2

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, ...

0 downloads 138 Views 1MB Size
Alameda-Contra Costa Transit District Retirement Board

Actuarial Experience Study January 1, 2007 through December 31, 2010

Karen T. Earley, FSA (484) 444-0271

Robert T. McCrory, FSA (206) 328-8628

Prepared August 2, 2011

EFI ACTUARIES | EFI ASSET/LIABILITY MANAGEMENT SERVICES, INC. The nation’s leader in plan-specific, interactive asset allocation optimization counseling WASHINGTON, DC

¤

PHILADELPHIA

¤

SEATTLE

¤

SAN FRANCISCO

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

Contents

Longevity and Promotion Pay Increases (AFSCME/NonRepresented)

20

1

Mortality (ATU / IBEW – Non-Disabled)

21

Purpose

1

Mortality (AFSCME / Non-Represented – Non-Disabled)

23

Retirement Rates

1

Mortality (All Groups – Disabled)

25

Termination Rates

1

Economic Assumptions

26

Disability Rates

2

Introduction

26

Longevity and Promotion Pay Increases

2

Inflation

26

Mortality Rates

3

Investment Return

27

Economic Assumptions

3

Payroll Growth

30

Summary of Experience and Impact on Plan Costs

3

Prior Experience Studies

4

Purposes of the Experience Study

31

Organization of Report

4

Importance of Reliable Assumptions

31

6

Prior Experience Studies

31

6

Scope of Report

32

Executive Summary

Active Decrements Service Retirement (ATU/IBEW)

Methodology

31

Service Retirement (AFSCME / Non-Represented)

10

Methodology (Demographic Assumptions)

33

Termination (ATU/IBEW)

12

Methodology (Economic Assumptions)

34

Termination (AFSCME/Non-represented)

14

Disability (ATU/IBEW)

15

Disability (AFSCME/Non-Represented)

17

Longevity and Promotion Pay Increases (ATU/IBEW)

18

ii

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

Executive Summary Purpose The purpose of this Actuarial Experience Study is to review the actuarial experience of the AC Transit Employee’s Retirement Plan (the Plan) during the period from January 1, 2007 through December 31, 2010. The Plan’s demographic experience – observed rates of retirement, withdrawal, termination, disability, and death – is compared with the experience expected under the actuarial assumptions currently used to determine Plan liabilities and cost, and revised assumptions are recommended as appropriate. In addition, the plan’s economic assumptions were reviewed. The economic assumptions include the assumed rates of inflation, investment return, and active payroll growth. The purpose of this Section of the Study is to give the reader a quick summary of the major conclusions that have been reached. Details are presented in later sections of this Report.

Retirement Rates Over the past four years, actual rates of retirement have been significantly lower than current actuarial assumptions would predict for the ATU/IBEW male members, while actual rates of retirement for the ATU/IBEW females have been slightly higher than expected.

Actual rates of retirement for AFSCME/Non-Represented members were slightly lower than expected in total, with more retirements than expected at ages 55-59 and fewer retirements than expected at ages 65-69. New sets of retirement rates are proposed for both groups, bringing assumptions into line with experience. For the ATU/IBEW members, we propose the use of separate rates for males and females. There are no changes proposed to the current retirement rates for females. For ATU/IBEW males, the retirement rates have been reduced at ages 55-61 and 65-69. The proposed AFSCME/Non-Represented rates bring the age distribution of the expected retirements more in line with actual experience, reflecting an increase in expected retirements at ages 55-59 and a decrease in expected retirements at ages 65-69.

Termination Rates Termination rates measure the frequency with which members terminate employment for reasons other than retirement, death or disability. In total, there were fewer terminations than expected among ATU/IBEW members, with more terminations than expected at very low levels of service (0-1 years) and fewer terminations at higher service levels. Service-based termination rates were adopted for the first time in the prior study, based on findings that members with less service terminated at much higher rates. Experience from the past four years confirms that this pattern is

1

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

still evident. Therefore, only minor changes to the current rates are proposed to fine-tune this assumption. The termination experience for the AFSCME/Non-Represented members was close to expected, in both the number of terminations and their age/service distribution. Therefore, no changes are proposed to this assumption.

Disability Rates Among ATU/IBEW members, the number of actual disabilities has been lower than the number expected. Experience over the past four years was aggregated with that of the prior experience study to improve the credibility of the data. Based on the combined experience of the last seven years, we have proposed a 20% decrease in the disability rates for all ages, for both males and females. The rates of disability observed during this Study among AFSCME/Non-Represented members were in reasonable agreement with the expected rates. In addition, the amount of data is extremely limited, even after combining the experience from the prior study. Therefore, no changes have been proposed.

Longevity and Promotion Pay Increases The current actuarial assumption for ATU/IBEW members is that the pay of active members will increase by 3.2% per year from inflation, and an additional 0.50% to 6.00% for merit, longevity and promotion, depending on the service of the member (lower increases after eight years of service).

An analysis of the average pay of active ATU/IBEW members by service reveals that this general pattern of increases is still very much appropriate: Pay increases are steeper in the early years of employment. However, the data shows that the magnitude of these increases has gradually been declining over recent years. This is not surprising based on the current economic environment. Valuation assumptions for merit, promotion and longevity are based on long-term expectations of future pay increases; therefore it would not be appropriate to change the assumption based solely on the experience of the past few years. However, market expectations indicate that pay increases will likely level off at a rate that is lower than the current assumption. Therefore, new merit increases are proposed for ATU/IBEW that reflect a lower level of pay increases for the first seven years of employment. No changes are proposed to the assumed increases for those with more than eight years of service. For AFSCME/Non-Represented members, the current assumption is that merit and longevity increases will equal 0.50% during all years of service. In this case, these merit and longevity increases appear to have provided a reasonable fit to actual experience. In addition, the pattern of longevity and promotion pay increases is much more difficult to discern from the data for these smaller groups; therefore no changes to the rates are proposed.

2

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

Mortality Rates Mortality experience among active and retired male members and surviving spouses in this Study aggregated with the prior three years have indicated that mortality rates are slowly improving and members are living longer. This gradual increase in mortality is expected to continue into the future. Further, a recent study by the Society of Actuaries discovered that members with higher benefit amounts have lower rates of death than members with lower benefits. The combination of these factors points to the need for updated mortality tables for all groups. Therefore, we have proposed the use of the Retired Pensioners (RP) 2000 tables for all groups, with age adjustments applied to ATU/IBEW members. No changes are proposed to the rates for disabled members.

Economic Assumptions A review of the Plan’s economic assumptions based on the allocation of Plan assets and the history of the financial markets indicates that the current economic assumptions of a nominal 7.7% annual rate of return and a 3.2% annual rate of inflation continue to be reasonable expectations. However, we have performed additional analysis based on the near- to mid-term future expectations of the Plan’s investment consultant, as well as using the expectations of other investment consultants. We have also reviewed market expectations for inflation as revealed in the Inflation Curve published by the Federal

Reserve Bank of Cleveland, and we are familiar with the economic assumptions being adopted by pension plans nationwide. Based on this evidence, we believe the Retirement Board should consider reducing the assumed inflation rate from 3.2% to 3%, and similarly reducing the nominal return assumption from 7.7% to 7.5%.

Summary of Experience and Impact on Plan Costs In the table shown on page 6, we present a summary of experience, the new proposed assumptions where applicable, and the expected impact of the proposed assumption changes on the overall Plan cost as a percentage of payroll, as of January 1, 2011. Should all of the recommendations in this Report be adopted, an increase in the total actuarial cost of approximately 0.24% of payroll would be applied, which amounts to about $300,000 for the next fiscal year. This increase is primarily due to an increase in the unfunded actuarial accrued liability associated with the proposed change in economic assumptions. Updated administrative tables may result from adoption of new assumptions as well. This issue should be addressed after changes are adopted and implemented. The assumptions will not determine the ultimate level of employer contributions; instead, the required contributions will depend on the actual demographic and financial experience of the Plan. The goal of an experience study is to make our best

3

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

estimate of future conditions so that the Plan costs computed by the actuary will be as predictable as possible.

Prior Experience Studies This is the sixth experience study performed by EFI for the Plan. The fifth experience study performed by EFI covered calendar years 2004-2006. That study, dated August 14 2007, served as the basis for actuarial assumptions used in the actuarial valuations of the Retirement Plan as of January 1, 2007, 2008, 2009 and 2010. A complete history of prior experience studies for the Plan is included in the Methodology section of this Report.

Organization of Report The first section of the Report deals with decrements among active members and also includes consideration of the merit component of pay increases. The second section of the Report deals with mortality among active and inactive members. The third section of the Report concerns economic assumptions, and a final section presents methodological details. The report has been prepared in accordance with generally accepted actuarial methods and procedures. EFI will be happy to answer any questions from AC Transit Board or staff regarding its methodology or conclusions. Karen T. Earley, FSA (484) 444-0271

Robert T. McCrory, FSA (206) 328-8628

4

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

Assumption Service Retirement

Experience Significantly fewer retirements than expected for ATU/IBEW males. Retirements were higher than expected for AFSCME/Non-rep between ages 65-69, and less than expected between ages 55-59.

Termination

Termination experience among AFSCME/Nonrepresentative members was close to expected. Termination experience for ATU/IBEW members was less than expected, with more terminations than expected with low service. Fewer than expected disabilities occurred for ATU/IBEW. Disabilities were close to expected for AFSCME/Non-representative members. Fewer deaths than expected for ATU/IBEW, especially when measured on benefit-weighted basis. More deaths than expected for Clerical/Non-Contract members. Review of actual longevity and promotion increases for AFSCME/Non-representative members show gradual, steady increases over a participant’s career. Increases for ATU/IBEW show a pattern of steeper increases in early years, with gradual, steady increases in later years.

Disability

Mortality

Longevity and Promotion Pay Increases

Economic Assumptions

Current inflation assumption (3.20%) is high relative to market expectations. Return assumption (7.70%) is reasonable, but many plans are considering reducing to 7.50%.

Recommendation New sets of retirement rates are proposed for ATU/IBEW males. No changes are proposed for ATU/IBEW females. New rates are also proposed for AFSCME and Non-rep members for ages 55-59 and 65-69, bringing assumptions more in line with recent experience. Propose new termination rates for ATU and IBEW members that more closely match the age and service profile of recent terminations. No changes proposed to current assumptions for AFSCME/Non-representative members. Reduce rates by 20% for ATU/IBEW members. No changes proposed to current disability assumption for AFSCME/Non-representative members. Propose use of RP-2000 tables for all non-disabled groups, with age adjustments for ATU/IBEW members. No change in current assumption is proposed for disabled members. Propose new rates for ATU/IBEW to reflect slightly lower increases in early years. No changes are proposed for AFSCME/Non-representative members.

Reduce inflation and pay growth assumption to 3.00%. Reduce return assumption to 7.50%.

Total Change in Contribution Rate

5

Plan Cost $(698,806) (0.65)%

65,194 0.03%

(105,818) (0.10)% 206,342 0.15%

(216,463) (0.09)%

1,059,443 0.90% $309,892 0.24%

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

6

Active Decrements Service Retirement (ATU/IBEW) Current Assumption Summary of Experience versus Current Assumptions (Ages 55-70) Eligible Exposure

Actual Retirements

Expected Retirements

1,383

148

221.3

66.9%

523

81

66.3

122.2%

1,906

229

287.6

79.6%

Actual Average Age

Expected Average Age

Males

62.2

62.0

Females

59.1

59.7

Combined

61.1

61.5

Males Females Combined

Proposed Assumption Summary of Experience versus Proposed Assumptions (Ages 55-70)

Actual to Expected Ratio

Eligible Exposure

Actual Retirements

Expected Retirements

Actual to Expected Ratio

1,383

148

176.9

83.4%

523

81

66.3

122.2%

1,906

229

243.2

94.2%

Actual Average Age

Expected Average Age

Males

62.2

62.4

Females

59.1

59.7

Combined

61.1

61.6

Males Females Combined



ATU members are currently eligible to retire at age 55 with 8 years of  service and IBEW members are eligible to retire at age 55 with 5 years of service.

Somewhat modified rates are proposed for males, with slightly lower (and simpler) rates proposed. No changes are proposed to the rates for females.



Experience for IBEW is fairly limited. The two groups were combined  to improve the credibility of the data.



Prior to June 30, 2008, ATU members were eligible for service retirement upon attaining age 55 and completing 10 years of service. There is not yet enough data to determine if this change has affected retirement patterns. We will monitor experience closely over the  next few years to see if a new pattern emerges.

The expected retirements under the new assumptions are still higher than the actual number of retirements for males. This margin allows for the possibility of an increase in retirements due to the new eligibility requirements, and it makes an allowance for the increased retirement rates reported in early 2011.



There were approximately 20% fewer retirements than expected in  total. However, female retirements were actually higher than expected, while male retirements were significantly lower than expected.

We have confirmed that in the prior study, the number of male retirements was also overstated using the current assumptions. The expected average retirement age under the new assumptions remains close to the actual retirement age for males.

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010



Retirement rates were studied separately for each year in the Study. There was a surge of retirements in 2010, with retirement rates about 50% higher than the average during the term of the Study. This increased rate continued into early 2011 according to staff, and has since subsided. The overall average retirement rate was not materially affected.

7

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

Service Retirement Rates (ATU/IBEW) – Current (Male and Female)

*

Service Retirement Rates (ATU/IBEW) – Proposed (Male only)

Age *

Rate

Age

Rate

55

11.7%

55

7.0%

56

10.4%

56

7.0%

57

9.1%

57

7.0%

58

9.1%

58

7.0%

59

7.8%

59

7.0%

60

15.0%

60

9.0%

61

15.0%

61

9.0%

62

30.0%

62

30.0%

63

15.0%

63

15.0%

64

15.0%

64

15.0%

65

54.0%

65

45.0%

66

40.0%

66

30.0%

67

40.0%

67

30.0%

68

40.0%

68

30.0%

69

40.0%

69

30.0%

70+

100.0%

70+

100.0%

With 8+ Years of Service for ATU members, 5+ Years of Service for IBEW members

8

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

Chart 1: Comparison of Retirement Rates, ATU/IBEW

Chart 2: Comparison of Retirements, ATU/IBEW 120

Expected Retirements Actual Retirements

100

Proposed Retirements 80 60 40

20 0 55 - 59

60 - 64 AGE

65 - 69

9

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

10

Service Retirement (AFSCME / Non-Represented) Current Assumption

Proposed Assumption

Summary of Experience versus Current Assumptions (Ages 50-70)

Summary of Experience versus Proposed Assumptions (Ages 50-70)

Eligible Exposure

Expected Retirements

Actual to Expected Ratio

Eligible Exposure

Actual Retirements

Expected Retirements

Actual to Expected Ratio

Males

360

49

50.7

96.6%

Males

360

49

46.6

105.2%

Females

232

26

27.7

93.9%

Females

232

26

27.7

93.9%

Combined

592

75

78.4

95.7%

Combined

592

75

74.3

100.9%

Actual Average Age

Expected Average Age

Males

61.2

62.4

Females

60.5

61.5

Combined

61.0

62.1

f



Actual Retirements

Experience for AFSCME and Non-Represented members is fairly limited. The two groups were combined to improve the credibility of  the data.



AFSCME members are currently eligible to retire at age 55 with 5 years of service and Non-represented members are eligible to retire at age 50 with 5 years of service.



Retirement experience for AFSCME and Non-Represented members was very close to expected over the four year period.



Retirement eligibility for Non-Represented members was changed to age 50 with 5 years of service as of July 1, 2005. Experience for pre55 retirements was not included in the previous study. The current assumption of 3% per year for ages 50-54 appears to be reasonable based on the retirement experience from 2007-2010.

Actual Average Age

Expected Average Age

Males

61.2

61.9

Females

60.5

60.8

Combined

61.0

61.5

New assumptions are proposed at ages 55-59 and 65-69. This brings the age distribution of the expected retirements more in line with actual experience, as shown in the graphs on the next page. In particular, experience over the current and prior studies does not support a spike in the retirement rate at age 65, so the rate at that age was reduced from 50% to 25%.

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

Chart 3: Comparison of Retirement Rates, AFSCME/Non-Represented

Chart 4: Comparison of Retirements, AFSCME/Non-Represented 35

Expected Retirements 30 25

Actual Retirements Proposed Retirements

20 15 10 5 0 50 - 54

55 - 59

60 - 64

AGE

Age*

Rate

Age*

Rate

50-54

3.0%

50-54

3.0%

55-59

8.0%

55-59

10.0%

60-64

15.0%

60-64

15.0%

65

50.0%

65

25.0%

66-69

26.3%

66-69

25.0%

70+

100.0%

70+

100.0%

* Assuming service requirement is met. Rates at ages 50-54 are only applied to the Non-Represented members.

11

65 - 70

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

12

Termination (ATU/IBEW) Current Assumption

Proposed Assumption

Summary of Experience versus Current Assumptions

Summary of Experience versus Proposed Assumptions

Eligible Exposure

Actual Withdrawals

Expected Withdrawals

Actual to Expected Ratio

Males

3,104

80

99.1

80.7%

Females

2,279

79

77.6

101.8%

Combined

5,383

159

176.7

90.0%

Actual Average Age

Expected Average Age

Actual Avg. Service

Expected Avg. Service

Males

42.8

43.5

5.9

6.7

Females

38.5

40.9

5.1

Combined

40.7

42.4

5.5

Actual Withdrawals

Expected Withdrawals

Actual to Expected Ratio

Males

3,104

80

94.2

84.9%

Females

2,279

79

73.2

107.9%

Combined

5,383

159

167.4

95.0%

Actual Average Age

Expected Average Age

Actual Avg. Service

Expected Avg. Service

Males

42.8

43.3

5.9

6.5

6.2

Females

38.5

40.8

5.1

6.0

6.5

Combined

40.7

42.2

5.5

6.3



Termination rates reflect the frequency with which active members  leave employment for reasons other than retirement, death or disability. Currently, service-based termination rates are assumed. A single table is used for both sexes.



The analysis performed in the prior study found that termination rates for ATU/IBEW are strongly related to service, decreasing as  service increases. This pattern is still evident, as can be seen in the chart on the following page.  The current assumptions underestimated the number of terminations by about 10%.



Eligible Exposure

A new set of termination rates is proposed, with slightly higher rates for those with less than two years of service, and lower rates for those with two to four years of service. Service-based rates remain appropriate based on the pattern of correlation between the amount of service and likelihood of termination. The proposed assumptions bring the experience closer to the current rates. Currently, a single table is used for males and females. The data from the last four years indicates that experience for males and females may be beginning to diverge and follow different patterns. We are not proposing separate tables at this time; however, we will monitor experience to see if this pattern continues.

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

Chart 5: Comparison of Termination Rates, ATU/IBEW

13

Chart 6: Comparison of Terminations, ATU/IBEW 80 70 Expected Terminations 60

Actual Terminations Proposed Terminations

50

40 30 20 10 0 0-1

Service

All Ages

0

8.0%

1

8.0%

2

7.0%

3

6.0%

4

5.0%

5–9

3.0%

10+

1.5%

2-3

4-9 SERVICE

10 - 19

Service

All Ages

0–1

10.0%

2–3

5.0%

4–8

3.0%

9+

1.5%

20 - 34

No terminations are assumed for participants who are eligible for retirement.

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

14

Termination (AFSCME/Non-represented) Current Assumption

Proposed Assumption

Summary of Experience versus Current Assumptions (2004-2010)



Eligible Exposure

Actual Withdrawals

Expected Withdrawals

Actual to Expected Ratio

2004-2006

580

21

23.2

90.7%

2007-2010

680

33

28.5

115.8%

Combined

1,260

54

51.7

104.4%

Actual Average Age

Expected Average Age

2004-2006

42.7

45.6

2007-2010

44.0

44.2

Combined

43.5

44.7



The currently assumed termination rates for AFSCME and NonRepresented members vary by service as shown in the table below. A single table is used for males and females.



The amount of termination experience is fairly limited; only 33 terminations have occurred amongst the AFSCME and Nonrepresented members over the past four years. To improve the credibility of the data, we have aggregated the experience of the past three years with that of the prior experience study (2004-2006).



The number of actual terminations was close to expected over the seven year period. The average age and service profile of the terminations was also reasonably close to expected.

No changes in this assumption are proposed at this time.

Current Termination Rates (AFSCME/Non-Represented) Service

All Ages

0

12.0%

1

7.0%

2-9

5.0%

10+

1.5%

No withdrawals are assumed for participants who are eligible for retirement.

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

15

Disability (ATU/IBEW) Current Assumption

Proposed Assumption

Summary of Experience versus Current Assumptions (2004-2010)

Summary of Experience versus Proposed Assumptions (2004-2010)

Eligible Exposure

Actual Disabilities

Expected Disabilities

Actual to Expected Ratio

Eligible Exposure

Actual Disabilities

Expected Disabilities

Actual to Expected Ratio

Males

4,584

23

32.6

70.6%

Males

4,584

23

26.0

88.5%

Females

2,324

36

41.4

87.0%

Females

2,324

36

33.1

108.8%

Combined

6,908

59

74.0

79.7%

Combined

6,908

59

59.1

99.8%

Actual Average Age

Expected Average Age

Actual Average Age

Expected Average Age

Males

50.6

51.2

Males

50.6

51.1

Females

50.6

50.3

Females

50.6

50.3

Combined

50.6

50.7

Combined

50.6

50.7



Currently, disability rates are based on age, with separate tables used for males and females. The disability decrement is applied only after eligibility for disability benefits.





Over the aggregated period, the number of disabilities occurring was less than the number assumed for both males and females; therefore we are recommending a 20% decrease in the disability rates for all ages.

The amount of disability experience is fairly limited; only 31  disabilities have occurred amongst the ATU and IBEW members over the past four years. To improve the credibility of the data, we have aggregated the experience of the past four years with that of the  prior experience study (2004-2006).

Actual average disability age is very close to the expected age under current assumptions.



One-third of future disabilities are assumed to be total and permanent disabilities, with the remaining two-thirds assumed to be occupational



The disability decrements are not applied to members eligible to retire.

Over the aggregated period, approximately 27% of the disabilities that have occurred have been total and permanent disabilities; therefore no change to the assumption that one-third of future disabilities will be total and permanent is recommended.

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

Current Representative Assumed Rates (ATU/IBEW)

Proposed Representative Assumed Rates (ATU/IBEW)

Age

Male

Female

Age

Male

Female

27

0.000%

0.000%

27

0.000%

0.000%

32

0.195%

0.439%

32

0.156%

0.351%

37

0.330%

0.743%

37

0.264%

0.594%

42

0.570%

1.283%

42

0.456%

1.026%

47

0.900%

2.025%

47

0.720%

1.620%

52

1.275%

2.869%

52

1.020%

2.295%

57

0.375%

0.844%

57

0.300%

0.675%

62

0.450%

1.013%

62

0.360%

0.810%

16

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

17

Disability (AFSCME/Non-Represented) Current Assumption

Proposed Assumption

Summary of Experience versus Current Assumptions (2004-2010) Eligible Exposure





Expected Disabilities

Actual to Expected Ratio

Males

775

0

1.5

0.0%

Females

496

2

1.8

111.1%

1,271

2

3.3

60.6%

Combined



Actual Disabilities

Experience from the prior study has been aggregated with the current data to improve the adequacy of the exposures. The analysis now includes seven years of experience: 2004-2010. Currently, disability rates are based on age, with separate tables used for males and females. The disability decrement is applied only after eligibility for disability benefits. Only two disabilities occurred over the seven year period. One of these disabilities was a total and permanent disability, and the other was an occupational disability.



We do not recommend any changes to this assumption. Considering the small number of disabilities, the actual disability experience is reasonably close to expected.

Current Representative Assumed Rates (AFSCME/Non-Represented) Age

Male

Female

27

0.000%

0.000%

32

0.098%

0.146%

37

0.165%

0.248%

42

0.285%

0.428%

47

0.450%

0.675%

52

1.638%

0.956%

57

0.047%

0.070%

62

0.056%

0.084%

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

18

Longevity and Promotion Pay Increases (ATU/IBEW) Salary increases consist of three components: Increases due to cost of living maintenance (inflation), increases related to non-inflationary pressures on base pay (such as productivity increases), and increases in individual pay due to merit, promotion, and longevity. Increases due to cost of living and non-inflationary base pay factors are addressed in a later section of this report. Each of the following pages contains a description of the current assumptions and proposed changes, as well as a graph of how the proposed rates compare to the current pay data. Only increases due to merit (promotion and longevity) are considered here. In the graphs, the average pay of the active members of the bargaining unit as of January 1, 2011 is plotted against service. A curve is then fitted to the average pay data, and this curve is used to determine a pay increase due to merit. While some units have a wide spread of salaries by each service level, we are trying to capture the average increase to be able to project future pay based benefits. This is a transverse study of longevity and promotion pay increases: Salaries are examined at one point in time (the valuation date), as opposed to being observed over a number of years (a longitudinal study). For a more detailed description of this type of study and its advantages, see the Methodology section at the end of this report.

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

Current Assumption



19

Proposed Assumption

Years of Service

Assumed Increase

Years of Service

Assumed Increase

Less than 8

6.00%

Less than 8

4.00%

8+

0.50%

8+

0.50%

Only increases due to merit (promotion and longevity) are  considered here.

Although the current assumptions fit the pattern of merit pay increases reasonably well, the data indicates that members are likely to receive smaller pay increases than previously assumed for the first eight years.



Therefore, new merit increases are proposed which contain reduced merit increases for the first seven years of employment.



The proposed assumptions fit the average pay curve better than the current rates.

ATU/IBEW Members Average Pay vs. Years of Service

ATU/IBEW Members Average Pay vs. Years of Service $90,000

$90,000

$80,000

$80,000

$70,000

$70,000

$60,000

$60,000 Actual

$50,000

Actual

$50,000

4.0%(8), 0.5%->

6.0%(8), 0.5%->

$40,000

$40,000

$30,000

$30,000

$20,000

$20,000

$10,000

$10,000

$0

$0 0

5

10

15

20 Years of Service

25

30

35

0

5

10

15

20 Years of Service

25

30

35

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

20

Longevity and Promotion Pay Increases (AFSCME/Non-Represented) Current Assumption AFSCME/Non-Represented Members Average Pay vs. Years of Service

Years of Service

Assumed Increase

All Years

0.50%

$140,000

$120,000





Only increases due to merit (promotion and longevity) are considered here. For AFSCME/Non-Represented members, the pattern of longevity and promotion pay increases is much more difficult to discern from the data. This is partly due to the smaller amount of data, as well as a large variety in the job descriptions and career paths of these members. The prior four experience studies have confirmed the reasonableness of a 0.50% per year assumption for salary growth due to merit. We recommend retaining this assumption.

$100,000

Average Earnings



$80,000

Actual

0.5%->

$60,000

$40,000

$20,000

$0 0

5

10

15

20

25

Service

30

35

40

45

50

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

21

Mortality (ATU / IBEW – Non-Disabled) Current Assumptions

Proposed Assumptions

Summary of Experience versus Current Assumptions (2004-2010)

Summary of Experience versus Proposed Assumptions (2004-2010)

Eligible Exposure

Actual Deaths

Expected Deaths

Actual to Expected Ratio

ACTIVE

Eligible Exposure

Actual Deaths

Expected Deaths

Actual to Expected Ratio

ACTIVE

Males

7,908

24

38.2

62.8%

Males

7,908

24

29.6

81.1%

Females

4,863

5

9.7

51.6%

Females

4,863

5

9.1

54.9%

12,771

29

47.9

60.5%

Total

12,771

29

38.7

74.9%

4,314

167

156.6

106.6%

975

24

23.4

102.6%

5,289

191

180.0

106.1%

75

1

2.7

37.0%

Total

RETIRED Males

RETIRED 4,314

167

159.8

104.5%

Males

975

24

23.7

101.3%

Females

5,289

191

183.5

104.1%

Total

Females Total

SURVIVING SPOUSES Males

SURVIVING SPOUSES 75

1

2.8

35.7%

Females

899

31

21.8

142.2%

Females

899

31

21.6

143.5%

Total

974

32

24.5

130.6%

Total

974

32

24.4

131.1%

12,297

192

189.0

101.6%

6,737

60

54.1

110.9%

19,034

252

243.1

103.7%

COMBINED Males Females Total

Males

COMBINED 12,297

192

200.8

95.6%

6,737

60

55.2

108.7%

19,034

252

256

98.4%

Males Females Total

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010



The current actuarial assumption for non-disabled active and retired  ATU/IBEW members is: o

Males: 1983 Group Annuity Mortality (GAM) Male table

o

Females: 1971 Group Annuity Mortality (GAM)





In the prior study, mortality was updated for male members to the  1983 GAM Male table. Female mortality remained unchanged.



Experience has been aggregated with the prior study to ensure adequate exposure, and to assist in the analysis of the underlying trends. The analysis now includes seven years of experience (2004 – 2010).



Actual deaths among active members are usually below actuarial assumptions. Active members frequently become disabled or retire when they are in poor health, so these deaths are reported in the inactive categories.

22

We propose the use of the RP 2000 Combined Healthy Tables, setforward one year for both males and females. This is a fairly small adjustment to assumed mortality rates. The RP 2000 Tables are the most current tables generally used for pension funding. A positive margin between the actual number of deaths and the predicted number of deaths (i.e. an actual to expected ratio greater than 100%) is desirable for two reasons: o

Overall mortality has historically improved, and is expected to improve in future years.

o

The RP 2000 Tables are generally designed using benefitweighted (rather than participant-weighted) data. This is because members with larger benefits tend to have lower mortality rates, at least at younger ages. Applying the tables on a participant basis, while accurately predicting the number of deaths, may underestimate the impact on liabilities.

o

To confirm the reliability of this assertion, the RP-2000 tables were applied to benefit-weighted data for ATU/IBEW for the period 2007-2010. Actual to expected ratios using this data were as follows:

o



Retired (Male/Female): 98.9%



Surviving Spouses (Male/Female): 110.7%

These ratios are well below the same ratios calculated using participant-weighted data. For this reason, we have proposed assumptions that continue to provide a margin between the number of actual deaths and the number expected.

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

23

Mortality (AFSCME / Non-Represented – Non-Disabled) Current Assumptions

Proposed Assumptions

Summary of Experience versus Current Assumptions (2004-2010)

Summary of Experience versus Proposed Assumptions (2004-2010)

Eligible Exposure

Actual Deaths

Expected Deaths

Actual to Expected Ratio

ACTIVE Males

Actual Deaths

Expected Deaths

Actual to Expected Ratio

ACTIVE 1,267

3

6.4

46.9%

Males

901

2

2.2

90.9%

Females

2,168

5

8.6

58.0%

Total

Females Total

Eligible Exposure

RETIRED

1,267

3

5.5

54.5%

901

2

2.6

76.9%

2,168

5

8.1

61.7%

RETIRED

Males

970

35

29.1

120.3%

Males

970

35

29.7

117.8%

Females

284

4

2.8

142.9%

Females

284

4

3.5

114.3%

1,254

39

31.9

122.3%

Total

1,254

39

33.2

117.5%

7

0

0.1

0.0%

Total

SURVIVING SPOUSES Males

SURVIVING SPOUSES 7

0

0.1

0.0%

Females

288

10

7.6

131.6%

Females

288

10

8.7

114.9%

Total

295

10

7.7

129.9%

Total

295

10

8.8

113.6%

COMBINED

Males

COMBINED

Males

2,244

38

35.6

106.7%

Males

2,244

38

35.3

106.7%

Females

1,473

16

12.6

127.0%

Females

1,473

16

14.8

127.0%

Total

3,717

54

48.2

112.0%

Total

3,717

54

50.1

107.8%

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010





24

The current actuarial assumption is that non-disabled female active  and retired AFSCME/Non-represented will experience mortality in accordance with the 1983 GAM mortality, while male active and retired members will experience mortality in accordance with the 1994 GAM mortality.  Actual deaths among active members are usually below actuarial assumptions. Active members frequently become disabled or retire when they are in poor health, so these deaths are reported in the  inactive categories.

Although the current mortality assumption has predicted mortality experience reasonably well, we would prefer to move to a more modern table that is consistent with that proposed for the ATU/IBEW members.



Actual versus expected mortality experience was also reviewed using benefit-weighted data, which confirmed the overall pattern of higher mortality experience than expected.

Accordingly, we propose the use of the Retired Pensioner (RP) 2000 Combined Healthy Tables for AFSCME and Non-representative members (with no age adjustments). Experience has been aggregated with the prior study to improve the adequacy of the exposures, and to assist in the analysis of the underlying trends. The analysis includes seven years of experience (2004 – 2010).

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

25

Mortality (All Groups – Disabled) Current Assumption



The current actuarial assumption is that occupationally disabled members will experience mortality in accordance with the Mortality Table for Disabled Members Not Receiving Social Security Benefits published by the Pension Benefit Guaranty Corporation (PBGC). Total and permanently disabled members are expect to experience mortality in accordance with the Mortality Table for Disabled Members Receiving Social Security Benefits.



The amount of experience for the disabled members is relatively limited for the purpose of developing assumptions regarding mortality, even when the experience of the most recent four calendar years is aggregated with the experience of the prior experience study (2004-2006).



Experience from both periods has been aggregated to improve the adequacy of the exposures. The analysis includes seven years of experience: 2004-2010.



The current tables have done a reasonable job of predicting the overall number of disabled deaths.

Summary of Experience versus Current Assumptions (Aggregated) Eligible Exposure

Actual Deaths

Expected Deaths

Actual to Expected Ratio

OCCUPATIONAL DISABILITY Males (ATU/IBEW)

527

10

12.9

77.5%

Females (ATU/IBEW)

334

8

2.4

333.3%

Males (AFSCME/Non-Rep)

14

0

0.2

0.0%

Females (AFSCME/Non-Rep)

2

0

0.0

0.0%

877

18

15.6

115.4%

Total

TOTAL AND PERMANENT DISABILITY Males (ATU/IBEW)

312

20

25.4

78.7%

Females (ATU/IBEW)

161

10

6.1

163.9%

Males (AFSCME/Non-Rep)

41

3

3.4

88.2%

Females (AFSCME/Non-Rep)

6

0

0.2

0.0%

520

33

35.0

94.2%

Occupational

877

18

15.6

115.4%

Total & Permanent

520

33

35.0

94.2%

1,397

51

50.6

100.8%

Total

ALL DISABILITIES

Total

Recommendation 

Given the limited nature of the experience available, the current disability mortality tables have done a reasonably accurate job of predicting deaths among disabled members. Accordingly, no changes to the disability mortality rate assumptions are recommended.

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

Economic Assumptions Introduction Economic assumptions used in the development of actuarial liabilities and costs for a defined benefit plan include: 

The inflation assumption;



The real investment return assumption; and



The real growth in pay relative to inflation.

26

which underlies each of the elements of economic assumptions listed above. Chart E-1 below shows the average rate of inflation over 30-year periods, with the earliest such period ending in 1955 and the latest ending in 2010. We note in the chart that average inflation seemed to be increasing steadily until the 1990’s when it leveled off and began to decrease.

We look to the past for indications of future economic behavior, but we must also consider how the future may be expected to be different. In order to reflect the long-term nature of defined benefit plan funding in the development of these economic assumptions, it is appropriate to review long term trends.

Inflation The rate of inflation is an important assumption because inflation underlies the growth of Plan liabilities. Over time, we expect the wages paid to Plan members to increase with the cost of living and with productivity increases. Benefits – and therefore liabilities – are directly proportional to wages, so they grow with inflation and productivity as well. While historical trends are not entirely indicative of the future, they do often serve as a useful guide in determination of assumptions. However, there are elements of the future economic environment that may differ from the past due to structural changes. An important and fundamental case in point is the rate of inflation,

Chart E-1: Average Past Inflation An examination of historical inflation could lead to the assumption that inflation is likely to be quite high, perhaps as high as 4%. However, there are a number of reasons to believe that future inflation levels will not be as high as Chart E-1 would seem to suggest. 

The average compound rate of inflation since 1926 is 2.99%.



An important reason for some of the high rates of inflation in the graph above is the nine-year period 1973-81 when

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

inflation averaged 9.2% per year. 

The years 1973-81 featured unprecedented levels of household formation. The demand for new houses, cars, office space and equipment caused by the maturation of the post-war baby boom may have largely been responsible for the inflation during these years. Since 1981, increases have been in the range 0.1% to 4.6% with one exception (6.1% in 1990), averaging 3.16% per year.



The population of the United States is aging, which implies a greater likelihood of low inflation in the future. This has been observed in other countries with aging populations, such as Japan.

27

inflation to be in future years. Various securities, such as Treasury inflation-protected securities (TIPS), provide the necessary data for these analyses. As an example, a recent publication by the Federal Reserve Bank of Cleveland attempts to incorporate some of this market data, as shown in Chart E-2. The expected inflation is just 1.86% for the next decade. The Plan’s investment consultant, NEPC, has projected a rate of inflation of 3.0% per year over the next 5-7 years. The inflation assumption currently in use in the Plan’s actuarial valuations is 3.2% annually. The predictions of future inflation by experts are not unanimous. Some commentators note that the large current and expected future deficits increase the likelihood of higher levels of inflation in the future. Therefore, the current assumption of 3.2% remains within a reasonable range.

Chart E-2: Expected Inflation (Source: Cleveland Federal Reserve website. As of May 1, 2011)



The Federal Open Market Committee has policies in place to control inflation, making future levels more likely to remain relatively low.



Financial markets offer evidence of what investors expect

However, we would recommend bringing the actuarial assumption for inflation into better alignment with inflation levels currently being experienced, levels included in financial market pricing of securities, and the inflation forecasts of the Plan’s investment consultant. Accordingly, we recommend decreasing the assumed inflation rate from 3.2% to 3.0% annually.

Investment Return The investment return assumption depends on the anticipated average level of inflation and the anticipated average real rate of return. The real rate of return is the investment return in excess

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

of underlying inflation. The expected average real rate of return is heavily dependent on asset mix: The portion of assets in stocks, bonds, and cash.

According to this model, the probability of achieving a 7.7% compound return is 53.3%.

In Chart E-3 below, we have simulated the return derived using the asset mix of the AC Transit Fund. The projected returns are derived by simulation, using the following algorithm: 1. The expected returns, standard deviation and correlation matrix for each asset class were taken from the capital market assumptions employed by the Plan’s investment consultant. These assumptions were supplied by NEPC with a 30-year time horizon. 2. 10,000 simulation trials for repeated ten year periods were run, and the mean compound return was computed for each.

Chart E-3: Summary of Return Simulations (30 Year)

3. Given the distribution of returns, we have created a chart that shows the likelihood of the geometric mean return for a specific trial exceeding a specified assumption over a ten year period, assuming that the return rates are net of investment and administrative expenses. The mean return from this simulation was 7.9%, for a real return of 4.9%. Note that the curve crosses the 50% likelihood threshold near 7.9%, meaning that chances are around 50/50 that a 7.9% return will be achieved over a ten year period. A lower return assumption would represent a more conservative position, and would result in a higher likelihood of achieving the expected return.

28

Chart E-4: Summary of Return Simulations (5 – 7 Years)

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

A somewhat different picture emerges when we consider the NEPC assumptions for the next five to seven years. For this shorter time frame, the NEPC return assumptions are materially lower, with a compound average return of 6.2%. As a result, the graph of the likelihood of achieving the expected return is much different, shown in Chart E-4. In Chart E-4, the compound average return is just 6%, for a real return of 3%. The probability of meeting or exceeding a 7.7% return is only about one-third. There are a number of factors that suggest that a small reduction in the assumed rate of return would be timely. 



We tend to assign more importance to the near- and mid-term return projections than to the 30-year projections. Fund performance is usually measured over five to ten years; longer measurement periods are often considered less relevant because of changes in the economy and in the investment markets. When we use the AC Transit asset allocation and the capital asset assumptions from another investment consulting firm (Pension Consulting Alliance) we compute an average return of 7.1%.

29



NEPC projects a 6.2% return for the next five to seven years for the current AC Transit asset allocation.



A number of retirement plans are reducing their return and inflation rates. CalPERS is now assuming 3% inflation, and their actuary has recommended – though the Board has voted to hold this in abeyance for now – a return assumption of 7.5%. Given the advantage of scale enjoyed by CalPERS in investment expenses, the recommendation would be equivalent to between 7% and 7.25% for AC Transit. Other plans have or are planning a similar reduction.

Therefore, we recommend a reduction in the assumed return on Plan assets. While the current 7.7% continues to be reasonable, for the reasons cited above we recommend that the assumption be decreased to 7.5% for the January 1, 2011 actuarial valuation. In addition, we recommend that the Board and staff conduct at least a brief discussion of this assumption annually, in consultation with the Plan’s actuary and investment consultant, to determine if further changes are appropriate.

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

Payroll Growth Under certain actuarial funding methods, the unfunded actuarial accrued liability is amortized as a level percentage of future member payroll. In such cases, the assumption concerning the growth in covered payroll determines the payments on the unfunded actuarial accrued liability under the amortization schedule. Unlike a mortgage, which is amortized with level dollars over the period of the mortgage, the traditional approach for amortizing a public sector plan’s unfunded actuarial liability is to allocate costs as a level percentage of aggregate underlying payroll. By way of illustration, if aggregate payroll were expected to increase by 4%, then contributions to finance the unfunded actuarial accrued liability should be growing by 4% each year. Early payments on a sloping schedule of this kind are often less than interest on the unfunded actuarial accrued liability, causing this liability to grow before the dollar amount is amortized. If the percentage of payroll growth assumption underlying the amortization schedule is too steep, insufficient amortization payments will be made in each successive year – allowing the unfunded actuarial accrued liability to grow. Components of the payroll growth assumptions are: 

Inflation



30

Other payroll increases not offset by salary reduction caused by replacement of terminating employees by new entrants. Such increases are often attributed to productivity gains.

Usually we recommend that long range gains due to productivity, the collective bargaining process or other pressures should be assumed to be zero or minimal. While productivity tends to increase in many sectors of the economy, any long-term assumption of salary growth beyond inflation carries with it an assumed improvement in relative standard of living. For transit employees in particular, such pay increases beyond the rate of inflation have not been observed. Therefore, the current assumption of 0% is reasonable. Accordingly, EFI recommends maintaining a payroll growth assumption equal to the assumed inflation rate.

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

Methodology Purposes of the Experience Study The first goal of this Experience Study is to review the recent past demographic experience of the Plan. We seek to understand the behavior of the participating members so that we can recommend actuarial assumptions concerning future demographic experience. The second goal of this Study is to recommend economic assumptions to be used in computing liabilities and costs. These economic assumptions include the expected rate of return on Plan assets and the anticipated rate of increase in the Consumer Price Index (CPI). These assumptions are determined based on the investment strategy adopted by the Plan and on the past behavior of the capital markets and the CPI, and on future expectations. Once adopted, the assumptions recommended by this Study will be used to determine future liabilities and costs and for purposes of evaluating prospective changes in benefits, eligibility conditions, and other aspects of the Plan’s operations.

Importance of Reliable Assumptions The liabilities and costs calculated in actuarial valuations and cost studies are based on a projection of future conditions. The actuary makes assumptions concerning the rates of retirement, withdrawal, termination, disability, and death among plan members. In addition, the actuary must project future earnings on plan assets, inflation, and growth in the pay of active members.

31

The actuary sets his assumptions based on past experience and future expectations. In setting demographic assumptions, such as rates of retirement, the past experience of the covered group of employees is often the best predictor of future behavior. When establishing economic assumptions, such as the expected return on plan assets, the historical behavior of the investment markets can serve as a guide. Actuarial funding methods are designed so that, if the actuarial assumptions are met, plan costs will generally be a predictable percentage of member pay from year to year. If actual economic or demographic experience varies from our assumptions, plan costs will rise or fall accordingly. Therefore, it is worth the effort to make our best estimate of future conditions so that the plan costs computed by the actuary will be as stable and predictable as possible.

Prior Experience Studies An actuarial experience study for the years 1980-82 was performed by the Wyatt Company as Plan actuary at that time. The Wyatt Company performed another experience study covering the years 1983-88, but these rates were never adopted by the District or used in an actuarial valuation of the Plan. In 1994, Coates Kenney, Inc. prepared an actuarial gain and loss study for the plan year from July 1, 1992 through June 30, 1993. The purpose of such a study is to determine whether the various actuarial assumptions were in reasonable accord with actual experience during the prior year. A gain and loss study will determine the gross financial impact of variations from assumed experience during the year, but it does not provide the detailed

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

guidance in setting new assumptions that an experience study provides. After the passage of California Proposition 162, EFI Actuaries was retained by the Retirement Board as its consulting actuary. In this role, EFI performed an experience study dated May 6, 1995, covering experience during calendar years 1991-94. Actuarial valuations from July 1, 1994 through January 1, 1997 were based on the results of this study. A second experience study was performed by EFI covering calendar years 1995-97. This study, dated April 8, 1998, served as the basis for actuarial assumptions used in the actuarial valuations of the Retirement Plan as of January 1, 1998, 1999, and 2000. A third experience study was performed by EFI covering calendar years 1998-2000. This study, dated June 26, 2001, served as the basis for actuarial assumptions used in the actuarial valuations of the Retirement Plan as of January 1, 2001, 2002, and 2003. A fourth experience study was performed by EFI covering calendar years 2001-2003. This study, dated August 29, 2004, served as the basis for actuarial assumptions used in the actuarial valuations of the Retirement Plan as of January 1, 2004, 2005, and 2006. A fifth experience study was performed by EFI covering calendar years 2004-2006. This study, dated August 14 2007, served as the basis for actuarial assumptions used in the actuarial valuations of the Retirement Plan as of January 1, 2007, 2008, 2009 and 2010.

32

Scope of Report Demographic assumptions relate to all behavioral characteristics of the group. Behavioral characteristics do not include the assumptions concerning future inflation, the real rates of return of the investments in the trust fund, or the anticipated growth in the underlying payroll of the members. Demographic assumptions include the following: 

Probability of retirement from active service,



Probability of termination of employment prior to retirement (with the member receiving a deferred vested benefit),



Probability of disability among active employees (either occupational or total and permanent),



Probability of death among active employees, and



Rates of mortality among retired and disabled members and their beneficiaries.

In addition, demographic assumptions include the merit (longevity and promotion) component of individual pay increases. This does not include the inflationary element in pay increases. For example, if inflation is 3.2% and the employee receives a 4.7% pay increase, 1.5% of this increase is deemed "merit". Economic assumptions include the rate of increase in the cost of living (inflation), which is a part of the overall pay increase assumption discussed above. In addition, a crucial economic

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

assumption is the real rate of return on plan assets -- the return on assets above the rate of inflation.

Methodology (Demographic Assumptions) One goal of this Study is to compute the probability of death, disability, retirement, withdrawal, or termination leading to a vested benefit at each age for active members and the probability of death at each age for inactive members. To this end, we proceed as follows: 

We count the number of members leaving for each cause during the term of the Study. This is the number of decrements.



We count the number of members who could have left for each cause during the Study. This is the exposure.





When the exposure is sufficient, we divide the number of decrements by the exposure at each combination of age and service for an employee group to determine the probability of leaving due to the cause in question. Where feasible, experience has been examined separately by gender. In some cases, experience has been combined when male and female experience is similar or when there is insufficient data to produce reliable rates by sex.

A unique challenge is presented by the somewhat large group of members on leave as of the date of each annual valuation. These members have an uncertain status each year, since their applications for retirement or disability are pending. They are

33

included in each actuarial valuation based on the best estimate of their ultimate status, provided for each individual by Plan staff. For purposes of this Study, these members are included in the total exposures and are recognized as a decrement based on the final resolution of their status when their applications for disability or retirement have been fully adjudicated. When there is insufficient exposure to derive statistically reliable rates by age and service, we may combine exposures and decrements for groups of ages and service. Alternatively, we may compare the total number of actual decrements with the total number of decrements predicted by a standard actuarial table, and adopt a table that predicts decrements, in total, reasonably close to those that have been observed. Where the rate of decrement is low and the underlying causes of the decrement in question are not expected to change significantly with time (for instance, for disability rates), we may combine the most recent experience with data from prior experience studies. For the study of the merit (longevity and promotion) components of individual pay increases, we generally choose to use a transverse study. A reliable way to assess average increases in pay due to merit is to analyze average pay versus service for the current active members of a plan. With a homogeneous group of any size at all, the pattern of promotions and longevity increases during the career of an average employee is clearly visible in this analysis. This is a transverse study of longevity and promotion pay increases: The data is taken as of a particular point in time.

Alameda-Contra Costa Transit District Retirement Board Actuarial Experience Study January 1, 2007 through December 31, 2010

Longitudinal studies, which use changes in pay collected over several years, are often unreliable when used on a stand-alone basis due to the effects of inflation, collective bargaining, and management decisions during the term of the study. However, we have reviewed the data using this method based on input from the System regarding year-by-year base pay increases in order to confirm that the conclusions drawn from the transverse study are reasonable and complete.

Methodology (Economic Assumptions) The Plan’s economic assumptions are critically important in computing actuarial liabilities and costs. A careful determination of these assumptions requires an analysis of the past performance of the capital markets and the Plan’s future investment outlook. To this end, we proceed as follows: 

Based on a detailed analysis of recent past history and reasonable expectations for the future, a long term projection of the rate of inflation is determined.



Based on the Plans’ investment strategy and historical rates of return on various asset classes, the long term real rate of return on assets is projected. This is the return on assets in excess of inflation.



The projected rate of inflation is combined with the assumption concerning merit pay increases to project future members’ pay.



The rate of inflation is combined with the estimated real return on assets to determine the overall return on assets.

34

Any estimate of future inflation and asset returns is difficult. Over time, there will be actuarial gains and losses as experience deviates from our assumptions. As past and recent capital market experience has shown, these gains and losses can have a substantial impact on cost volatility.