Session 122 PD, Lessons Learned: Two Years of Three Rs. Moderator: Shyam Prasad Kolli, FSA, MAAA

Transcription

1 Session 122 PD, Lessons Learned: Two Years of Three Rs Moderator: Shyam Prasad Kolli, FSA, MAAA Presenters: David M. Dillon, FSA, MAAA Andrew Ryan Large, FSA, CERA, MAAA SOA Antitrust Disclaimer SOA Presentation Disclaimer

2 2016 SOA Annual Meeting & Exhibit Two Years of the Three Rs Moderator: Presenters: Shyam Prasad Kolli, FSA, MAAA Milliman, Inc. Dave Dillon, FSA, MAAA Lewis & Ellis, Inc. Andy Large, FSA, CERA, MAAA Milliman, Inc.

3 Disclaimer This presentation is intended for educational purposes only and does not constitute financial, accounting, legal, or tax advice of any kind. It does not replace independent professional judgment. Statements of fact and opinions expressed are those of the presenter individually and, unless expressly stated to the contrary, are not the opinion or position of the participant s employer, Society of Actuaries, its cosponsors or its committees. No warranty is given, nor is any responsibility assumed by the presenter or the SOA for the content, accuracy or completeness of the information presented. Attendees should note that the sessions may be audio-recorded and may be published in various media, including print, audio and video formats without further notice. 2

4 Primary Topics Membership & Metal Tiers Pent Up Demand Risk Corridor Risk Adjustment Transition Policies Transitional Reinsurance Additional Considerations R 3

5 Marketplace Membership Distributions Office of the Assistant Secretary for Planning and Evaluation (ASPE) Age Nationwide 2016 State Variability Age Age Min 25% ile 50% ile 75% ile Max < % 9.0% 8.0% < % 6.5% 9.0% 11.0% 23.0% % 10.0% 11.0% % 8.0% 9.0% 11.0% 14.0% % 16.0% 17.0% % 16.0% 17.0% 18.5% 37.0% % 16.0% 16.0% % 15.0% 15.0% 17.0% 21.0% % 21.0% 21.0% % 19.0% 20.0% 21.5% 24.0% % 27.0% 26.0% % 24.0% 27.0% 31.0% 38.0% % 1.0% 1.0% % 0.0% 1.0% 1.0% 2.0% 4

6 Marketplace Membership Distributions Enrollment Type Nationwide 2016 State Variability Enrollment Enrollment Min 25% ile 50% ile 75% ile Max New 27.0% 39.0% New 6.0% 33.0% 39.0% 44.5% 99.0% Re-enrollee 73.0% 61.0% Re-enrollee 1.0% 55.5% 61.0% 44.8% 94.0% Financial Assistance Financial Nationwide Financial 2016 State Variability Assistance Assistance Min 25% ile 50% ile 75% ile Max Yes 85.0% 82.0% 83.0% Yes 6.0% 77.3% 84.0% 87.0% 91.0% No 15.0% 18.0% 17.0% No 9.0% 13.0% 16.0% 23.0% 94.0% 5

7 Marketplace Membership Distributions Metal Tier Nationwide 2016 State Variability Tier Type Tier Type Min 25% ile 50% ile 75% ile Max Bronze 20.0% 22.0% 23.0% Bronze 8.0% 20.0% 23.0% 28.5% 46.0% Silver 65.0% 68.0% 68.0% Silver 31.0% 60.0% 67.0% 70.5% 86.0% Gold 9.0% 7.0% 6.0% Gold 3.0% 5.0% 7.0% 11.0% 21.0% Platinum 5.0% 2.0% 2.0% Platinum 0.0% 0.0% 0.0% 1.0% 16.0% Catastrophic 2.0% 1.0% 1.0% Catastrophic 0.0% 1.0% 1.0% 1.5% 6.0% 6

8 Modeling Pent-Up Demand Difficult to Identify Until No Longer Present Marketwide morbidity change Enrollment issues Late joiners Transitional Policies Benefit changes

9 The Legend of Pent-Up Demand Potential Evidence: CCIIO Risk Pool Analysis (8/11/16) 0.1% decline in adjusted* PMPM claim costs Coding practices likely to increase reported risk scores 5% average drop for 10 states with highest enrollment growth Counterpoint: Sustained High Costs Young and healthies still not present Less healthy enrollment adverse selection Rate increases death spiral Voluntary market withdrawals

10 Risk Corridor Impact Loss Ratios by State

11 Risk Corridor Impact Loss Ratios by State

12 Risk Adjustment The PLRS Guessing Game Primary Contributing Factors Age Metal Level HCC / Diagnosis Coding Additional Potential Correlations Transitional Policies Medicaid Expansion

13 PLRS 2014 vs Difficulty in Comparing 2014 vs Startup Issues Partial Data Poor Coding Efforts ICD10 Previous Uninsureds Plan Switchers / Late Enrollers Multiple Year Claims Marketplace Entry / Exit

14 Risk Adjustment: 2014 & 2015 Results Society of Actuaries Research Paper An Examination of Relative Risk in the ACA Individual Market Risk measures show that the average measure of risk increased from 2014 to Increased risk scores may be a combination of identification through better coding as well as well as a measure of the actual population health. Some states and plans had significant swings, in both directions. State variability due to underlying market characteristics means that overall measures may not adequately represent the experience. 13

15 Risk Adjustment: 2014 & 2015 Results There was a fairly strong correlation between the 2014 risk score and the 2015 risk score. The correlation appears to be tighter for the lower scoring states, but the dispersion grows above

16 Risk Adjustment: 2014 & 2015 Results The increase in risk score was not uniform and there is a non-uniform shift in the shape of the distribution from something nearly normal to one that is somewhat skewed to the higher end

17 Risk Adjustment: 2014 & 2015 Results Changes in risk scores were not as geographically correlated as were the base 2015 scores. Lower scored areas from 2014 were not necessarily the states that had the highest increase in average score. 16

18 Risk Adjustment: 2015 Results Risk scores by state. The western states and New England have lower scores in general than the central and southern states. 17

19 Risk Adjustment: 2015 Results Variation in risk scores by population. This shows how widely dispersed the reported risk is from state to state, and how it is not dependent on the size of the pool. 18

20 Risk Adjustment: 2015 Results Variation in risk scores by population, excluding TX, CA, and FL. 19

21 Risk Adjustment: 2015 Results Do higher premiums correlate with higher risk scores? The answer appears to be no which is not an unexpected result since premiums reflect more than relative illness burden, e.g. the cost of services and physician practice patterns within a state. 20

22 The PLRS Guessing Game Projecting Forward to 2017 and Beyond Difficult to draw conclusions Results may vary Zero-sum How much coding is enough? Potential Changes Inclusion of Rx data Duration adjustments Loss absorption Might we see 6-buckets?

23 Risk Exposure Mitigation Extreme volatility, especially for small carriers 0.01 change in PLRS can equate to $3 transfer payment Anecdotally, 40% of premium

24 Mitigation Strategies Static populations may be more stable PLRS does not follow insureds Scenario modeling / Simulation studies Alabama DOI Push for EDGE data accuracy Completion / Seasonality analysis Coding vendors? Winning vs. avoiding a loss

25 Transition Policy Impact Do transitional plans may have lower risk profiles than the ACA Individual pool? That is, do states with no transitional enrollment have a lower risk score? 24

26 Transition Policy Impact Changes in risk scores were not particularly damped or heightened by the state decisions on Transition Policies. 25

27 Transitional Reinsurance Unlike risk scores, there is no particular geographic pattern. The graph excludes the two outlier states of Wyoming ($89) and Alaska ($164), in order to highlight the differences in the other states. 26

28 Transitional Reinsurance Reinsurance PMPM is not as well correlated with premium PMPM as would be expected. It might be expected that for pools as large as the ACA individual pool, the presence of very large claimants would not differ as much as is evident 27

29 Transitional Reinsurance 2014 Accrued: Collected: Paid: Carried Forward: $6.9 billion $9.7 billion $7.9 billion $1.8 billion 2015 Accrued: Collected: From 2014: Paid: $7.4 billion $6.0 billion $1.8 billion $7.0 billion Source: SUMMARY REPORT ON TRANSITIONAL REINSURANCE PAYMENTS AND PERMANENT RISK ADJUSTMENT TRANSFERS FOR THE 2015 BENEFIT YEAR & 2016 SOA Health Meeting Session 132 L, Financial analysis of ACA health plans 28

30 Additional Considerations Liquidity Modeling Transitional Policy Adjustments Can we estimate via public data? Adjusting Expectations for Market Insolvencies Qualified Opinions? November 8, 2016

31 Contact & Questions Contact: Andy Large Dave Dillon Questions: