This is a follow-up to the GMWB I  blog of a few weeks ago, where I also described the characteristics of typical GMWB products. Here I will compare various decumulation options with the hypothetical GMWB product described earlier and various parameters. This will get a little technical, but please bear with me as the conclusions are valuable, especially if you are thing of getting one of these Guaranteed Minimum Withdrawal Benefit products. The conclusion is that when things look too good to be true, they usually are.

The assumption in the Monte Carlo analysis used is still that the distributions are normal (Gaussian). While it is known that returns of securities are not normal, especially if consider exposure to Black Swans, this is a reasonable assumption to explore alternative scenarios and get some insight into the likely outcomes. The decumulation options compared are: (1) GMWB, (2) draw 5% of previous year-end assets, (3) annual draw of 5% indexed to inflation, (4) 5% fixed (CD/GIC) and (5) joint annuity of 5.9% annual pay (un-indexed). The comparison will be done both from a remaining assets and annual draw perspective, all in real dollars (i.e. dollars adjusted for the assumed 3% annual inflation).

I have received un-indexed joint annuity quotes for a 60 year old couple (no minimum payment period), from AAA U.S. source of 5.9% and a (less than AAA) Canadian source of 5.6%. I have used the 5.9% as the reference point, but the conclusions would hardly be different if the lower figure was used. I have chosen a joint annuity, rather than a male and/or female one with higher payout, as one of the benchmarks because if the couple annuitizes all or a significant portion of their assets, the income stream must last until the second one of the couple dies.

Also I’ll assume throughout that we are looking over a 40 year horizon (60 years old couple expecting that at least one will live to 100). The couple is working with a $100,000 portfolio. Inflation is 3% per year and, unless stated otherwise, capital markets deliver on the couples’ portfolio a Return=7.5% with Standard-Deviation= 7.5%. For simplicity taxes are assumed to be zero, even though this assumption may somewhat disadvantage or advantage some of the products (e.g. annuities are disadvantaged, GIC/CDs are advantaged, etc)
Note that, while the graphs are labeled as ‘median’, some of the cases under consideration have no variation around the ‘median’ (e.g. 5% CD/GIC, 5.9% joint annuity), while at times we will be discussing 25-or-75 percentile performance, when so indicated.

Scenario #0
This is the very simple case comparing taking a 5.9% un-indexed joint annuity (with zero residual value at all times) and a 5% GIC/CD renewable every 5 years.

Result #0- Both income streams erode significantly with inflation, but by taking about 15--20% lower annual income with the CD/GIC the couple still has the original $100,000 at all times (for the estate or emergencies), though significantly eroded by inflation after 40 years. So you may wish to trade off the income vs. the residual depending on your preferences. The other consideration would be if inflation accelerates significantly over the next 5 years, The GIC/CD solution may give you the option to reinvest at higher interest rates in 5 years.

 

Scenario #1
-Joint annuity of 5.9%
-GMWB 5% minimum annual payment, ratcheted up every 3 years (if current asset value is higher than the base used 3 years ago). Annual fees =3.5%, so average returns on assumed portfolio reduce from 7.5% to 4.0%.
-draw of 5% of the original $100,000 indexed annually with inflation from the same portfolio. Annual management fee assumed to be 0% (using ETFs we can get to 0.3% or less depending on the portfolio composition but conclusions will not be substantially different)
-Assume that you can get 5% interest on a CD/GIC (on May 30, 2008 you can get 4.7%)
-Each year draw 5% of the portfolio’s previous year end value.

Result #1: A draw of 5% of the portfolios previous year end value is the best median income stream (and never runs out of money! In fact median assets are expected to be over $75,000 of the original dollars) , except for 5% indexed annual draw from the couples’ portfolio (but with a very high probability of running out of money; see the rapid decrease of the median remaining portfolio value). You will also note that the 5.9% joint annuity consistently delivers higher income stream than the GMWB product (but of course the residual portfolio value is zero at all times.)

 

Scenario #2
Let’s now look at the same cases as in Scenario #1, except let’s look at the 25 and 75 percentile outcomes (rather than the median or 50 percentile).

Result #2- Note that the annuity beats (or is about the same as) the income stream of the 75 percentile line of the GMWB at all ages (though the portfolio has zero value). While the 25-to-75 percentile spread of the GMWB is quite narrow, as compared to the draw of 5% of previous year end value, the latter is always higher than the GMWB. By the way, the residual portfolio value is also higher than that of the GMWB.

 

 

 

Scenario #3
Consider now if the market performance is superior (or if we chose a much more aggressive portfolio). Let’s assume that the Return=9.5% with a Standard-Deviation= 15%. Of course these higher portfolio returns will be applicable for both the GMWB (i.e. after the 3.5% fee the return is reduced now to 6%) and the case where we draw 5% of previous year end asset value.

Result #3- The ‘median’ GMWB income stream beats the 5.9% annuity after about the first 10 years (but not by a significant amount). However drawing 5% of previous year end asset value results in a far superior median annual income stream and residual assets over the full 40 years, essentially inflation indexed (or better)!

 

Scenario #4
This is the same as scenario #1 except we reduced the annual fee on the GMWB from 3.5% to 1.5%, resulting in a Return=6% instead of the market offered return of 7.5% per year. The standard deviation stays at 7.5%.

Result #4- GMWB 25-percentile income stream performance is significantly below the annuity, while the 75-percentile performance outpaces annuity after about 7 years. However the performance still lags the 5% draw of previous year end asset value significantly and with a widening gap as time goes on (and fees take their toll).

 

 

 

Conclusions
The Guaranteed Minimum Withdrawal Benefit (GMWB) products with a 3.5% management fee have nothing to offer to even conservative investors, even when they are compared to annuities or GIC/CDs. At 1.5% management fee, GMWBs may become interesting to individuals who would otherwise invest in an annuity or GIC/CDs. But over a 40 year retirement, use of GMWB/ Annuities/ GIC/CD exclusively would lead to a relentless erosion of the investor’s purchasing power. In fact over 40 years the purchasing power would be reduced by 70% given an annual 3% inflation.

The most promising approach appears to be 5% of the previous year end value of the portfolio draw. This requires taking a little more market risk than just using CD/GICs in order to reduce inflation risk. The payoff in both likely income and residual assets is too great to ignore. I suspect, that even for a very conservative investor who is leaning toward annuities, they should be used only for part (50%?) of the assets, and even then one may want to do that only in stages and as late in retirement as tolerable.

In the real world, no plan can/will be set into motion and be expected to remain unchanged over a 40 year horizon (except annuities). In the real world there are forces that continuously push you off-course and course corrections will be required. Higher/lower than expected inflation, higher/lower expected returns and volatility, illness, death, unforeseen major expenses and higher/lower expected longevity will guarantee that. You will be required to review even a very good plan every year to insure that you are still on track.

But none of the conclusions should really come as a surprise:
-the market offers the same returns to everybody and the same inflationary headwinds
-if you move up on the risk curve (even a little bit), you can expect (though are not guaranteed) to get rewarded for the risk taken
-if you buy a product with a punitive 3.5% annual fee, you deserve the punishment that you get and what value do you receive for that fee?

So if it looks too good to be true, it usually is!