Call to Order and Roll Call
The meeting of the Public Pension Oversight Board was held on Tuesday, May 27, 2014, at 1:00 PM, in Room 154 of the Capitol Annex. Senator Joe Bowen, Co-Chair, called the meeting to order, and the secretary called the roll.
Present were:
Members: Senator Joe Bowen, Co-Chair; Representative Brent Yonts, Co-Chair; Senators Jimmy Higdon and Gerald A. Neal; Representatives Brian Linder and Tommy Thompson; Robyn Bender, Tom Bennett, Jane Driskell, James M. "Mac" Jefferson, and Sharon Mattingly.
Guests: Representative Dennis Horlander; Lowell Reese, Kentucky Roll Call, Frankfort, Kentucky; Mary Helen Peter, Kentucky Retirement Systems Trustee, Louisville, Kentucky; Bradley Mackin and Carolyn Purcell, Legislative Research Commission, Program Review and Investigations Committee; among others.
LRC Staff: Brad Gross, Terrence Sullivan, and Marlene Rutherford.
Approval of Minutes
Mr. Bennett moved that the Minutes of April 28, 2014, be approved, which was seconded by Mr. Jefferson. The minutes were approved without objection.
2014 Kentucky Retirement Systems Actuarial Experience Study Results
Mr. Thielen, Executive Director, Kentucky Retirement Systems (KRS), introduced the presenters: Tom Cavanaugh, Consulting Actuary, Co-Founder and Chief Executive Officer, Cavanaugh Macdonald Consulting, LLC; Alisa Bennett, Senior and Consulting Actuary who performs much of the health actuarial projections; and Todd Green, Senior and Consulting Actuary. Mr. Thielen said that Cavanaugh Macdonald Consulting has been the actuarial consultant for the KRS since 2006.
Before presenting and discussing the experience study results, Mr. Cavanaugh provided background information as to the role of an actuary. In reviewing a pension plan there are three overriding policies: 1) a benefit policy; 2) an investment policy; and 3) a funding policy. The benefit policy reflects the statutory requirements that indicate who receives a benefit, when they receive it, and how much the benefit will be. The investment policy deals with the asset classes that are invested in by the plan and the proportion of each asset class to the total investment. The funding policy, overseen by an actuary, is a policy that indicates how much in funds will be contributed to the plan and when the contributions will be made. A truism with regard to any pension plan, regardless of plan structure, is that the dollars in, or the sum of the contributions made by the employees and employers, plus any investment income generated by the assets, must equal the dollars out, or the benefits paid to members and any expenses paid from the fund. An actuary makes projections or assumptions about future events based on the members in a pension system, the likelihood the members will receive a benefit, how long the member will receive the benefit, and the level of benefit, while taking into account the demographics of the system and the economics of the investments over time. These assumptions are combined in the evaluation of the plan annually, as of June 30 each year, based on data provided by KRS on the membership of each system, including both active and retired members. The benefit structure and assumptions are programmed from the data to project the system members through their working lifetime and through their actual lifetime to obtain the benefits that will be paid over that period, which is then discounted using an interest rate assumption to get a total liability of the system, or the total present value of the benefits to be paid. The total present value can be compared with the assets, and through an actuarial funding method or budgeting method, the actuary can determine how the plan should be budgeted on a logical basis over the years so that the funds are obtained that are needed to pay out the projected benefits.
He said the presentation would show contribution levels set for the system now and the levels that might need to be implemented if assumptions are changed as recommended by the experience study. The evaluation is performed on an annual basis and assumptions are made as to what may occur in the future, and then this needs to be reviewed periodically to decide whether or not the actual experience of the system is being followed. Experience studies are performed every five years to determine what has actually occurred with the members and to compare the experience to what was assumed would happen, and then make recommendations for adjustments, if necessary, to align the assumptions better with what is actually occurring with the members, with one example being changes in assumptions of mortality rates. Mr. Cavanaugh said that an actuary works on the liability side of a balance sheet, not the assets side, and does not deal with the investments themselves.
Mr. Green discussed the demographic assumptions and how the recommended changes were determined for the hazardous and non-hazardous plans in the Kentucky Employees Retirement System (KERS) and County Employees Retirement System (CERS), and in the State Police Retirement System (SPRS). He said the key findings on the demographic side resulted in recommendations for adjustment to the withdrawal and retirement rates, and the mortality decrements for all systems, as well as an adjustment to the disability decrements for the KERS hazardous and non-hazardous plans and the CERS non-hazardous plan to better match the experience. They also recommended adjustment for certain coverage assumptions for retiree healthcare benefits to better match the experience over the last five years. The key findings on the economic assumption resulted in recommended changes on price inflation, based on the Consumer Price Index (CPI), the investment rate of return, and wage inflation, which is the basis for determining salary scales since the retirement plans are pay based systems and salaries are projected from present day to retirement for all active members. The demographic assumptions used to support the recommended changes were: 1) rate of withdrawal, i.e. the withdrawal of funds prior to eligibility for retirement, death, disability, or the time when a member decides to retire when they become eligible for retirement; 2) rate of mortality, because once a benefit is triggered it is paid until the member dies; and 3) rate of salary increases.
Mr. Green indicated that the process for the experience study is labor intensive and compares the data files received for the last five years and used to perform annual evaluations with what occurs from one year to the next over the five year period. If that actual experience is different from the anticipated experience the actuary makes a recommendation for change. Mr. Green did not that the actuary does take into consideration that certain behaviors may be different from the assumption because of forces that may not repeat and therefore may not justify a change, such as more withdrawals than were anticipated because of a short term change in the economy. In that case, the actuary may not recommend a change immediately, but rather will monitor the experience for a longer period of time to determine if a recommended change is later needed.
The illustration for explaining the review of demographic assumptions was the KERS non-hazardous plan. The rates of withdrawal are used to determine the expected number of separations from active service that will occur prior to attaining the eligibility requirement for a retirement benefit. The current assumption utilizes a service based approach for the first five years of service and then an age based approach for years of service beyond five years. However, the assumption was changed to purely service related because withdrawal is more consistently correlated with service, rather than age. Based on the assumption for the experience period, a proposed rate increase is recommended for the withdrawal rate to better reflect the actual experience, because there were 12,930 expected terminations compared to 19,870 actual terminations during the study period. The recommended assumption rate was not increased completely to the experience rate, but rather was moved about halfway toward the actual experience rate and, in particular, the withdrawal rates were increased for the first fourteen years of service to better reflect when withdrawals were more likely to occur.
Responding to a question by Representative Yonts, Mr. Green said that the lack of salary increases for several years and the fact that employees are more mobile in their jobs was taken into consideration based on the last five years of experience with the retirement systems. He also said that previously, using the assumption of a service-based approach for only the first five years and an age based approach thereafter, it was determined that by switching to the service-based approach for all years of service going forward that the assumption will require very few changes because withdrawal is more correlated with service than age.
In reviewing the pre-retirement mortality assumption, a recommendation was made to update the pre-retirement mortality to the RP-2000 Combined Mortality Table because when comparing the actual versus expected deaths prior to retirement there were less actual deaths during active service than was expected. This recommendation will be consistent with the mortality assumption change being recommended for deaths after retirement. Mr. Cavanaugh explained that the RP-2000 is a title on a published mortality table that the Society of Actuaries updates based on experience. Mr. Green said the active mortality was much lower than expected, and it is recommended that the pre-retirement mortality rate be decreased to 50 percent of the RP-2000 rate for both males and females.
The disability retirement assumptions were reviewed for all the plans, and except for the SPRS, there were far fewer disability retirements than expected. The disability rates for KERS non-hazardous, hazardous, and CERS non-hazardous were therefore recommended to be lowered; however, the CERS hazardous and SPRS rates were unchanged because the comparison of actual versus expected disability retirements was not sufficient to recommend a change.
In reviewing service retirement assumptions (i.e. once an individual becomes eligible for retirement benefits how soon do they retire), there were fewer actual retirements than expected for the non-hazardous groups and more than expected for the hazardous groups. Therefore, a slight reduction in rates to reflect the experience was recommended at each retirement age.
Chair Bowen asked the panel to indicate when they are reviewing the assumptions what the impact of the recommended changes will be on the systems and whether the change will increase or decrease the liabilities of the systems.
Mr. Green indicated that the lowering of withdrawal rates would reduce the costs to the systems because fewer individuals would live to receive the benefit. However, lowering the pre-retirement mortality and the disability rates would increase the costs because more individuals would receive a retirement benefit. He also noted that a slight reduction in the retirement rate would also reduce costs slightly.
The post-retirement mortality rate assumption projects the percentage of retirees who are expected to die in a future year, which typically has the most significant impact on liability projections. Again, a comparison of actual versus expected retiree deaths was reviewed and the actual deaths exceeded the expected deaths over the five year period. Mr. Green indicated that the current assumption uses two different mortality tables, the 1983 Group Annuity Mortality table for members who retired prior to June 30, 2006 and the 1994 Group Annuity Mortality table for all other members retiring after that date. There is a recommended change in the mortality rate assumption to use the RP-2000 Combined Mortality table, projected to the year 2013 and using some setbacks, for example a one year setback for females, to improve the assumption. This same process was used for the mortality for members retiring on disability, using a four year setback for males. Experience indicated that overall more members died than expected during the study period at younger ages, while fewer members died than anticipated during the study period at older ages. Mr. Green indicated that the recommended changes to the post retirement mortality rate for both male and female, including disabled, would act to increase costs because the improved mortality table anticipates that people will live longer.
In response to a question by Chair Bowen, Mr. Green stated that the recommendations made are based on the experience study just completed for the last five years, 2008-2013.
Mr. Green then reviewed the salary scale assumptions. As part of the process for making these assumptions a projection for current salaries and future salaries is required. In the evaluation, the salary scale assumption was compared to what actually occurred, and the experience for all groups showed that the actual salary increases were less than expected. However, the actuary determined that the experience was influenced by unusual economic conditions, including a very low inflationary environment, and therefore no change to the merit component of the salary scale was recommended. Mr. Green stated that this assumption is composed of three components: 1) inflation; 2) productivity, or real wage increases; and 3) merit and promotion. The inflation and productivity components are combined to produce the assumed rates of wage inflation. The rate represents the across the board average annual increase in salaries shown in the experience data. The merit component includes the additional increases in salary due to performance, seniority, and promotions.
Responding to comments and questions by Representative Yonts concerning salaries and the actuarially required contribution (ARC) budgeted by the General Assembly and the fact that there are approximately 4,000 fewer employees currently than in prior years, Mr. Green stated the experience study could not determine if the freezes, cuts, and furloughs because of the economic environment were permanent. Further, Mr. Green stated that if the next experience study reflects this same trend a recommended change would be made in the merit component of the salary scale, and he noted that the recommended reduction in the CPI inflation rate translated to an overall lower salary scale. Mr. Green said that lowering the inflation component of the salary scales would reduce the ARC, or the liability. Mr. Cavanaugh pointed out that data for the last fifty years was evaluated, not just a two or four year time period.
Responding to Chair Bowen, Mr. Cavanaugh indicated that recommendations on the assumptions are made with the experience study, and each valuation looks at the impact of the actual compared to expected experience in the different categories, such as salary scale, retirements, withdrawals, disabilities, which are included in the annual valuation report. He said that the funding method or process utilized in recommending the actuarially required contribution (ARC) rate is a self-correcting mechanism, and that if the assumptions are too liberal losses will result that will increase the contribution rate, and if the assumptions are too conservative actuarial gains will result that will lower the contribution rate, so the methodology self-corrects from year to year, and once every five years the experience is reviewed to determine if any adjustments are needed in any of the assumptions.
In response to Representative Yonts concerning the recommended assumption change for wage inflation globally from 4.5 percent to 4 percent and the effect to the ARC, Mr. Cavanaugh indicated that the recommended decrease of the wage inflation alone would lower the ARC. He explained that this is the case because salaries for each individual is 0.50 percent lower than it would have been if the change was not made, so the projected salary in the future is lower, the benefits to be paid out would be lower, which lowers the liability and the ARC. Mr. Green also said that the analysis of salary increases yielded an actual versus expected ratio of 97 percent and that a ratio less than 100 percent indicates that salary increases were less than anticipated by the current assumption.
Mr. Green discussed the impact of the assumption changes from a demographic point of view. By changing the assumptions, mainly the withdrawal rate, it is projected that both males and females will leave the systems at a higher rate than under the current assumptions. When viewed in terms of mortality, the proposed change to the retirement tables used show that both males and females will die at a lower rate than previously assumed, corresponding to longer lives for both groups.
Mr. Cavanaugh stated that the demographic factors are reviewed for the pension side of the plans, and particularly the salary increase assumption impacts the pension plans only, but he noted that all the assumptions factor into the healthcare portion of the liabilities as well. There are other assumptions made that are healthcare specific, which were discussed by Ms. Bennett.
She said the recommended economic assumptions were developed in accordance with the Actuarial Standard of Practice No. 6 issued by the Actuarial Standards Board that is specific to retiree group benefit obligations or insurance plans. Two assumptions reviewed specifically to healthcare are the healthcare trend assumption, which is the overall rate of increase of healthcare spending, and the participation assumption, which is important because for those members participating before 2003 they are offered a healthcare plan with a subsidized premium when they retire.
The short term healthcare trend rates are set on an annual basis based on information and data with an ultimate trend rate of 5 percent and that there appeared to be nothing unusual about the KRS medical plans that would cause a long-term trend that is higher or lower than what is used for the calculation. Each year it is assumed that healthcare will cost more than the previous year and this assumption is reviewed each year and the trend assumption reset if needed. There is no change recommended and the actuary will continue to update healthcare trends annually and base the healthcare trends on the KRS experience and demographics, while taking into account the projected trend from external sources.
The current assumption for KERS non-hazardous is that 90 percent of members who retire and are eligible for a health benefit will take the benefit. However, current refinement to the data showed that those who declined coverage were typically those members who had less years of service because the subsidy is less for those members. Proposed healthcare participation rates are as follows: 1) for those with service under ten years, reduced to 50 percent; 2) for those with ten to fourteen years of service, reduced to 75 percent; 3) for those with fifteen to nineteen years of service, remained the same at 90 percent; and 4) for those with twenty years or more service, increased to 100 percent. Similarly, the CERS non-hazardous plan, which is currently assumed to have an 85 percent participation rate for all employees, were recommended to be changed to a similar sliding scale based on years of service. There were also proposed decreases from 100 percent to lesser percentages for the KERS and CERS hazardous plans, except for those members with twenty plus years of service, which remained at 100 percent. There is no recommended change to the assumed participation rate for the SPRS, which is set at 100 percent. She also said there were no recommended changes in participation rates for other service retirement tiers, and the 100 percent assumption for these groups would continue. These recipients are those who were hired or became participants after 2003 and who receive flat dollar benefits based on years of service, and the assumption is that the individual will take whatever benefit is earned by the time they retire.
She also discussed the deferred vested member healthcare participation rate for members hired before July 1, 2003. These are individuals who left with a vested right to a benefit. For the KERS non-hazardous members the current assumption is 90 percent and for hazardous duty the assumption is 100 percent. The actual data reflected that for individuals coming back to get the benefit at a later date only about half are taking the benefit, and therefore a reduction of the assumption to 50 percent was recommended for both the non-hazardous and hazardous plans.
Ms. Bennett said that hazardous duty covered retirees have the option of electing spouse coverage, and the current assumption was that 100 percent of the retirees would elect to cover a spouse. However, in reviewing the actual data, recommendations are made to lower the assumption for the KERS hazardous plan to 50 percent, and for the CERS and SPRS plans to 75 percent, and that lowering the assumptions lowers the liabilities because there are fewer spouses covered under the plan.
Mr. Cavanaugh then discussed the economic assumptions used in performing the actuarial valuation for the three plans. Three assumptions were reviewed: 1) price inflation; 2) investment return; and 3) wage inflation. He noted that the real rate of return is a component of the overall investment return assumption, and no recommended change is made for that assumption, but a recommendation is being made to the price inflation assumption from 3.5 percent to 3.25 percent and also a reduction in the real wage growth assumption from 1 percent to 0.75 percent so that the overall wage inflation goes down from 4.5 percent to 4 percent. He said the reduction of 50 basis points in wage inflation affects the individual salary scale and is also used because the unfunded accrued liability that is generated in each year’s valuation is financed as a level percent of payroll over a specific period of time, and that Senate Bill 2 re-established the period of time to 30 years with the 2013 valuation and goes down one year every year. In financing the unfunded accrued liability, the assumption is that payroll will grow at the wage inflation assumption, currently at 4.5 percent per year and recommended to be 4 percent per year.
Responding to a question by Chair Bowen, Mr. Cavanaugh indicated that the impact on the contributions and liabilities do reflect healthcare inflation and that this trend rate is reviewed each year and adjustments are recommended on an annual basis.
In response to comments and questions by Mr. Jefferson concerning the amortization of the unfunded liability and the percentage of payroll method, Mr. Cavanaugh indicated that if there was a flat or negative payroll growth it would depend on what caused the flat or negative payroll growth. If there is flat payroll growth because of a reduction in members but employees received pay increases, the liabilities will still increase as assumed but the payroll base is not there to finance the unfunded liability and the rate must increase in order to collect the same amount of dollars to pay off the unfunded liability. He also said that if there are still the same number of active members but there were no pay increases, the liabilities do not grow as assumed so there is a reduction in liability from what was predicted because there were no pay raises but that is offset somewhat by the fact that the payroll does not grow and that puts upward pressure on the rate. He said that over the last five years there has been a combination of reduction in members but also gains each year in salaries, so that when the annual gain/loss analysis is performed gains have been reflected, or reductions in the liability each year, because employees are not receiving pay raises. However, because the overall payroll is not increasing the rate is getting pushed up. He said that if the payroll is not growing the way it is assumed it will grow, if the rate gets higher but the payroll does not increase, the total dollars will not increase.
The price inflation assumption is 3.5 percent and is recommended by the actuary to be reduced to 3.25 percent. In analyzing the data, the average rates of inflation were determined by measuring the compound growth rate for the CPI for all urban consumers (CPI-U) over various time periods. Over the shorter historic periods, the average annual rate of increase in the CPI-U has been below 3 percent and the years of high inflation from 1973 to 1982 had a significant impact on the averages. A reasonable range for price inflation assumption is between 2 percent and 4 percent. Since 1926, the average annual rate of inflation was 2.99 percent. Mr. Cavanaugh stated that looking over long periods of time the average has been about 4 percent and that because there are a lot of inflation pressures built into the economy now, they are not comfortable reducing the long term inflation assumption, so that the real rate of return stays at 4.25 percent, which results in a 7.5 percent investment return assumption going forward.
Responding to a question from Mr. Jefferson concerning price inflation and various points of data that suggest a lower rate of inflation than the 3.25 percent that is recommended, Mr. Cavanaugh stated that a recommendation has been made to the KRS Board that the economic assumption be reviewed every two years rather than waiting five years, and that the decrease in the price inflation assumption from 3.5 percent to 3.25 percent took into consideration the review every two years to determine if an adjustment would need to be made. He said there are articles relating to the bond market that lend concern dealing with heavier inflation over longer periods of time, and with the Social Security actuary looking at a reasonable range of 1.8 percent to 3.8 percent that he was comfortable recommending a reduction to 3.25 percent with the expectation that this assumption would be reviewed every two years, which would allow the recommendation to be lowered again if warranted. He said the investment rate of return has not been good over the last five years and on average the rate of return has been under 5 percent and that the short term rate is not indicative of what would be expected in the long term results.
When performing the analysis for the investment return assumption, the actuary uses the capital market assumptions used by the KRS investment consultant and the policy of target asset allocations to develop stochastic projections of returns and groups them by percentiles over different time spans. The capital market assumptions used to develop the reasonable range for the real rate of return are net of investment expenses therefore a separate assumption of investment expenses is not necessary. The real rate of return between the twenty-fifth and seventy-fifth percentile is between 2.63 percent and 4.96 percent, with the fiftieth percentile being at 3.79 percent real return. Combining the real rate of return, the inflation assumption, the administrative expenses that are paid separately, and the capital market assumption returns from investment consulting the net investment return ranges from 5.88 percent to 8.21 percent, with the fiftieth percentile at 7.04 percent. He said normally a recommendation would be at or around the fiftieth percentile results; however, there are some mitigating issues that lead to a slightly upward adjustment to the recommendation. Specifically, the capital market assumptions are based on a ten year time period versus the system’s lifetime, and some investment consultants that have produced capital market assumptions on longer timeframes project higher returns than when viewed on a ten year time horizon. Also, historical returns have been higher than the fiftieth percentile number suggests, and the capital market assumptions do not include any added return due to active management or other asset deployment strategies. Taking into account the historical analysis, the peer group analysis, and the capital market assumption analysis, it is recommended that the investment return assumption be set at 7.5 percent, which is slightly greater than the fiftieth percentile when compared to the peer groups for the plans, putting the KERS non-hazardous plan at the 64th percentile, and the KERS hazardous, CERS non-hazardous and hazardous, and the SPRS plans at the 61st percentile of the peer groups. Additionally, the actuary makes the further recommendation that the investment return assumption be reviewed every two years so that it is established before the budgeted rates are set for the state.
In response to questions and comments from Mr. Jefferson concerning the investment return assumption recommendation being above the 50th percentile, especially given the real tangible limitations on the funded status of the plans, Mr. Cavanaugh indicated that the least well funded plan, the KERS non-hazardous plan, has a more conservative asset allocation and that a 7.25 percent investment return assumption could have been recommended, leaving the 7.5 percent assumption for the other plans, but it was felt that this was reflected in the asset allocation. In response to a follow-up question, Mr. Cavanaugh stated that the most significant of the economic assumptions is the real rate of return assumption because, for example, if the inflation assumption was 2.5 percent rather than 3.25 percent, the investment return assumption of 7.5 percent would decrease to 6.75 percent for the rate of return. This would impact the retiree liability because there is no offset and the benefits are discounted at a lower interest rate, so the liability would increase for retirees. However, for the active members there is an offset because the salary scale will go down by the 75 basis points and salaries will be projected at a lower rate and then will be discounted at a lower rate, and so in this way the inflation assumption is not as sensitive as the real rate of return assumption. In other words, the real rate of return assumption has an impact on the calculation for all liabilities, even without any change to any other assumption, whereas the other assumptions often are offset by another resulting change. Mr. Cavanaugh again emphasized that the recommendation to review the economic assumption every two years is very important so that these issues can be addressed more frequently than every five years.
Mr. Cavanaugh also discussed the wage inflation assumption. The current assumption of 4.5 percent in the wage inflation number is 1 percent above the base price inflation assumption. Comparing wage inflation with price inflation the annual real wage growth has averaged 0.62 percent. The social security actuary bases the 75 year costs projections on a national wage growth assumption of 1.1 percent greater than the price inflation assumption of 2.8 percent. Therefore, to be consistent with historical data, and considering the activity in wages in the public sector, a lower number of 0.75 percent real wage growth was reasonable, and therefore the percent above price inflation is recommended to be reduced to 0.75 percent, with a recommendation that the 4.5 percent of payroll growth assumption be lowered to 4 percent, of which 25 basis points is from the price inflation reduction and 25 basis points is from a reduction in the real wage growth. He said this impacts both the salary scale itself, meaning the individual salary increases would be lowered by the 50 basis points, which lowers the liability, and the payroll growth is not as big, and therefore the contribution rate as a percentage of that payroll has to be higher to finance the unfunded accrued liability.
Mr. Cavanaugh stated the financial impact of the recommended changes for each plan was reviewed, and the unfunded accrued liability, funded ratio, and the employer contribution rate is shown both before the recommended changes to the assumptions and after, combined, from a pension and insurance standpoint. So for example, for the KERS non-hazardous plan, the contribution rate for this biennium that is built into the budget is 38.77 percent of payroll, and after the changes the employer rate would be 41.36 percent of payroll. Similarly, for the KERS hazardous plan the employer rate as a percentage of payroll before the changes is 26.34 percent and after is 26.9 percent, and he noted that this modest change in the rate is due to the changed assumptions in the health insurance participation rate. The CERS non-hazardous employer rates have a modest increase from 18.1 percent to 18.8 percent, and the CERS hazardous rate decreases from 35.7 percent to 32.03 percent. The SPRS employer contribution rate increases from 75.76 percent to 83.2 percent, with all numbers reported as a percentage of payroll.
Responding to a comment and question by Representative Yonts relating to the investment return assumption in the context of the unfunded liability and why a lowering of the assumption is recommended, Mr. Cavanaugh indicated that the supporting data and recommended changes would indicate that there needs to be an increase in the contribution rate, but that this increase would not be effective until the next biennium. Mr. Thielen pointed out that action on the recommendations as a result of the experience study report were deferred by the KRS board and will probably be taken up at the September meeting, with the understanding that any new assumptions would not take effect until 2017. He said there was concern expressed among some of the board members that the budget had been set for the next two years and cannot be changed and that any change if the assumptions were adopted would not have any impact until fiscal 2017. In other words, the board considered the economic and political ramifications in making a change now after the General Assembly had funded 100 percent of the actuarially required contribution and understood that adopting the changes at this time could be problematic.
Chair Bowen asked Mr. Thielen to explain his reference to political ramifications on the KRS board’s decision to defer action on the recommendations. Mr. Thielen clarified that this was a personal statement on his part and that it was his opinion that the board has considered the pressures placed on the General Assembly to increase its commitment to the retirement funds and since any action will not impact the current budget it still has time to review the recommendations and make any changes and provide those changes to the General Assembly in the 2016 session. Chair Bowen indicated that decisions by the General Assembly need to be made in an expeditious manner, and he expressed concern of the KRS board deferring action on the recommendations.
Responding to a question by Mr. Jefferson relating to the funding impact for the CERS plans, Mr. Thielen stated that the KRS board would continue to set the contribution rate for the CERS plans each year based on the valuation from the previous fiscal year, so if the board adopts changes in the future it will not affect the contribution rate adopted in November.
Mr. Peden, the Interim Chief Investment Officer at KRS, stated that one of the factors in the KRS board’s decision to defer taking action at the May meeting was to afford the Investment Committee an opportunity to engage RVKuhns during the asset liability study and to conduct some preliminary asset allocation work, so that if there were any changes in asset allocations going forward for the next five years that were significantly different from the past five years that this would lead the conversation on the expected return.
In response to a question by Representative Yonts as to when the assumptions were last changed, Mr. Thielen stated that an experience study was done from July 1, 2000 through June 30, 2005 by Siegel and Company, the actuary at the time, and that study was followed up by an experience study covering the period July 1, 2005 through June 30, 2008. However, because of the 2008 recession and the fact that the economic environment had changed considerably, KRS wanted Cavanaugh Macdonald to re-do the experience study over the shorter period, which was done for a three-year period covering July 1, 2005 through June 30, 2008. During the last thirteen years there have been three experience studies. In response to a follow-up question as to how recommendations might have changed following those studies, Mr. Thielen indicated that the assumed rate of return was 8.25 percent until 2006 and was changed as a result of the 2000-2005 experience study to 7.75 percent. Also, the mortality tables were changed for retirees and the cost inflation rate was changed as a result of past experience studies.
Chair Bowen asked Mr. Thielen to compare the difference in the returns and viability between the Kentucky Teachers Retirement System (KTRS) and the KRS. Mr. Thielen stated the two systems have different programs and that KRS has built an investment portfolio designed for the long term, and the thirty year return on the KRS pension funds has been 9.55 percent, which is above the 7.75 percent assumption and is better in the long term than the KTRS return. He also noted that the KTRS program has different cash flow needs, different complexities, and different healthcare benefits from the KRS plans, and that a one year comparison does not reflect the full picture. Mr. Peden stated that the KRS Investment Committee works with the general consultant to develop or perform an asset liability study, which matches up asset cash flows with liability cash flows, and the investment return, is a function of the asset allocation set by the Investment Committee every three to five years. He noted that asset allocation drives the investment performance more than any other factor, and that the portfolio is built to average a 7.75 rate of return without taking additional price risks – including volatility or standard deviation – and that will stand up to difficult market periods. He also pointed out that CERS has 20 percent of the portfolio in U. S. equity and 20 percent in non U. S. equity, and that last calendar year U. S. public equity performed between 30 percent and 35 percent. By contrast, he noted that KTRS has up to 50 percent or more in that asset class, and that might be good for that fund, but that he is building a portfolio to meet the goals and constraints of the KRS. Mr. Thielen pointed out that KRS is not in the environment to be the best public pension plan for investment return, but to match the assumed rate of return or better so that obligations can be met based on the specific needs of the plans and at a low risk, noting that this strategy may not give the KRS plans the highest returns but may protect the plans from having the lowest returns in bad years.
In response to questions from Representative Yonts concerning investment strategies, Mr. Peden indicated that the asset allocation is set every three to five years, and Mr. Thielen stated that an asset liability modeling study has been done at least every five years and the external investment consultant works with the actuaries to determine what the allocation should be, which is a very expensive process. He also noted, with assistance from Mr. Peden, that a liability modeling study was completed in 2010 for KERS and 2011 for CERS. Mr. Peden indicated that staff has the ability to make a recommendation to the Investment Committee if investments are not performing well or there is an event that staff is uncomfortable with, and that it would be highly unlikely to completely exit an asset class, but it could be recommended that it be pared back. Mr. Peden stated the Investment Committee meets quarterly or more often if needed, and if changes are made that the full KRS Board would need to ratify the decisions made by the Investment Committee at a quarterly board meeting.
Monthly Investment Update
Mr. Peden said that the audited numbers for April reflect that the pension plan was up 0.47 percent versus the benchmark of 0.53 percent, and for the fiscal year it is up 12.11 percent versus its benchmark of 11.54 percent, and the pension plan performance over ten years was 6.75 percent versus the benchmark of 6.76 percent, with the ten year numbers including the recent financial crisis. The insurance asset allocation performance was up 0.46 percent versus the benchmark of 0.53 percent for the month, and for the fiscal year is up 11.54 percent versus the benchmark of 11.53 percent, and the ten year return is 6.31 percent versus the benchmark of 6.49 percent.
Chair Bowen asked that this information be provided to the PPOB in the materials folder each meeting. Representative Yonts also asked that in addition to the monthly report that a comparison be included for the last six and twelve month periods.
Understanding that it has been necessary to cash out assets to pay benefits, Representative Yonts asked how expensive it is to the system to cash out assets. Mr. Peden indicated that the KRS attempts to time the cash out with when the benefit payments to retirees are sent, which is around the fourteenth of the month, and assets are invested as long as possible. He also noted there is always a cash target that is built into the asset allocation and depending on the system the demand for cash is still needed, especially in the KERS non-hazardous plan. Conversely, CERS has not had to raise as much cash over the last twelve months because of good investment returns, and the cash flow has improved in the insurance funds as well. Mr. Thielen noted that at one time about $75 million in assets had to be sold in the KERS non-hazardous plan monthly to make benefit payments.
Responding to a question by Representative Linder as to the makeup of KERS hazardous employees and why spouse coverage is included in the hazardous plan’s health benefit, Mr. Thielen indicated that spousal coverage has existed for many years and that it is a rich benefit, but he did not have knowledge as to how or when the decision was made to include the benefit in statute. He also stated that certain employees in the Department of Fish and Wildlife, any law enforcement personnel not covered by the SPRS, and corrections or prison personnel that have contact with inmates are in the hazardous plans and there are approximately five thousand total members in the plan.
Chair Bowen stated that the PPOB would not meet again until August 28 and at that meeting the board would be reviewing the KRS board’s semi-annual review of investments. In the September meeting the board would review the retiree health benefits. In October, the board would discuss legislative recommendations. In the November meeting the board would have an overall review of the discussions to date, and in December the board would be reviewing the 2014 KRS annual audit.
Mr. Thielen advised the board that the KRS board has expressed a willingness to meet with the PPOB at a meeting in the future.
With no additional business to come before the board, the meeting adjourned at approximately 3:00 p.m.