly mistitled, thread at viewtopic.php?t=433101 has ended up comparing two approaches directly (with an implicit comparison of a third) for one example (a single female aged 65), with one metric (the amount of income, in real terms, 35 years later, aged 100 - effectively a measure of how good the longevity insurance actually is), and one target income (4% of the initial portfolio). The methods have been tested for annualised inflation values from 3% to 7%. I note that over 30-40 year rolling periods, historically since 1930, US inflation lay between 4% (median) and 5% (worst case). The three methods are:
1) Purchase of an SPIA at retirement (this is implied in the colour graphs presented later in that thread, post viewtopic.php?p=7913455#p7913455 onwards, for a deferral or delay of 0)
2) Purchase of a DIA at retirement with income provided by a TIPS ladder (assuming a yield of 2% in all but one post) during the deferral period for deferral periods of between 0 and 25 years (90 being the oldest quote I have found). Income from the annuity in excess of the target income is stored in a rolling TIPS ladder which is then withdrawn from when the real annuity income falls below the target
3) Delay the purchase of an SPIA by between 0 and 25 years. A TIPS ladder is constructed to provide the income during the delay period and a single, duration matched, TIPS is held to provide the premium. A rolling TIPS ladder is also used.
In comparing the methods 2) and 3), it would appear that, for a TIPS yield of 2%, Method 3 (i.e., a delayed SPIA) is significantly more robust against inflation than than a DIA (Method 2). However, at TIPS yields of 1%, the robustness against is (as expected) much less and similar for the two methods. The immediate purchase (i.e., Method 1) gives a worse performance then the best deferral for TIPS yield of both 1% or 2%.
A proper comparison would also consider other alternatives, for example
1) Stock/bond portfolio. Using the real returns in the Simba database, historically a 50/50 portfolio (TSM/TBM) with a 4% inflation adjusted withdrawal survived for 35 years 97% of the time, with the first failure occurring after 32 years or so (i.e., in the worst historical cases, the income after 35 years would have been zero). Of course, this runs market (and inflation) risk since future values of SWR may (or may not) be less than those found historically
2) TIPS ladder. For yields of 2%, a TIPS ladder would support 4% withdrawals for about 34 years (ignoring any interest rate and reinvestment risk from 30 years onwards), while at yields of 1%, a TIPS ladder would support 4% withdrawals for 28.6 years. In both cases, the income after 35 years would be zero.
Furthermore, there is one further consideration (the one you have already mentioned), namely how much does a retiree wish to give up to the insurance company. Taking your $1m example and the methods I have explored (assuming TIPS yields of 2%),
1) For an immediate purchase, $1m is handed over to the insurance company, and slowly returned from the outset.
2) For a DIA, and a 15 year deferral (i.e. 70 to 85), would entail using $763k to construct the TIPS ladder to provide income during the first 15 years and $237k to purchase the DIA.
3) For a SPIA delayed for 15 years, the TIPS ladder still takes $764k to construct (at 2% yields) and the premium available is $237k that is then placed, for example, in a duration matched TIPS.
In the event of death earlier than the delay/deferral period, the legacy would consist of the unspent TIPS ladders plus the premium (for delayed SPIA) or zero (DIA).
cheers
StillGoing
1) Purchase of an SPIA at retirement (this is implied in the colour graphs presented later in that thread, post viewtopic.php?p=7913455#p7913455 onwards, for a deferral or delay of 0)
2) Purchase of a DIA at retirement with income provided by a TIPS ladder (assuming a yield of 2% in all but one post) during the deferral period for deferral periods of between 0 and 25 years (90 being the oldest quote I have found). Income from the annuity in excess of the target income is stored in a rolling TIPS ladder which is then withdrawn from when the real annuity income falls below the target
3) Delay the purchase of an SPIA by between 0 and 25 years. A TIPS ladder is constructed to provide the income during the delay period and a single, duration matched, TIPS is held to provide the premium. A rolling TIPS ladder is also used.
In comparing the methods 2) and 3), it would appear that, for a TIPS yield of 2%, Method 3 (i.e., a delayed SPIA) is significantly more robust against inflation than than a DIA (Method 2). However, at TIPS yields of 1%, the robustness against is (as expected) much less and similar for the two methods. The immediate purchase (i.e., Method 1) gives a worse performance then the best deferral for TIPS yield of both 1% or 2%.
A proper comparison would also consider other alternatives, for example
1) Stock/bond portfolio. Using the real returns in the Simba database, historically a 50/50 portfolio (TSM/TBM) with a 4% inflation adjusted withdrawal survived for 35 years 97% of the time, with the first failure occurring after 32 years or so (i.e., in the worst historical cases, the income after 35 years would have been zero). Of course, this runs market (and inflation) risk since future values of SWR may (or may not) be less than those found historically
2) TIPS ladder. For yields of 2%, a TIPS ladder would support 4% withdrawals for about 34 years (ignoring any interest rate and reinvestment risk from 30 years onwards), while at yields of 1%, a TIPS ladder would support 4% withdrawals for 28.6 years. In both cases, the income after 35 years would be zero.
Furthermore, there is one further consideration (the one you have already mentioned), namely how much does a retiree wish to give up to the insurance company. Taking your $1m example and the methods I have explored (assuming TIPS yields of 2%),
1) For an immediate purchase, $1m is handed over to the insurance company, and slowly returned from the outset.
2) For a DIA, and a 15 year deferral (i.e. 70 to 85), would entail using $763k to construct the TIPS ladder to provide income during the first 15 years and $237k to purchase the DIA.
3) For a SPIA delayed for 15 years, the TIPS ladder still takes $764k to construct (at 2% yields) and the premium available is $237k that is then placed, for example, in a duration matched TIPS.
In the event of death earlier than the delay/deferral period, the legacy would consist of the unspent TIPS ladders plus the premium (for delayed SPIA) or zero (DIA).
cheers
StillGoing
Statistics: Posted by StillGoing — Mon Jun 17, 2024 3:32 am — Replies 1 — Views 238