CPI Indexed Group Self-Annuities Versus Individual Drawdowns in High Inflation Environments

This is the second article^[1] using the 65-years from June 1959 to June 2024 of the Austmod database to perform 141 historical 30-year investment scenarios with quarterly start dates from June 1959 to June 1994.

The analysis suggests it is possible to produce individual drawdown strategies and GSAs that have a high likelihood of success in producing CPI indexed income streams for periods of 30 years and beyond, provided there is an appropriate matching of investment strategy with first year payment levels. The data suggests that a balanced fund investment mix may not be the best strategy for supporting CPI indexed payments.

These runs include the high inflation period of the 1970s and 1980s (as well as the 1987 Stock Market Crash, the Global Financial Crisis and COVID-19). The runs compare the performance of pooled Group Self-Annuities (GSAs) for a 70-year-old male and 65-year-old female versus an individual drawdown using a (i) a Balanced fund, (ii) Shares, (iii) Cash, (iv) a 50/50 mix of Cash and Shares, and (v) a simple, rule-based dynamic investment strategy, in delivering CPI indexed income streams for targeted investment gaps to CPI of 0%, 3% and 4.5%.  

The Australian Bureau of Statistics (ABS) Australian Life Tables 2019-2021 (ALT 2019-2021) have been used as the basis of mortality. These are used for comparative purposes and are not intended to be representative of annuity tables used in practice.

The results show that the pooled GSA outcomes that incorporate mortality yields have significantly higher first-year payments and higher residual surpluses at the 30-year point (which provide funding for longevity beyond 30-years) versus the individual drawdowns. 

For GSAs, the difference between the actual mortality yield versus the average mortality yield flows through projections each year and this has a significant impact on the success of maintaining CPI-indexed payments over the life of the GSA income stream.

As with the previous article, the dynamic investment strategy (which switches between Shares and Cash and a 50/50 mix) outperforms the static investment mixes.

For illustration, the comparison includes a 65-year-old female and a 70-year-old male. A 65-year-old female represents a common scenario which has generally lower mortality rates, and a 70-year-old male represents a common scenario that has generally higher average mortality rates. The mortality yields are limited to 30-years throughout these results; this is also consistent with the 30-year time horizons for the projections. This doesn’t have any material impact on the analysis.

The GSA initial pensions are calculated by combining the l_x values (i.e., number of lives at exact age x) from ALT 2019-2021 with the targeted investment gap to CPI to determine the initial pension payment^[2]. 

A similar method to calculate the average ‘30-year’ mortality yield by comparing calculations that include and exclude mortality and calculating the discount rate needed to equate both. This suggests the ‘average’ 30-year mortality yield for a 65-year-old female (from ages 65 to 95) is approximately 2.2% per annum and 4.7% per annum for a 70-year-old male (from ages 70 to 100).

Chart 1 shows how the difference between the average mortality yield and actual mortality yields changes over time for a 65-year-old female versus a 70-year-old male for the first 15 years of the projection. In the early period, the gap between actual and average mortality is negative, and this becomes positive beyond 12-14 years for older ages.

Chart 1: Comparison of gaps between average and actual mortality

For a male aged 70, there is a negative mortality gap for the first 12-years of the projections. For a female aged 65, the negative gap from mortality is lower, by comparison, in the first ten years but persists for almost 14 years before the mortality gaps make positive contributions.

These negative mortality gaps need to be made up by the investment markets to ensure sustainability of the fund assets supporting the income stream. If these negative mortality gaps coincide with negative investment gaps (i.e., periods of high inflation and weak investment markets), it compounds the negative effects and can lead to worse outcomes than individual drawdowns with no mortality pooling.

Conversely, if the negative investment gaps occur in later periods, then the effects are mitigated by positive mortality gaps, and they result in better outcomes than for individual drawdowns.

CPI indexed GSA pensions versus individual drawdowns

The projections use a $500,000 initial investment, and the initial annual CPI indexed pension payments are calculated incorporating the mortality yields for a 65-year-old female and a 70-year-old male respectively.

Table 1 compares the targeted investment gap to CPI for individual drawdowns with no mortality pooling to the equivalent first year pension payments that incorporate the mortality yields for a 65-year-old female and a 70-year-old male.

Table 1: Initial GSA payments versus initial individual drawdowns

Table 1 shows that the inclusion of mortality yields significantly increases the first-year payments for any given level of investment risk.

Table 2 analyses the ratios of the mortality-based payments to the individual drawdowns to show the percentage increase in first year income payments that can be achieved through pooling, by comparison to the individual drawdowns priced at similar investment return to CPI gaps.

Table 2: Increase for GSA initial payments versus individual drawdowns

These tables highlight the significant income benefit that can be derived from mortality pooling by comparison to individual drawdowns. However, the situation is not quite so straightforward because of how the mortality gaps flow into the results and the choice of investment strategy to support the income stream. 

Comparison of investment strategies for GSAs and individual drawdowns

The examination now turns to how various investment strategies performed in delivering 30-year CPI indexed income streams for GSAs versus Individual drawdowns. 

As in the previous article, the investment strategies compared are:

• 100% Cash,
• 100% Australian Shares,
• a 50/50 asset mix,
• a Balanced fund, and
• the dynamic investment strategy.

Table 3 has a summary of the results for how the GSAs compared to individual drawdowns for the different investment strategies. It shows the percentage of the 141 runs that reached the 30-year target in providing CPI indexed income streams and the average residual surplus recorded (expressed as a multiple of the 30th-year CPI indexed payment). It also shows the percentage of runs that did not achieve the 30-year target and the average durations that were recorded for the failed runs.

Table 3: Summary results for different investment strategies

The results show that for a 0% investment gap to CPI, and an initial payment of $22,051, all the GSA runs for a female aged 65 had a 100% success rate in delivering CPI indexed pensions for 30 years. It is worth noting that the 50/50 mix and dynamic strategy for individual drawdowns had 100% success rates for this level of drawdown. This raises a point that the pricing of the GSA must be relatively more attractive than an individual drawdown strategy to make it compelling for retirees.

A 0% gap, with initial payments of $29,742 for a male aged 70, had 100% success rates in reaching 30-years of CPI indexed payments for GSAs in Cash, a 50/50 mix and the dynamic strategy.  

However, the Balanced GSA failed on 12 occasions for quarters commencing June 1968 to December 1969 and March 1972 to March 1973. This occurred because large negative investment gaps to CPI occurred in the early years which compounded the negative mortality gaps and could not be overcome in later years. The Shares GSA failed on three occasions for quarters commencing December 1969, June 1972 and December 1972 for similar reasons. Yet, the success rates for these GSAs were still well above the comparable individual drawdown success rates at this level of payment.

The 100% Cash GSA produced a 100% success rate for all 141 runs at the $29,742 initial payment level because it didn’t suffer from the large negative investment gaps in the early periods while the mortality gaps were negative, and later benefited from large positive mortality yields combining with high cash rates that exceeded CPI increases. 

For the higher investment gaps of 3% and 4.5% to CPI, none of the runs produced 100% success rates, although the dynamic GSAs (using a simple rule-based switching strategy) had the highest success rates for both 65-year-old females and 70-year-old males producing over 90% success rates at the 3% investment gap level and over 70% success rates at the 4.5% investment gap.

All the GSAs had significantly higher success rates than the corresponding individual drawdowns for all the ‘like-for-like’ investment strategies, highlighting the benefits of mortality pooling for any specific investment strategy.

Key findings:

• For all targeted investment gaps, the ‘like-for-like’ GSA generally has higher success rates compared to individual drawdown strategies in delivering 30-year CPI indexed income streams.
• For targeted investment gaps of 3% and 4.5% there were no results that produced 100% success rates in achieving the 30-year targets, with highest success rates being over 90% at the 3% gap and over 70% at the 4.5% gap.
• Mortality yields can produce substantial increases to individual drawdowns for a given investment gap to CPI, however, the choice of investment gap to CPI and investment mix are critical in producing results that can sustain CPI indexed investment streams over 30-year periods and beyond.
• The actual mortality experience causes negative strains for up to 12-14 years after which strong positive contributions flow into the results.
• If poor investment markets and high inflation rates occur in the early periods, combined with the negative mortality strain, it can cause CPI indexed investment streams to fail within the first 30-years, and in some instances perform worse than the CPI indexed individual drawdowns in reaching the 30-year period.
• A dynamic investment strategy can significantly outperform the static investment mixes in supporting CPI indexed investment streams and help mitigate the negative mortality effects during the initial period.

Concluding analysis     

• The 65-year historical data period which incorporates high inflation over the 1970s and 1980s, and poor investment markets in the 1970’s, the stock market crash of 1987, the Global Financial Crisis and the COVID-19 pandemic provides a robust testing environment.
• The results suggest it is possible to create 30-year CPI indexed individual drawdowns provided initial drawdown calculations are consistent with the investment strategy being applied.
• GSAs can provide significantly higher income streams than individual drawdowns for any given level of investment risk, however they are subject to negative mortality gaps for the first 12-14 years which can compound the impact of negative investment gaps that may occur in the early years of a projection.

Further work could consider a mixture of different cohorts for GSAs and apply Austmod simulations to examine a portfolio of risks. It would be expected that numerous different cohorts with different starting dates and ages will produce even higher overall success rates than the single age cohort GSAs shown in this article.

References

CPD: Actuaries Institute Members can claim two CPD points for every hour of reading articles on Actuaries Digital.