HOUSE BILL NO. 79 "An Act relating to participation of certain peace officers and firefighters in the defined benefit and defined contribution plans of the Public Employees' Retirement System of Alaska; relating to eligibility of peace officers and firefighters for medical, disability, and death benefits; relating to liability of the Public Employees' Retirement System of Alaska; and providing for an effective date." 9:41:14 AM REPRESENTATIVE CHUCK KOPP, BILL SPONSOR, indicated the hearing was the second for HB 79 in the House Finance Committee. He explained that the since the prior hearing he had met with individuals from the Division of Retirement and Benefits and the Attorney Generals office, which resulted in a committee substitute version. He also had received an actuary report from the state's actuary, Buck Consultants. In addition, the states Chief Investment Officer, Bob Mitchell, Department of Revenue, was available to testify. He indicated that Mr. Mitchell had performed modeling on the current effectiveness of the plan for the peace officers and firefighters. The brief presentation would put into perspective why HB 79 was important. He wanted to also hear from the states actuary. 9:43:10 AM AT EASE 9:45:23 AM RECONVENED HOUSE BILL NO. 79 "An Act relating to participation of certain peace officers and firefighters in the defined benefit and defined contribution plans of the Public Employees' Retirement System of Alaska; relating to eligibility of peace officers and firefighters for medical, disability, and death benefits; relating to liability of the Public Employees' Retirement System of Alaska; and providing for an effective date." Co-Chair Wilson invited Representative Kopp to the table. 10:02:10 AM AT EASE 10:04:16 AM RECONVENED BOB MITCHELL, CHIEF INVESTMENT OFFICER, DEPARTMENT OF REVENUE, began the PowerPoint Presentation titled  Target Date Fund Simulation Exercise. He explained that in July 2006, the state moved from a defined benefit plan to a defined contribution plan. He turned to slide 2 titled Background: In 2017, the Department of Administration requested that the Department of Revenue build a stochastic model that simulates the experience of defined contribution employees enrolled in the Alaska Target Date Retirement Trusts. The purpose of the model was to test the likelihood that 30-year employees will have sufficient assets to last 30 years into retirement. Four cases were tested: PERS with SBS, Police/Fire with SBS, TRS, and TRS with 6.13% deferred compensation contributions. In 2019, the ARMB requested an update to this analysis at the upcoming June defined contribution committee meeting. Also in 2019, Representative Kopp requested an update to this analysis, incorporating additional occupational scenarios, including Police/Fire without SBS, 25-Year Career Police/Fire with SBS, 25-Year Career Police/Fire without SBS. Mr. Mitchell delineated that he built a Monte Carlo simulation model. The model calculated, based on the participants income each month, what the employee and employer contributions were and how the investments grew over time. The department used the target date fund simulation exercise, which took a default investment fund based on participants age and retirement date where the asset allocation adjusted automatically. The division built the model and ran it 10,000 times, called trials. The trials were ranked from the lowest terminal asset value to the highest. He indicated that if there was a positive number, the trial was a success. The analysis occurred in 2017. 10:08:50 AM He moved to slide 3 titled Target Date Fund Glide Path. He elucidated that the slide showed a graphic depiction of the target date glide path and was what the asset allocation would look like over time beginning 30 years prior to retirement. He noted that it began as 90 percent equity and 10 percent fixed income allocation that became more conservative as retirement approached; at retirement equity was 30 percent. He highlighted slide 4 titled Callan 2019 Return and Risk Assumptions. The slide contained a chart of capital market assumptions, which were estimates of returns and their volatility and the correlation of the performance of each of the asset classes. The information compiled by Callan and Associates was used as the engine of investment performance. Mr. Mitchell outlined slide 5 titled "Assumptions." A blend of Callan's 2019 10-year & long-term capital market assumptions were used. 10-year assumptions were assumed during the first 10 years, were scaled linearly to the long-term assumptions during the next 10 years, and the long-term assumptions were assumed thereafter. Callan's generic fixed income assumption was used in place of the specific fixed income mix employed by the target date funds. Inflation was set at 2.25%/year, with employee salaries assumed to grow at 2.75%/year. Initial consumption in retirement was set at 70% of earnings at retirement. Consumption was assumed to grow with inflation thereafter. 10,000 trials were used for each simulation. The trials were rank-ordered, and the simulations that represented the 25th-, 50th-and 75th-percentiles of the distribution of outcomes are displayed. A summary of the scenarios examined can be found in the table below. Mr. Mitchell explained that given the model focused on replacement income, the model results was not overly sensitive to the salary assumption. 10:10:38 AM Vice-Chair Ortiz asked how important the accuracy of the 2.5 percent estimated inflation rate was and its impact. Mr. Mitchell answered that 2.5 percent had been chosen because it was the inflation number used by Callan. He delineated that the rate was important because it made the model internally consistent. He expected that if inflation was materially different, asset class performance would move in a similar direction over a long period of time. The actual number should not overly impact the results. Co-Chair Wilson cited the 2.75 percent salary assumption amount and wondered whether it reflected the contractual amount or were the steps incurred over the years. Mr. Mitchell explained that the 2.75 percent was the number discussed with the Division of Retirement and Benefits and was not connected to salary schedules. He noted that given that retirement consumption was a percentage of final salary, if changes were made over time it would not be directionally different than what was experienced. He stated that the definition of success was a percentage of final salary. He did not think there would be a very different result using different numbers. Co-Chair Wilson surmised that unlike the prior tiers, the outcome was accumulated over years and was not based on the earnings of the participants top 3 or 4 years. Mr. Mitchell responded that she was correct. 10:13:27 AM Representative Carpenter suggested that if a random number was used such as 2.75 percent and an employee's salary was much higher than what their rate had been over their lifetime of employment, a spike in liability would occur at the time of retirement. He deduced the 2.75 percent was not reflective of what would happen at retirement for the employee. Co-Chair Wilson asked Mr. Mitchell if Representative Carpenter's statement was accurate. Mr. Mitchell responded in the affirmative. He indicated that a model was simplification of reality. He provided an example of an employee with compensation higher than the assumption over the several years of employment. The question became whether the 70 percent was still a relevant number of the much higher income. He voiced that a model had difficulty with the scenarios. The utility of a model was to provide a directional sense of the effectiveness of the program. Representative Carpenter ascertained that that the plan incentivized maximizing earnings in the final years before retirement in order to gain the most amount out of retirement. He thought it was counterproductive when forecasting costs and risks from the state's perspective. 10:15:39 AM Representative Kopp did not believe Representative Carpenter was correct. He clarified that the model was based on Tier 4 and compared it to what a new program would look like. He suggested that the new model would disallow the spiking that occurred in Tier 4. The model was designed to prevent income spiking because it was spread out over a longer time period and had to be averaged. Representative Carpenter was only referring to the slide and information presented to the committee presently. He wanted to point out the fallacy to the thinking. Mr. Mitchell pointed to the table at the bottom of slide 5. He noted that the table reflected the information requested by Representative Kopp. He pointed out that for each case the information portrayed the level of employee and employer contribution as a percentage of the participants salary that was summed on the bottom line. He viewed the data as a reference to determine the percent of salary that was invested for retirement ranging from 25 percent to 13 percent. 10:18:00 AM Mr. Mitchell moved to slide 6 titled Illustration of Simulated Outcomes. The slide showed the simulated initial retirement balance and the balances over time for 250 cases. He pointed to the very high cases and viewed them as unrealistic. He offered that for that reason he ranked the outcomes at the fiftieth percentile, the twenty-fifth percentile, and seventy-fifth percentile as guiding data. 10:19:02 AM Mr. Mitchell addressed slide 7 titled Results. The graph reflected a PERS and SBS retirement 30-year retirement. The participant had an estimated $1.86 million in retirement assets, which grew to about $2.5 million over 30 years and was the median case. The lower green line showed the twenty-fifth percentile and the higher green line represented the seventy-fifth percentile. He noted that not all the trial outcomes were successful. The amount of contribution based on percentage of income and the number of years were the variables of all the plans. Vice-Chair Ortiz asked Mr. Mitchell to define what it meant to be unsuccessful. Mr. Mitchell responded that success was defined as a participant consuming at 70 percent of their final income that grew with inflation over a 30 year period. Vice-Chair Ortiz asked whether "unsuccessful" meant the money ran out. Mr. Mitchell replied in the affirmative. 10:21:38 AM Mr. Mitchell described slide 7 that showed the total contributions at 25 percent. He moved to slide 8 that depicted the 30-Year Police/Fire Plan with SBS. He explained that the success rate was similar, but the dollar figures were slightly different. Directionally it was the same from a success perspective. He discussed slide 9 that showed the 30-Year Police/Fire Plan without SBS. He recalled that PRS plus SBS comprised 25 percent and SBS represented 12 percent of the amount. The slide depicted the contributions at 13 percent resulting in the assets being depleted over time in the median case depicted by the dark blue line. He added that even the 75th percentile showed a declining trend. Co-Chair Wilson was concerned that if a new tier was introduced what guaranteed that the state would not end up with a huge unfunded liability. She mentioned concerns with negative effects on bonding if the state implemented a new tier. Mr. Mitchell thought it would be better to direct her question to the ARM board and the states actuary. The key distinction of the defined benefit plan was that the risk for paying the benefits rested with the employers. The risk in the defined contribution plan was born by the employees. There was no additional unfunded liability as a result of the defined contribution component. He deferred to the actuary for detailed data. 10:25:10 AM Co-Chair Wilson understood that the state placed a certain amount of funds to offset the prior Tiers liability. She asked if the obligation was the same with the proposed plan. Mr. Mitchell could not directly speak to the question. He offered that the proposed plan contained an assessment on employers that was a function of total payroll, which was comprised of defined benefit and defined contribution employees and was an assessment on the employer and the total number of employees. He noted that the defined contribution employees were not subsidizing the defined benefit employees. Representative Knopp wondered what aspects had the potential to negatively impact the model. He asked if age or the number of participants were risk factors in terms of success. 10:26:49 AM Representative Kopp responded that the current presentation only looked at the effects of Tier 4 and depicted that over time the result was only slightly better than a retirement based solely on Social Security. He advised that the state's actuary could comment on the proposed new tier and any risk for unfunded liability. He indicated that the presentation was focused on the current Tier 4 and how it impacted the class of employees. He was not proposing to open the proposed plan to other employees. The bill proposed the new plan for a small number of employees. He furthered that the plan was a hybrid and had a lot of parallels with the defined contribution plan that kept the liability with the employees and contained self-correcting levers to ensure the fund would not go unfunded. 10:28:32 AM Vice-Chair Ortiz asked what slide 9 showed. Mr. Mitchell observed that the assets were insufficient at retirement to sustain a participant for 30 years. Co-Chair Wilson asked how the investments were chosen. Mr. Mitchell responded that the default plan was used, and different investments would provide different outcomes. Representative LeBon asked how self-correcting levers would protect the state with the proposed plan and how had they failed the state in Tiers 1 to 3. 10:29:57 AM Representative Kopp replied that in the proposed bill used annual true-ups to determine whether the employee contributions should increase, and the employer contributions could change. In addition, the post-pension retirement adjustments and COLA would be eliminated if the actuary showed that the plan was not maintaining a high funding standard. He noted that the elements were totally new and had never been implemented in any plan. He recalled earlier testimony from Washington state with a very similar program that was currently funded at 110 percent. Representative LeBon asked whether the employee bore the brunt of the obligation to self-corrected the plan. 10:31:27 AM KENT TRUITT, STAFF, REPRESENTATIVE CHUCK KOPP, responded that the employer contribution was a variable lever. He indicated that if the ARM board found that the plan was accruing liability, they would have the ability to increase the employer contribution. Currently, the total employer contribution was 22 percent. In the proposed CS [not introduced] the total employer contribution was 22 percent and about 13 percent went to the plan, while the remainder went to the unfunded liability of the prior unfunded tiers. Mr. Mitchell continued to slide 10. The slide showed TERS for Tier 4 with similar conclusions as the prior slide but at approximately 15 percent (versus 13 percent in prior slide) of the total contributions. Vice-Chair Ortiz conveyed that teachers no longer participated in Social Security. Mr. Mitchell understood that many members of the TRS opted out of Social Security and did not participate in SBS, but most state employees had access to SBS. Vice-Chair Ortiz noted that the term opt out presumed that a teacher had a choice, but teachers did not have the option to choose Social Security. Co-Chair Wilson relayed that Vice-Chair Ortiz's statement was accurate. Representative Kopp had worked for a municipality for many years that offered neither SBS nor Social Security. He added that many municipalities operated in the same manner. 10:35:38 AM Co-Chair Wilson provided an example of a teacher working in the summer and receiving Social Security. Mr. Mitchell responded that the analysis did not take other employment into consideration. Co-Chair Wilson wanted more information regarding how outside employment would affect the plan. Representative Knopp asked if most of the public employees had bargaining units. He wondered if they participated in retirement plans through the bargaining units. Representative Kopp responded that all bargaining units were different, it depended on the municipality. Representative Knopp wondered the extent of who qualified for the proposed plan. Representative Kopp responded that volunteer firefighters would not be included in the plan. epresentative Knopp asked about militia members. Representative Kopp responded that militia members did not fall under the definition of a peace officer. Representative LeBon asked whether a participant in a defined contribution plan who subsequently secured employment accruing social security could receive both benefits. He relayed his own experience where he was entitled to both benefits. Representative Kopp responded in the affirmative. He added that it was difficult to maintain outside employment as a police officer and at retirement age many police and firefighters had limitations due to disabilities that limited other job opportunities. Mr. Mitchell moved to slide 11 that graphically depicted 30-Year TRS + 6.13% Deferred Compensation teachers with a deferred compensation amount that the employee would otherwise pay if they participated in SBS. He relayed that there was a match with SBS from 15 percent to 21 percent. The outcome was materially better than slide 10 without SBS. 10:42:12 AM Mr. Mitchell reported that the following slides portrayed a variation; working 25 years with or without SBS. He briefly continued to slide 12 that graphically depicted 25-Year Police/Fire + SBS and slide 13 illustrating 25-Year Police/Fire w/o SBS. He noted that the shorter career and lower contribution as a percentage of income demonstrated unsuccessful outcomes. He emphasized that length of career and contribution rate as a percentage of income were the largest factors in the success of the plan. Co-Chair Wilson asked where the 25 years came from. Representative Kopp responded that Tier 3 with 25 years of continuous service entitled the participant to full retirement including medical benefits. Co-Chair Wilson asked whether it was realistic to expect that most police officers would make it to 25 years of employment. Representative Kopp replied that 25 years could be a hard lift for some officers. However, the cost of the plan had to remain manageable and affordable. 10:45:33 AM Vice-Chair Johnston asked if retirement was 25 years or the age 55 to receive the medical benefits in tier 3 and tier 4. Representative Kopp answered that 20 years of work qualified tier 2 for full retirement benefits and the tier 3 qualifier was 25 years or age 60. Mr. Mitchell returned to slide 13 and highlighted that in plans where a relatively low proportion of the participants compensation was invested for retirement the probability that their assets would last for 30 years after retirement was below 50 percent. 10:47:39 AM Mr. Mitchell turned to slide 14 titled Probability of Success: 30-Year PERS + SBS = 69% 30-Year Police/Fire + SBS = 69% 30-Year Police/Fire w/o SBS = 22% 30-Year TRS w/o SBS = 31% 30-Year TRS + 6.13% Deferred Comp. = 56% 25-Year Police/Fire + SBS = 43% 25-Year Police/Fire w/o SBS = 6% Success = retirement assets surviving 30 years into retirement, assuming initial consumption level of 70% of final take-home pay, increasing with inflation. Representative Josephson referred to the top 2 bullets on the slide. He surmised that the employee would not be able to draw down their plan in order to achieve the results. Mr. Mitchell answered in the affirmative. He added that adjustments were not made for personal circumstances. 10:49:39 AM Representative Carpenter asked why 70 percent of final take- home pay was used in the model. Mr. Mitchell indicated the number had been provided by the Division of Retirement and Benefits as a reasonable goal. HB 79 was HEARD and HELD in committee for further consideration.