NCPERS Pension Funding Forum

Transcription

1 NCPERS Pension Funding Forum Gene Kalwarski, FSA, CEO Sandy Matheson, Director Maine PERS

2 Topics for Discussion Just What is Stress Testing? Why Stress Testing is Good for DB Plans? How Does Stress Testing Work? Is There a Stress Testing Requirement? Real Life Example - Maine PERS 1

3 What is Stress Testing? A way to measure the risk of a financial system s capacity to meet (or not meet) its future obligations. 2

4 Background Who Uses Stress Testing? Insurance Industry Requires Solvency Testing Using Dynamic Financial Analysis since 1980 to forecast ruin probabilities Banking Industry Dodd-Frank - established stress tests as a key component for regulators to gauge the banking industry health Social Security Trustees Report Optimistic, pessimistic, and best estimate 3

5 Stress Testing Is a Projection It is Not a solution - It is a measurement It is Not a decision - It is information necessary to make better decisions 4

6 How to Think About Stress Testing 5

7 Traditional Approach - What is Missing? Summary of Principal Results 7/1/2015 7/1/2016 Participant Counts Actives Participants with Deferred Benefits Retirees and Disableds Beneficiaries Total 1,821 1,801 Annual Salaries of Active Members $ 23,265,360 $ 21,492,384 Annual Retirement Allowances $ 39,365,610 $ 40,392,906 Assets and Liabilities Actuarial Accrued Liability $ 525,930,801 $ 522,542,498 Actuarial Value of Assets 376,332, ,071,975 Unfunded Actuarial Liability / (Surplus) $ 149,598,025 $ 147,470,523 Funding Ratio (Actuarial Value of Assets) 71.56% 71.78% Funding Ratio (Market Value of Assets) 70.02% 67.09% Present Value of Accrued Liability $ 509,442,056 $ 508,437,615 Market Value of Assets 368,266, ,581,323 Unfunded FASB Accrued Liability $ 141,175,982 $ 157,856,292 Accrued Benefit Funding Ratio 72.29% 68.95% Actuarial gain/(loss) Liability gain/(loss) $ 5,093,044 $ 1,608,028 Actuarial asset gain/(loss) (2,016,676) (6,626,834) Net Transfer gain/(loss) 24,173 (297,812) Total actuarial gain/(loss) $ 3,100,541 $ (5,316,618) Annual Contribution Requirement at End of Year FYE 2016 FYE 2017 Normal Cost (with interest and expense load) $ 1,960,945 $ 1,825,059 UAL Amortization (with interest) 17,915,982 18,523,911 Annual Contribution Requirement $ 19,876,927 $ 20,348,970 6

8 Why Is Stress Testing Important? Without projections you have no idea if funding works Baseline projections are never right Uncertainty increases over time Most pension plans today are mature Mature plans have negative cash flows Negative cash flow plans are most vulnerable Forward looking potential outcomes are more important to decision making than what has already happened 7

9 How Does Stress Testing Work? First, start with a projection of expected results (best estimate assumptions) Then test what happens if you are wrong (project alternative scenarios) Finally, you decide what results you can t live with (decide what your risk tolerances are) 8

10 Baseline Projection: Not A Stress Test 9

11 A Simple Stress Test 10

12 A Robust Stress Test 11

13 Stochastic Stress Testing S

14 Lesson - How to Not to Do Stress Testing 13

15 Is There a Pension Stress Testing Requirement? No but it is a very good, if not critical, practice for all pension plans 14

16 Is There a Pension Stress Testing Requirement? 15

17 Is There a Pension Stress Testing Requirement? 16

18 Types of Actuarial Analysis Most Systems Do Annual Actuarial Valuations Compares current year s results to prior year Asset/liability Studies Focuses mostly on asset allocation Experience Studies Again, backward looking analysis Stress Testing Not a common practice 17

19 MainePERS Where It Worked MainePERS has a multiple-employer costsharing plan for local governments The Plan was 86% funded in 2016 following two years of 1.5% returns Most economic projections in the Spring of 2016 predicted returns could hover at 4% for 4 years before slowly climbing back to 8% 18

20 So We Tested What Would Happen 19

21 Live Stress Testing Demonstration Assumed Historical Cost Sharing 1 Cola adj Funded % 1928 ER EE Contr cap 30/90 Provision N % 89% Choose Run 3 intial ben #N/A #N/A max 99.0% Early Ret Fact N 40/ % 95% 2.25% COLA cap future g/l #N/A #N/A min 12% Discontinue for all? N % 96% reduce cola? N % 91% % 79% % 69% 4.720% 67% % 65% % 69% % 76% % 74% 1.360% 75% % 76% % 74% % 70% % 70% % 72% % 76% % 84% 2.500% 88% 4.940% 90% % 93% % 97% % 106% % 115% selection= ### #REF! #REF! #REF! #### #REF! #REF! #REF! #REF! 0% 8.080% 30% 25% 20% 15% 10% 5% FY FY Employer Contrib Rate Baseline Employer Contrib. Rate FY FY FY FY FY Member Contrib. Rate Baseline Member Contrib. Rate FY FY FY FY FY FY

22 What Did We See? If rates more than doubled, we knew Benefits would have to be reduced COLAs frozen or eliminated Employers would drop the plan This combination could create a last-man standing situation and probable demise of the plan 21

23 Why Did This Create Success We created a new framework that makes the plan sustainable into the future We changed the way risk is shared Made both employer and employee contribution rates variable each year based on market returns Employer 55%; employees 45% Created rate caps and minimums for both Excess contribution requirements are amortized into the COLA 22

24 The Result Employers know how to budget in bad times because they know the maximum their rates can be Employees get to share in the good times rather than continue to pay but get benefit reductions Retiree COLAs caps do not get reduced nor do annual COLAs get frozen 23

25 Live Stress Testing Demonstration 24

26 Other MainePERS Plan with Long-Term Stress Testing Success 25

27 Stress Testing 26