Session 101 PD - Methods to Evaluate Retirement Plan Designs. Moderator: Dylan Porter, FSA, EA, FCA, MAAA

Transcription

1 Session 101 PD - Methods to Evaluate Retirement Plan Designs Moderator: Dylan Porter, FSA, EA, FCA, MAAA Presenters: Rowland Davis, FSA Marc Des Rosiers, FSA, FCIA SOA Antitrust Compliance Guidelines SOA Presentation Disclaimer

2 SOA Annual Meeting: October 17, 2017 Session 101: Methods to Evaluate Retirement Plan Designs SOA project: Quantitative Evaluation Framework Presenter: Rowland Davis, FSA

3 Overview Project started early 2016; completion by year-end Rowland Davis is lead researcher - Strongly engaged Project Oversight Group Goal: create a platform for consistent evaluation and comparison of different retirement programs, using quantitative measures with special emphasis on risk-sharing features Key features - Stochastic economic scenarios - Standardized assumptions - Broad set of risk and reward metrics - Standardized presentation of metric outcomes - Separate analysis for the accumulation phase and the payout phase 2

4 Standardized Assumptions: Demographic Career length / retirement age - Accumulations / benefit accruals start at age 30 - Retirement is at age 67 Career characteristics / dynamics - Results are a blend of full and fragmented careers, based on SOA 2003 turnover study Baseline uses medium turnover (approx. 10% aggregate); high and low t/o options available for sensitivity Full career = all years within a single plan (40% weight, using medium turnover rates) Fragmented = for medium turnover rates, three separate periods of plan participation (60% weight, using medium turnover rates) Only has impact where accruals are not uniform at each age (e.g. standard final pay DB plan) - Pay progression designed to represent typical median earner Age 30 pay = $40K; age 67 pay = $54K (current dollars) Projections based on stochastic wage inflation plus merit / promo factor Age 67 Social Security replacement ratio (in 2053) is assumed to be 40% of final pay 3

5 Standardized Assumptions: Economic Economic simulation model - Price inflation, wage inflation, bond yields Linked dynamics in all projections Starting values based on an economy in a long-term equilibrium state - Returns for bond portfolio, TIPS portfolio and risk asset portfolio (two versions) Assumptions and results - Mean / median values anchored by long-term expectations from economist surveys and publicly available assumption sets from Mercer and AonHewitt investment consulting practices - Provision for sensitivity testing through parameter set 4

6 Methodology Separate, but linked, models for accumulation phase and payout phase based on theoretical fungibility for design options Accumulation phase - To support comparison of various design choices, all results are anchored by assuming the same dollars are flowing into each plan design For DC plans, a constant 10% of pay each year For DB plans, and any other plan with variable costs/contributions, the mean value of the cost is set at 10% of pay (based on assumed level funding over career); the full range of the cost distribution is also developed - At retirement, the lump sum value of the accumulation is converted to lifetime income through purchase of a life annuity (at market interest rate, 2% fixed COLA, unisex mortality rates projected to/through 2053, assumed 5% load for group annuity pricing) - Most of the benefit metrics are based on the distribution of total replacement ratios = income per above / final pay at age % assumed for Social Security Payout phase - Start with age 67 accumulations from 10% of pay contributions into a typical TDF investment - Create a baseline set of yearly payouts (convert using a fixed rate, no load annuity) - Compare yearly payouts from design under analysis to the baseline 5

7 Payout Phase Sneak Peek Standard 4% Rule payouts (as % of baseline) Range of income benefits as percent of baseline*: *"Baseline" benefits are based on conversion of age 67 balance to lifetime income using a fixed-price lifetime annuity, interest rate = 6.1%, no load, full CPI COLA, no death benefits. Age sub-group values are weighted averages, using deaths at each age as the weighting factor. 200% 180% 160% 140% 120% 100% 80% 60% 40% 20% 0% Income Benefits as % of Baseline (Mean value, and inter-quartile range) % 180% 160% 140% 120% 100% 80% 60% 40% 20% 0% 6

8 Payout Phase Sneak Peek, contd. Standard 4% Rule accessible wealth Level of accessible wealth as percent of final pay (inflation adjusted) (Note that the age 67 value is before the purchase of any annuities.) 1000% 900% 800% 700% 600% 500% 400% 300% 200% 100% 0% Accessible Wealth as % of Final Pay (infl. adjusted) (Mean value, and inter-quartile range) % 900% 800% 700% 600% 500% 400% 300% 200% 100% 0% 7

9 Payout Phase Sneak Peek, contd. Combo of Longevity Annuity (12%) and Structured Withdrawal Plan (88%) payouts (as % of baseline) Range of income benefits as percent of baseline*: *"Baseline" benefits are based on conversion of age 67 balance to lifetime income using a fixed-price lifetime annuity, interest rate = 6.1%, no load, full CPI COLA, no death benefits. Age sub-group values are weighted averages, using deaths at each age as the weighting factor. 200% 180% 160% 140% 120% 100% 80% 60% 40% 20% 0% Income Benefits as % of Baseline (Mean value, and inter-quartile range) % 180% 160% 140% 120% 100% 80% 60% 40% 20% 0% 8

10 Payout Phase Sneak Peek, contd. Combo of Longevity Annuity (12%) and Structured Withdrawal Plan (88%) accessible wealth Level of accessible wealth as percent of final pay (inflation adjusted) (Note that the age 67 value is before the purchase of any annuities.) 1000% 900% 800% 700% 600% 500% 400% 300% 200% 100% 0% Accessible Wealth as % of Final Pay (infl. adjusted) (Mean value, and inter-quartile range) % 900% 800% 700% 600% 500% 400% 300% 200% 100% 0% 9

11 Accumulation Phase Full set of benchmark results for both standard DB and DC plan - Allocations to risk assets from 0% to 100% in 10% increments - DB plan does not have backloading Key metrics for total replacement ratios (incl. SS) - Full percentile distribution range - Mean value = reward - Conditional tail expectation (bottom and top quintiles) = risk / upside opportunity - Shortfall risk relative to target replacement ratio (both 70% RR and 60% RR) Key metrics for cost - Full percentile distribution range - Mean value = reward - Conditional tail expectation (bottom and top quintiles) = savings opportunity / risk - Adjusted cost metrics (both reward and risk) required to achieve a 70% total replacement ratio With 50% confidence With 70% confidence With 90% confidence 10

12 Metric Display: DB Plan with 70% Risk Asset Allocation Replacement Ratio (Benefit / Final Pay) (Includes Social Security at 39%) Cost (Percent of Pay) Risk 71.4% Avg. 77.6% 70% Reward 84.8% 80% 90% 100% 110% 120% Risk / Reward Metrics Avg. = mean value for full distribution of results Reward 5.7% 4% 6% Avg. 10.0% 8% 10% 12% 14% 16% Risk 15.7% 60% 50% 130% 140% Risk = mean value for worst quintile of results (i.e. for benefits, the lowest 20% of outcomes; for cost, the highest 20% of outcomes) 2% 0% 18% 20% Percentile values: 99th 92.6% 90th 82.8% 75th 80.5% 50th 77.0% 25th 74.8% 10th 72.6% 1st 69.4% Reward = mean value for best quintile of results (i.e. for benefits, the highest 20% of outcomes; for cost, the lowest 20% of outcomes) Percentile values: 99th 21.4% 90th 14.7% 75th 11.8% 50th 9.5% 25th 7.4% 10th 5.9% 1st 3.9% Other Metrics Benefit Shortfall Risk Benefits by Career Pattern Replacement Ratio Shortfall Turnover assumption = Baseline / medium Target: Probability Plan: Traditional career pay 70%+ 2.4% Average replacement ratio: # Employers Weight Soc. Sec. Plan Ben. Total Full career with single employer 1 40% 39.0% 38.6% 77.6% 60%+ 0.0% Multiple employers over career 3 60% 39.0% 38.6% 77.6% Blended average 39.0% 38.6% 77.6% Stress Index for Accumulation Path Adjusted Cost for a 70% Repl. Ratio Target: Measures reflect all years during career where account, or accrued benefit, experiences an investment-related loss, averaged over all scenarios. Loss Frequency Average Loss (% pay) Cummulative Career Losses (% pay) Index Cost to achieve a 70%+ replacement ratio, with a specified level of confidence. Confidence target: Average Cost (% pay) Best / smoothest possible path 0.0% 0.0% 0.0% 0.0 (i.e. no losses, such as DC w/ cash or SV investments) 50% confidence 8.4% 13.1% Result for DC w/100% fixed income (core style) 18.7% -5.6% -38.4% 0.9 Result for this plan 0.0% 0.0% 0.0% % confidence 8.9% 14.0% Result for DC w/ typical TDF (90/50) 26.3% -24.4% % 5.1 Worst / most volatile possible path 30.0% -41.7% % % confidence 9.5% 14.8% (i.e. DC plan w/ 100% risk asset allocation) Cost Risk Metric (% pay) 11

13 Metric Display: DC Plan with 70% Risk Asset Allocation Replacement Ratio (Benefit / Final Pay) (Includes Social Security at 39%) Cost (Percent of Pay) Avg. 82.6% Reward 108.7% Reward 10.0% Avg. 10.0% Risk 10.0% Risk 64.0% 60% 50% 70% 80% 90% 100% 110% 120% 130% 140% Risk / Reward Metrics Avg. = mean value for full distribution of results Risk = mean value for worst quintile of results (i.e. for benefits, the lowest 20% of outcomes; for cost, the highest 20% of outcomes) 2% 0% 4% 6% 8% 10% 12% 14% 16% 18% 20% Percentile values: 99th 136.2% 90th 105.1% 75th 90.0% 50th 79.8% 25th 71.2% 10th 65.1% 1st 57.3% Reward = mean value for best quintile of results (i.e. for benefits, the highest 20% of outcomes; for cost, the lowest 20% of outcomes) Percentile values: 99th 10.0% 90th 10.0% 75th 10.0% 50th 10.0% 25th 10.0% 10th 10.0% 1st 10.0% Other Metrics Benefit Shortfall Risk Benefits by Career Pattern Replacement Ratio Shortfall Turnover assumption = Baseline / medium Target: Probability Plan: Defined Contribution 70%+ 22.3% Average replacement ratio: # Employers Weight Soc. Sec. Plan Ben. Total Full career with single employer 1 40% 39.0% 43.6% 82.6% 60%+ 3.7% Multiple employers over career 3 60% 39.0% 43.6% 82.6% Blended average 39.0% 43.6% 82.6% Stress Index for Accumulation Path Adjusted Cost for a 70% Repl. Ratio Target: Measures reflect all years during career where account, or accrued benefit, experiences an investment-related loss, averaged over all scenarios. Loss Frequency Average Loss (% pay) Cummulative Career Losses (% pay) Index Cost to achieve a 70%+ replacement ratio, with a specified level of confidence. Confidence target: Average Cost (% pay) Best / smoothest possible path 0.0% 0.0% 0.0% 0.0 (i.e. no losses, such as DC w/ cash or SV investments) 50% confidence 7.6% 7.6% Result for DC w/100% fixed income (core style) 18.7% -5.6% -38.4% 0.9 Result for this plan 26.3% -24.4% % % confidence 9.2% 9.2% Result for DC w/ typical TDF (90/50) 27.0% -22.6% % 4.9 Worst / most volatile possible path 30.0% -41.7% % % confidence 11.9% 11.9% (i.e. DC plan w/ 100% risk asset allocation) Cost Risk Metric (% pay) 12

14 Creating a Risk-Reward Space : Benefit Risk Chart 5: Risk = RR risk metric 80% DB 75% M O R E R I S K Risk = RR risk metric 70% 65% 60% DC 55% Higher risk asset allocation 50% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100% Reward = Avg. RR 13

15 Creating a Risk-Reward Space : Cost Risk Chart 7: Risk = Cost risk metric 20% DB 18% 16% M O R E R I S K Risk = cost risk metric 14% 12% 10% 8% 6% 4% DC 2% Higher risk asset allocation 0% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100% Reward = Avg. RR 14

16 Creating a Risk-Reward Space : Shortfall w/ 70% Target RR 100% 90% 80% Chart 6: Risk = shortfall risk metric M O R E R I S K Risk = shortfall % 70% 60% 50% 40% 30% DC 20% 10% DB 0% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100% Reward = Avg. RR Higher risk asset allocation 15

17 Using the Adjusted Cost Metric Cost (% pay) 18% 16% 14% 12% 10% 8% 6% 4% 2% 0% DC vs DB Cost Analysis 70% RR with 90% Confidence 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Risk Asset % DB Risk DC DB Avg 16

18 Completed Analyses ( Studies ) Name Description Study A001 Risk vs Reward Explores basic risk and reward relationships for standard DC plans, and standard DB plans. Introduces various graphics that may be used to define "efficient frontiers" using the QE Framework metrics. Study A002 Risk Sharing Plans General discussion of how the QE Framework may be used to evaluate and compare alternative plan structures that incorporate risk sharing features. Study A003 TDF Analysis Compares the DC Plan results using Target Date Funds (TDF's) with benchmark results using fixed allocations to risky assets. Also explores the impact of additional fees that may be charged for TDF's. Study A004 Portability & Backloading Describes the metric used for portability, and illustrates level of backloading in selected DB designs. Study A005 Mean Reversion Analysis Explores impact of removing the mean reversion factor that is used in the baseline simulation model. Includes a complete set of DC and DB benchmark results using the alternative (lognormal) model. Study A006 Assumption Sensitivity Illustrates how model can be run with alternative assumptions, using assumption sets with lower levels of inflation, yields and returns. Shows impact on selected metrics. Includes a complete set of DC and DB benchmarks for one set of alternative assumptions. Study A007 Cash Balance: Fixed Credit Rate Shows results for various Cash Balance plans using fixed "Investment" crediting rates. These results indicate that this type of plan is virtually identical to standard DB plans in terms of all economic results. Study A008 Study A009 Cash Balance: Variable Credit Rate Cash Balance Plan: Participating Credit Rate Shows results for various Cash Balance plans using variable "Investment" crediting rates. These results indicate that this type of plan is virtually identical to both fixed rate Cash Balance plans, and to standard DB plans in terms of all economic results. Shows results for various Cash Balance plans using participating "Investment" crediting rates, based on actual fund investment results. Efficient frontier graphs used to show risk-sharing nature. Portability issues reviewed. Study A010 Return Guarantees Shows results for a traditional DC system with various return guarantees purchased at market price, applied to the full career accumulation. Includes some sensitivity analysis relative to the risky asset model used, the equity risk premium assumption, and the risk free rate used for pricing the guarantees. Discusses other ways that return guarantees may be included in a plan. Study A011 DC Plan with Tracker Support Shows results for DC plans that also incorporate a variable contribution which varies with the tracking error of the accumulation vs a target accumulation path. Efficient frontier graphs used to show risk-sharing nature. Evaluation is made by comparison with equivalent combo DC+DB benchmark plans., and implementation ideas are discussed. 17

19 Evaluation of Hybrid Risk-Sharing Designs Using the risk-sharing space concept - The hybrid plan results would be developed at 30%, 50% and 70% risk asset allocation - The results can be plotted as three points on one of the graphs (e.g. the one using the benefit risk metric) - At each of these points there also exists a combo DB+DC arrangement with the same results - Can then compare other metrics (e.g. the cost risk metric) between the hybrid design and the equivalent combo DB+DC arrangement as a benchmark - If for example, the cost risk metric for the hybrid design is lower than that for the equivalent combo DB+DC, then there is some efficiency in managing the cost risk Using the adjusted cost metric - Develop the hybrid plan adjusted cost metrics required for the plan to meet a 70% target replacement ratio (including SS) with, for example, 90% confidence - Compare results directly Against any other hybrid plan Against benchmark combo DB+DC arrangements 18

20 Two Sample Hybrid Risk-Sharing Plans Kentucky Plan - Implemented in 2014 for new public employees - Typical of plan structure promoted by Laura and John Arnold Foundation - Participating cash balance plan structure Defined benefit plan with full employer exposure to liabilities Floor credit rate = 4% 75% participation in returns over 4% (5-yr. moving average) for retiring participants only Tracker Plan - Proposal from Retirement 20/20 paper by Rowland Davis - DC plan with variable contributions Establish target path for accumulation Monitor tracking error for each cohort using standardized sample employee If negative tracking error exceeds certain limits, employer makes supplemental contributions - Will move accumulation back towards the target path - Fixed cap on the amount of supplemental contributions so no employer exposure to liabilities 19

21 Kentucky Plan With 70% Risk Asset Allocation CHART A: Results with a 70% allocation to risk assets 16% 14% 12% 16% 14% 12% Results: Based on expected average cost, the KY Plan matches the outcomes from a combo DB/DC arrangement with about a 50/50 blend -- so the KY Plan shows significant risk sharing, packaged within a single hybrid DB structure. 10% 8% 6% Orange = adjusted cost risk 10% 8% 6% However, the cost risk metric for the KY Plan exceeds the level for a 50/50 combo arrangement -- so there is some loss of efficiency in terms of risk management for the sponsor. 4% 2% Blue = avg. adjusted cost 4% 2% 0% DB/DC 100/0 75/25 50/50 25/75 0/100 Blend < Risk Sharing Spectrum > KY Plan 0% 20

22 Tracker Plan 21

23 Tracker Plan With 70% Risk Asset Allocation CHART A: Results with a 70% allocation to risk assets 16% 14% 12% 16% 14% 12% Results: Based on expected average cost, the Tracker Plan matches the outcomes from a combo DB/DC arrangement with about a 25/75 blend -- so the Tracker Plan shows noticeable risk sharing, packaged within a single hybrid DC structure. 10% 8% 6% Orange = adjusted cost risk 10% 8% 6% The cost risk metric for the Tracker Plan slightly exceeds the level for a 25/75 combo arrangement -- so there is some minor loss of efficiency in terms of risk management for the sponsor. However, the Tracker Plan has no DB-type liabilities, while the 25/75 combo arrangement would have some. 4% Blue = avg. adjusted cost 4% 2% 2% 0% DB/DC 100/0 75/25 50/50 25/75 0/100 Blend < Risk Sharing Spectrum > Tracker Plan 0% 22

24 Next Steps Finish technical review (payout phase) Documentation SharePoint website - Documentation - Excel workbooks - Completed Studies Further studies, with Pension Section oversight - Accumulation schemes with inter-generational risk sharing Target benefit DB plans Notional account DC plans (e.g. SAFE Plan from Center for American Progress) - Payout schemes Variable annuities - Insured products - Collectives with inter-generational risk sharing Combo arrangements with partial annuitization (e.g. longevity insurance products) Outreach - Actuarial community: articles, podcast, webinar? - Policy makers, think tanks: webinar? other? 23

25 2017 SOA Annual Meeting & Exhibit MARC DES ROSIERS, FSA, FCIA Session 101, Methods to Evaluate Retirement Plan Designs October 17, 2017

26 SOCIETY OF ACTUARIES Antitrust Compliance Guidelines Active participation in the Society of Actuaries is an important aspect of membership. While the positive contributions of professional societies and associations are well-recognized and encouraged, association activities are vulnerable to close antitrust scrutiny. By their very nature, associations bring together industry competitors and other market participants. The United States antitrust laws aim to protect consumers by preserving the free economy and prohibiting anti-competitive business practices; they promote competition. There are both state and federal antitrust laws, although state antitrust laws closely follow federal law. The Sherman Act, is the primary U.S. antitrust law pertaining to association activities. The Sherman Act prohibits every contract, combination or conspiracy that places an unreasonable restraint on trade. There are, however, some activities that are illegal under all circumstances, such as price fixing, market allocation and collusive bidding. There is no safe harbor under the antitrust law for professional association activities. Therefore, association meeting participants should refrain from discussing any activity that could potentially be construed as having an anti-competitive effect. Discussions relating to product or service pricing, market allocations, membership restrictions, product standardization or other conditions on trade could arguably be perceived as a restraint on trade and may expose the SOA and its members to antitrust enforcement procedures. While participating in all SOA in person meetings, webinars, teleconferences or side discussions, you should avoid discussing competitively sensitive information with competitors and follow these guidelines: Do not discuss prices for services or products or anything else that might affect prices Do not discuss what you or other entities plan to do in a particular geographic or product markets or with particular customers. Do not speak on behalf of the SOA or any of its committees unless specifically authorized to do so. Do leave a meeting where any anticompetitive pricing or market allocation discussion occurs. Do alert SOA staff and/or legal counsel to any concerning discussions Do consult with legal counsel before raising any matter or making a statement that may involve competitively sensitive information. Adherence to these guidelines involves not only avoidance of antitrust violations, but avoidance of behavior which might be so construed. These guidelines only provide an overview of prohibited activities. SOA legal counsel reviews meeting agenda and materials as deemed appropriate and any discussion that departs from the formal agenda should be scrutinized carefully. Antitrust compliance is everyone s responsibility; however, please seek legal counsel if you have any questions or concerns. 2

27 Presentation Disclaimer Presentations are intended for educational purposes only and do not replace independent professional judgment. Statements of fact and opinions expressed are those of the participants individually and, unless expressly stated to the contrary, are not the opinion or position of the Society of Actuaries, its cosponsors or its committees. The Society of Actuaries does not endorse or approve, and assumes no responsibility for, the content, accuracy or completeness of the information presented. Attendees should note that the sessions are audio-recorded and may be published in various media, including print, audio and video formats without further notice. 3

28 Background

29 5

30 Background A framework to evaluate the value and effectiveness of a DC plan Used to compare DC programs and highlight strengths and weaknesses Considers quantitative and qualitative features 6

31 7

32 Principles governing framework Evaluation of plans, not an individual Each feature compared against range of existing possibilities Range of features applicable to particular plan size/industry Measure of ongoing plan success Shared responsibility between member and sponsor/employer Importance of auto features (auto-enrollment and autoescalation) 8

33 Objective Function

34 Objective Function Assigns a value between 0% and 100% to a DC plan Weights for each criterion (or subcriterion) add up to 100% Plan value is sum of the product of each criterion s weight times its value 10

35 Objective Function Has Two Versions Based on plan terms only, without regard to existing participant experience Plan value = (Provisions) w 1 + (Adequacy) w 2 + (Other criteria) w 3 Based on both plan terms and existing participant experience Plan value = (Provisions) w 1 + (Adequacy) w 2 + (Other criteria) w 3 + (Plan success) w 4 where w i are weights assigned to each of the main criteria 11

36 Overview of Model

37 Other Models BrightScope Quantitative and qualitative details Highlights plan strengths and weaknesses Nonmonetary features makes it comprehensive Benefit adequacy Watson Wyatt study Measures benefit adequacy, plan success and investment efficiency BrightScope and PLANSPONSOR Data sources to determine range of plan features in the market 13

38 14

39 Plan Provisions Subcriteria 15

40 Plan design Criteria Value Employer contributions min(employer contribution rate, 9%)/9% Matching formula min(employer matching percentage, 100%) Availability of Roth contribution option Available: 100%; Not available: 0% Employee contributions Available: 100%; Low: 50%; None: 0% 16

41 Investment options Criteria Value Fees Qualitative assessment of 0%, 25%, 50%, 75% or 100% Efficiency of investment options Qualitative assessment of 0%, 25%, 50%, 75% or 100% Diversification of options menu Qualitative assessment of 0%, 25%, 50%, 75% or 100% Retirement income solutions Qualitative assessment of 0%, 25%, 50%, 75% or 100% 17

42 18

43 Enrollment design Criteria Value Vesting Qualitative assessment of 0%, 25%, 50%, 75% or 100% Eligibility Qualitative assessment of 0%, 25%, 50%, 75% or 100% Auto-enrollment Qualitative assessment of 0%, 25%, 50%, 75% or 100% Auto-escalation Qualitative assessment of 0%, 25%, 50%, 75% or 100% 19

44 Communications Criteria Value Plan information Qualitative assessment of 0%, 25%, 50%, 75% or 100% Education and tools (investor profile, online Qualitative assessment of 0%, 25%, 50%, 75% or 100% planning) Plan adviser services and support Qualitative assessment of 0%, 25%, 50%, 75% or 100% Effectiveness of education and communication approach Qualitative assessment of 0%, 25%, 50%, 75% or 100% 20

45 21

46 Plan Adequacy 22

47 Plan Adequacy Value for plan adequacy = Expected total replacement ratio Target replacement ratio over full career Expected total RR = (Social security RR) + (Other employer-provided RR) + (Plan RR) Social security RR = Average social security RR based on income level Other employer-provided pension RR = Replacement ratio provided by another employersponsored pension plan over full career Plan RR = Accumulated assets at retirement as a multiple of real pay Annuity certain to end of life expectancy Target RR = Target replacement ratio required to provide adequate retirement income Based on employee and employer contribution accumulations, includes auto-escalation 23

48 24

49 Governance & Other 25

50 Governance criteria Investment monitoring and review process Employee committee representation Risk management framework and compliance Transparency Qualitative assessment of 0%, 25%, 50%, 75% or 100% Value is average of all criteria 26

51 Other Criteria Loan provisions Other retirement programs with employer Hardship withdrawal provisions Fee equalization policy Qualitative assessment of 0%, 25%, 50%, 75% or 100% Value is average of all criteria 27

52 28

53 Plan Success 29

54 Participation Rate Participation = Actual participation rate Expected participation rate Actual participation rate = (Number of plan members)/(number of eligible employees) Expected participation rate = Estimated participation rate for plan size or industry 30

55 Investment Efficiency Investment efficiency = (Actual percentage of diversified equities) (Optimal equity level) Investment efficiency =100%- (Optimal equity level-actual percentage of diversified equities)/(optimal equity level) Actual percentage of diversified equities = Plan assets invested in diversified equities, excluding company stock Optimal equity level = 100% - Participant's average age /

56 32

57 Objective Function Results

58 Top Level Function 34

59 Objective Function Weights Criteria Weights without existing plan Weights with existing plan Value for plan provisions 34% 25% Value for plan adequacy 56% 41% Value for governance and other provisions 10% 7% Value for plan success (existing plans only) N/A 27% Total 100% 100% 35

60 36

61 37

62 Example (Appendix B of Report) Criteria Value B1. Base case: 5% employer contributions 72% B2. Base case but with alternate formula taking into account plan success 77% B3. Base case but with 8% employer contributions 85% B4. Base case but with auto-enrollment and auto-escalation 76% 38

63 Analytic Hierarchy Process

64 Analytic Hierarchy Process Weights are determined using the Analytic Hierarchy Process (AHP) Structured technique for organizing and analyzing complex decisions A branch of operations research, invented by mathematician Thomas L. Saaty Method to ensure importance of each criterion are consistent with each other 40

65 Principles of AHP Each criterion rated in terms of its importance relative to other criteria. A method to evaluate each criterion relative to each other in a consistent manner Based on linear algebra concepts eigenvectors Converts values in a two-dimensional matrix to vectors to get objective function weights Google PageRank search engine algorithm uses eigenvectors! 41

66 Pairwise Comparisons AHP uses pairwise comparisons to establish a ranking hierarchy for each criterion Qualitative judgment on a scale of 1 to 9 between each two alternatives. Comparing each one to the others: six pairwise comparisons 42

67 Pairwise Comparisons with Four Nodes 43

68 AHP Value Judgment Scale Intensity of Importance Definition Explanation 1 Equal importance Two elements have the same value 3 Moderate importance One element is moderately better 5 Strong importance One element is significantly better 7 Very strong importance One element is greatly better 9 Extreme importance One element is better than the other at the highest possible degree 44

69 Value judgments for Each Pairwise Comparison Plan provisions 5 Plan adequacy 3 Plan adequacy slightly more important than actual plan provisions Plan provisions 5 Governance and other 1 Plan provisions such as employer contributions, vesting and enrollment significantly more important than other criteria Plan provisions 1 Plan success 1 For an ongoing arrangement, plan provisions as important as participation levels and investment efficiency Plan adequacy 5 Governance and other 1 Plan adequacy significantly more important than governance and other criteria Plan adequacy 1 Plan success 1 Plan adequacy just as important as plan success Governance and other 1 Plan success 3 Plan success somewhat more important than governance and other criteria 45

70 Reciprocal Matrix For each pairwise comparison, the number representing the greater weight is transferred to the cell that intersects in the matrix Reciprocal of that number is put into the cell of the other intersection, working horizontally Criteria Plan provisions Plan adequacy Governance and other Plan success Plan Plan Governance Plan Priority Provisions Adequacy and Other Success 1 1/ /5 1/5 1 1/

71 "Priority" is the Weight The priority is the normalized value obtained by this formula: Priority for criterion i = Sum of normalized values for row / Number of rows where: Normalized value for cell [i, j] = value in cell [i,j]/sum of values in column j 47

72 Consistency Index and Consistency Ratio Method to verify if results are consistent Consistency index using as the lambda max a measure of the maximum eigenvalue of the matrix L max = λ max. Consistency Index (CI) = (λ max n) / (n 1) Consistency ratio (CR) = CI / RI where RI is the Random Index, the CI value obtained from randomly generated matrices Lambda max Consistency index Assessment Very consistent (<10%) Consistency ratio

73 49

74 50

75 References

76 References Used to Derive Ranges Aon Hewitt, 2011 Trends and Experience in Defined Contribution Plans, 2011 Deloitte / International Foundation of Employee Benefit Plans, Annual Defined Contribution Benchmarking Survey, 2014 Vanguard Institutional Investor Group, How America Saves 2014, 2014 Michael Clingman, Kyle Burkhalter, and Chris Chaplain, Replacement Rates for Hypothetical Retired Workers, Actuarial Note Number , Office of the Chief Actuary, Social Security Administration, July 2015 BrightScope / Investment Company Institute, The BrightScope/ICI Defined Contribution Plan Profile: A Close Look at 401(k) Plans, December 2014 Jack Van Derhei and Lori Lucas, The Impact of Auto-enrollment and Automatic Contribution Escalation on Retirement Income Adequacy, Employee Benefit Research Institute Issue Brief, no. 349, November 2010 PLANSPONSOR, 2014 DC Survey: Plan Benchmarking, January 2015, 52

77 Quick references to the report Section 3.2, Objective Function Benchmarking criteria: Plan Provisions (Section 3.3) Plan Adequacy (Section 3.4) Other Criteria (Section 3.5) Plan Success (Section 3.6) Appendix A: Using/Modifying the Excel Model Spreadsheet Appendix B : Examples 53

78 Summary

79 A System to Evaluate and Compare DC Plans Rational approach to quantify plan features, based on: Contribution levels Fees Investment options Auto-enrollment, auto-escalation Eligibility, vesting Replacement ratio adequacy Plan participation and investment efficiency Nonmonetary features (e.g., income solutions, communications, etc.) 55

80