HOUSE BILL NO. 78 "An Act relating to the Public Employees' Retirement System of Alaska and the teachers' retirement system; providing certain employees an opportunity to choose between the defined benefit and defined contribution plans of the Public Employees' Retirement System of Alaska and the teachers' retirement system; and providing for an effective date." 1:44:00 PM DAVID KERSHNER, PRINCIPAL CONSULTING ACTUARY, GALLAGHER, introduced the PowerPoint presentation "State of Alaska Retirement Systems" dated April 29, 2025 (copy on file). The presentation detailed the findings of the fiscal note analysis related to HB 78. He explained that Buck had been the plan actuary for the state's retirement system since 2006, and that he had personally been involved since 2015. He stated that in 2023, Buck was acquired by Arthur J. Gallagher and Company, which was a global risk management, insurance, and consulting firm. He added that the Buck name was officially retired in July of 2024 and the firm was now known as Gallagher. Mr. Kershner continued to slide 2 and explained that he would provide an overview of how the Public Employees' Retirement System (PERS) and the Teachers' Retirement System (TERS) were funded under Alaska statute. He noted that the subtleties were sometimes lost when trying to understand why costs were projected to increase. He relayed that he would then describe how the assumptions were set for the annual evaluations adopted by the Alaska Retirement Management Board (ARMB) for the contribution rates. He would then discuss the assumptions specific to HB 78 and how the assumptions related to the ongoing assumptions for the valuations. He stated that he would touch briefly on the risk-sharing provisions within HB 78 and other economic considerations. He would continue to summarize key information from the March 24, 2025, fiscal note letter (copy on file), which provided specific numbers related to HB 78. If time allowed, he would present an example beginning on slide 31 that would help illustrate why costs were projected to increase. Mr. Kershner continued to slide 4 and explained that PERS and TRS were funded though mandates in Alaska statutes. He stated that there were two main sources of funding for the retirement systems. The first source was investment income on the invested assets, and the second source was contributions. He clarified that there were three types of contributions: employer contributions, member contributions, and state contributions. Mr. Kershner described the employer contributions under PERS. He explained that non-state employers contributed 22 percent of total payroll, which included both defined benefit (DB) and defined contribution (DC) retirement members. He noted that under PERS, the state as an employer contributed not only the 22 percent, but the full actuarial rate based on the total payroll of its employees, which was just under 50 percent of the total PERS payroll. He relayed that under TRS, employers contributed 12.56 percent of the total payroll. Mr. Kershner reviewed the member contributions. He explained that under PERS, peace officers and firefighters contributed 7.5 percent of pay, while all other PERS members contributed 6.75 percent of pay. He added that TRS members contributed 8.65 percent of pay. He explained that the actuarially determined contribution for the DB plan consisted of two components. The first was the normal cost, which was the actuarial cost of benefits expected to accrue in the upcoming year as active members earned one more year of service. He stated that the second component was the past service cost, which represented paying down the unfunded liability over a 25-year period as defined in statute. He emphasized that the two components were calculated separately for the pension trust and the health care trust before being combined, which would be shown later in the presentation. 1:49:11 PM Mr. Kershner continued on slide 5 and shared that the normal costs for the DB plan were covered by member contributions and employer contributions. He explained that employers also contributed the costs for the DC retirement plan, which included four benefits: occupational death and disability benefits, health care benefits, employer defined contribution amounts of 5 percent for PERS members and 7 percent for TRS members, and 3 percent health reimbursement account contributions. Mr. Kershner stated that a portion of the employer contribution also went toward the DB past service cost, which paid down the unfunded liability. He explained that the employer contribution portion was calculated as the total employer contribution minus the amount paid toward the normal cost and minus the DC retirement plan costs. He noted that the employer contribution portion was important because it determined how much the state would have to pay to make up the difference. Mr. Kershner explained that the portion of the DB past service cost not paid by employers was paid by the state through additional state contributions. He added that additional state contributions could increase for two reasons: if the cost of the underlying benefits changed, or if there were changes in the distribution of employer contributions. He noted that the second reason was a more subtle factor that was sometimes not fully understood and could have a significant impact on projected state contributions. Mr. Kershner continued to slide 6 and stated that the past service cost paid down the unfunded liability. He explained that the unfunded liability was the difference between the actuarial accrued liability and the actuarial value of assets. The actuarial accrued liability was the present value of future benefits expected to be paid for current members, both active and inactive, that was attributable to service as of the date of calculation. The actuarial value of assets was a smoothed, market-related value. He noted that the market value could fluctuate significantly, as had been seen over the last few weeks. If the market value of assets were used to determine contributions, then as the market value moved up and down, contributions would also rise and fall. He explained that an actuarial value of assets was used to mitigate volatility. He reiterated that the actuarial value of assets was a smoothed market value in which gains and losses were recognized over a five-year period. Each year that there was a market gain or loss, 20 percent of it was recognized in the smoothed value, and the remaining 80 percent was recognized over the subsequent four years. Mr. Kershner explained that annual valuations always resulted in unexpected changes in the unfunded liability, which could arise from five sources. The first was asset returns for the year and how the returns compared to the expected return, which was currently 7.25 percent. The second was changes in the data that drove the liabilities. Each year, updated data was received and compared against prior assumptions, which produced the liability experience. A third source was contributions that were greater or less than the actuarially determined contribution. For example, the state contributed $3 billion to PERS and $2 billion to TRS in 2015. Since the contributions exceeded the actuarially determined contribution, the contributions produced a contribution gain, which reduced the unfunded liability for the year. Mr. Kershner stated that actuarial assumptions also affected the unfunded liability and any change in assumptions impacted the accrued liability, either positively or negatively. Plan provisions could also affect the unfunded liability. Although most changes were prospective, such as creating a new tier of benefits, some changes that affected service prior to the date of change also altered the unfunded liability. Mr. Kershner continued to slide 7 and explained that the current funding methodology for PERS and TRS had been established in statute in 2014. One key element was that the amortization method for the unfunded liability changed from a level dollar approach to a level percentage of pay approach. The level dollar approach was similar to a home mortgage, where the outstanding principal was paid down in equal amounts over 15 or 30 years. Under level percentage of pay, amortization amounts were assumed to increase as payroll was projected to increase. He explained that when compared, level percentage of pay amounts were lower in the early years, crossed over after about ten years, and then became greater than the level dollar amounts. On a present value basis, both methods equaled the unfunded liability, but the patterns differed. Mr. Kershner explained that in 2014, the amortization period was reset to 25 years, meaning that the PERS and TRS trusts were expected to be fully funded by 2039. He noted that another change was that contribution rates were no longer set using the most recent valuation but instead were based on a two-year lag. For example, the FY 26 contributions that were going into effect in July of 2025 were set based on the 2023 valuations, which had been adopted in September of 2024. He added that the actuarial value of assets was reset to market value in 2014, with the five-year smoothing method implemented on a prospective basis. He noted that the 20 percent market value corridor was eliminated. 1:56:38 PM Mr. Kershner continued to slide 8 and shared that in 2018, ARMB changed the amortization method and layered amortization, which meant that every year's change in the unfunded liability was amortized over a separate 25-year period. Each unexpected change in the unfunded liability created a new layer that was then amortized over 25 years. He emphasized that the only reason for making the change was to help control volatility of the state's contributions because the original 25-year period would have become shorter over time. Mr. Kershner reported that the effect of introducing layered amortization was that the original date of 2039, which was when the plans had been expected to be fully funded, was extended further into the future. He indicated that because every new layer restarted the 25-year clock, the most recent valuations projected that the pension trusts for both PERS and TRS would be 100 percent funded by 2048. He added that the health care trusts for both PERS and TRS were already over 100 percent funded and were expected to remain at a similar funding level going forward. Mr. Kershner noted that the next two slides were not related to HB 78 and were provided for informational purposes. He remarked that in recent House Finance Committee and Senate Finance Committee meetings, there had been many questions as to why the unfunded liabilities had continued to grow since the $3 billion contribution. He explained that the team at Gallagher had prepared charts to present to ARMB and believed the charts would be helpful for the committee as well. Mr. Kershner moved to slide 9, which showed the changes in the unfunded liability for the PERS pension since 2014. He noted that slide 10 showed the same information for TRS. He explained that the first column [column A] reflected the gains and losses on the market value of assets. For example, if the assumed return was 7.25 percent and the plan market return was 10 percent, there had been a gain because the plan earned more than expected. Mr. Kershner elaborated that column B showed the associated smoothed actuarial value gains and losses, based on the market gains and losses from column A. He noted that the year ending June 30, 2021, was exceptional as the plans achieved returns above 30 percent, which was highly unusual. However, there was a $1.6 billion loss in 2022, which eliminated nearly all of the prior year's gains. He explained that the change illustrated how volatile market value gains and losses could be from year to year. Mr. Kershner stated that column C showed the gains and losses on the liabilities due to unexpected changes in participant data from one year to the next. He noted that column D reflected the gains and losses due to contributions that were either greater or less than the actuarially determined contributions. For example, PERS had an $835 million contribution gain in 2015 due to the $1 billion contribution. Mr. Kershner added that column E showed the impact of assumption changes. He noted that an experience study was conducted every four years and the assumptions were potentially updated in response to the study's findings. He clarified that the impact on the unfunded liability from the assumption changes appeared in column E. He explained that the final column showed the net increase or decrease in the unfunded liability based on the results from columns B through E. The PERS pension unfunded liability had increased by $325 million over ten years, which was more than expected. He reported that a large portion of the increase was due to the assumption changes in column E in which over $750 million in liability had been added, primarily because of the lowering of the expected return assumption. In 2014, the assumption had been 8 percent, in 2018 it had been lowered to 7.38 percent, and in 2022 it had been further reduced to 7.25 percent. He explained that when the assumed return was lowered, liabilities increased because lower earnings on invested assets were expected. Mr. Kershner continued on slide 10, which contained similar information for TRS. He noted that in 2015, column D showed a $1.5 billion contribution gain. He explained that $1.7 billion of the $2 billion contributed in 2015 to TRS had gone to the pension trust and about $300 million had gone to the health care trust. He explained that the $1.5 billion gain was the difference between the $1.7 billion contributed and the amount required based on the actuarial calculations. 2:03:01 PM Representative Stapp asked if Mr. Kershner could describe the process of layered amortization compounding multiple times. He explained that he understood the math in theory. He relayed his understanding that layered amortization began in 2015 and the 25-year clock began, but a new layer of amortization had been added when the assumptions changed, which pushed the liability to being fully funded by 2048. He expressed concern that the number could not simply be pushed to 2095 since participants would not live that long. Mr. Kershner replied that the analogy he considered most useful was to think of the unfunded liability like a first mortgage on a home. The initial unfunded liability had been created in 2014 and was being amortized over 25 years. He remarked that each subsequent assumption change was like taking out a new home equity line of credit, with each new line funded over a separate 25-year period. He explained that each new line of credit created a new layer of amortization. He relayed that there was a substantial discussion when ARMB considered adopting layered amortization about whether the amortization period should be shorter than 25 years, particularly because DB PERS and TRS plans were closed to new entrants. The legal opinion had been that because the statutes specifically required unfunded liabilities to be amortized over 25 years using level percentage of pay amortization, the board should avoid conflict with the statutes by keeping the 25-year period. Mr. Kershner relayed that the next experience study would be conducted the following year, and discussions had already occurred at board meetings about possibly shortening the amortization period. He reiterated that the 25-year period helped reduce volatility. For example, if layered amortization was not adopted and the system faced a $1 billion loss in 2038, the loss would need to be funded immediately because only one year would have remained in the amortization period. He stressed that layered amortization was introduced to help mitigate contribution volatility to the state. Representative Stapp commented that it made sense to him why the layer amortization method would be used. He explained that his concern stemmed from the fact that the plan was closed, meaning that there were participants at the end receiving cash outflows, which meant that the state had liabilities. He was unsure about extending an assumption change out to 2070 because the participants would not be alive to receive the cash outflows by 2080. He argued that since the plan was closed, the amortization period should be shorter. Otherwise, he felt it would be as though a liability was being carried into years when no payments would exist. He acknowledged that the analogy of a house made sense, but a house could exist indefinitely while individuals had finite lifespans. Mr. Kershner responded that discussions about shortening the amortization period had already taken place at ARMB meetings. He anticipated that if the DB plans remained closed at the next experience study, the most likely outcome was that the amortization period would be reduced to 15 years. He explained that the majority of the unfunded liability would already be fully funded by 2039, because the 25-year period introduced in 2014 had not been changed by the layered amortization. He noted that 21 years remained in the 25-year schedule in 2018 and the schedule would be fully funded by 2039. He clarified that it was the successive layers added after 2018 that each carried their own 25-year period. 2:08:59 PM Representative Stapp remarked that the explanation made sense. He compared the layering to a pyramid where each new layer became smaller as it was added on top. He understood that the formula was the total employer contribution minus the employer portion of the DB normal cost. He asked how payroll growth was modeled in the actuarial work. He remarked that payroll growth seemed to be a significant factor in the formula considering that it was based on employer contributions minus normal costs. Mr. Kershner responded that the payroll growth assumption historically had been set at the inflation rate plus a margin. He noted that prior to 2018, the margin had been 50 basis points. For example, when inflation was 3.12 percent in 2014, the payroll growth assumption had been 3.62 percent. He added that the margin had been reduced to 25 basis points in 2018. With the current inflation assumption of 2.5 percent, the payroll growth assumption was 2.75 percent. He noted that the board was considering changes to the current inflation assumption. He explained that projections were currently being prepared for the June 2025 meeting to evaluate the impact of lowering payroll growth on future state contributions. He observed that actual payroll growth during the past decade had been much lower than the assumption. He reported that PERS payroll had grown by about 2 percent per year, while TRS payroll had grown by about 1 percent per year. He reiterated that the board was considering lowering the assumption, which would accelerate funding into earlier years. The lowest payroll growth assumption available was 0 percent, which would essentially create a level dollar approach. He noted that if the assumption were reduced to 1 percent instead of 2.75 percent, some of the contributions would shift from future years into earlier years and result in a less steep increase in future contributions. Representative Stapp explained that the reason he asked was because the committee had just received information about a salary study [Salary Study (2025) conducted by the Segal Company]. He remarked that the study made it clear that state employees had not been appropriately compensated over the past decade. He expressed concern that the liability would only be shifted if the committee considered increasing salaries for employees while ARMB lowered projections. Mr. Kershner responded that the salary increase assumption was separate because retirement benefits under the plans were based on the average pay at the time of retirement. He clarified that the benefit projections were based on the salary increase assumption, which was expected to increase in the future. He emphasized that an increase in the salary increase assumption would raise projected benefits and therefore also raise projected liabilities. He continued that the speed at which the unfunded liability was paid down was tied to the payroll growth assumption. He noted that the payroll growth assumption was generally independent of the salary increase assumption because payroll growth considered the total payroll, which included current members as well as new entrants replacing retiring members. He stressed that liabilities were tied to the salary increase assumption, while the rate of paying down the unfunded liability was tied to the payroll growth assumption. Representative Stapp recalled that Mr. Kershner had stated that the change in participant data demonstrated in column C on slide 9 led to the losses, and he asked what the assertion meant. He asked if the change was due to a higher inflation rate that increased the cost of living adjustment (COLA) payments or if it was related to investment returns. Mr. Kershner responded that it was "all of the above" other than the expected return on assets that was reflected in columns A and B. He clarified that changes in the data referred to the information received each year on actively employed individuals. He explained that some employees were terminated, some retired, and others remained active. He noted that peace officers and firefighters could retire with an unreduced benefit after 20 years of service, which was factored into assumptions. He explained that the inflation component affected post-retirement pension adjustments, which were linked to the Anchorage consumer price index (CPI). Mr. Kershner emphasized that each year, new data revealed whether more people retired, died, or terminated employment than had been assumed. He explained that comparing the actual data with the expected assumptions created liability gains or losses. He clarified that more early retirements than expected resulted in a loss because benefits were then paid earlier and for longer periods of time. The opposite effect occurred if participants who were already in pay status died sooner than expected under the mortality assumption. He suggested that while it was a loss for the individual, it was a gain to the plan because benefits were no longer being paid unless a survivor benefit applied. He summarized that liability gains or losses were the result of all such potential changes. Representative Stapp remarked that although it was somewhat morbid, he understood. 2:16:52 PM Representative Bynum directed attention back to slide 4. He understood that the slide illustrated "who paid and how much." He remarked that non-state employers contributed 22 percent of the total payroll. He asked if the state would assume the additional portion for non-state employers if the actual cost exceeded 22 percent. Mr. Kershner responded in the affirmative. He explained that the fourth bullet on slide 5 indicated that the past service cost not paid by employers was covered by additional state contributions. He added that the State of Alaska as an employer paid the full rate beginning in FY 22. He clarified that the state previously contributed at the 22 percent level, but it now paid the full rate. Representative Bynum recalled that there had been discussion about potentially shortening the 25-year amortization period, even though statute required that it be 25 years. He asked what the overall financial impact would be if the amortization period were reduced. Mr. Kershner replied that shortening the amortization period increased annual payments. He compared the reduction to the difference between a 15-year and a 30-year mortgage, where the shorter term required larger payments but resulted in lower total costs because less interest accrued. He emphasized that the same concept applied to amortizing unfunded liability. Representative Bynum remarked that any increased cost would ultimately shift to the state because of the 22 percent cap for non-state employers. Mr. Kershner confirmed that Representative Bynum's understanding was correct. Representative Bynum understood that the investment return assumption had decreased from 8 percent in 2014 to 7.25 percent in 2022. He asked what the interim year and amount had been. Mr. Kershner responded that the assumption had been lowered from 8 percent to 7.38 percent in 2018, and in 2022 it had been reduced further to 7.25 percent. He added that when the return assumption was 8 percent in 2014, the inflation rate was 3.12 percent. In 2018, the inflation rate was lowered to 2.5 percent, which was where it remained. He explained that several other assumptions were updated in 2018, including mortality, salary increase rates, and retirement rates. He noted that while the numbers presented on slides 9 and 10 reflected combined impacts, the single assumption with the largest effect on liabilities was always the expected return assumption. 2:21:18 PM Representative Hannan stated that she had a question related to slide 6 and the data on liabilities. She wondered what pieces of data were included, but she had realized that it referred to payroll increases, which projected how much a future retiree would earn and would change over time. She turned to slide 9 and asked if a gain for liabilities was considered unfavorable and a loss was favorable, since a loss brought the plan closer to stability. Mr. Kershner responded that a gain on liabilities was the goal. He explained that a gain or loss was the difference between what was expected to occur and what actually occurred. He clarified that when the result was favorable to the plan, it was recorded as a gain, and when it was unfavorable, it was a loss. For example, if the plan assumed a 7.25 percent return but earned 10 percent, it was considered an asset gain. Conversely, if the plan earned only 5 percent, it was an asset loss. Representative Hannan noted that she was interested in the liabilities. Mr. Kershner explained that a 2.5 percent inflation rate was assumed but if actual inflation was higher, the post- retirement pension adjustments granted based on CPI changes would be above the assumption. For example, there were significant liability losses in 2022 and 2023 because inflation was high and the benefits that were paid out were higher than expected. He added that liability changes also came from demographic differences such as more or fewer retirements, deaths, or terminations than anticipated. All of the changes combined were reflected in column C. Representative Hannan referenced the liability results for 2018 through 2020. She noted that 2018 showed a large triple-digit gain, 2019 dropped to a third of that amount, and 2020 stayed at the lower level before increasing again in subsequent years. She asked if there had been an analysis of the data to explain what happened to the liabilities in 2019 and 2020. She wondered if more people had died or fewer people had retired. Mr. Kershner replied that the calculation was comprehensive and involved many factors. He explained that the executive summary section of the Fiscal Note Analysis conducted by Gallagher (copy on file) contained tables outlining the sources of liability gains and losses. The sources usually included 10 to 12 categories such as demographic experience, retirements and terminations, post-retirement pension adjustments due to inflation, and salary increases above or below expectations. He noted that in any given year, some of the factors could produce gains while others produced losses, which were combined in aggregate to produce the reported totals. He added that in 2022 and 2023, unusually high inflation increased pension adjustments, but lower than expected salary increases created an offsetting gain. Representative Hannan remarked that the numbers generally made sense to her, but the wages did not. She asked how the reports accounted for the fact that no new members were entering the DB system. She understood that liabilities were based on wages from current employees, but many individuals whose benefits were being paid were no longer contributing to the system. She suggested that perhaps it was not an element that could be captured in the reports and that she should trust Gallagher to make accurate predictions. Mr. Kershner responded that there were two factors that could cause a gain or a loss in liabilities for people who were already retired. The first factor was whether the retiree received a post-retirement pension adjustment that was either greater or less than anticipated. The second factor was mortality. He explained that if there were 10,000 retirees and it was assumed that 1,000 of them would die in a year, but the data showed that only 500 had died, then the plan would have 9,500 retirees instead of 9,000. That outcome would be considered a loss. Conversely, if more retirees died than expected, it would be a gain from the plan's perspective because there would be fewer beneficiaries receiving payments. He clarified that salary increases had no impact on people who had already retired because their benefits were fixed and already known. The only factors that affected retirees were inflation and mortality. 2:28:06 PM Representative Bynum remarked that the health care trust was currently funded at more than 100 percent, but the presentation did not indicate the current funding rate for the pension. He asked why the health care trust was overfunded while the retirement portion of the plan was significantly underfunded. Mr. Kershner replied that more details would be provided later in the presentation. He offered reassurance that he would discuss the FY 26 pension and health care contribution rates. He relayed that there were three key reasons for the health care trust's overfunding. The first reason was that the Department of Revenue's (DOR) Division of Retirement and Benefits (DRB) had implemented the Employer Group Waiver Program (EGWP), a federal subsidy received by the plan based on retiree benefits. He explained that the subsidy functioned as an additional contribution to the trust and reduced liabilities. In 2018, the introduction of EGWP reduced the overall health care liabilities for PERS and TRS by nearly $1 billion. Mr. Kershner continued that the second reason was a change in plan administration in 2019. The new administrator's efficiency in processing claims resulted in lower claim payments than had been observed in previous years. The third reason was that for several years, actual claims paid out to retirees and beneficiaries had been lower than projected, creating health care gains. In 2024, the trend had reversed and actual claims exceeded expectations, creating a health care loss. However, the trusts remained overfunded, primarily due to the implementation of EGWP in 2018. Representative Bynum asked whether the EGWP subsidy was ongoing or had been a one-time injection into the plan. Mr. Kershner responded that the EGWP subsidy was likely ongoing. He explained that it would eventually end at some point in the future, but all indications suggested that it would continue for the foreseeable future. He added that the Inflation Reduction Act (IRA) of 2022 had made changes to the EGWP subsidies, which had already been reflected in the liability calculations. He emphasized that the program was being monitored but was considered to be an ongoing subsidy that reduced future health care liabilities for the plan. Representative Bynum remarked that the reason he raised the issue was because health insurance had been one of the largest drivers of costs for state government as well as for municipal governments that funded schools. He observed that health care costs seemed to be out of control across the board. He thought it was unusual that the health care portion of the retirement program remained overfunded and was projected to continue being overfunded. He asked what percent of the plan was currently funded. Mr. Kershner responded that he could follow up with the information. Representative Bynum was amenable to receiving the information later. Mr. Kershner explained that he found the information and as of June 30, 2024, the PERS health care trust was funded at approximately 132 percent, and the TRS health care trust was funded at approximately 139 percent. He clarified that the figures meant the assets exceeded the liabilities. He relayed that the numbers represented a snapshot in time that was subject to change each year. Representative Bynum asked what the funded percentages for the retirement portion of the PERS and TRS plans were. Mr. Kershner replied that as of June 30, 2024, the PERS pension funded ratio was 68 percent, while the TRS pension funded ratio was just under 78 percent. 2:34:41 PM Mr. Kershner continued on slide 12 and explained that the FY 26 contribution rates had been adopted by ARMB in September of 2024 and would take effect on July 1, 2025. He stated that he would describe the mechanics of how the rates were calculated and noted that the same process was used to evaluate the costs of HB 78. He explained that the slide contained the contribution rates for the DB plans. The first three columns on the slide represented PERS, while the next three columns represented TRS. He emphasized that pension and health care were calculated separately before being aggregated. He noted that the statutes set contribution rates at 22 percent for PERS non-state employers and 12.56 percent for TRS. All the percentages shown on the slide were calculated as percentages of total pay for both DB and DC members. He explained that over time, the DC payroll represented the majority of total pay because the number of active DB members continued to decline. He directed attention to line 1 on the slide, which showed the normal cost rate. He explained that the normal cost rate reflected the cost of benefits expected to accrue in the upcoming year. Mr. Kershner noted that line 1a represented the total rate, from which the member rate was subtracted. For PERS, the member rate reflected the contribution rates of peace officers and firefighters, along with all other contribution rates, multiplied by their pay and divided by total pay. The calculation averaged out to 1.71 percent. As a result, the employer was responsible for paying 2.14 percent of pay toward the normal cost in FY 26. For pension and health care combined, the employer contribution was 1.97 percent. He noted that there were no member contributions for health care benefits, which explained why line 1b was zero for health care. He relayed that the past service cost rate represented the sum of all amortization layer payments divided by total pay. The rate for the PERS pension was 18.63 percent and 21.12 percent for the TRS pension. The rate for both health care trusts was zero because the trusts were overfunded. Mr. Kershner explained that adding the employer rate from line 1c to the past service rate produced the actuarially determined contribution rate. Including both pension and health care, the PERS contribution rate was 22.74 percent of pay, and the TRS contribution rate was 25.48 percent. He stated that line 4 showed the adopted contribution rates. He shared that ARMB had elected not to contribute the health care normal cost for the past three years because it would only increase the existing surplus in the overfunded health care trusts. By foregoing the contribution, the employer portion could instead be dedicated more heavily to the pension trust. He noted that ARMB had recently chosen to apply the original 25-year amortization schedule that was established in 2014, which accelerated contributions into the earlier years. For example, the board's adopted rate for the PERS pension was 21.43 percent, compared to the actuarially determined 20.77 percent, while the health care contribution rate was set at zero. Mr. Kershner noted that the next two slides provided the same information for the DC retirement plans, which covered employees hired on or after June 2006. He moved to slide 13 and explained that the DC plans included four components: occupational death and disability, retiree medical or health care, the DC employer rate, which was 5 percent for PERS and 7 percent for TRS, and a 3 percent health reimbursement account contribution. He stated that the actuarially determined past service cost rate for occupational death and disability and retiree medical was zero because the trusts were overfunded. He noted that there were no past service costs for the DC and health reimbursement accounts because the percentages were fixed. Mr. Kershner continued to slide 14 and explained that line 3 of the DC retirement plan tables combined the two relevant components that were then divided by pay for DC retirement plan members. He explained that a projection was made and the DC retirement plan rates on line 7 were converted to percentages of total pay, which were consistent with how the DB rates were expressed. He pointed out that the total employer contribution rate for PERS DC plan members was 6.9 percent of total pay, while the total rate for TRS DC plan members was 7.65 percent. Mr. Kershner moved to slide 15 and indicated that the figures were consolidated for PERS and TRS. He noted that line 1 reflected the DB rates, carried forward from slide 12. Line 2 showed the DC plan rates from slides 13 and 14. The rates were then added together to produce the total rate on line 3, which was 28.33 percent for PERS and 31.33 percent for TRS. He relayed that the totals were then compared to the statutory employer contribution rates, which were 22 percent for PERS non-state employers and 12.56 percent for TRS. The excess rate was paid by the state and shown on line 5a as the difference between the total rate from line 3 and the statutory rate on line 4. Mr. Kershner noted that line 5b made an adjustment for interest because employer and member contributions were paid throughout the year with each payroll, while additional state contributions were generally paid at the beginning of the fiscal year. The timing difference meant an extra half-year of interest was earned on the state contributions, and line 5b accounted for the adjustment. Multiplying by the projected pay on line 6 produced the dollar amount of additional state contributions. He reported that that the PERS additional state contribution for non-state employers was $79.8 million beginning July 1, 2025, while the TRS contribution was just under $139 million. 2:42:48 PM Representative Galvin noted that the adopted contribution rate on slide 12 was 21.43 percent. She asked how long Gallagher projected it would take for the pension to be fully funded for PERS and TRS at the 21.43 percent rate. She recalled that it was previously projected to be fully funded by 2039. Mr. Kershner responded that the 21.43 percent rate adopted by the board reflected the original 25-year amortization period that ended in 2039. He relayed that the 22.74 percent rate on line 3 on slide 12 was based on the layered amortization approach and it projected full funding by 2048. He clarified that the 22.74 percent rate reflected the layered amortization method, while the 21.43 percent was tied to the fixed 25-year schedule. The pension trust would be fully funded by 2039 if the board continued adopting rates consistent with line 4. Representative Stapp expressed confusion about why the layered amortization approach had not been adopted. He thought that ignoring layered amortization seemed like it would push liability payments further into the future and increase long-term costs. He asked Mr. Kershner to explain the reasoning. Mr. Kershner responded that Representative Stapp had the information opposite. He explained that layered amortization extended the timeframe for full funding, since each year's unfunded liability was amortized over 25 years. For example, the unfunded liability created in 2024 would be paid down through 2048, and the new 2025 layer would extend to 2049. By contrast, the rate adopted by ARMB on line 4 assumed that layered amortization had never been implemented, and instead continued the fixed amortization schedule ending in 2039. The approach accelerated contributions in the near term and reduced the burden later, because funding was concentrated within the remaining 14 to 15 years of the original schedule. Representative Stapp asked if the losses from the plan were being incorporated into the state contribution, and whether that was why state contributions were increasing. He clarified that he understood that ARMB was absorbing the plan's projected losses through additional state contributions to ensure the liability was paid off by 2039. He asked if the additional costs and changes in assumptions were driving the increased state contribution. Mr. Kershner reiterated that the process was similar to making extra mortgage payments on a 30-year mortgage. Instead of using layered amortization, which would extend to 2048, the board assumed that the plan would be funded as if the schedule would end in 2039. He explained that it would be similar to paying an extra mortgage payment each year and paying down the mortgage faster. Representative Stapp asked for more information about the disparity between total employer contributions for pensions and the contributions provided to current employees under the DC plan. He asked whether such a disparity was standard. He noted that many state workers were in Tier IV and often talked about how the current system with a 6.9 percent contribution rate was inadequate. He suggested that it would be hard to disagree with the Tier IV workers. Mr. Kershner responded that the difference depended on the nature of the benefits. He explained that if HB 78 were to pass, members would receive a lifetime pension with post- retirement adjustments, which were more valuable benefits than those under the current DC plan. Consequently, the costs for DB members were higher than for DC members due to the greater value of the benefits. 2:48:52 PM Representative Stapp acknowledged that higher benefits naturally entailed higher costs. He asked for clarification that the Supplemental Benefit System (SBS) was not included in the calculations on slides 13 and 14. Mr. Kershner confirmed that SBS was entirely outside of the calculations. Representative Tomaszewski asked how it was justified that DB plans were superior to DC plans. He pointed out that in a DC plan, members retained their principal and could pass it on to their children. He asked for clarification on why the DB plan was considered the better system. Mr. Kershner responded that that the investment risk was taken on by individual members under a DC plan, which meant that actual benefits could exceed projections if investment returns surpassed the assumed 7.25 percent. However, market losses could also reduce benefits. In contrast, a DB plan provided a fixed, guaranteed lifetime benefit based on years of service and final average pay, often including survivor benefits. The DB plan ensured predictable retirement income regardless of investment performance and also included survivor benefits that allowed members to pass their benefits to a beneficiary upon their death. Mr. Kershner continued that DB members also received post- retirement pension adjustments, which were valuable and costly because the adjustments provided an automatic inflation-adjusted benefit, which was particularly significant in periods of high inflation. He emphasized that based on the level of benefits in the plan, the DB plan was more valuable and more costly than the DC plan. He added that if employer contributions to the DC plan were increased from 5 percent to 10 percent, the comparison would differ. However, the DB plan carried higher costs under current assumptions and benefit levels. Representative Tomaszewski commented that investment outcomes could influence the value, but he appreciated the explanation. He asked whether the benefits under HB 78 would be discussed, including survivor benefits. Mr. Kershner confirmed that the topic would be covered in the upcoming slides. 2:52:55 PM Representative Bynum noted that Mr. Kershner had described the DB plan as more valuable and more costly. He asked if the characterization was based on the fact that the comparison was not "apples to apples" and if the higher cost reflected the value of the benefits rather than equal contributions. Mr. Kershner responded that projected costs were determined based on a series of assumptions. He explained that if HB 78 were to pass, members in the DB plan would receive benefits that cost more than members' current contributions under the DC plan. He noted that the comparison was illustrated on slides 13 and 14. Representative Bynum commented that market volatility affected the DC plan depending on when members withdrew funds and how the investments performed. He understood that previously, the employee was not impacted and only the plan's liability was affected. He asked whether market volatility could negatively impact employees, retirees, and employers if HB 78 were to pass. Mr. Kershner explained that investment risk was borne entirely by the plan sponsor in a DB plan. He noted that if the statutory contribution limits did not exist, employers would pay more if investment earnings fell short of expectations. Conversely, if earnings exceeded assumptions, contribution rates would decrease. He emphasized that plan members were unaffected by investment risk in a DB plan, while investment risk was entirely borne by the members under a DC plan. Representative Bynum noted that under the current PERS model, retirees were not impacted by market volatility. He understood that HB 78 implemented a shared risk model, which meant that the employee and employer shared in the risk. He observed that the 22 percent statutory contribution cap still existed under the current assumptions and the liability shifted to the state when the percentage was exceeded. He asked if employees shared in the risk under HB 78. Mr. Kershner responded in the affirmative. He explained that Gallagher's projections assumed that all future experience matched the current assumptions and that none of the risk-sharing provisions would be triggered. He added that HB 78 included two risk-sharing provisions: ARMB could reduce post-retirement pension adjustments for HB 78 members, and the board could increase the member contribution rate from 8 percent up to 12 percent if fund levels fell below certain thresholds. He noted that based on the projections, the fund was not expected to dip below the threshold, but the built-in provisions would activate if adverse experience affected unfunded liabilities. Representative Bynum asserted that that the term "shared risk" suggested equal distribution of potential risk. He asked for clarification that the risk under HB 78 was tiered. He understood that the first layer of risk fell on the state due to the 22 percent cap, the second layer of risk could fall on retirees depending on the board's decisions, and a third layer of risk could potentially fall on employees. He asked if risk sharing was based on "trigger points." Mr. Kershner responded in the affirmative. He explained that the trigger was a 90 percent funded level. If HB 78 were to pass and the funded levels of the trust fell below 90 percent, the triggers would activate and allow the board to adjust member contributions and reduce post-retirement pension adjustments. He added that if the funded status decreased from 100 percent to 95 percent, the resulting risk would be borne entirely by the employers because the triggers would not activate. Representative Bynum highlighted that market volatility posed a risk depending on whether a member was in a legacy PERS plan or a plan established after the potential passage of HB 78. He clarified that the risk could be shared under HB 78 but the risk would be taken on by the employee under the current DC plans. Co-Chair Foster asked if committee members were prepared to stay late to hear the remainder of the presentation. He added that Mr. Kershner had a flight to catch in the evening. Representative Hannan noted that she had two constituent appointments scheduled, but she would return to the committee meeting afterwards. Co-Chair Foster received confirmation from committee members that they were willing to stay. 3:02:34 PM Representative Allard asked for more information about retirement benefits that might extend to a spouse or to a child who is disabled and under state care. She asked what the average retirement age of Alaska state employees was. Mr. Kershner responded that the average retirement age differed by benefit level and employee group. He explained that the earliest retirements tended to occur among peace officers and firefighters within PERS, due to more subsidized early retirement benefits. For example, a peace officer or firefighter could retire at any age with at least 20 years of service, whereas non-peace officer PERS members could only retire after 30 years of service. He noted that teachers tended to retire later, with many continuing to work into their 70s. Representative Allard asked what the average life expectancy of an Alaskan was. She asked how long the state might be paying benefits for peace officers that retired between the ages of 40 and 45 if the officer began working at around age 21. Mr. Kershner responded that retirements at such a young age were rare. For example, a peace officer or firefighter retiring at age 50 might live approximately 30 more years, reaching age 80. Representative Allard suggested that the average life expectancy was 74 or 75 years old. Mr. Kershner clarified that average life expectancy at birth was around 74 to 75 years, but once an individual reached age 50, their life expectancy was longer. He added that a 65-year-old male had a life expectancy of roughly 17 to 18 more years and a 65-year-old female had a life expectancy of approximately 19 to 20 more years. Representative Allard asked if the benefits would go to a deceased member's surviving spouse and whether a form must be completed to ensure that the benefits were passed on. Mr. Kershner responded that it depended on the individual's situation. When a member retired in the DB plan, the member had the option of selecting from a variety of types of benefits. The standard benefit was a single life annuity paid for the member's lifetime, after which no further payments were made. Alternatively, the member could elect a joint and survivor benefit, which meant that a percentage would be paid to a designated beneficiary after the member's death. The selection of a survivor benefit was at the discretion of the retiree. Representative Allard asked if a DB member could designate whoever they wanted to receive a portion of their benefits if their spouse was deceased, such as their children. Mr. Kershner deferred the question to DRB. 3:08:07 PM KATHY LEA, DIRECTOR, DIVISION OF RETIREMENT AND BENEFITS, DEPARTMENT OF ADMINISTRATION, responded that under DB plans, a child could only be designated as the survivor if the child was totally incapacitated. If the spouse was still alive, the spouse would need to waive their rights to the benefits. If a retiree had chosen the joint and survivor option and the spouse was deceased, the only payment to beneficiaries would be the last paycheck or any remaining balance in the contribution account. Representative Allard asked for clarification that only fully incapacitated children could receive benefits. Ms. Lea confirmed that the child had to be completely incapacitated and unable to support themselves to receive benefits. Representative Allard shared that she had a conversation with a state worker who had thanked her for asking certain questions. She relayed that the worker had received nothing upon their parent's death and all of the benefits returned to the state. She asked for confirmation that the if the deceased person did not receive the money, the money would return to the state. Ms. Lea assumed the parent in the example had not retired. Representative Allard clarified that the parent had retired. Ms. Lea explained that for a retired member, benefits were first paid from their contribution account. Once the contributions were exhausted after about two years, remaining payments came from employer contributions and fund earnings. If the retiree had exhausted their contributions and passed away, only the final paycheck was paid to beneficiaries, and the remainder reverted to the state. Representative Allard asked if a DB member passed away within two years of retiring and had, for example, $900,000 in their account, would the remaining funds revert to the state upon the member's death. Ms. Lea responded that it would be highly unusual for a PERS member to have $900,000 in their account if they had been retired for at least two years. She relayed that $900,000 exceeded the average participant account balance. She noted that the worker with whom Representative Allard had a conversation might have been referring to a different plan, possibly the SBS plan. She clarified that once a participant died, the remaining assets in their plan were paid to either their survivor or the beneficiaries. The only circumstance in which no one received the funds was if there was no beneficiary designation on file, in which case the matter would have to go to court. Representative Allard asked what would happen to the retirement money if a DB retiree passed away and had no spouse and no incapacitated children. She stressed that she wanted to know what would happen to the money if there was no one to inherit it. Ms. Lee responded that it would return to the trust. Representative Allard understood that the money would return to the trust, which was owned by the state. Ms. Lea responded that the state did not own the trust, but the members of the trust owned the trust. She explained that if a DB retiree had no survivor and no beneficiaries, the remaining assets reverted to the PERS or TRS trust and did not go back to the state. She emphasized that the state could not access the money. 3:14:07 PM Representative Allard asked for clarification regarding DC accounts. She asked if children could inherit a deceased DC member's account if the spouse was deceased. Ms. Lea responded that a DC member could leave their money to any designated beneficiary. If no beneficiary designation existed, the account was distributed first to the spouse, then to children, and then to the parents, continuing in that order until the account was exhausted. Co-Chair Foster asked if Ms. Lea was available to stay late. He suggested that she could return to the committee at a later date if necessary. Representative Allard pointed out that under DB plans, if a member had no surviving spouse or incapacitated children, the money reverted to the trust and was not available to the family. Under DC plans, the money could be inherited by children if the spouse was deceased. She questioned why DB plans were considered better for the individual when DC plans allowed funds to remain with family members. She noted that while DB plans might reduce state costs, the benefits did not necessarily extend to the member's family and might result in payments continuing for decades after retirement without providing long-term family security. Co-Chair Foster asked if Ms. Lea was available to return the following afternoon. Representative Allard suggested that she could provide Ms. Lea with her questions ahead of time. Co-Chair Foster recommended that Ms. Lea return to the committee the following day if possible. 3:17:19 PM Co-Chair Foster asked if Representative Galvin had a question. Representative Galvin commented that she would hold her questions until the presentation reached slide 28. Co-Chair Josephson recalled that approximately 15 minutes earlier, Mr. Kershner had noted that the risk in HB 78 was borne by the state. He noted that Representative Bynum had asked whether the plan was actually a shared risk, and Mr. Kershner had responded that it was. However, Mr. Kershner had also stated that it was unlikely that the risk would be shared because the funds would not dip beneath the 90 percent solvency level. He found the observation vitally important and he thought it was arguably the key to the bill. He asked if Mr. Kershner was indicating that the bill had been written to ensure that the triggers would likely not be needed. Mr. Kershner responded that all of Gallagher's projections assumed that future experience would match the assumptions. He noted that Gallagher was not projecting the 90 percent threshold to be reached based on current assumptions. He clarified that the trust would begin at 100 percent funded and was projected to remain well above 90 percent funded. He acknowledged that it was possible that events could occur that would trigger the 90 percent threshold, such as a terrible year in the markets that caused asset values to decline by a significant percentage, but it was not likely. Mr. Kershner suggested that Gallagher could conduct an example sensitivity analysis for the committee and could analyze the impact of three successive years of poor asset returns followed by recovery to 7.25 percent. He noted that such a scenario could have a significant impact on the DB plans and potentially the state. If a similar scenario caused the funding levels of the HB 78 trust to fall below 90 percent, then not all of the risk would fall directly to the state. He explained that some of the risk would be borne by the trust members because the board had the authority under HB 78 to increase member contributions from 8 percent to as high as 12 percent. He explained that such an increase might not eliminate the entire adverse effect and the state could still be impacted. When the DB plans were closed in 2006, the state was essentially mitigating some of its future risks. He clarified that placing the post-2006 members in a DC environment shifted the risk to the member. If the members were to be placed back into a DB environment, some of the risk would be reversed as well. Some of the risk was controlled by the risk-sharing provisions, but it was not completely eliminated. Co-Chair Josephson asked if the cost to the state under HB 78 related to its effectiveness. He noted that Gallagher had anticipated the bill would be effective. For example, the state was not effective at hiring eligibility technicians for the Supplemental Nutrition Assistance Program (SNAP), which resulted in a $12 million penalty to the state by the federal government. He questioned if Gallagher was projecting that it would be easier to hire people under Tier V because retirement benefits were more attractive than in Tier IV. Mr. Kershner responded that the expectation was that if HB 78 were to pass and employees were able to join DB plans, future retention would be improved. He explained that the bill was expected to lower turnover because employees would be incentivized to stay employed in order to receive retirement benefits. He continued that more people were expected to remain employed if HB 78 were to pass than if the current DC plan remained in place. He relayed that more members in the future meant more benefits, more liabilities, and therefore more contributions. He affirmed that higher retention would likely lead to higher costs. He clarified that the analysis was strictly focused on the actuarial impact to the state contributions and the analysis did not include potential economic benefits such as reduced training and recruiting costs. He stressed that other factors needed to be considered alongside the actuarial findings to determine the total economic benefit to the state. He emphasized that Gallagher's analysis was restricted to the actuarial impact on future contributions to PERS and TRS. 3:24:45 PM Co-Chair Foster remarked that the most critical portion of the presentation appeared to be slides 26 through 36. He requested that questions be held until the end and suggested that Mr. Kershner could be brought back for follow-up questions via teleconference if needed. Mr. Kershner stated that the next slides covered actuarial assumptions, many of which had already been addressed through questions from committee members. He moved to slide 17 and explained that Alaska statutes required an analysis of plan experience at least every four years through an experience study. He clarified that assumptions had last been set in 2022 and the next study would be conducted in 2026. He stated that an experience study involved reviewing the previous four years, comparing actual outcomes with assumptions, and determining whether adjustments were needed. Mr. Kershner cautioned that recent experience had to be interpreted somewhat subjectively. During the COVID-19 pandemic, termination rates had been unusually low and salary increases were somewhat lower than expected. He explained that it would not be appropriate to assume that the patterns would persist indefinitely. He likened the process to driving a car, noting that while one glanced in the rearview mirror, the focus remained on the road ahead. He emphasized that while past experience informed the process, the primary focus was on long-term trends. Mr. Kershner noted that statute required that there be an annual review of health care assumptions. He explained that the health actuary evaluated health care costs on a per capita basis and reviewed trend rates, which reflected assumed future increases in the costs. He remarked that actuaries were guided by standards of practice, which required the use of assumptions representing the best estimate of long-term outcomes. He relayed that debate had occurred regarding what termination rates should be applied to members if HB 78 were to pass. He stated because there was an expectation of higher retention under a DB plan, it would be inappropriate to assume DB members would experience the same high termination rates as DC members. He explained that lower termination rates were consistent with higher retention and that the expectations were incorporated into the assumptions. Mr. Kershner continued to slide 18 and emphasized that no single correct answer existed for assumptions. He explained that assumptions could fall within a reasonable range. For example, while a return assumption of 12 percent would be unreasonable, both 7.25 percent and slightly lower alternatives such as 7 percent or 6.75 percent could be considered reasonable. He added that the current assumptions had been established using Alaska-specific data for turnover, retirement patterns, and mortality, based on the four years of experience through 2021. He confirmed that Alaska member experience was incorporated to the extent that the data was statistically credible. He stated that mortality assumptions were based on standardized national tables encompassing hundreds of thousands of lives and deaths, since individual plan data might not always be sufficient on its own. 3:30:02 PM Mr. Kershner moved to slide 20 and explained that the current assumptions used in the actuarial evaluations had been applied. He clarified that the model assumed 75 percent of the DC retirement rates and 25 percent of the DB retirement rates. He stated that the approach was consistent with Sections 4 and 60 of HB 78, which relied on the most recent assumptions with the exception of retirement rates. He stated that the previously mentioned fiscal note letter prepared by Gallagher had acknowledged that there was uncertainty about the future termination experience of members if HB 78 were to pass. He relayed that it was projected that the experience of members after the potential passage of HB 78 would more closely resemble that of DB members. Mr. Kershner continued to slide 21 and noted that some termination rates had been set many years earlier, when the DB plan had a larger active membership base. If HB 78 passed, Gallagher would be able to analyze over time the actual turnover experience of members. He explained that assumptions specific to the members under HB 78 would be developed once credible experience was available. Mr. Kershner advanced to slide 23 and noted that some of the bill's risk-sharing provisions depended on the funded status of the trust. He reiterated that the triggers would not be activated based on the assumption that future experience would match existing assumptions. He remarked that if HB 78 were to pass and assumptions reflected adverse experience that reduced funding below 90 percent, there would be a potential increase in the member contribution rate or a possible reduction in post- retirement pension adjustments. He continued to slide 24 and emphasized that the potential economic benefit to the state outside of funding PERS and TRS also had to be considered when reviewing cost projections to evaluate the total economic impact to the state if HB 78 were to pass. Representative Bynum asked if it would have been more conservative to use a blend of more modern termination rates rather than relying on the older DB rates. He asked if using modern rates would have provided a more conservative evaluation given that HB 78 involved a different plan with different benefits. Mr. Kershner responded that such an approach could have been taken. He explained that if more moderate termination rates had been applied, the cost estimates would have been lower than the rates that were projected. He confirmed that using a different set of assumptions would indeed have produced different results. 3:35:53 PM Mr. Kershner continued to slide 26 and detailed the fiscal analysis. He explained that the remaining slides came directly from the March 24, 2025, fiscal note letter. He reported that slides 26 and 27 summarized the potential impact to state contributions. He noted that slide 26 showed the additional state contributions, and slide 27 showed the state as an employer contributions under PERS. Mr. Kershner clarified that the calculations assumed that 100 percent of active members in the DC plan would elect to transfer to the DB plan if the bill were to pass. He acknowledged that if some active members chose to remain in the DC plan, the cost estimates would be lower. He stated that the model did not necessarily reflect the worst-case scenario because higher costs were possible, but it represented a conservative scenario by assuming all non- retired DC members would transfer to DB. Mr. Kershner relayed that the first three columns on the chart on slide 26 displayed the additional state contributions for the non-state employers under PERS through FY 2039. He noted that the first year shown was FY 27 because the contribution rates for FY 26 had already been adopted by ARMB. He summarized that the figures reflected the current baseline projections of additional state contributions for PERS, followed by the projections if HB 78 were implemented and all members elected to transfer. He relayed that the figures in the 2030 row would be explained in detail in a few slides. He added that the next columns contained the same information for TRS, and the final columns presented the combination of the PERS and TRS additional state contributions. The chart was essentially projecting the potential impact of HB 78 based on the stated assumptions. He indicated that additional state contributions could increase by as much as $467 million through FY 39. Mr. Kershner advanced to slide 27, which displayed essentially the same information, but for the contributions by the state as an employer under PERS. He clarified that the state paid the full rate. The slide showed the state contributions for the DB plan and what the contributions would be under HB 78. He explained that the middle columns reflected the projected contributions for the current DC retirement plan and the HB 78 column represented the projected contribution rate. Mr. Kershner noted that a handful of people had already retired under the DC plan and the retirees would continue to receive their DC benefits. He elaborated that because the DC trusts were overfunded, the actual rate for DC retirees would be zero in the projections. He added that the DC costs under the HB 78 column represented the health reimbursement account contributions, which equaled 4 percent for PERS peace officers and firefighters and 3 percent for all others. He pointed out that the state would actually see a decrease in DC contributions. When the increase for the DB plan was combined with the decrease in the DC contributions, the result was reflected in the last column on the right. He reported that the state as an employer could potentially contribute an additional $687 million through FY 39. Mr. Kershner advanced to slide 28 and explained that there were two reasons why state contributions were projected to increase. The first reason was that the underlying benefits that HB 78 members would receive were more valuable from an actuarial perspective. He observed that some individuals might prefer DC benefits; however, the DB provisions that HB 78 members would be entitled to receive would cost more than the current DC benefits. He reported that members of the DC plan received occupational death and disability benefits, health care benefits, an employer contribution to the DC account of 5 percent of pay for PERS members and 7 percent of pay for TRS members, and an additional 3 percent of pay contribution to the health reimbursement accounts. Mr. Kershner explained that if the bill were to pass, members of the DB plan would receive a lifetime pension unless they elected a survivor benefit at retirement. He added that the normal benefit was a lifetime pension. He specified that members would remain eligible to receive the same health care benefits as current DC members. The new DB members would also be entitled to death and disability benefits that were more valuable than those available under the DC plan. He continued that members would also be eligible for post-retirement pension adjustments linked to annual changes in the Anchorage CPI. Additionally, members would continue to receive employer contributions to the health reimbursement accounts, which equaled 4 percent of pay for PERS peace officer and firefighter members and 3 percent of pay for all others. 3:41:31 PM Mr. Kershner continued to slide 29 and explained that the underlying health care benefits that members would receive under 78 were the same as the benefits that members currently received. He clarified that more individuals were expected to remain employed and receive health care benefits because future retention was projected to improve under HB 78. He emphasized that there would be higher actuarial costs for the same benefits when members participated in the DB plan compared to the DC retirement plan. The table on the bottom of the slide showed the projected costs for FY 30 as a percentage of pay for each group. He relayed that the FY 26 DC costs had been converted to a total pay basis. He noted that the costs presented on the slide were projected as a percentage of DC members' pay only. He stated that for PERS DC members, the current employer cost as a percentage of the PERS DC pay was 9.17 percent. The figure included less than 1 percent for occupational death, disability, and health care, 5 percent for defined contribution, and 3 percent for health reimbursement accounts. Mr. Kershner continued that for TRS members, the total was employer cost was 10.74 percent. He explained that the next section on the right side of the slide illustrated the projected costs for the same members as participants in the DB plan under HB 78. He relayed that the pension costs, net of the 8 percent member contributions, were about 6.5 percent of pay. He added that the health care costs were shown as about 2 percent of pay. He underscored that the amounts were higher than current levels because more people were projected to remain employed and receive benefits. He reiterated that more participants led to more benefits paid, which created higher liabilities and, therefore, higher costs. The health reimbursement account cost of 3.16 percent represented a blended rate of 4 percent for peace officers and firefighters and 3 percent for all others. He explained that the total column in both sections clearly showed that the HB 78 costs were higher. He stressed that more valuable benefits resulted in higher costs. Mr. Kershner moved to slide 30 and relayed that the second reason for the higher costs was that the additional costs under HB 78 would shift directly to the state and lead to higher state contributions. He explained that because employer contributions were fixed at 22 percent for non- state PERS employers and 12.56 percent for TRS, the actuarially determined contribution left the excess costs to be covered by additional state contributions. He stressed that any increase to the actuarial rate would also fall to the state. He noted that the impact would be demonstrated through an example on the next slide. 3:44:56 PM Mr. Kershner advanced to slide 31 and stated that the example applied to PERS, although TRS would be similar. He explained that the chart on the left expressed the current cost as a percentage of total payroll. He noted that the blue segment represented the DB cost for current DB members, while the green segment showed the cost for DC members. He clarified that the costs for the DC members matched the values presented on the previous slide, but were converted to a total pay basis. He reminded the committee that slide 29 had shown a percentage of 9.17, while slide 31 reflected 7.67 percent, because 7.67 percent represented a percentage of all pay rather than just DC pay. The total current cost for DB and DC combined was 29.16 percent of total pay. The chart on the right illustrated the change in costs, and the cost for the DB plan decreased. He pointed out that the green section representing the cost for DC was eliminated, leaving a value of zero, and the yellow section showed the cost for members under HB 78. He reiterated that slide 29 showed that the cost for the PERS members under HB 78 was 11.78 percent of pay. He explained that when that 11.78 percent was converted to total pay, it became 10.09 percent. He summarized that when the 20.13 percent for DB was combined with zero for DC and 10.09 percent for HB 78, the total cost increased to 30.22 percent. Mr. Kershner continued to slide 32 and explained that all contribution rates were expressed as a percentage of total pay. The projected payroll for PERS DC members in FY 30 was expected to increase due to anticipated higher retention and fewer terminations, which meant that more of the current group of employees would remain in FY 30 as opposed to new hires entering at lower pay. He emphasized that although the underlying cost for the DB members did not change, the percentage went down because the cost was divided by a higher payroll figure. He clarified that if the total cost was divided by $2,853,980,000 versus the current cost divided by $3,008,228,000, the percentage decreased because the denominator was higher. He reiterated that the state as an employer contributed the full actuarial rate. Mr. Kershner explained that the total projected FY 30 rate was calculated by multiplying the combined DB and DC current cost of 29.16 percent of total pay by the projected payroll of $2,853,980,000, then multiplying it by the percentage of the total payroll attributable to the PERS state employees, which was just under 50 percent. He stated that the result was $414,945,000. He explained that under HB 78, the DB total rate was 30.22 percent multiplied by the higher payroll and by the same percentage, which equaled $453,271,000. He noted that the same figures appeared for FY 30 in the tables of numbers through FY 39 that had been included in previous slides. Mr. Kershner explained that the portion of the FY 30 DB contribution rate paid by non-state employers for both PERS and TRS was projected to decrease from 14.33 percent to 11.91 percent. He clarified that the 14.33 percent represented the 22 percent statutory rate minus the rate that employers paid for DC members, which was 7.67 percent. He explained that under HB 78, employers would pay the 22 percent statutory rate minus the DC rate, which was zero, minus the HB 78 rate of 10.09 percent. He shared that the total equaled 11.91 percent. If the bill were to pass, employers would continue to pay 22 percent but would contribute less toward the unfunded liability, and the state would be required to make up the difference. Mr. Kershner advanced to slide 33 and explained that the FY 30 additional state contribution rate was the total DB rate of 21.49 percent minus the portion paid by non-state employers of 14.33 percent, plus the half-year interest adjustment, which equaled 6.91 percent. He added that under HB 78, the FY 30 additional state contribution rate was the DB rate of 20.13 percent minus the portion paid by non- state employers of 11.91 percent, which equaled 7.94 percent. Mr. Kershner explained that the final bullet on slide 33 multiplied the percentages from the previous bullet [6.91 percent and 7.94 percent] by the total payroll and by the portion of pay attributable to non-state employers. He stated that the additional state contribution for FY 30 was projected to increase from just under $99 million to just under $120 million. He emphasized that there were many moving parts in the calculations, such as the underlying costs themselves and the amount of the unfunded liability that was paid by the employers. As the unfunded liability decreased, the state covered the shortfall through additional contributions. He added that the same three slides could be presented for TRS, which followed a similar structure. He remarked that was moving quickly due to time constraints. 3:52:13 PM Representative Galvin thanked Mr. Kershner for his presentation. She relayed that she had calculated the numbers on slide 32 on her own and her calculations totaled $38.3 million, possibly more, in FY 30 for the state as the employer contributions. She added that she calculated $20.8 million for the FY 30 PERS amount. She noted that slide 28 had clarified the role of post-retirement pension adjustments and she understood that adjustments could be made if circumstances shifted. She asked Mr. Kershner to further explain the adjustment feature. She noted that she also had questions about the multiplier effect. Mr. Kershner responded that post-retirement pension adjustments were often referred to as cost-of-living adjustments [COLA], which were offered in most other states. He stated that the purpose was to provide purchasing power protection to retirees. For example, if a retiree received a $1,000 monthly benefit, the total would remain fixed for life if there was no adjustment, despite inflation diminishing its value. He explained that COLAs were intended to preserve purchasing power and were linked to the Anchorage CPI changes each year. He stated that under age 65, members received 50 percent of the CPI increase, and over age 65, members received 75 percent. Mr. Kershner explained that members of the DB plan who were current residents of Alaska also received a separate 10 percent COLA, which was a flat 10 percent of the benefit and not linked to inflation. He clarified that under HB 78, members would not be eligible for the 10 percent COLA, and COLA costs were not included in the HB 78 analysis. He noted that current DB members hired prior to 2006 would continue to receive the benefit. Representative Galvin commented that the HB 78 plan seemed to be an entirely new product. She understood that the presentation had not covered other benefits that would be valuable to the state as a result of the bill, such as the lower employee training costs. She relayed that she had started writing a list of the other benefits and asked if any assessment had been completed to discover other potential advantages. Some of the benefits on her list were lower training costs, retention costs, hiring bonus costs, lowering overtime costs, and lowering contractual costs when positions could not be filled. Representative Galvin continued that another cost to the state involved education. She observed that when teachers remained in their positions for longer periods, higher academic results were achieved, which reduced the need for costly academic interventions. She asserted that consistent hiring and retention of Alaska's public service workers saved the state money in multiple ways. She emphasized that the benefit of education was important. She explained that she wanted to better understand the overall value. 3:58:14 PM Mr. Kershner responded that the points Representative Galvin raised were important. He referred to slide 24, which indicated that additional economic factors must be considered. He clarified that as actuaries for the retirement systems, Gallagher's qualifications extended only to projecting potential cost impacts to the retirement plans' funding. He explained that an economist or other professional would need to evaluate the types of issues Representative Galvin mentioned. He stated that once cost estimates or savings were determined, the figures should be combined with the actuaries' cost projections to assess the overall economic benefit to the state. He emphasized Gallagher was not qualified to provide the analysis. Representative Galvin asked if Mr. Kershner could provide any general statements regarding the effect of not offering DB plans, based on his experience in reviewing DB plans in other states and his awareness of the plans' overall impact. Mr. Kershner suggested that he could share his personal opinion. He believed that DB plans were beneficial to both individuals and employers. He noted that the plans also provided job security for actuaries because actuaries were not needed for DC plans. He commented that DB plans were generally considered preferable to DC plans. However, more valuable benefits usually came at a higher cost, which had been presented in the analysis. He relayed that DB plans offered many positive outcomes and benefits to both the state and individuals, but the plans also involved potential risks and costs, which the presentation had sought to illustrate. Representative Galvin asked about slide 4 of the presentation, which compared PERS and TRS employee contributions. She relayed that she was confused as to why PERS contributions were set at a higher percentage than TRS contributions. She commented that the difference seemed significant and asked if Mr. Kershner could help her better understand the distinction. She remarked that it seemed related to peace officers, firefighters, and similar roles. Mr. Kershner responded that, unfortunately, he did not know the history. He explained that he was aware that the statutory rates were set in 2008 but he was not involved at that time and did not know the reasoning for why PERS was 22 percent and TRS was 12.56 percent. Representative Galvin asked whether there was any general understanding about how other states handle the rates, or if it was just the way the rates had been negotiated. 4:02:09 PM Mr. Kershner replied that the total DB contribution rates were shown on slide 12 and was 22.74 percent for PERS and 25.48 percent for TRS, before any employer contribution. He noted that the total rates were in the same vicinity, but the way the costs were shared among employers in Alaska was unique. He reiterated that he did not know how the statutory split had been decided many years ago. Representative Galvin remarked that she understood that Dr. Teresa Ghilarducci had conducted an economic analysis for the Alaska State Senate, which suggested that there could be potential savings of approximately $76 million due to retention and recruitment effects. Representative Stapp commented that the disparity between TRS and PERS was likely because TRS employees did not have SBS, which resulted in a lower employer contribution. He asked why the state would assume additional liability given the projected reduction in the non-state employer contribution rate, instead of maintaining the non-state employer contribution at 22 percent until the existing unfunded liability was cleared. Mr. Kershner responded that under HB 78, non-state employers would continue to contribute 22 percent. He explained that because the actuarial rate would increase, any amount above the capped employer contribution would fall to the state as additional state contributions. Representative Stapp understood that when the 22 percent rate was originally applied to municipal employees, the expectation had been that once the unfunded liability was paid off, municipalities would not continue at 22 percent. He asked whether the current expectation was for the 22 percent rate to remain fixed indefinitely. He asked whether the state would ever incur an unfunded liability under HB 78 if all Gallagher's assumptions were met. Mr. Kershner responded that Gallagher's projection was that the state would not incur an unfunded liability. However, he clarified that a small unfunded liability existed because the FY 26 contribution rates had already been adopted and HB 78 would initially be unfunded. The contributions in the first year would not be enough to cover the cost of the HB 78 benefits, but it would not fall below the 90 percent threshold. Mr. Kershner added that after FY 39, the projected contribution rate dropped precipitously because the large unfunded liability would have been fully paid. For example, the PERS DB rate dropped from approximately 20 percent down to around 2 or 3 percent. He added that employers would not be expected to continue contributing 22 percent for current PERS participants because the total rate would be less than 22 percent after FY 39, regardless of whether the bill passed. Representative Stapp thought that if an unfunded liability occurred and the rate remained at the 22 percent level, the state could revert unfunded liability to the employers if HB 78 were to pass. He added that slide 9 indicated that since 2000, more liability to the state occurred than was projected. He noted that each time projections were revised since 2000, the plans had performed worse than expected, requiring downward adjustments of the assumed rate of return. He recalled that the initial thought during the bill's crafting was to stress test the projections at 6.75 percent, and that Dr. Ghilarducci had recommended testing at 6.25 percent or 7.5 percent. He observed that the current assumption remained 7.25 percent. He wondered if Mr. Kershner would reimburse the state in the event the projections were off by $555 million. He argued that if the projections were favorable, the state would return funds to the system. Mr. Kershner replied that he would not reimburse the state. He noted that in 2021, asset returns were strong and the unfunded liability decreased by $579 million. He explained that projections in 2021 had been more favorable compared with the prior year's projections and he emphasized that outcomes depended on actual plan experience. He clarified that he did not control asset performance and expected returns had been higher in prior years. He added that all public sector plans had reduced assumed investment returns because future expected returns on equities and other investments were lower than the returns were 10 to 20 years ago. Mr. Kershner continued that assumptions had to reflect best estimates of future experience. He stated that using an unrealistic assumption such as an 8.5 percent investment return would lower projected contribution needs artificially and would not reflect current realities. He explained that 20 years ago, such an assumption may have been reasonable, but it was no longer plausible given the current asset structure and investment allocations. He reiterated that assumptions were updated regularly through experience studies every four years to keep projections current and prevent outdated assumptions from skewing results. 4:11:15 PM Representative Stapp remarked that he had learned that if all historical public sector DB plans were aggregated, the average rate of return was 6.85 percent. He questioned why a number that was likely to be revised downward again would be used to make projections. He expressed confusion about why the comparison would not be made to at least the average rate of return throughout the lifetime of DB plans. Representative Stapp continued that he expected that in another five to seven years, regardless of the plan in place, the assumed rate of return would decline from 7.25 percent to 7 percent, or from 7 percent down to 6.85 percent. He asserted that when the inevitable decline occurred, the cost to the state would increase because the plans would not be earning the same amount. He asked whether it would be challenging to show what the outcome would look like if the plan only returned 6.85 percent, or an approximately 50 basis point difference from the current projection. Mr. Kershner responded that it would not be difficult to run such an analysis. He explained that if the analysis were calculated at 6.75 percent instead of 7.25 percent, the current projections would all increase because the assumption would be for less investment income. He clarified that contributions would therefore need to increase to make up the difference. He added that both the current baseline numbers and the HB 78 numbers would increase, but the difference would remain similar to what was currently projected. He noted that while the totals would all be higher, the delta between the current plan and the plan under HB 78 would remain in roughly the same range. He confirmed that it was possible to run all the projections at a different rate, but the net changes would likely remain of the same magnitude. He repeated that if the analysis were run at 6.75 percent, all of the numbers would increase, but the difference between the two would likely remain about the same. 4:14:12 PM Representative Tomaszewski directed attention to slide 23 regarding risk sharing provisions. He observed that the slide indicated that the board could reduce post-retirement pension adjustments and increase member contributions from 8 percent to 12 percent. He asked whether the provision was unique to the bill or whether other states or plans included similar provisions. Mr. Kershner replied that some other states had risk sharing provisions in which the funded status of the plans was stronger because the risk was shared, which meant less exposure for both the state and the plan sponsor. He emphasized that including risk sharing provisions was not unique to HB 78. He added that the approach was becoming more common in the public sector, though not yet widespread among all states. He noted that a number of states had either a hybrid DB and DC arrangement or some type of risk sharing where employees shared more of the risk compared to a fixed rate. Representative Tomaszewski asked if other state boards were required to take action if funding for the trust declined or if action was optional in other states. He inquired if there were states where the contribution rate would be automatically increased if a trust fund declined to the 90 percent threshold. He asked whether any of the other state plans included automatic increases or limitations. Mr. Kershner responded that he could not provide all of the details off the top of his head, but he was aware that some plans were not subjective and did not involve active decision making by the governing board, as occurred in Alaska. He explained that in some plans, adjustments could be triggered not necessarily by the funded status of the plan, but if the rate of return for a particular year fell below a certain level. He added that each state with risk sharing provisions had different mechanisms, and some of the mechanisms were automatic rather than discretionary, unlike in Alaska where ARMB had the ultimate authority. Representative Tomaszewski asked whether it would be more reasonable to make the plan more fluid by having automatic adjustments to keep the plan afloat, rather than subjecting changes to the discretion of the board. He asked whether automatic triggers that could maintain the fund at 100 percent funded would improve the plan. Mr. Kershner responded that he believed that automatic adjustments would provide additional protection. For example, if the trust fund fell below 90 percent, the board could currently increase the HB 78 member contribution rate. However, if there were an automatic trigger that mandated the rate increase to a defined level rather than leaving the decision to the board, the automatic adjustment would remove other discretionary factors. He clarified that the intent of risk sharing provisions was to provide protection. An automatic adjustment would ensure that the provisions activated without reliance on the board's judgment, which could be influenced by political or other pressures. 4:18:32 PM Co-Chair Foster relayed that he would have his staff reach out to representatives who were not present during the entire meeting to determine if they had questions or desired a follow-up with Mr. Kershner. HB 78 was HEARD and HELD in committee for further consideration. Co-Chair Foster reviewed the meeting agenda for the following day.