-----------------------------------------
The Problem of locking up assets and "cheating" bequestees
I've been told, quite assertively and quite often by an advisor
friend of mine, despite my (formerly) shallow, naïve, amateur insights on the value
of longevity pooling, to never ever annuitize…ever. Ever!!! Trying to press on this issue, even for some perspective
or to get more info, is a little like a wall-head-banging excercise. So I
don't. Or at least not anymore. My
presumption is that, in the absence of a reasoned and calm discussion, the
rationale for her position is that: A) there are products and strategies that
can provide longevity risk hedging and income flooring similar to a traditional
annuity but without the irreversibility of giving your capital to an insurance
company "forever" -- and here she used a particular company's GLWB
rider with step-ups and minimum withdrawals as an exemplar, and B) not so unrelatedly, the idea is that irreversibly giving your capital away to the evil insurance company is both "unfair"
to potential bequestees if you die the day after signing the contract as well
as risky if you happen to have spend shocks in the future (the latter part of
"B" I will mostly ignore). This is part of the fabric of the annuity
paradox, by the way. This is also something I want to understand. Let's hit both A and B and see what happens.
On "A" - that there are products or strategies that
can somehow square the circle of longevity-pooling-value vs. allowing for a bequest
To my mild chagrin, she is probably more right than I gave
her credit for on this point. Some of the new products out there are in fact worthy of attention if one is careful about the analysis. On the other hand it is still not a simple problem and I
am still not all that sanguine on things like GLWB riders. It is a much more
complex problem than I expected and I still think that it (GLWB) does not
unlock the true power of "the pool" for fully supporting
lifestyle in real terms at later ages like "real" annuitization can [1]. Joe Tomlinson,
a quantish advanced advisor that I totally trust, did a good cover on the GLWB vs
Annuity question back in 2013 at Advisor Perspectives here (Do Annuities ReduceBequest Values? ).
I am not yet an expert on GLWB
features so let's at least start this discussion by using a wise average-joe version of a thought
on GLWB (random internet quote on a bogleheads bulletin board):
It's not like the insurance co rubs their hands together and creates money you're otherwise not entitled to. Monte Carol analysis shows that these annuities turn out better than a conventional portfolio something like 6% of the time. --see appendix A
Now let's also quote Moshe Milevsky (Life Annuities: An Optimal Product for Retirement Income, Moshe Milevsky CFA institute 2013) because, well because he
is ubiquitous in ret-fin, unconscionably prolific, blindingly bright, and has
written on this a number of times with continuous math in hand:
Whether the GLWB is better than the
life annuity from the consumer’s perspective depends on a complex relationship
between the pricing of the guarantee, the retiree’s optimal consumption
strategy, and the existence of bequest motives.
The common denominator of all these
insurance riders is that they contain an implicit put option on financial
markets plus some form of longevity insurance, akin to a pure life annuity. Of
course, using the concept of put–call parity, they can also be viewed as call
options to annuitize at some variable strike price. The (anecdotal) “sales
pitch” for these products revolves around the idea that the guarantees permit
investors to take on more investment risk than they would without the
guarantees.1
Here is the bottom line: To the
naked eye, the VAs with GLWB might appear to have all the benefits of a life
annuity—guaranteed income, risk pooling—but without the costs associated with
illiquidity and irreversibility. However, although the GLWB product has merits,
especially considering the research evidence that it was initially underpriced,
it is not a substitute for pure life annuities because of its lower yields. For
example, whereas a life annuity might pay 6% to a 65-year-old, the GLWB rate
under the same market conditions would be in the vicinity of 4%
So, is GLWB a magic bean that allows one to displace a
traditional annuity so that one can both hedge out longevity risk AND keep
access to capital for some reason (bequest, spend shocks, etc.)? A little bit
yes and a little bit no. They do what they say they will do (and keep in mind, it appears as if fewer and fewer offerings
are on the table these days) but they sag a bit when trying to deliver real (vs. nominal)
returns in the very late game, there is a distinctly lower payout (vs annuities)
due to whatever the insurer thinks it needs to keep to be able to make the
promise (those fees that help compensate for longevity risks, admin costs, maybe market volatility risk, etc etc), and there are maybe more problems
in the details if one were to dig into the contract terms[2]. There is probably a default risk factor in
there too. I'd say the jury is out but getting more biased by the minute.
For some additional links and excerpts and thoughts see Appendix A below.
On "B" - that it is somehow unfair to "someone other
than me" to irreversibly throw my capital away into an insurer annuity
hole…because I might die tomorrow and all that money will go to the pool (incorrectly described as the insurance company) rather than someone in my circle.
We can talk about the technical pros and cons of products
and strategies all day long but I think this "B" thing is quite a bit
more interesting than "A." We'll use me as part of a case study. But first, let's start
with a baseline set of assumptions that includes the idea that I think I am
already in some optimal equilibrium and that I have a no-big-bequest
strategy. That means that [Wealth - PV(LS+SS+SA+SR+MC+O)
= 0] [3] which also means I am in perfect equilibrium
and that that equilibrium might just include a big future allocation to tradition
SPIAs[4] if I decide that that allocation will optimize my lifetime utility of
consumption. So who is this then, exactly, that insists that I not seek my
optimal welfare the way I think it should be sought? A third party? They have
no stake! A child? Perhaps but I have stated "no bequest." This is not
their call to make, it is mine and I have already factored their future welfare
into my equation. If there is to be a disequilibrating
event, like switching strategies to suit someone's interests other than mine,
then there should also be fairness. That question of fairness brings me to the beginning of what we might call my
"mini moral theory of bequest."
B.1 - A question of symmetry
This, by the way, is my (possibly over-wrought) purpose for
the post. I think that if I am going be induced by a third party to change from
one optimal strategy to another less optimal one (my unexamined strawman here was
that a SPIA might be part of the optimal thing and anything else will be less so) that
there should be some symmetry: the negative cost I have to incur to revise my strategy
to "their" liking (benefit?) should have some symmetry with some positive thing
that I receive from them in exchange. So,
let's define my switching costs as maybe framed in these terms: a) a general loss
of optimality, b) probably a substantial increase in unfunded superannuation
risk if I were to seek "no longevity hedge at all" (call that "no hedge" a
systematic withdrawl plan or SWP) in order to favor bequests, and c) there will be a presumed
smaller cost, given my various biases, to swapping a traditional longevity pool
strategy for a GLWB longevity rider strategy, two strategies that some people put in the same bucket even if I don't.
My big question then is "where is the other side of the deal?" A
third party of any kind, child or otherwise -- if they want to make a claim on the
bequest-value of my estate if I die tomorrow after I would have handed a check
to New York Life or Ohio National, a proposition they do not seem to like -- should at least be willing to pay for some of the
value of that incremental benefit (their benefit, my new risk) that they would propose to receive (that is, if I
am willing to sell it (and I wasn't before) and if I am convinced of my current
plan's optimality). Really? They get a benefit if I croak early but aren't there for me if I don't and I run out of money because they talked me out of annuitizing!? That is a terrible deal that I call "not fair." I wan't to be compensated for my new risk. Overthinking? Of course it is and yes I realize I am
exaggerating everything here but let's take a look at this anyway because it's fun. This is fun, right? The question, I suppose, is now about how to value the symmetry.
B.2 - Valuing the fairness and symmetry - a map
Let's say for the sake of argument that there are two types of switch: 1) from a full immediate
annuitization strategy to an uninsured systematic withdrawal plan SWP, and 2) from
a full immediate annuitization strategy to a GLWB strategy. There are others,
perhaps. In any case, I should try to evaluate the fairness (impossible?)[5]
and cost (possible) of these switches. First a map:
For what it's worth, if one were to want to have a more formal
and less amateur evaluation of GLWBs one might look at Wade Pfau's piece at
Advisor Perspectives on this… GLWBs: Retiree Protection or Money Illusion?
B.2.1 - Swap 1: Swapping out SPIA
for a systematic withdrawal plan
First the hard part. What is fair? I have no idea but what I hear from others is
that if one dies the day after contract signing and all the money is gone and
(importantly) one does not understand the theory of longevity pooling, that that's
not "fair." If one understands longevity pooling it is another story entirely. On the other hand, if one were to die at age
253, well then, it was fair or you at least beat the insurance company (forgetting about
the theory of risk pooling again). The
point where it flips from unfair to fair is somewhere in between tomorrow and age 253 and is 100%
opinion, especially if risk pooling is not understood. We'll run some numbers though. The way I have seen this done in the
past by someone somewhere is to look at something like:
a. fair = when the cumulative PV of payments is at least = the
original annuity price
b. fair = when the IRR passes the discount rate
c. fair = when payments continue past when a SWP would have run out
Some bad assumptions first. There is no: inflation; company defaults;
insurance loads; fees; worries about rate reversion and low current rates;
taxes; knowledge of advanced topics like Annuity Equivalent Wealth, wealth
depletion time, real option value of not annuitizing; or free boundary of
annuitization, etc. Human capital is depleted. There may be more...
Let's say we are 59, male, have liquid wealth = W = $1,798,528,
mean annuitant longevity is ~86 depending on how you do it, and we and want to
lock in C=100k in lifetime cashflow for our optimal plan. We kinda backed into W by way of the 100k by using this formula:
where "l" (load) is zero and the tPx comes from a
SOA table using G2 extensions to 2018 for a 59yo male and where the annuity discount
rate is ~3.01% for now. [6] If we double check this idea on
AAcalc.com with the proper assumptions we get an annuity price est. of 1,807,806
(3/15/2018 ). Close enough
for a reality check. I could try harder but why? With respect to our a b c:
a. The cumulative PV (npv) of the 100k cash flow discounted at the
same discount (this analysis is not shown) hits W at ~84.
b. the IRR of the annuity cash flow passes the discount rate
at around age 85 if I did it right.
c. If I run a net wealth process with the same W and C and
return = discount, it runs dry, not surprisingly, at ~84. This would be a
stochastic process in real life so the distribution of "run dry"
would be relatively wide depending on volatility and the sequence of returns in individual paths of the random return
generating process.
Let’s say that a reasonable person, unaware of longevity pooling theory, would conclude that the
annuity would be "morally fair" if death occurs at around 84-85.
Let's give that some more room for the unmodeled variability and risk of
"c" (the SWP) and say that 75-80 might be fair too.
We're 59 so let’s call it 79, so 20 years. This is debatable but not totally insane. Many people might lean a little later than
that to try to get a bigger "fair" share of the pie…recalling that
they might not know about how risk pooling works. Have I mentioned risk pooling yet?
B.2.1.1 - Swap 1: Fairness valuation - a minimum price
Given the above, if someone wants to call an SPIA "unfair" (or
suboptimal) and wants me to switch to a SWP so they can have access to my
estate if I die the day after the contract, they would have to at least offer
to pick up the tab for the age 79 -> ∞ risk if it unfolds, because I just
threw my longevity risk "hedge" to the wind...for their benefit. That would be
fair, right? Using the following:
the price would be 385,389 and is in practice -- and
unsurprisingly -- no more no less than a DIA price for a 59 year old looking for
100k cash flow in 20 years. A quick check at AA calc shows a 391,620 price for
a DIA price for a given set of assumptions.
Close enough. So if they want me to ditch my optimal consumption plan
and longevity security blanket, they'd have to at least cover my superannuation risk
for ~390k. Plus given the stochasticity of both returns and spending and the
possibility I might run out of $ before 79, it might be more more than that. This is a minimum price.
B.2.1.2 - Swap 1: Fairness valuation - a max
"threshold" price
The max price we might consider is another matter altogether. I suspect it involves some kind of real option
valuation thing. I can't do that (yet). But! We have been given a new tool by a ret-fin god. Moshe Milevsky just wrote a paper on the utility
and valuation of longevity pooling. In
that paper (you can read it yourself here) he gives a specific way to value the exact price of the longevity pooling we
are being asked to give up for someone else's benefit. Using the principles and math in that
paper we can uncertainly say that the price could range anywhere from:
on the high side, with unrealistic assumptions for returns and longevity, to maybe
on the lower side (and if we were to use more realistic
assumptions for the mortality curve, the presence of income like Social
Security or pensions, and risk aversion).
The delta value, as I mentioned in a past post, is defined as:
"Basically, its DELTA percent of the money you have available to annuitize at retirement. So, if you have $100,000 and decide not to purchase the longevity insurance -- and take the chance of living longer than life expectancy -- you would need to be compensated with $100,000 times DELTA." -- Prof. Milevsky
Given some amateur hack assumptions that you can play with
in his paper on your own, we can say that the "delta" can range from
something like .279 (see where I came up with that in something I did below) to .649 of W per equation 25.
That would mean that the delta -- or the true "price" for the
value of my involuntarily forgoing the pool (as always, my disclaimer is: "if I have
that right") -- might range from 500,963 (eq 33) to 1,167,245 (eq 25). In other words, if someone wants me to forgo
my own rationalized, longevity hedged, and possibly optimal strategy for something that I know is not to my benefit but
might plausibly be to their benefit…then they better bring a checkbook.
B.2.2.1 -- Swap 2: Swapping a SPIA for GLWB product - using
annuity pricing first
I am going to assert by way of what a former boss called
"proof by intimidation" that GLWB products are inferior over the long
run for a retiree when compared to full SPIA products (or I can at least lean on the
Milevsky quote above for now). That means I assert there is a real cost to ditching
annuities for a GLWB. If you disagree, write it up and send it here. Otherwise we have to come up with a price for
swap number 2. To do that in this
section I'll see if I can come at it in a way similar to what I did above.
For fun, and for the sake of argument, let's say my advisor
friend is precisely correct (GLWB is always >= annuity especially in the
presence of a bequest motive) and that Milevsky is also correct (ignoring
bequest, GLWB is always <= annuity and produces, on average, a payout on the order of
something like 1/3 less than the annuity; that average thing is important, btw, but ignored for now).
That gives us an idea that maybe we can look at a net negative 1/3
payout as a fairness factor, all else being equal which of course isn’t really
the case. This kind of thing is highly debatable of course but let's roll with it.
One problem though is that I do not know when GLWB income can start for
a 59 year old. I'll assume it can be somewhere
between 59 and 79, 79 being the arbitrary "fair" age we chose above. In practice I don't think it can start right
away because of blackout periods, but I might be wrong and in any case it does not really matter here in this post. Then we can say that the price of swap 2 might be
the following after admitting that the notation is totally "amateur" (it's a good thing these posts can't get red pen marks):
where fg(a) is basically, for today anyway, something like
the 1/3 reduction in payout from going from annuity to GLWB or we'll say the
incremental fair value price for switching is the sum of 1/3 of my planned, probability weighted, discounted, 100k
payout over some interval of remaining age, whether 59+ or 79+ or whatever. If
you want something more complex, you are on your own. If you buy in, then we are in business. In that case we can say that if the
diminution or vitiation of my preferred optimal annuity cash flow is run from a
start age of 59 then the cost/price is 598,910. If it is run from age 79 it is
128,334. The point here is that if one believes, like I do, that the net GLWB
benefit ex bequest is <= annuity, then the cost of asking me to switch is not zero, it
is something. As above, if you want me
to bequest something to you, bring a checkbook. Or convince me that my bequest plan in
morally flawed…but that is not a math problem. That is a moral debate.
B.2.2.2 -- Swap 2: Swapping a SPIA for GLWB product - using the pooling-delta method
[ I had some conceptual errors in this section in the 2nd paragraph that I tried to fix in a later post here in this later post: Correction - My post on bequests and fairness...esp wrt the delta concept . Even without the correction, recall that one has to take what an amateur blogger says with a grain of salt in the first place.]
In effect the question at hand is really one of the tradeoffs on
the margin between risk aversion and bequest motive. Milevsky's choice of
mathematical object for "delta" is
which shows that the value of longevity pooling is a strictly a function
of risk aversion and the shape of the mortality curve (from page 23 of the link I provided: "Note
that the results are not as sensitive to interest rates. The δ benefits really
are driven by risk aversion γ and by the assumed mortality rates.") to
which I would add the bequest consideration too. So, I'd say: if bequest motive is zero and no
one is pestering you about getting a piece of the estate then maybe stick with
the annuity if it optimizes your lifetime utility of wealth and consumption
(that's a different topic). If the bequest
motive is > 0 then maybe we can view the GLWB as a class of partially
defective but liquid annuity that leaves at least some longevity risk and/or
consumption disutility on the table. That means we still have to evaluate the
nudge from annuity to "slightly-less-than" annuity.
In Milevsky's table 1, case B in the link I mentioned (before
he gets to eq 33; this is the sub-table with exponential but moderate
mortality, moderate risk aversion, and a row with the presence of 99% prior annuitization;
this is my proxy for the defective annuity idea, btw), delta is still .078 which
would make the delta against the ~1.8M grubstake we mentioned = $140,285. So maybe that is a lower bound for a price
for our bequestee, on the margin, when moving from 100% annuity (SPIA) to 99%
annuity (again: proxy for GLWB).
For an approximate "upper bound"[7] we can try to use Milevskys's equation
33. In eq 33, a(x) = a normal annuity
factor for the annuitant and a(x*) is an annuity factor with an economic age substituted in for a real age where the economic age is one that is adjusted for risk aversion
or in the paper adjusted like this: x* = x - b* ln(gamma). b in this case is
the mortality dispersion and gamma is the risk aversion coefficient. Here is an example. In this chart below I first fit a curve using
Gompertz-Makeham math to the SOA data I used above; actually I set the dispersion
to 8.4 to partially fit the curve and then I solved for the mode that gave me the same annuity factor
(1.7985) that I used before and that I validated against AAcalc.com. The PDF
plotted with the newly fit GM math is the red line where the dotted line is the curve
based on the original SOA table. Then, adjusting the economic age to 53.2 using the x* method I mentioned, the blue line is the
adjusted mortality PDF. That means that
the new annuity factor a(x*) is 18.629.
For an illustrative and arbitrary risk aversion coefficient of "2," plugged into eq
33, that gives us a delta of .279. That
delta, if thrown against W, would be a price of 501k for our would be bequestee.
This range that we got in this section of 140k to 501k, using
delta, is not radically different from 128-599k in the previous section using
annuity pricing. I might be wrong but I
want to call this more or less a validation of the idea that there is probably
a price of some kind for being nudged towards a bequest motive by "unlocking" the
annuity strategy even in the presence of a "full but defective annuity" like a GLWB (or maybe call it unfilled residual hedging). That's only if we consider a GLWB to be a type
of defective annuity. But recall that
this concept of defective annuity hinges on the quote from Milevsky in the
"A" section above. If you don't believe that then it is this post that
is defective rather than the GLWB.
Conclusion …and a question on their value
proposition….
All this above was fun (or it is if one has an odd sense of fun like RH). And, if it were to be real, it would go a
long way towards immunizing me from the risk you were asking me to take in a
fair, symmetrical way. That's only for me, though. This is only my side of the table. While I now feel
fairly dealt with, the new question might be whether you think this is a fair deal for you. That is not my
problem, though and this is starting to feel like a game theory kind of thing…a second move in a two
player game. It would be interesting to
see what the math is on the other side. By that I mean: is the price (or one of
the prices) that you paid above worth the probability that my estate will not
be decimated by my spending and that you will get some "fair" value for
what you spent in order to get access to an option on my estate by asking me to ditch my
SPIA. Option pricing is hard enough but there is also now some kind of odd
moral hazard because I am going to spend like crazy to blow out your option
value if I can… unless, of course, you are my children, in which case I will spend like a
miser -- my "policy" notwithstanding --so that not only do you get every last penny that I don't spend but every
word of this post will be mooted in the most embarrassing way.
[2] For example, here is some one insurers fine print
"No additional increases will be applied to the rider’s value once the
contract’s value is reduced to $0 or after the Index Anniversary immediately
following the Annuitant’s 95th birthday." That's dispositive of nothing
but it does show that living to 300 will not have any accidental benefits that
accidentally accrue to the 300 year old.
One is betting against a casino here.
[3] LS = lifestyle plan, SS = spend shocks, SA =
superannuation risk, SR = sequence risk, MC = modest contributions to the
health and welfare of my children…up to a point, O = other, because there is
always other.
[4] Or DIAs or VAs or TIPs ladders. TBD but the allocation and rationale should be opaque and uninteresting to all but me.
[5] I just read a great book on probability that makes the
case that wide swaths of probability theory originated with questions coming
from issues related moral fairness (how to split a gambling pot
after n games have been played, m games remain and the game
has been interrupted and so forth). See Book Mention - Ten Great Ideas About Chance.
[6] This is a work around. Gordon Irlam at AAcalc uses the
whole Treasury yield curve with interpolation within and beyond the data. I hope I'm not giving away secrets
but it's more or less like this from an email:
"The yield curve is
interpolated using Python's PchipInterpolator, par rates are converted to spot
rates for every 6 months. For rates beyond 30 years the average forward rate
from spot rates with a duration of 15 years or longer is used to compute the
spot rates. The appropriate spot discount is used for each survival weighted
probability." -- G. I.
My discount rate is a weak proxy for all of this but works ok. I'm
using annual payouts and a money's worth ration of 100% btw.
[7] here we are completely ignoring the difference between
the annuity proper and the GLWB. This is just the pure delta value the price
for abandoning annuitization.
Appendix A -- Links and Excerpts
The Utility Value of Longevity Risk Pooling: Analytic Insights Moshe A. Milevsky1 and Huaxiong Huang 10 March 2018
Life Annuities: An Optimal Product for Retirement Income,
Moshe Milevsky CFA institute 2013
- Here is the bottom line: To the naked eye, the VAs with GLWB might appear to have all the benefits of a life annuity—guaranteed income, risk pooling—but without the costs associated with illiquidity and irreversibility. However, although the GLWB product has merits, especially considering the research evidence that it was initially underpriced, it is not a substitute for pure life annuities because of its lower yields. For example, whereas a life annuity might pay 6% to a 65-year-old, the GLWB rate under the same market conditions would be in the vicinity of 4%.
- Whether the GLWB is better than the life annuity from the consumer’s perspective depends on a complex relationship between the pricing of the guarantee, the retiree’s optimal consumption strategy, and the existence of bequest motives.
Financial valuation of guaranteed minimum withdrawal benefits, Milevsky and Salisbury , http://www.math.nus.edu.sg
2005
- The GMWB rider/feature is being subsidized by the basic insurance fee. This is consistent with the over-pricing of standard features in VA policies. Thus, a 50 basis point fee for the rider together with a baseline 120 basis point fee might add-up to 170 basis points, which is enough to cover the risk.
- we show how the product can be bifurcated into a type of Quanto Asian Put (QAP) plus a generic term-certain annuity
- A recent innovation in this market is the abovereferenced GMWB ‘rider’. In contrast to all the other bells and whistles3 it contains absolutely no life insurance component and is thus well within the domain of analysis of financial economics
- Most insurance companies charge for this downside protection by deducting an ongoing fraction of assets as opposed to an up-front fee. These unique features differentiate the pricing of this derivative security from the standard Black and Scholes (1973) approach where the option premiums are paid up-front and in advance. This fact introduces subtle hedging issues
- we find that under a stylized product specification which guarantees a 7% withdrawal, and assuming a forward-looking investment volatility of σ = 20%, the cost of providing a GMWB ranges from 73 to 160 basis points of assets per annum… we find the recent GMWB products that have been introduced in the market are only charging 30–50 basis points, even though the underlying annuity subaccounts contain high-volatility investment choices
Optimal Portfolio Allocations with GMWB Annuities, Blanchett
in JFP
- A GMWB annuity compares much more favorably against an immediate fixed annuity, which is a more relevant proxy for generating guaranteed income for life.
- while GMWB annuities may appear to be relatively “inefficient” at an individual product level, a GMWB annuity can improve the overall efficiency of a retirement portfolio and better help a retiree generate sustainable income for life for the right situations.
- while the GMWB annuity may be viewed as “expensive” when compared against a traditional portfolio, the cost is on par with an immediate fixed annuity.
- When viewed from a total portfolio framework, where the goal is to maximize a utility function based on income replacement, the appeal of a GMWB annuity increases considerably, especially for retirees with lower levels of pension income, more conservative withdrawal rates, and more conservative equity allocations during retirement.
The Best Solution for Protecting Retirement Portfolios: Putand Call Options versus GLWBs, Joe Tomlinson, Advisor Perspectives 4/2013
- Another way to provide downside protection is to purchase a variable annuity with a guaranteed lifetime withdrawal benefit (VA/GLWB), or a similar product, the stand-alone living benefit (SALB), which combines mutual funds with a GLWB. This is quite different from using an options strategy, but these products achieve a similar result.
- The level of fees makes a big difference in average bequest values. The low-cost GLWB is competitive with the option strategies in terms of average bequest, and it has the advantage of reducing the failure risk to zero. It accomplishes this by pooling longevity risk.
- The higher-cost VA/GLWB also makes use of mortality pooling and provides this same protection against failure. But the bequest values are no longer competitive with the option strategies. Pay attention to costs in VA products.
- For downside protection in a retirement portfolio, I would stay away from options strategies and choose an annuity-like approach providing guaranteed lifetime income. In today's environment, the Vanguard VA/GLWB is a strong contender. I would also compare financial outcomes for the Vanguard product with projected outcomes for low-cost versions of other types of annuities, such as immediate annuities and deferred-income annuities.
The Next Generation of Income Guarantee Riders: Part 1 - TheDeferral Phase. Pfau at Advisor Perspectives, 2012
GLWBs: Retiree Protection or Money Illusion? by Wade Pfau, 12/13/11 Advisor Perspectives
- examining historical data, I have found that those riders carry a cost that will not be readily apparent to retirees: their cash flows rapidly decrease on an inflation-adjusted basis.
- Moreover, GLWB riders, which are also known as guaranteed minimum withdrawal benefits (GMWBs), compare unfavorably to the cash flows a retiree would obtain using a systematic withdrawal plan.
- As long as one does not exceed the allowed withdrawal amounts, guaranteed withdrawals never decrease (in nominal terms) even if the account balance falls to zero. In this regard, GLWBs share similarities with immediate annuities, though a GLWB contract can be terminated and remaining assets can be returned.
- I remain concerned that prospective retirees may be overvaluing the guarantees in their mind because they are not properly considering how inflation will erode their real value over time. In behavioral economics, this bias is known as money illusion. People can logically understand the effects of inflation, but their emotional responses and decisions remain attached to nominal values.
- The
guarantees are not inflation-adjusted and would have been worth little in
rolling periods of
- Benefiting
from GLWB guarantees would require worse circumstances than any historical
stretch so far. It’s important to note, however, that
- Retirees may find “peace of mind” from the guarantees a GLWB provides. Perhaps the guarantee would stop retirees from panicking and selling stocks after a market drop. Moreover, if retirees do have a tendency to reduce their spending as they age, it may not be necessary for them to maintain a fully inflation-adjusted income. GLWB guarantees may also come in handy if people live beyond a 30-year retirement horizon. Perhaps most importantly, I have conducted research that suggests that withdrawal rates may fall dramatically below 4% for recent retirees. Should that prove to be the case, retirees could possibly benefit from the guarantees.
- the nominal withdrawal amount guaranteed by a GLWB can become quite small in real terms, and money illusion biases may prevent clients from properly appreciating this drawback.
- In
the event that someone replicating GLWB withdrawals on his or her own runs
out of wealth, an outcome that no investor has faced historically, then
GLWB owners will surely worry about the risk of default for the insurer.
- Higher fees would erode the account balance more quickly, allowing for fewer step ups, more exposure to inflation, and making it even easier for systematic withdrawals to match GLWB withdrawal amounts.
- It is
possible that purchasing a fixed single-premium immediate annuity (SPIA)
would generally enable retirees to obtain a guaranteed income source more
cheaply. Financial planner Joe Tomlinson refers to GLWBs as “actuaries
gone wild,” due to their complex combination of downside protection and
upside potential.
Understanding Variable Annuities with GMWBs by Robert
Huebscher, 3/1/11 Advisor
Perspectives
Income Annuities versus GLWBs: A Product Comparison, Joe Tomlinson, Advisor Perspectives 2012
- VA/GLWBs are often costly, and the typical purchaser has few tools with which to assess the costs.
- But when examined more closely, there are in fact more similarities than differences. [VA/GLWB vs income annuity]
- The VA/GLWB is a more popular product than the income annuity, and it is often touted as providing stock market participation with downside protection, like "having your cake and eating it too." This is largely a mischaracterization, because the fees one pays to obtain the GLWB’s guarantees take a substantial bite out of the equity returns. Given the dictates of financial markets, it is not reasonable to expect to both hedge downside risk and enjoy equity-like returns.
- Under the particular pricing structures for these products, the VA/GLWB provides a larger expected bequest than the income annuity. This result is sensitive to VA/GLWB fees, and I'll show an example below where the results are reversed.
- Guaranteeing an income that lasts for life requires a significant sacrifice of expected bequest value. Advisors and clients need to give serious thought to whether guarantees are worth the cost.
- The results are quite clear: At this higher fee level, the VA/GLWB becomes a more costly way to provide guarantees than the income annuity.
- Income annuities will meet the needs of those who want maximum retirement income, while the VA/GLWB will work well for those who can give up some income for more liquidity and flexibility.
- I do not see much appetite for sales forces to give up commissions in favor of improved customer value.
Annuities and their Derivatives: The Recent Canadian Experience,
Milevsky and Shao, SSRN 2010
Optimal Initiation of a GLWB in a Variable Annuity: No Arbitrage Approach Milevsky Huang, Salisbury ;
SSRN 2013
- Our main practical finding is that given current design parameters in which volatility (asset allocation) is restricted to less than 20%, while guaranteed payout rate (GPR) as well as bonus (roll-up) rates are less than 5%,GLWBs that are in-the-money should be turned on by the late 50s and certainly the early 60s.
- We conclude by suggesting that much of the non-initiation at older age is irrational (which obviously benefits the insurance industry.)
Why It Rarely Pays To Wait On Taking Withdrawals From A Variable Annuity GLWB Rider – A Case Study, Kitces, 2014
- a deeper analysis reveals that in the end, most retirees with a GLWB rider may do little more than pay a lot in annuity costs to receive a guarantee to just spend their own original contributions and nothing more.
- GLWB riders extract any withdrawals against the policyowner’s own cash value first, often the time horizon it takes to actually reach the point where the policyowner has worked through their own cash value and into the insurance company’s pocket via the guarantee is actually longer than the client’s own life expectancy! In other words, most clients are just receiving their own money back and may literally die before ever seeing a dime of the insurance company’s money under a GLWB rider!
- it turns out that for those who do have annuities with GLWB riders, the best course of action may actually be not to wait, and instead to tap the annuity for permitted withdrawals as soon as possible (to the maximum allowed under the contract).
- the benefit increases just aren’t enough to offset the shortened time window of life expectancy that results from waiting in the first place! Instead, the best course of action is to actually try to deplete the annuity as quickly as possible, in an attempt to reach the point where the policyowner can get a “return” from the insurance company
- if the value of the rider is to guarantee a floor while taking withdrawals, then the goal should be taking withdrawals to try to hit the floor… and the older the client, the better it works!
- Some early GLWB riders don’t actually guarantee payments for life, but only a more limited period of time. Most GMIB contracts (and the occasional withdrawal rider) require a holding period (where the benefit base grows) before withdrawals can begin, and/or the benefit base can be annuitized. Perhaps most important, contracts vary dramatically about the consequences of taking withdrawals in excess of the guaranteed growth amount of the benefit base; some are relatively favorable, while others may terminate the rider entirely for taking even $1 over the limit! To say the least, the devil is in the details,
- The two broad categories are a Guaranteed Lifetime Withdrawal Benefit (GLWB) and a Guaranteed Minimum Income Benefit (GMIB); the key distinction being the words in the middle, “lifetime withdrawal” versus “minimum income.” Both riders rely on the use of a withdrawal or income benefit base, which is a phantom amount against which the lifetime withdrawal or minimum income guarantee applies.
- The notable thing about the benefit base, though, is that it is not liquid; when a variable annuity states that it will grow by a guaranteed 5%, 6%, or 7%, it is not the liquid cash value that is growing, but the benefit base.
Do I Really Get A 7.2% Return Forever From That GuaranteedLifetime Withdrawal Benefit (GLWB)? Annuity123.com 2013
- “Lifetime withdrawal benefits” are very popular these days; most of the indexed or variable deferred annuities sold today are bought by consumers who paid extra for this “rider”. Regrettably, all too many of those buyers believe that, for that extra cost, they will earn a guaranteed “investment return equal to the “rollup rate” of the annuity. They won’t. They will get a guaranteed amount of income, but that’s not the same thing.
- A Guaranteed Lifetime Withdrawal Benefit is not an investment feature. Its value (a set percentage of a benefit base that is guaranteed to increase at a set rate for 10 years or until he begins withdrawals) is an insurance feature. And insurance features don’t perform like investments – because they’re not. No insurance feature, of any policy on the planet, will ever “pay off” on average.
- If the average buyer of an insurance policy profits from buying it, the insurance company will soon go broke. The true value of an insurance feature is not its IRR (or ROI, or any investment measure), but the fact that it guarantees an outcome, no matter what.
- the “rollup rate” and the “payout percentage” of an annuity with a lifetime income benefit is not an investment return and the buyer will not get that percentage, plus his original principal. Those rates are simply factors producing an insurance benefit (in this case, the sure and certain income to Joe of over $10,000 per year, one tenth of his original investment, even if everything goes wrong.
- For some consumers, this guarantee will not be worth the cost (which can run anywhere from 0.3% to over 1.3% of the benefit base each year; for others, it will. But in any case, no buyer should expect to get a “return on principal” equal to the “rollup rate”, or even the “payout percentage”.
What You Should Know About Lifetime Withdrawal Guarantees,
Randy Myers, wsj online.
- Let’s
suppose that you’re reasonably sure you’ll never want to take out more
than your guaranteed minimum withdrawal, that you don’t care whether you
leave any account balance for your heirs, or that you’re convinced the
financial markets aren’t going to do very well during your retirement. In
all of those cases, opting for the highest rate might make sense. But
here’s the rub. By choosing and using a higher payout rate, you guarantee
that you’re going to burn through your actual account value faster than
you would with a lower rate. If you ever do need to make an emergency
withdrawal, you’ll have a smaller lump sum available to you in your
account balance. You’d also leave a smaller account balance to your heirs
in the event of your death. And you would lessen the chances of enjoying
an increase in your actual account value that would trigger a step-up in
your benefit base, which could be important in terms of keeping pace with
inflation as you get older.
Pros/Cons for GMWB (Guaranteed Minimum Withdrawal Benefit),
bogleheads discussion. Circa 2013
- I don't see that there are any pros, none. VAs with such features have such high costs that portfolio growth is so hampered that you're all but forced to take out one of the withdrawal options. Given that outcome, you (someone) is better off investing in an entirely different vehicle from the start. If the question is to get into one in the first place or not, that's pretty obvious. If someone already has one and the market has just fallen, then there's some calculation needed. It's not like the insurance co rubs their hands together and creates money youre otherwise not entitled to. Monte Carol analysis shows that these annuities turn out better than a conventional portfolio something like 6% of the time.
No comments:
Post a Comment