Public Plan Investment Performance, 2001-2016 JP Aubry Center for - - PowerPoint PPT Presentation

JP Aubry Center for Retirement Research at Boston College Director of State and Local Research Public Pension Management and Asset Investment Review Commission Harrisburg, PA September 20, 2018

Public Plan Investment Performance, 2001-2016

CRR assesses plan performance in two ways.

1

1. A comparison of investment returns across plans: Observed differences are the result of both differences in asset allocation and/or asset class performance. 2. A comparison of each plan’s investment return to its

• wn benchmark:

Performance relative to benchmark focuses on each plan’s ability to execute its own strategy.

The long-term (2001-2016) investment return varies greatly among public plans.

2

Distribution of Plans by Long-term (2001-2016) Annualized Return

Source: Author's calculations using the Public Plans Database (2001-2016).

0% 3% 6% 9% 12% 15% <4.0% 4.5% 5.0% 5.5% 6.0% 6.5% 7.0%+ Percent of Plans Annualized Return, 2001-2016

PA SERS - 5.5% Bottom Quartile 2nd Quartile 3rd Quartile Top Quartile PA Schools - 5.3%

The difference in returns accounts for much

• f the variation in today’s funded status.

3

2016 Market Funded Ratios under Various Return Assumptions, by Quartile

62.9% 79.6% 88.5% 63.6% 0% 25% 50% 75% 100% Bottom quartile Top quartile Actual returns Alternative returns

Source: Jean-Pierre Aubry, Anqi Chen, Alicia H. Munnell, and Kevin Wandrei. 2018. “What Explains Differences in Public Pension Returns since 2001?” State and Local Plans Issue in Brief 60. Chestnut Hill, MA: Center for Retirement Research at Boston College.

At a high level, the asset allocation of most public plans is quite similar.

4

Asset Allocation for State and Local Pension Plans, 2016

42.3% 48.5% 51.9% 43.8% 22.7% 23.5% 22.3% 22.7% 33.2% 26.5% 23.5% 31.9% 0% 25% 50% 75% 100% Bottom Second Third Top Quartile Alternatives Fixed income Equities

Source: Jean-Pierre Aubry, Anqi Chen, Alicia H. Munnell, and Kevin Wandrei. 2018. “What Explains Differences in Public Pension Returns since 2001?” State and Local Plans Issue in Brief 60. Chestnut Hill, MA: Center for Retirement Research at Boston College.

But the top-quartile plans outperformed

• thers in most asset classes.

5

Annualized Asset Class Returns by Quartile, 2001-2016

Asset Class Top Third Second Bottom Public Equities 6.7% 5.3% 5.4% 4.7% Fixed Income 6.3% 6.3% 5.8% 5.8% Alternatives Private Equity 9.7% 8.9% 7.1% 9.3% Hedge Funds 4.1% 5.6% 7.5% 6.2% Real Estate 10.1% 8.8% 8.4% 7.2% Commodities 8.1% 3.1% 0.2% 3.9%

Source: Jean-Pierre Aubry, Anqi Chen, Alicia H. Munnell, and Kevin Wandrei. 2018. “What Explains Differences in Public Pension Returns since 2001?” State and Local Plans Issue in Brief 60. Chestnut Hill, MA: Center for Retirement Research at Boston College.

So, for most plans, asset class returns - not allocation - explain the difference from the top quartile.

6

Role of Allocation and Returns on the Difference from Top Quartile

0.38%

• 0.10%
• 0.12%

0.04% 1.16% 1.05% 0.68% 0.93% 1.54% 0.96% 0.57% 0.98%

• 1%

0% 1% 2% Bottom Second Third Average Allocation Asset class returns Total

Source: Jean-Pierre Aubry, Anqi Chen, Alicia H. Munnell, and Kevin Wandrei. 2018. “What Explains Differences in Public Pension Returns since 2001?” State and Local Plans Issue in Brief 60. Chestnut Hill, MA: Center for Retirement Research at Boston College.

In general, plans have shifted away from traditional stocks and bonds to alternatives.

7

Asset Allocation for State and Local Pension Plans, 2001-2015

0% 20% 40% 60% 80% 100% 2001 2003 2005 2007 2009 2011 2013 2015 Traditional equities Fixed income & cash Alternatives

Source: Jean-Pierre Aubry, Anqi Chen, and Alicia H. Munnell. 2017. “A First Look at Alternative Investments and Public Pensions.” State and Local Plans Issue in Brief 55. Chestnut Hill, MA: Center for Retirement Research at Boston College.

All plans have made the shift away from traditional bonds in relatively similar fashion.

8

Allocation to Fixed Income by Quartile of Returns, 2001-2016

35% 29% 23% 20% 30% 40% 50% 2001 2003 2005 2007 2009 2011 2013 2015 Bottom quartile Second quartile Third quartile Top quartile

Source: Jean-Pierre Aubry, Anqi Chen, Alicia H. Munnell, and Kevin Wandrei. 2018. “What Explains Differences in Public Pension Returns since 2001?” State and Local Plans Issue in Brief 60. Chestnut Hill, MA: Center for Retirement Research at Boston College.

However, after the crises, bottom quartile plans made the largest shift out of equities….

9

Allocation to Traditional Equities by Quartile of Returns, 2001-2016

59% 42% 52% 54% 40% 50% 60% 70% 2001 2003 2005 2007 2009 2011 2013 2015 Bottom quartile Second quartile Third quartile Top quartile

Source: Jean-Pierre Aubry, Anqi Chen, Alicia H. Munnell, and Kevin Wandrei. 2018. “What Explains Differences in Public Pension Returns since 2001?” State and Local Plans Issue in Brief 60. Chestnut Hill, MA: Center for Retirement Research at Boston College.

…and into alternatives.

10

Allocation to Alternatives by Quartile of Returns, 2001-2016

33% 6% 24% 12% 0% 10% 20% 30% 40% 2001 2003 2005 2007 2009 2011 2013 2015 Bottom quartile Second quartile Third quartile Top quartile

Source: Jean-Pierre Aubry, Anqi Chen, Alicia H. Munnell, and Kevin Wandrei. 2018. “What Explains Differences in Public Pension Returns since 2001?” State and Local Plans Issue in Brief 60. Chestnut Hill, MA: Center for Retirement Research at Boston College.

Specifically, they shifted more heavily into hedge funds and commodities…

11

Percentage of Plan Holdings in Selected Alternative Asset Classes by Quartile of Returns, 2016

9.1% 8.6% 7.9% 10.4% 8.3% 10.6% 7.0% 8.9% 10.2% 6.1% 4.9% 8.7% 5.1% 2.0% 2.1% 3.1% 1.6% 0.9% 0% 10% 20% 30% 40% Bottom Second Third Top Quartile Other Commodities Hedge funds Real estate Private equity

Source: Jean-Pierre Aubry, Anqi Chen, Alicia H. Munnell, and Kevin Wandrei. 2018. “What Explains Differences in Public Pension Returns since 2001?” State and Local Plans Issue in Brief 60. Chestnut Hill, MA: Center for Retirement Research at Boston College.

..during a period when these asset classes dramatically underperformed others.

12

Returns from Alternative Asset Classes and Traditional Equities, FY 2001-2016

Asset class 2000-2007 2008-2009 2010-2016 Alternatives Private equity (before fees) 8.1%

• 13.0%

25.0% Hedge funds (after fees) 10.7%

• 10.9%

1.3% Real estate (before fees) 14.5%

• 6.3%

12.1% Commodities (after fees) 16.2%

• 4.1%
• 3.0%

Traditional equity 2.7%

• 21.3%

14.9%

Source: Jean-Pierre Aubry, Anqi Chen, Alicia H. Munnell. 2017. “A First Look at Alternative Investments and Public Pensions.” State and Local Plans Issue in Brief 55. Chestnut Hill, MA: Center for Retirement Research at Boston College. Note: Returns based on Thomson Reuters Private Equity Buyout Index, Hedge Fund Research Global Hedge Fund Index, NCREIF Property Index, S&P GSCI Index, and Wilshire 5000 Index (Total Return).

As a result, allocation played some role in the lower returns of the worst-performing plans

13

Role of Allocation and Returns on the Difference from Top Quartile

Source: Jean-Pierre Aubry, Anqi Chen, Alicia H. Munnell, and Kevin Wandrei. 2018. “What Explains Differences in Public Pension Returns since 2001?” State and Local Plans Issue in Brief 60. Chestnut Hill, MA: Center for Retirement Research at Boston College.

0.38%

• 0.10%
• 0.12%

0.04% 1.16% 1.05% 0.68% 0.93% 1.54% 0.96% 0.57% 0.98%

• 1%

0% 1% 2% Bottom Second Third Average Allocation Asset class returns Total

CRR assesses plan performance in two ways.

14

1. A comparison of investment returns across plans: Observed differences are the result of both differences in asset allocation and/or asset class performance. 2. A comparison of each plan’s investment return to its

• wn benchmark:

Performance relative to benchmark focuses on each plan’s ability to execute its own strategy.

Most plans beat their benchmark for traditional investments, but only about half beat their benchmark for alternatives.

15

Percentage of Plans that Outperformed Their Asset-Class Benchmark from 2001-2016

Source: Author's calculations using the Public Plans Database (2001-2016).

72% 92% 55% 0% 20% 40% 60% 80% 100% Equities Fixed Income Alternatives

Currently, the portfolio benchmark for most plans reflects the plan’s asset allocation.

16

Distribution of Plans, by Type of Portfolio Benchmark, 2016

Source: Jean-Pierre Aubry and Caroline V. Crawford. “How Do Fees Affect Plans’ Ability to Beat Their Benchmarks?” State and Local Plans Issue in Brief 61. Chestnut Hill, MA: Center for Retirement Research at Boston College.

Weighted avg.

• f asset class

benchmarks, 65% Peer universe performance, 19% Expected ROR, 7% Index (+premium), 9%

About a third of plans did not meet their portfolio benchmark over the long term.

17

Distribution of the Gap between Portfolio Performance and Benchmark from 2002-2016

Underperformed, 32% 0-50 bps, 31% > 50 bps, 35%

Source: Jean-Pierre Aubry and Caroline V. Crawford. “How Do Fees Affect Plans’ Ability to Beat Their Benchmarks?” State and Local Plans Issue in Brief 61. Chestnut Hill, MA: Center for Retirement Research at Boston College.

Plans that fell short of their benchmark were more likely to be bottom-quartile plans.

18

Percentage of Plans that Were in the Bottom Quartile, by Performance Relative to Benchmark

38% 22% 0% 10% 20% 30% 40% 50% Underperformed Overperformed

Source: Author's calculations using the Public Plans Database (2001-2016).

What about fees?

19

The data suggest that fees have a limited role in the relative performance of plans.

20

Quartile Ranking by Gross Returns Compared to Quartile Ranking by Net-of-fee Returns

16% 14% 70% 0% 20% 40% 60% 80% 100% Higher gross return ranking Lower gross return ranking No difference in ranking Percent of Plans

Source: Author's calculations using the Public Plans Database (2001-2016).

Plans that fell short of their benchmark did pay higher fees across all asset classes.

21

Average Expense Ratios from 2011-2016, by Plan Performance Relative to Benchmark

30 20 93 31 25 111 40 80 120 Equities Fixed income Alternatives Basis points Outperformed Underperformed

Source: Jean-Pierre Aubry and Caroline V. Crawford. “How Do Fees Affect Plans’ Ability to Beat Their Benchmarks?” State and Local Plans Issue in Brief 61. Chestnut Hill, MA: Center for Retirement Research at Boston College.

But dramatic fee cuts would have been required to help most underperformers meet their benchmark.

22

Percentage Reduction in Fees Required to Achieve Benchmark Returns

11% 6% 11% 44% 28% 0% 10% 20% 30% 40% 50% 0-15% 15-33.3% 33.3-50% 50-100% Cannot close gap with fees Percent of Underperforming Plans Percentage reduction in fees that is necessary to meet benchmark

Source: Author's calculations using the Public Plans Database (2001-2016).

Conclusion

23

• The observed differences in long-term investment performance

among plans are meaningful.

• For most, the difference is due to asset class returns. But, for the

worst-performing plans, allocation to hedge funds and commodities has played a role.

• While most plans outperform their benchmarks, plans that

underperformed were more likely to have bottom-quartile investment returns.

• Plans that underperformed their benchmark also paid higher fees

(although, in many cases, moderate fee reduction would not have resulted in outperformance of their benchmark).

Appendix

24

1. A simpler allocation: 60% Wilshire 5000 + 40% Barclay’s 2. Use of leverage by public pension plans 3. Fair value of investments 4. Unfunded commitments to alternative investment funds

The benefits of a simpler investment approach depend on the period in question.

25

Percentage of Plans that Outperformed a Simple 60/40 Stock/Bond Portfolio

Source: Public Plans Database (2001-2016). .

95% 56% 0% 20% 40% 60% 80% 100% 2001-2017 2010-2017 Percent of Plans

SERS and PSERS have underperformed a simpler portfolio recently, but outperformed

• ver the long-term.

26

Plan Returns Relative to a Simple 60/40 Stock/Bond Portfolio

0.5%

• 0.3%

0.9%

• 0.4%
• 1.0%
• 0.5%

0.0% 0.5% 1.0% 1.5% 2001-2017 2010-2017 Return Relative to 60/40 Portfolio Pennsylvania SERS Pennsylvania Schools

Source: Author's calculations using the Public Plans Database (2001-2016), Dow Jones Wilshire 5000, and Barclays US Aggregate Bond Index.

The explicit use of leverage is rare among public plans.

27

Major State and Local Plans that Report an Explicit Use of Leverage, 2017

Plan Name Type of Leverage Percentage of Portfolio That is Levered Massachusetts SRS Uses leverage for real estate investments 1.7% Massachusetts Teachers Uses leverage for real estate investments 1.7% Missouri State Employees Uses leverage to achieve a beta balanced portfolio 52.1% Ohio Police & Fire Policy to leverage fixed income portfolio 2x 20.0% San Francisco City & County Uses leverage for real estate investments 0.0% Virginia RS Uses leverage in its real assets portfolio 3.6% Wisconsin RS Policy to leverage in fixed income portfolio 10.0% Sacramento County ERS Uses leverage for real assets portfolio 0.8% San Diego City ERS Uses leverage for real estate investments 1.8% Pennsylvania PSERS Uses leverage in fixed income portfolio 17.30%

Source: Author's calculations using the Public Plans Database (2001-2016).

The majority of pension plan assets are classified as Level 1 and/or valued at NAV.

28

Source: Author's calculations using the Public Plans Database (2001-2016).

Percent of Assets, by Fair Value Measurement, 2017

42% 36% 53% 19% 7% 16% 5% 1% 5% 35% 57% 26% 0% 20% 40% 60% 80% 100% PPD Average PSERS SERS Percent of Assets Level 1 Level 2 Level 3 NAV

Future capital calls may limit the investment flexibility of plans.

29

Source: Author's calculations using the Public Plans Database (2001-2016).

Distribution of Plans by Unfunded Commitments as a Percent of Assets, 2017

10% 7% 10% 9% 20% 12% 8% 6% 4% 4% 3% 1% 1% 5% 0% 5% 10% 15% 20% 25% Percent of Plans Unfunded Commitments as a Percent of Assets

PA SERS - 11.8% PA Schools - 15.7%