AC Transit...The recommendation to change mortality assumptions has the largest impact on contribution rates. The recommended change to retirement rates also would increase contribution

AC Transit

Employees’ Retirement Plan

Actuarial Experience Study for

January 1, 2011 through

December 31, 2014

Produced by Cheiron

August 2015

TABLE OF CONTENTS

Section Page

Transmittal Letter i

Section I Executive Summary ..............................................................................................1

Section II Certification ..........................................................................................................3

Section III Economic Assumptions ........................................................................................4

A. Price Inflation........................................................................................................4

B. Wage Inflation ......................................................................................................7

C. Discount Rate ........................................................................................................8

Section IV Demographic Assumptions .................................................................................12

A. Merit Salary Increases.........................................................................................12

B. Retirement Rates .................................................................................................14

C. Termination Rates ...............................................................................................22

D. Disability Rates ...................................................................................................26

E. Mortality Rates....................................................................................................30

F. Other Demographic Assumptions .......................................................................34

Appendices

Appendix A Summary of Current Assumptions......................................................................35

Appendix B Summary of Proposed Assumptions ...................................................................42

August 5, 2015

Retirement Board of

Alameda – Contra Costa Transit District Retirement Plan

1600 Franklin Street, 5th Floor

Oakland, CA 94612

Dear Members of the Board:

The purpose of this report is to present an Actuarial Experience Study of the Alameda – Contra

Costa Transit District Retirement Plan (AC Transit Employees’ Retirement Plan, the Plan)

covering actuarial experience from January 1, 2011 through December 31, 2014. The report

includes analyses and recommendations of economic and demographic assumptions to be used

beginning with the January 1, 2015 actuarial valuation.

If you have any questions about the report or would like additional information, please let us

know.

Sincerely,

Cheiron

Robert T. McCrory, FSA, CERA, FCA, EA, Timothy S. Doyle, ASA, MAAA

Principal Consulting Actuary Associate Actuary

ALAMEDA - CONTRA COSTA TRANSIT DISTRICT RETIREMENT PLAN

EXPERIENCE STUDY AS OF DECEMBER 31, 2014

SECTION I

EXECUTIVE SUMMARY

1

Actuarial assumptions (economic and demographic) are intended to be long-term in nature, and

should be both individually reasonable and consistent in the aggregate. The purpose of this

experience study is to evaluate whether or not the current assumptions adequately reflect the

long-term expectations for AC Transit, and if not, to recommend adjustments. It is important to

note that frequent and significant changes in the actuarial assumptions are not typically

recommended, unless there are known fundamental changes in expectations of the economy, or

with respect to AC Transit’s membership or assets that would warrant such frequent or

significant changes.

SUMMARY OF ECONOMIC ASSUMPTION ANALYSIS

The specific economic assumptions analyzed in this report are price inflation, wage inflation, and

the discount rate. These assumptions have a significant impact on the contribution rates in the

short-term and the risk of negative outcomes in the long-term.

The current economic assumptions adopted by the Retirement Board include a 7.25% long-term

rate of return on Plan assets, an annual increase in prices measured by the Consumer Price Index

(CPI) of 3%, and annual wage increase equal to price increases.

These assumptions are consistent with the long-term (30-year) capital market assumptions of the

Plan’s investment consultant, NEPC; their 30-year projections include an average annual return

on investments of 7.18%, with 3% assumed annual inflation. Other data presented in this report

indicate that the current assumptions adopted by the Retirement Board continue to be reasonable,

and that there is no compelling reason for changing them.

However, NEPC projects much lower returns for the next five to seven years, averaging 5.8%. If

NEPC’s projections are realized, the Board can expect a pattern of actuarial losses in the near

term. Therefore, the Board may want to consider decreasing its return and inflation assumptions

to reflect short-term NEPC expectations.

SUMMARY OF DEMOGRAPHIC ASSUMPTION ANALYSIS

This experience study specifically analyzes and makes the following recommendations for the

demographic assumptions.

Merit salary increases – No changes recommended.

Retirement rates – Higher rates for longer service members and lower rates for shorter

service members who are eligible for retirement.

Termination rates – Changes in rates for members with less than ten years of service.

Disability rates – No changes recommended.

Mortality rates – Adjusted Blue Collar tables for ATU and IBEW members and White

Collar tables for AFSCME and Non-Represented members with generational

improvement for all members.



SECTION I

EXECUTIVE SUMMARY

2

The recommendation to change mortality assumptions has the largest impact on contribution

rates. The recommended change to retirement rates also would increase contribution rates while

the change to termination rates would reduce contribution rates. Further information about these

changes to contribution rates can be found in Cheiron’s cost letter to AC Transit, dated August 5,

2015.

The recently completed mortality study by the Society of Actuaries found that mortality rates had

improved faster than previously anticipated and recommended future projections of mortality

improvement commensurate with recent experience in the short-term tapering to a long-term

expected rate of improvement by 2027. The recommended change to mortality rates for AC

Transit reflects both the improvement in mortality since the last experience study, which was

greater than anticipated, and the application of the recommended higher rates of improvement

projected in the future.

The body of this report provides additional detail and support for our conclusions and

recommendations.



SECTION II

CERTIFICATION

3

The purpose of this report is to provide the results of an Actuarial Experience Study of the

Alameda – Contra Costa Transit District Retirement Plan (AC Transit) covering actuarial

experience from January 1, 2011 through December 31, 2014. This report is for the use of the

AC Transit Retirement Board in selecting assumptions to be used in actuarial valuations

beginning January 1, 2015.

In preparing our report, we relied on information (some oral and some written) supplied by AC

Transit. This information includes, but is not limited to, the plan provisions, employee data, and

financial information. We performed an informal examination of the obvious characteristics of

the data for reasonableness and consistency in accordance with Actuarial Standard of Practice

No. 23.

To the best of our knowledge, this report and its contents have been prepared in accordance with

generally recognized and accepted actuarial principles and practices that are consistent with the

Code of Professional Conduct and applicable Actuarial Standards of Practice set out by the

Actuarial Standards Board. Furthermore, as credentialed actuaries, we meet the Qualification

Standards of the American Academy of Actuaries to render the opinion contained in this report.

This report does not address any contractual or legal issues. We are not attorneys and our firm

does not provide any legal services or advice.

This report was prepared for the AC Transit Retirement Board for the purposes described herein.

This report is not intended to benefit any other party, and Cheiron assumes no duty or liability to

any such party.

Robert T. McCrory, FSA, CERA, FCA, EA Timothy S. Doyle, ASA, MAAA

Consulting Actuary Associate Actuary



SECTION III

ECONOMIC ASSUMPTIONS PRICE INFLATION

4

The economic assumptions used in actuarial valuations are intended to be long-term in nature,

and should be both individually reasonable and consistent with each other. The specific

assumptions analyzed in this report are:

Price inflation – used indirectly as an underlying component of other economic

assumptions.

Wage inflation – across the board wage growth used to project benefits and to amortize

the unfunded liability as a level percentage of expected payroll.

Discount rate – used both to project long-term asset growth and to discount future cash

flows in calculating the liabilities and costs of the Plan.

In order to develop recommendations for each of these assumptions, we considered historical

data, both nationally and for the Plan, and expectations for the future, as expressed by the Plan’s

investment consultant and the Board.

PRICE INFLATION

Long-term price inflation rates are the foundation of other economic assumptions. In a growing

economy, wages and investments are expected to grow at the underlying inflation rate plus some

additional real growth rate, whether it reflects productivity in terms of wages or risk premiums in

terms of investments.

Historical Data

Chart III-1 below shows inflation for the U.S. by individual year since 1950.

Chart III-1

Over the 50 years ending March 2015, the geometric average inflation rate for the U.S. has been

about 4.1%, but this average is heavily influenced by the high inflation rates in the 1970s and

early 1980s. Over the last 30 years, the geometric average inflation rate has been 2.7%.



SECTION III


5

Future Expectations

A measure of the market consensus of expected future inflation rates is the difference in yields

between conventional treasury bonds and Treasury Inflation-Protected Securities (TIPS) at the

same maturity. Table III-1 shows the yields on both types of bonds and the break-even inflation

rate as of April 30, 2015. Break-even inflation is the level of inflation needed for an investment

in TIPS to “break even” with an investment in conventional treasury bonds of the same maturity.

Table III-1

Data Source Federal Reserve, Constant Maturity Yields, Monthly Series

The Federal Reserve Bank of Philadelphia publishes a quarterly survey of professional economic

forecasters that includes their forecasts of inflation over the next 10 years. The survey for the

first quarter of 2015 shows a median inflation forecast of 2.10%; a minimum forecast of about

1.4% and a maximum forecast of 3.1%.

Boston College’s Center for Retirement Research maintains a database on over 150 large public

plans. For 2013, the inflation assumptions used by the plans in the database ranged from 2.50%

to 5.00%. These assumptions tend to be based on time horizons that are longer than 10 years.

Chart III-2 on the next page shows the distribution of the current 10-year forecasts for CPI-U

from the professional survey published by the Federal Reserve Bank of Philadelphia compared to

the 2013 assumptions used by the large public pension plans in the Center for Retirement

Research’s Public Plans Database.

Break-Even Inflation

Based on Treasury Bond Yields

Time to Conventional Break Even

Maturity Yield TIPS Yield Inflation

5 Years 1.35 -0.26 1.61

10 Years 1.94 0.08 1.86

20 Years 2.33 0.42 1.91



SECTION III


6

Chart III-2

Finally, NEPC, the Board’s investment consultant, uses an inflation assumption of 3.0% for the

next five to seven years, and 3.25% for the next 30 years.

Based on all of these considerations, we believe a reasonable range for long-term price inflation

for use in the Plan’s actuarial valuations is between 2.0% and 3.5%. Therefore, the current

assumption can be retained.

Minimum 1.40% 2.50%

25th Percentile 2.00% 3.00%



Maximum 3.10% 5.00%

0.0%

1.0%

2.0%

3.0%

4.0%

5.0%

6.0%

EconomicForecasters

Public PlansDatabase

Survey of CPI Assumptions

Min to 25th 25th to 50th

50th to 75th 75th to Max



SECTION III

ECONOMIC ASSUMPTIONS WAGE INFLATION

7

WAGE INFLATION

Wage inflation can be thought of as the annual across-the-board increase in wages. Individuals

often receive salary increases in excess of the wage inflation rate, and we study these increases

as a part of the merit salary scale assumption. Wage inflation generally exceeds price inflation by

some margin reflecting the history of increased purchasing power.

Wage inflation is used in the actuarial valuation as the minimum expected salary increase for an

individual and, for purposes of amortizing the unfunded actuarial liability, the rate at which

payroll is expected to grow over the long term, assuming a stable active member population.

Chart III-3 shows the increase in national average wages (as reported by the Social Security

Administration) compared to inflation from 2003 through 2013.

Chart III-3

Over this period, national wage inflation averaged approximately 2.77% compared to annual

price inflation of 2.33%, making wage increases less than 0.5% above inflation. Note the

significant drop in 2008 and 2009 as well as the recent decline in national average wage growth

in 2013, the latest year for which data is available.

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.

For example, over the last four years, pay increases for salaried members averaged about 1.5%

per year, while for hourly employees the annual increase was less than 1%. Therefore, the

current assumption of no increases in wages over inflation continues to be reasonable.



SECTION III

ECONOMIC ASSUMPTIONS DISCOUNT RATE

8

DISCOUNT RATE

The discount rate assumption is generally the most significant of all the assumptions employed in

actuarial valuations. The discount rate is based on the long-term expected return on plan

investments. In the short-term, a higher discount rate results in lower expected contributions.

However, over the long term, actual contributions will depend on actual investment returns and

not the discount rate (or expected investment returns). If actual investment returns are lower than

expected, contribution rates will increase in the future. It is important to set a realistic discount

rate so that projections of future contributions for budgeting purposes will not be biased,

particularly to be too low.

Other Large Public Retirement Plans

Based on the Public Fund Survey, developed by the National Association of State Retirement

Administrators (NASRA) covering most of the largest public retirement systems in the country,

there has been a general movement over at least the last decade to reduce the discount rate used

in actuarial valuations. Chart III-4 shows the change in the distribution of assumptions since

2001. The median assumption is now 7.75% and the number of plans using a discount rate of

7.5% or lower has increased significantly.

Chart III-4



SECTION III


9

Target Asset Allocation and Future Expectations

The discount rate 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

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.

In Chart III-5 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.

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 expenses.

Chart III-5: Distribution of Returns Using NEPC 30-Year Assumptions



SECTION III


10

The first simulation, using the capital market assumptions from NEPC for the next 30 years, is

shown above in Chart III-5. The simulated probability is shown for each 10-year compound

return from -5% to 20%. The mean return from this simulation was 7.18%, for a real return of

4.18%. The probability of achieving a return of 7.25% or higher, as an average over the next 10

years, is 47.7%, a little less than half. This suggests that there is no significant inconsistency

between the Plan discount rate and the long-term capital market assumptions of the Plan’s

investment consultant.

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 5.82%. As a result, the graph of the distribution of expected

returns is much different, shown in Chart III-6.

Chart III-6: Distribution of Returns Using NEPC Five- to Seven-Year Assumptions

In Chart III-6, the compound average return is just 5.82%, for a real return of 2.82%. The

probability of meeting or exceeding a 7.25% return is only about one-third.

It is important to note the wide range of outcomes shown in Chart III-5, based on the NEPC 30-

year assumptions. Recall that we are showing the distribution of average returns over a 10-year

period; these are not annual returns. Analysis of the simulation data shows there is a 50%



SECTION III


11

probability of the 10-year average return being either lower than 4.5% or higher than 9.5%, and a

50% chance of the average return being between those two numbers. Another way of looking at

the distribution is that the Plan has a roughly 50/50 chance of being within 2.5% of its assumed

return over a 10-year period.

While the current discount rate is consistent with the NEPC long-term capital market

assumptions, there are a number of factors that suggest that the near-term expected rate of return

should be discussed.

Many investment consultants expect poor rates of return in the immediate and near-term

future. They reason that there is little in the way of yields on fixed income, and that the

equity markets are fully valued.

If NEPC and much of the investment community are correct in their projections, we can

expect returns below the 7.25% assumed rate for a number of years. This will result in

actuarial losses and increases in employer contribution rates.

We believe that near- and mid-term return projections should be considered along with long-

term projections. Fund performance is usually measured over five to ten years; longer

measurement periods are often considered less relevant because of the potential for changes

in the economy and in the investment markets.

Therefore, we recommend considering a reduction in the assumed return on Plan assets. While

the current 7.25% continues to be reasonable, for the reasons cited above we recommend that a

reduction to 7.0% be considered. In addition, the prospect of several years of actuarial losses, in

line with the NEPC assumptions, and the resulting increases in District contribution rates should

be communicated to the District staff and Board for their use in planning.

In addition, we recommend that the Board and staff continue to 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.



SECTION IV

DEMOGRAPHIC ASSUMPTIONS MERIT SALARY INCREASES

12

Demographic assumptions are used to predict membership behavior, including rates of

retirement, termination, disability, and mortality. These assumptions are based primarily on the

historical experience of AC Transit, with some adjustments where future experience is expected

to differ from historical experience and with deference to standard tables where AC Transit

experience is not fully credible and a standard table is available. For purposes of this study, merit

salary increases are also considered a demographic assumption because the assumption is based

primarily on AC Transit’s historical experience.

MERIT SALARY INCREASES

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 were addressed in an earlier section of this

report.

Charts IV-1 on the next page compares the current pay patterns for ATU and IBEW members

with current pay data. Chart IV-2 compares the current pay patterns for AFSCME and Non-

Represented members with 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 AC

Transit as of January 1, 2015 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.

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). This type of study serves as a reliable way to assess average increases in pay

due to merit. The analysis begins with a plot of 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.

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.

For ATU and IBEW members the current assumed pay increases due to merit are shown by the

red line in Chart IV-1. The purple diamonds represent the average pay at each year of service.

The current assumption of 4% increases for the first 8 years of service and 0.5% thereafter is

close to the observed increases in merit pay, and no change in the assumption is recommended at

this time.

For AFSCME and Non-Represented members the current assumed pay increases due to merit are

shown by the red line in Chart IV-2. The purple diamonds represent the average pay at each year

of service. The current assumption of 0.5% increases for all years of service is close to the

observed increases in merit pay, and no change in the assumption is recommended at this time.



SECTION IV

DEMOGRAPHIC ASSUMPTIONS MERIT SALARY INCREASES

13

Chart IV-1

Chart IV-2



SECTION IV

DEMOGRAPHIC ASSUMPTIONS RETIREMENT RATES

14

ANALYSIS OF OTHER DEMOGRAPHIC ASSUMPTIONS

For all of the remaining demographic assumptions, we determined the ratio of the actual number

of decrements for each membership group compared to the expected number of decrements (A/E

ratio or actual-to-expected ratio). If the assumption is perfect, this ratio will be 100 percent.

Otherwise, any recommended assumption change should move from the current A/E ratio

towards 100 percent unless future experience is expected to be different than the experience

during the period of study.

We also calculate an r-squared statistic for each assumption. R-squared measures how well the

assumption fits the actual data and can be thought of as the percentage of the variation in actual

data explained by the assumption. Ideally, r-squared would equal 100 percent although this is

never the case. Any recommended assumption change should increase the r-squared compared to

the current assumption making it closer to 100 percent unless the pattern of future decrements is

expected to be different from the pattern experienced during the period of study.

In addition, we calculated the 90 percent confidence interval, which represents the range within

which the true decrement rate during the experience study period fell with 90 percent confidence.

(If there is insufficient data to calculate a confidence interval, the confidence interval is shown as

the entire range of the graph.) We generally propose assumption changes when the current

assumption is outside the 90 percent confidence interval of the observed experience. However,

adjustments are made to account for differences between future expectations and historical

experience, to account for the past experience represented by the current assumption, and to

maintain a neutral to slight conservative bias in the selection of the assumption. For disability

and mortality rates, we compare AC Transit’s experience to that of a standard table, and only

adjust the standard table to the extent AC Transit’s experience is large enough to be credible in

the case of disabilities. For mortality, we adjust the standard table to bring the proposed

assumption closer to an A/E ratio of 100.

RETIREMENT RATES

The current retirement rates vary by age and are applied to all members who are eligible to retire.

As a result, a member who is age 60 with 10 years of service, for example, is assumed just as

likely to retire as a member who is age 60 with 30 years of service. In reviewing the data for AC

Transit, we found that at any given age, members with more service are generally more likely to

retire than members with fewer years of service. AC Transit isn’t large enough to justify

assumptions for each age and service combination, so we recommend separate assumptions by

age for each of the following two service groups for ATU/IBEW members:

Members with less than 20 years of service,

Members with 20 or more years of service.



SECTION IV


15

The AFSCME/Non-Represented member population is not large enough to recommend different

retirement assumptions before and after 20 years of service.

Table IV-R1 shows the calculation of actual-to-expected ratios and the r-squared statistic for

male ATU and IBEW members with less than twenty years of service. Charts IV-R1 shows the

information graphically along with the 90 percent confidence interval.

The data shows much lower actual retirement rates than expected under the current assumption.

The proposed assumption decreases the aggregate assumed rate of retirement and increases the

aggregate A/E ratio from 52 percent to 91 percent. The r-squared also increases from 0.54 to

0.66.

Table IV-R1

Retirement Rates - Males <20 Years of ServiceRetirements Actual to Expected Ratios

Age Exposures Actual Current Recommended Current Recommended

55-59 327 10 23 10 44% 102%

60-64 190 13 28 15 47% 86%

65-69 85 19 30 21 64% 89%

Total 602 42 81 46 52% 91%

R-squared 0.5376 0.6591



SECTION IV


16

Chart IV-R1



SECTION IV


17


male ATU and IBEW members with twenty or more years of service, and Chart IV-R2 shows

the information graphically along with the 90 percent confidence interval.

The data shows somewhat higher actual retirement rates than expected under the current

assumption. The proposed assumption lowers the overall assumed rate of retirement and

decreases the aggregate A/E ratio from 137 percent to 108 percent. The r-squared also increases

from 0.54 to 0.92.

Table IV-R2

Chart IV-R2

Retirement Rates - Males 20+ Years of ServiceRetirements Actual to Expected Ratios


55-59 394 28 28 28 102% 102%

60-64 284 73 44 64 168% 114%

65-69 43 19 17 19 115% 98%

Total 721 120 88 111 137% 108%

R-squared 0.5449 0.9196



SECTION IV


18


female ATU and IBEW members with less than twenty years of service. Charts IV-R3 shows the


The data shows lower actual retirement rates than expected under the current assumption. The

proposed assumption decreases the aggregate assumed rate of retirement and increases the

aggregate A/E ratio from 66 percent to 111 percent. The r-squared also increases from 0.32 to

0.44.

Table IV-R3

Chart IV-R3

Retirement Rates - Females <20 Years of ServiceRetirements Actual to Expected Ratios


55-59 130 8 13 7 61% 123%

60-64 65 7 12 7 59% 98%

65-69 15 6 7 5 90% 114%

Total 210 21 32 19 66% 111%

R-squared 0.3159 0.4379



SECTION IV


19


female ATU and IBEW members with twenty or more years of service, and Chart IV-R4 shows

the information graphically along with the 90 percent confidence interval.

The data shows higher actual retirement rates than expected under the current assumption. The

proposed assumption lowers the overall assumed rate of retirement and decreases the aggregate

A/E ratio from 143 percent to 102 percent. The r-squared also increases from 0.80 to 0.83.

Table IV-R4

Chart IV-R4

Retirement Rates - Females 20+ Years of ServiceRetirements Actual to Expected Ratios


55-59 210 30 21 29 144% 102%

60-64 80 22 14 21 159% 106%

65-69 8 3 4 4 80% 83%

Total 298 55 38 54 143% 102%

R-squared 0.7957 0.8261



SECTION IV


20


AFSCME and Non-Represented members, and Chart IV-R5 shows the information graphically

along with the 90 percent confidence interval.

The data shows higher actual retirement rates than expected under the current assumption,

particularly under age 55. The proposed assumption increases the assumed rate of retirement and

decreases the aggregate A/E ratio from 120 percent to 106 percent. The r-squared also increases

from 0.35 to 0.49.

Table IV-R5

AFSCME/Non-Represented Retirement Rates - All MembersRetirements Actual to Expected Ratios


50-54 47 9 1 7 638% 128%

55-59 187 14 19 14 75% 100%

60-64 151 27 23 26 119% 102%

65 19 7 5 6 147% 123%

66-69 27 8 7 8 119% 99%

70-74 13 4 13 5 31% 88%

Total 444 69 67 66 103% 105%

R-squared 0.3487 0.4924



SECTION IV


21

Chart IV-R5



SECTION IV

DEMOGRAPHIC ASSUMPTIONS TERMINATION RATES

22

Termination rates reflect the frequency at which active members leave employment for reasons

other than retirement, death or disability. Currently, there is one set of service-based termination

rates for ATU and IBEW members, and another set for AFSCME/Non-Represented members.

For each service group, we determined the ratio of the actual number of terminations at each age

compared to the expected number of terminations (A/E ratio). If the assumption is perfect, this

ratio will be 100 percent. In addition, we calculated the 90 percent confidence interval, which

represents the range within which the true termination rate during the experience study period

fell with 90 percent confidence. (If there is insufficient data to calculate a confidence interval,

the confidence interval is shown as the entire range of the graph.) We generally propose

assumption changes when the current assumption is outside the 90 percent confidence interval of

the observed experience. However, adjustments are made to account for differences between

future expectations and historical experience, to account for the past experience represented by

the current assumption, and to maintain a neutral to slight conservative bias in the selection of

the assumption.

Table IV-T1 shows the calculation of actual-to-expected ratios and the r-squared statistic for

ATU and IBEW members, and Chart IV-T1 shows the information graphically along with the 90

percent confidence interval.

The data shows higher actual termination rates than expected under the current assumption. The

proposed assumption increases the assumed rates of termination and decreases the aggregate A/E

ratio from 108 percent to101 percent. The r-squared also increases from 0.67 to 0.92.



SECTION IV


23

Table IV-T1

Chart IV-T1

Termination Rates - All ATU/IBEW MembersTerminations Actual to Expected Ratios

Service Exposures Actual Current Recommended Current Recommended

0 206 41 21 31 199% 133%

1 290 19 29 20 66% 94%

2 148 5 7 5 68% 97%

3 172 1 9 6 12% 17%

4 211 7 6 7 111% 95%

5 281 15 8 10 178% 153%

6 259 8 8 9 103% 88%

7 197 1 6 7 17% 15%

8 157 4 5 5 85% 73%

9 159 5 2 6 210% 90%

10 243 10 4 9 274% 118%

11-14 1,223 24 18 18 131% 131%

15+ 1,087 11 16 16 67% 67%

Total 4,633 151 139 150 108% 101%

R-squared 0.6707 0.9216



SECTION IV


24

Table IV-T2 shows the calculation of actual-to-expected ratios and the r-squared statistic for

AFSCME and Non-Represented members, and Chart IV-T2 shows the information graphically

along with the 90 percent confidence interval. The amount of termination experience for

AFSCME and Non-Represented members is fairly limited; only 30 members terminated between

January 1, 2011 and December 31, 2014. To improve the credibility of the data, we have

aggregated the experience of the past four years with that of the prior experience study (2007-

2010).

The data shows higher actual termination rates than expected under the current assumption. The

proposed assumption increases the assumed rates of termination and decreases the aggregate A/E

ratio from 119 percent to 104 percent. The r-squared also increases from 0.89 to 0.96.

Table IV-T2

Termination Rates - All AFSCME/Non-Represented Members, 2007-2014 DataTerminations Actual to Expected Ratios

Service Exposures Actual Current Recommended Current Recommended

0 71 7 9 7 82% 99%

1-3 338 27 19 27 139% 100%

4-7 289 17 14 14 118% 118%

8-10 141 10 6 7 177% 142%

11+ 502 5 8 8 66% 66%

Total 1,341 66 56 63 119% 104%

R-squared 0.8870 0.9565



SECTION IV


25

Chart IV-T2



SECTION IV

DEMOGRAPHIC ASSUMPTIONS DISABILITY RATES

26

This section analyzes the incidence of disability by the age of the employee. There is one set of

unisex assumptions for ATU and IBEW members at each age, as well as another set of unisex

assumptions for AFSCME and Non-Represented members at each age. The disability decrement

is only applied after members are eligible for disability benefits, and the decrements are not

applied to members who are eligible to retire with 20 or more years of service.

The amount of disability experience is fairly limited; only 35 disabilities have occurred during

the last four years for ATU and IBEW members and there were no disabilities among AFSCME

and Non-Represented members. To improve the credibility of the data, we have aggregated the

experience of the past four years with that of the prior experience study (2007-2010).

Table IV-D1 shows the calculation of actual-to-expected ratios and the r-squared statistic for

male ATU and IBEW members, and Chart IV-D1 shows the information graphically along with

the 90 percent confidence interval.

The data shows disability rates that are close to the current assumption. We are not proposing

any change to the disability assumption for ATU and IBEW members. The current assumption

has an A/E ratio of 94 percent. The r-squared is 0.55.

Table IV-D1

Disability Incidence RatesAge Disabilities Actual to Expected Ratios

Band Exposures Actual Current Recommended Current Recommended

25 - 29 7 0 - - 0% 0%

30 - 34 288 1 1 1 131% 131%

35 - 39 717 2 3 3 65% 65%

40 - 44 1,109 6 8 8 76% 76%

45 - 49 1,598 7 18 18 39% 39%

50 - 54 2,322 37 35 35 107% 107%

55 - 59 817 8 3 3 243% 243%

60 - 64 486 4 2 2 168% 168%

65 - 69 162 1 - - 0% 0%

70 + 9 0 - - 0% 0%

Total 7,515 66 70 70 94% 94%

R-squared 0.5546 0.5546



SECTION IV


27

Chart IV-D1

Table IV-D2 on the next page shows the calculation of actual-to-expected ratios and the r-

squared statistic for male AFSCME and Non-Represented members, and Chart IV-D2 shows the


The data shows that the current disability rates are higher than the current assumption. In this

context, however, the 52 percent A/E ratio doesn’t mean much; there was only one disability

among all AFSCME and Non-Represented members in the last eight years, while we predicted

two disabilities. We are not proposing any change to the disability assumption for AFSCME and

Non-Represented members.



SECTION IV


28

Table IV-D2

Chart IV-D2

Disability Incidence RatesAge Disabilities Actual to Expected Ratios

Band Exposures Actual Current Recommended Current Recommended

25 - 29 - 0 - - 0% 0%

30 - 34 7 0 0 0 0% 0%

35 - 39 37 0 0 0 0% 0%

40 - 44 103 0 0 0 0% 0%

45 - 49 154 0 0 0 0% 0%

50 - 54 314 1 1 1 82% 82%

55 - 59 96 0 0 0 0% 0%

60 - 64 119 0 0 0 0% 0%

65 - 69 34 0 - - 0% 0%

70 + 4 0 - - 0% 0%

Total 868 1 2 2 52% 52%

R-squared 0.0774 0.0774



SECTION IV


29

The current assumption is that one third of disabilities are assumed to be total and permanent

disabilities, with the remaining two thirds assumed to be occupational disabilities. Over the

aggregated eight year period, 25 of 67, or 37%, of disabilities were total and permanent. We see

no need to change the assumed proportion of total and permanent disabilities.



SECTION IV

DEMOGRAPHIC ASSUMPTIONS MORTALITY RATES

30

Post-retirement mortality assumptions are typically developed separately by gender for both

healthy annuitants and disabled annuitants. Pre-retirement mortality assumptions are developed

separately for males and females. Unlike most of the other demographic assumptions that rely

exclusively on the experience of the plan, for mortality, standard mortality tables and projection

scales serve as the primary basis for the assumption.

The Society of Actuaries recently completed an extensive mortality study and issued a set of

mortality tables named the RP-2014 mortality tables and a mortality improvement projection

scale named the MP-2014 scale. We used these tables as the basis for our analysis.

The steps in our analysis are as follows:

1. Select a standard mortality table that is based on experience most closely matching the

anticipated experience of AC Transit.

2. Compare actual AC Transit experience to what would have been predicted by the

selected standard table for the period of the experience study.

3. Adjust the standard table either fully or partially depending on the level of credibility for

AC Transit experience. This adjusted table is called the base table.

4. Select an appropriate standard mortality improvement projection scale and apply it to the

base table.

Mortality assumptions are developed separately for active employees, healthy annuitants, and

disabled annuitants. Within each of these groups, mortality rates are developed separately for

males and females. Unlike most of the other demographic assumptions that rely exclusively on

the experience of the plan, for mortality, standard mortality tables are used with standard

modifications so that the aggregate experience matches the plan’s experience.

Historically we have proposed assumption changes when the Actual-to-Expected (A/E) ratio for

the current assumption is less than 100%. However, for this Study we are recommending a

change in this approach going forward, where the proposed assumptions are intended to track

closely to actual experience (i.e. an A/E ratio close to 100%, but with a ratio slightly less than

100% still being reasonable). However, as described below, this new approach also includes an

expectation that the assumed mortality rates will automatically become more conservative each

year, since the actual mortality rates are also expected to decrease over time.

We also historically recommended the same or a related table for active employees and healthy

annuitants, which has been the current practice for AC Transit. However, recent mortality studies

by the Society of Actuaries and others have shown significantly lower rates of mortality for

active employees versus those of the same age who are no longer working, therefore this year we

have suggested using separate tables for active versus retired members.



SECTION IV


31

In the prior study, AC Transit elected to continue using the following assumptions:

Healthy active members, retirees and beneficiaries

The Combined Healthy Retired Pensioners (RP) 2000 tables published by the Society of

Actuaries with a one year set-forward for both male and female ATU and IBEW

members.

The Combined Healthy Retired Pensioners (RP) 2000 tables published by the Society of

Actuaries with no set-forward for male or female AFSCME and Non-Represented

members.

Occupational Disabled members

The Mortality Table for Disabled Participants Not Receiving Social Security Benefits

published by the Pension Benefit Guaranty Corporation (PBGC).

Total and Permanent Disabled members

The Mortality Table for Disabled Participants Receiving Social Security Benefits

published by the Pension Benefit Guaranty Corporation (PBGC).

Since the prior study, the Society of Actuaries' Retirement Plans Experience Committee (RPEC)

has released a new mortality improvement scale, Scale MP-2014. The mortality improvements

included in the most commonly used current projection scale - Scale AA - were found to produce

some unsatisfactory results in projecting mortality. Scale MP-2014 reflects more up-to-date data,

approximately 20 years more current than that used in the development of Scale AA, and it was

reviewed against a significant amount of data drawn from California public plan experience.

MP-2014 represents the Society of Actuaries’ most advanced actuarial methodology in

incorporating mortality improvement trends with actual recent mortality rates, by using rates that

vary not only by age but by calendar year – known as a two-dimensional approach to projecting

mortality improvements. Scale MP-2014 was designed with the intent of being applied to

mortality on a generational basis. The effect of this is to build in an automatic expectation of

future improvements in mortality.

This is a different approach from building in a margin for conservatism in the current rates to

account for the expectation that the same rates will be applied in future years, when mortality

experience has improved. Recent reports issued by RPEC suggest that using generational

mortality is a preferable approach, as it allows for an explicit declaration of the amount of future

mortality improvement included in the assumptions.

RPEC has also recently released a new set of base mortality rate tables – the RP-2014 tables,

which are intended to replace the RP-2000 tables and are based on a recent study of US defined

benefit plan mortality experience. However, RPEC excluded all public pension plan data in the

construction of these tables - including a large amount of California public sector data - because



SECTION IV


32

there were significant differences between the private and public sector retirement experience,

and the new tables are expected to be used by private sector plans to meet accounting and federal

funding requirements specific to private plans.

Fortunately, there are alternative sets of assumptions that have been developed that may serve as

a logical basis for developing mortality assumptions for AC Transit. RPEC’s most recent study

showed that applying mortality improvement scale MP-2014 to the RP-2000 tables yielded

similar life expectancies as the RP-2014 tables. As such, we are recommending the following

assumptions:

Active members

RP-2000 Combined Healthy Blue Collar mortality with generational improvements using

Scale MP-2014 for ATU and IBEW members.

RP-2000 Combined Healthy White Collar mortality with generational improvements

using Scale MP-2014 for AFSCME and Non-Represented members.

Healthy retirees and beneficiaries

RP-2000 Combined Healthy Blue Collar mortality with generational improvements using

Scale MP-2014 for ATU and IBEW members, adjusted by 130%.

RP-2000 Combined Healthy White Collar mortality with generational improvements

using Scale MP-2014 for AFSCME and Non-Represented members, adjusted by 120%.

Disabled members

RP-2000 Disabled Annuitant mortality with no adjustments.

As shown in Table IV-M1 on the next page, our proposed mortality rates for healthy annuitants

are close to recent experience. As we have done in prior experience studies, we have combined

the experience of the past four years with that of the prior four year period in order to have a

more robust dataset to review. To perform our comparisons, the RP-2000 rates (without

projection) were projected from their base year (2000) to the midpoint of the combined eight-

year study period (2011).

The match between the actual and expected experience across all statuses (active, retired, and

disabled) is close under the proposed assumptions: 92.7%. We are comfortable that the ratio of

actual to expected deaths is less than 100% within some subgroups, since as described above, the

use of generational mortality assumptions will automatically result in mortality assumptions that

decrease over time. In particular, the number of deaths among the disabled members lower than

expected and the number of deaths among AFSCME and Non-Represented retirees and

beneficiaries is somewhat lower than expected, but these groups have the smallest amount of

overall experience.



SECTION IV


33

Table IV-M1

Mortality Experience (2008-2014)

Current Proposed

Total Actual Actual Expected Expected Current Recommended

Exposures Deaths Rates Deaths Deaths A/E Ratio A/E Ratio

Actives

AFSCME/Non-Represented Male 1,371 5 0.36% 6.2 4.6 80% 109%

AFSCME/Non-Represented Female 981 1 0.10% 3.2 2.5 31% 40%

ATU/IBEW Male 8,544 15 0.18% 35.4 32.8 42% 46%

ATU/IBEW Female 5,234 4 0.08% 11.2 9.6 36% 41%

Total Actives 16,130 25 0.15% 56.0 49.5 45% 51%

Retired and Surviving Spouse

AFSCME-Non-Represented Male 1,353 44 3.25% 38.2 32.3 115% 136%

AFSCME-Non-Represented Female 900 20 2.22% 17.6 16.6 113% 120%

ATU/IBEW Male 5,674 207 3.65% 215.4 219.7 96% 94%

ATU/IBEW Female 2,872 71 2.47% 67.5 70.2 105% 101%

Total Ret/Surv 10,799 342 3.17% 338.7 338.9 101% 101%

Disabled

AFSCME/Non-Represented Male 50 4 8.00% 2.1 2.3 191% 175%

AFSCME/Non-Represented Female 14 0 0.00% 0.0 0.2 0% 0%

ATU/IBEW Male 898 26 2.90% 70.6 41.5 37% 63%

ATU/IBEW Female 586 15 2.56% 21.2 12.2 71% 123%

Total Disabled 1,548 45 2.91% 94.0 56.2 48% 80%

TOTAL 28,477 412 1.45% 488.7 444.5 84% 93%



SECTION IV

DEMOGRAPHIC ASSUMPTIONS OTHER DEMOGRAPHIC ASSUMPTIONS

34

SICK LEAVE

The current assumptions increase the liability for retirement benefits for ATU and IBEW active

participants by 0.75% to account for the impact of unused vacation and sick leave. In addition,

the data provided by AC Transit includes the accrued sick leave for each active member. We

compared the total sick leave for ATU and IBEW active participants to their total service, and

the 0.75% increase remains appropriate

FAMILY COMPOSITION

The current assumption is that 100% of active AC Transit active participants have beneficiaries

eligible for pre-retirement death benefits and that male spouses are three years older than their

wives. AC Transit does not provide spouse information for retirees. Therefore we have made a

conservative assumption – that all active members are married.

PAY FOR BENEFITS

AC Transit provided several benefit calculations of members who retired during the 2014

calendar year. The final pay used in those calculations to determine retirement benefits was then

compared to the pay Cheiron uses in its present value calculations. The pay in the benefit

calculations was found to be consistent with Cheiron’s current practices. No change in the

assumption is recommended at this time.

GENERAL LEAVE

We could not find any data that would warrant changing the General Leave assumption for Non-

Represented Participants at this time.

PLAN EXPENSES

An allowance of $750,000 for Plan administrative expenses was included in the annual cost

calculation in the prior valuation. The Plan’s administrative expenses in during the last two years

have averaged approximately $828,000. We recommend changing the Plan’s assumed

administrative expenses for 2015 to $825,000, increasing each year at the assumed rate of

inflation.



APPENDIX A

SUMMARY OF CURRENT ASSUMPTIONS

35

The demographic assumptions were adopted by the Board based on an experience study covering

the period from January 1, 2007 through December 31, 2010.

1. Rate of Return

The annual rate of return on all Plan assets is assumed to be 7.25% net of investment

expenses.

2. Cost of Living

The cost of living as measured by the Consumer Price Index (CPI) will increase at the

rate of 3.00% per year.

3. Plan Expenses

An allowance of $750,000 for Plan administrative expenses has been included in the

annual cost calculated.

4. Increases in Pay

Assumed pay increases for active Participants consist of increases due to cost of living

adjustments and those due to longevity and promotion.

Based on an analysis of pay levels and service for the ATU/IBEW Participants, we

assume that pay increases due to longevity and promotion will be 4% per year for the first

eight years of service and 0.5% per year thereafter.

Based on an analysis of pay levels and service for the AFSCME/Non-Union Participants,

longevity and promotion increases are assumed to be 0.5% annually.

In addition, annual adjustments in pay due to the cost of living will equal the CPI, for an

additional annual increase of 3.00%.



APPENDIX A


36

5. Mortality Improvement

No mortality improvement is explicitly assumed; however we build a margin in our

mortality assumption between the actual and expected number of deaths in order to

assume some future mortality improvements. The experience study report for the period

covering January 1, 2007 to December 31, 2010 contains a full description of these

margins.

6. Active Participant Mortality

Rates of mortality for the ATU/IBEW active Participants are given by the Combined

Healthy Retired Pensioners (RP) 2000 tables published by the Society of Actuaries with a

one year set-forward for both male and female members.

Rates of mortality for the AFSCME/Non-Union active Participants are given by the

Combined Healthy Retired Pensioners (RP) 2000 tables published by the Society of

Actuaries with no set-forward for either male or female members.

7. Retired Participant Mortality

Rates of mortality for the ATU/IBEW retired Participants, spouses and surviving spouses

are given by the Combined Healthy Retired Pensioners (RP) 2000 tables published by the

Society of Actuaries with a 1 year set-forward for both males and females.

Rates of mortality for the AFSCME/Non-Union retired Participants, spouses and

surviving spouses are given by the Combined Healthy Retired Pensioners (RP) 2000

tables published by the Society of Actuaries with no set-forward for either males or

females.

8. Occupational Disabled Participant Mortality

Rates of mortality for all occupationally disabled Participants are given by the Mortality

Table for Disabled Participants Not Receiving Social Security Benefits published by the

Pension Benefit Guaranty Corporation (PBGC).



APPENDIX A


37

9. Totally and Permanent Disabled Participant Mortality

Rates of mortality for all total and permanent disabled Participants are given by the

Mortality Table for Disabled Participants Receiving Social Security Benefits published

by the Pension Benefit Guaranty Corporation (PBGC).

10. Family Composition

All Participants are assumed to have beneficiaries eligible for pre-retirement death

benefits. Male spouses are assumed to be three years older than their wives.

11. Sick Leave Conversion

Projected retirement benefits for active ATU/IBEW Participants are increased by 0.75%

to allow for the conversion of unused sick leave to deferred compensation contributions

(and therefore pay) in the years before retirement.

12. Pay for Benefits

In order to allow for leaves of absence among current active Participants, in projecting

future benefits each Participant’s future pay is based on the highest five-year average pay

earned during all years preceding the valuation date.



APPENDIX A


38

13. Service Retirement

Rates of service retirement among ATU/IBEW Participants eligible to retire are given by

the following table:

Age Male Female

55 7.0% 11.7%

56 7.0% 10.4%

57 7.0% 9.1%

58 7.0% 9.1%

59 7.0% 7.8%

60 9.0% 15.0%

61 9.0% 15.0%

62 30.0% 30.0%

63 15.0% 15.0%

64 15.0% 15.0%

65 45.0% 54.0%

66 30.0% 40.0%

67 30.0% 40.0%

68 30.0% 40.0%

69 30.0% 40.0%

70 100.0% 100.0%

Rates of service retirement among AFSCME/Non-Represented Participants eligible to

retire are given by the following table on the next page:



APPENDIX A


39

Age

AFSCME/

Non-Represented

50 – 54 3.0%

55 – 59 10.0%

60 – 64 15.0%

65 – 69 25.0%

70+ 100.0%

14. Termination

Rates of termination for all Participants from causes other than death, disability, and

service retirement are as follows:

Years of

Service

ATU/

IBEW

AFSCME/

Non-Union

0 10.0% 12.0%

1 10.0% 7.0%

2-3 5.0% 5.0%

4-8 3.0% 5.0%

9 1.5% 5.0%

10+ 1.5% 1.5%

Assumed termination rates do not extend past eligibility for retirement benefits.



APPENDIX A


40

15. Disability

Rates of disability for male Participants are given by the following table:

Age

AFSCME/

Non-Union

ATU/

IBEW

20 0.000% 0.000%

25 0.000% 0.000%

30 0.0413% 0.132%

35 0.0638% 0.204%

40 0.1163% 0.372%

45 0.1875% 0.600%

50 0.2813% 0.900%

55 0.0220% 0.282%

60 0.0263% 0.336%

65 0.000% 0.000%

70 0.000% 0.000%

225% of the above ATU/IBEW male rates are assumed for the ATU/IBEW female

Participants.

150% of the above AFSCME/Non-Union male rates are assumed for the AFSCME/Non-

Union female Participants.

One-third of disabilities are assumed to be total and permanent and the remaining two-

thirds are assumed to be occupational.



APPENDIX A


41

It is assumed that ATU occupational disabilities will be reduced by 15% to reflect

worker’s compensation awards.

Disabled Participants are assumed not to return to active service.

16. Employment Status

No transfers between AFSCME/Non-Union and ATU/IBEW are assumed.

17. General Leave

Retirement liabilities for Non-Represented Participants with employment contracts with

the District are increased by 20% to reflect the effect of accumulated General Leave.

Liabilities for death, disability and termination benefits for these Participants are

increased by 10% for General Leave.



APPENDIX B

SUMMARY OF PROPOSED ASSUMPTIONS

42

The recommended assumptions have not yet been adopted by the Board and are proposed for

discussion with the Board at their August 5, 2015 meeting. The demographic assumptions are

based on an experience study covering the period from January 1, 2011 through December 31,

2014.

1. Rate of Return

The annual rate of return on all Plan assets is assumed to be 7.25% net of investment

expenses. The prior assumption was 7.375%.

2. Cost of Living

The cost of living as measured by the Consumer Price Index (CPI) will increase at the

rate of 3.00% per year.

3. Plan Expenses

An allowance of $825,000 for Plan administrative expenses has been included in the

annual cost calculated.

4. Increases in Pay

Assumed pay increases for active Participants consist of increases due to cost of living

adjustments and those due to longevity and promotion.

Based on an analysis of pay levels and service for the ATU/IBEW Participants, we

assume that pay increases due to longevity and promotion will be 4% per year for the first

eight years of service and 0.5% per year thereafter.

Based on an analysis of pay levels and service for the AFSCME/Non-Union Participants,

longevity and promotion increases are assumed to be 0.5% annually.

In addition, annual adjustments in pay due to the cost of living will equal the CPI, for an

additional annual increase of 3.00%.



APPENDIX B


43

5. Mortality Improvement

Mortality is assumed to improve in future years in accordance with the MP-2014

generational improvement tables.

6. Active Participant Mortality

Rates of mortality for the ATU/IBEW active Participants are given by the Combined

Healthy Retired Pensioners (RP) 2000 Tables with Blue Collar Adjustments published by

the Society of Actuaries with generational improvements using Scale MP-2014 and no

adjustments for either male or female members.

Rates of mortality for the AFSCME/Non-Union active Participants are given by the

Combined Healthy Retired Pensioners (RP) 2000 tables with White Collar Adjustments

published by the Society of Actuaries with generational improvements using Scale MP-

2014 and no adjustments for either male or female members.

7. Retired Participant Mortality

Rates of mortality for the ATU/IBEW retired Participants, spouses and surviving spouses

are given by the Combined Healthy Retired Pensioners (RP) 2000 Tables with Blue

Collar Adjustments published by the Society of Actuaries with generational

improvements using Scale MP-2014 and a 130% adjustment factor for both male and

female members.

Rates of mortality for the AFSCME/Non-Union retired Participants, spouses and

surviving spouses are given by the Combined Healthy Retired Pensioners (RP) 2000

Tables with White Collar Adjustments published by the Society of Actuaries with

generational improvements using Scale MP-2014 and a 120% adjustment factor for both

male and female members.

8. Disabled Participant Mortality

Rates of mortality for all disabled Participants are given by the Retired Pensioners (RP)

2000 tables for Disabled Annuitants published by the Society of Actuaries with

generational improvements using Scale MP-2014 and no adjustments for either male or

female members.



APPENDIX B


44

9. Family Composition

All Participants are assumed to have beneficiaries eligible for pre-retirement death

benefits. Male spouses are assumed to be three years older than their wives.

10. Sick Leave Conversion

Projected retirement benefits for active ATU/IBEW Participants are increased by 0.75%

to allow for the conversion of unused sick leave to deferred compensation contributions

(and therefore pay) in the years before retirement.

11. Pay for Benefits

In order to allow for leaves of absence among current active Participants, in projecting

future benefits each Participant’s future pay is based on the highest five-year average pay

earned during all years preceding the valuation date.



APPENDIX B


45

12. Service Retirement

Rates of service retirement among ATU/IBEW Participants eligible to retire are given by

the following table:

Age Males Females

< 20 Years of

Service

20+ Years of

Service

< 20 Years of

Service

20+ Years of

Service

55 – 59 3.0% 7.0% 5.0% 14.0%

60 – 64 8.0% 22.5% 11.0% 26.0%

65-69 25.0% 45.0% 35.0% 45.0%

70+ 100.0% 100.0% 100.0% 100.0%

Rates of service retirement among AFSCME/Non-Represented Participants eligible to

retire are given by the following table:

Age

AFSCME/

Non-Represented

50 – 54 15.0%

55 – 59 7.5%

60 – 64 17.5%

65 – 69 30.0%

70+ 100.0%



APPENDIX B


46

13. Termination

Rates of termination for all Participants from causes other than death, disability, and

service retirement are as follows:

Years of

Service

ATU/

IBEW

AFSCME/

Non-Union

0 15.0% 10.0%

1 7.0% 8.0%

2-3 3.5% 8.0%

4-8 3.5% 5.0%

9-10 3.5% 5.0%

11+ 1.5% 1.5%

Assumed termination rates do not extend past eligibility for retirement benefits.



APPENDIX B


47

14. Disability

Rates of disability for male Participants are given by the following table:

Age

AFSCME/

Non-Union

ATU/

IBEW

20 0.000% 0.000%

25 0.000% 0.000%

30 0.0413% 0.132%

35 0.0638% 0.204%

40 0.1163% 0.372%

45 0.1875% 0.600%

50 0.2813% 0.900%

55 0.0220% 0.282%

60 0.0263% 0.336%

65 0.000% 0.000%

70 0.000% 0.000%

225% of the above ATU/IBEW male rates are assumed for the ATU/IBEW female

Participants.

150% of the above AFSCME/Non-Union male rates are assumed for the AFSCME/Non-

Union female Participants.

One-third of disabilities are assumed to be total and permanent and the remaining two-

thirds are assumed to be occupational.

It is assumed that ATU occupational disabilities will be reduced by 15% to reflect

worker’s compensation awards.

Disabled Participants are assumed not to return to active service.



APPENDIX B


48

15. Employment Status

No transfers between AFSCME/Non-Union and ATU/IBEW are assumed.

16. General Leave

Retirement liabilities for Non-Represented Participants with employment contracts with

the District are increased by 20% to reflect the effect of accumulated General Leave.

Liabilities for death, disability and termination benefits for these Participants are

increased by 10% for General Leave.

Documents

AC Transit...The recommendation to change mortality assumptions has the largest impact on contribution rates. The recommended change to retirement rates also would increase contribution