As baby boomers contemplate their retirement longevity, the variable annuity industry has shifted the focus of its guarantees from dying to living.
This is happening in a new generation of VA riders that promise periodic income benefits, the most popular of them being a guaranteed living benefit called the guaranteed minimum withdrawal benefit.
A GMWB is essentially a pure money-back guarantee on an original premium deposit. (In the language of financial engineering it is a systematic withdrawal plan with a complex sequence of put options.)
Under GMWBs, VAs continue providing a guaranteed income flow regardless of the performance of the VA’s underlying subaccounts and markets. This implicit downside protection is especially important in the early years of retirement, when portfolios are more vulnerable to the devastating impact of a bear market.
First-generation GMWBs do pose a critical concern–namely, that regardless of whether the VA promises bonuses or withdrawals of 5%, 6% and 7% of initial premium over 15-20 years, once the income flow actually starts, the payouts are not designed to keep up with the consumer price index and especially the inflation rate for retirees. A secondary concern is that even with GMWBs, retirees still face longevity risk once the guarantees have been exhausted and all the promised money has been returned; one or both members of the couple might still outlive their retirement resources.
Indeed, inflation is quite different and higher for typical retirees than for the general population. The U.S. Department of Labor has been tracking a unique inflation index for the elderly, the CPI-E, which consistently has outpaced the regular CPI. The CPI-E reflects, for example, the rising cost of medical care for the elderly, a cost that is more heavily weighted in the CPI-E.
This suggests that true living benefits should be structured to increase over time in a partial attempt to hedge these increasing expenditures.
Perhaps in response to the need for real vs. nominal income, a number of VA manufacturers have introduced GMWBs that step up the guaranteed base upon which the withdrawal benefits are computed. The step-ups occur on contract anniversaries ranging from quarterly to every five years. The hope and expectation is that, if the underlying net account value increases over time, the 5% (6% or 7%) withdrawal rate will be applied to a higher base and the income flow will trend upward over time. Some manufacturers even guarantee this stepped-up income flow for life, providing the additional benefit of longevity insurance.
Although many of these step-up riders and their supporting marketing material are being positioned as “sure things,” it remains unclear to what extent the income flow will actually step up and thus keep up with retirees’ unique inflation rate.
To investigate this, I conducted a series of Monte Carlo simulations that help shed light on the odds. The simulations generated thousands of different scenarios for the underlying subaccount values, based on hypothetical asset allocations. The process started with a single premium deposit of $100,000 and then took $5,000 withdrawals in each of years one, two and three. Then, if and only if the hypothetical account value was above $100,000, the guaranteed base was stepped up to this higher value and the new withdrawal amounts became 5% of the new base.
This process continued in three-year increments until the very end of the longevity curve.
Table 1 illustrates the results of this analysis for a portfolio of 80% equity-based and 20% bond-based investments. The table displays the algorithm’s computed median withdrawal amount after every third year, as well as the 75th and 25th percentile, providing an inter-quartile range of the possible outcomes.
The numbers in Table 1 can be benchmarked against the eroding power of an assumed 3% inflation rate. Notice that under this rate, the initial withdrawal or income flow of $5,000 must grow to $5,464 by year three of retirement to keep up with inflation and to $7,129 by the 12th year.
So, does the GMWB living benefit keep up? In the 80% equity and 20% bonds case, the answer is yes but barely. Although the median income does grow and steps up over time, it does not keep up with a 3% inflation rate. Note, however, that in one quarter of the scenarios, the step-up performed even better than a 3% inflation rate. So, it can be said that the GMWB has a decent chance of keeping up with a 3% retiree inflation rate.
For readers who would like to replicate the results themselves, the scenarios within Table 1 were simulated under the assumption that the average annual gross return will be 9% and net returns (after all insurance and management fees) will be 7% on, with a volatility (a.k.a., uncertainty) of 16%. Returns, withdrawals and portfolio balances were computed monthly, with the continuously compounded returns generated by a standard Normal distribution.
To contrast this under a moderately aggressive portfolio, I also generated scenarios and results for a 100% equity portfolio. This analysis assumes a more aggressive 8% net expected return with a volatility of 20%.
Table 2 shows the results using the same tabular format. Once again, the same cyclical process from the previous simulation continues for the life of the policyholder, or at the very least until the entire guaranteed amount has been returned to the policyholder.
Notice that in Table 2, under a 100% equity portfolio, the median income flow does, in fact, keep up with a 3% retiree inflation rate. For example, in year 12, where the inflation-adjusted value of the $5,000 income is now $7,129, the median income flow is projected to be $7,190 per year. In fact, even in the worst 25% of scenarios, income does start to increase by year No. 12.
From a broader perspective, some important lessons emerge from this kind of analysis regardless of the exact parameter values.
Although the retiree certainly can expect to receive a step-up and the median income flow does increase over time, there is at least a 25% chance that the policyholder will experience no step-ups during the first 10-12 years of the policy. If one is more skeptical about the equity premium, the probability of no step-up is even higher.
Thus, for the step-up feature actually to keep up with retiree inflation, a substantial equity exposure within the underlying portfolio is critical. Notice how the middle column in Table 1 lagged the 3% inflation rate, while the equivalent results in Table 2 just managed to keep up with this benchmark.
In sum, for a GMWB rider to deliver on its promise properly, I believe it must provide some form of benefit increase (a.k.a., step-up during the income phase) to keep up with the cost of living.
Second, it must contain some form of longevity insurance by extending the maturity from a fixed horizon to the life of the individual and his/her spouse.
Finally, the underlying portfolio must be linked to a growth-oriented allocation for a step-up option to become a stepped-up reality.