Tuesday, February 28, 2017

Lawyers, Guns, and Money 2

Warren Zevon always had an excellent ear for music, irony, and social commentary.  Zevon’s observation fits perfectly with today’s referral of House Bill 3013 to the House Business and Labor Committee, which motivates this post.  Much of the content has already been included in emails to the bill's chief sponsor, Rep Gene Whisnant, one of a trio of ALEC-supported,   Bend area legislators responsible for the spew of anti-PERS member legislation sclerosing the Legislative pipeline this year.  HB 3013 is a deceptively simple bill that effectively decouples the interest rate used for valuing the fund, setting earnings on investments for Tier 1 members and Employers, from the annuity rate used in actuarial tables to compute benefits for retiring members.  In short, the annuity rate would be approximately halved relative to the assumed rate, which would have the effect of severely reducing money match benefits, and all beneficiary forms of Full Formula and Formula plus Annuity.  There is but one exception to this bill left unstated, and it is to that exception that the rest of this post is devoted.

 

My major worry over previous suggestions to decouple the assumed rate from the pension earnings rate was that it would target only Money Match members.  I knew that this method would violate contract provisions, state statutes, Oregon Administrative rules, and very possibly the Internal Revenue Service's basis for qualifying the plan. Because PERS gives Tier 1 and Tier 2 members the "best of" comparison in determining their method of retirement (Money Match, Full Formula, and Formula + Annuity for the small number of members still eligible), those benefit comparisons must be based on "Actuarial Equivalency".  To change the annuity earnings rate only for Money Match would destroy the basis of actuarial equivalency, which would violate the contract, state statutes, and IRC code.  It appears that HB 3013 attempts to evade this problem by changing the annuity earnings rate for ALL forms of retirement methods (Money Match, Full Formula, Formula Plus Annuity).  That MAY take care of the statutory, contractual and IRC problems, but reveals a fatal flaw in the bill that would be exposed if the Actuaries were ASKED to consider employer savings for any range of scenarios instead of using only past behavior of PERS retirees.  Let me explain below.

 

Taking the bill's assumptions, its effects would be to virtually eliminate the Money Match comparison for all but a few PERS retirees (mostly the long inactive).  In effect, this bill would push the vast majority of future Tier 1 and Tier 2 retirees to the Full Formula.  This won't be by choice, of course; it is a natural consequence of devaluing the Money Match benefit compared to Full Formula.  And this is where the complication emerges.  The calculation of Full Formula depends on only 2 variables and 1 constant:  Final Average Salary (however it is computed), years of service, and a multiplier depending on class of service (General Service 1.67% per year; Police, Fire, and Legislature and Judges - 2.0% per year service).  What this calculation yields is the Option 1 benefit (no survivor option).  This is the highest Full Formula benefit any member can receive; any optional benefit forms all derive from this base benefit.  Notice that no mortality figures into this calculation, no interest rate, no assumed rate, no annuity formula.  Thus, nothing external to this Formula can change this, except for changes to the calculation of FAS and possibly the multiplier, but those can only be prospective while the previous rates are locked in statute.  For anyone retiring in the next year or so, the changes to the multiplier and computation of FAS will have minimal effect.  

 

So what you might say.  Weigh that against the stark reality that there are more than 70,000 active and inactive members currently eligible to retire.  What this proposal does not anticipate, and the actuaries probably haven't seriously considered is that human behavior can play a considerable role in tilting the odds against some or most of the possible savings of this bill.  Recall the Option 1 benefit (no survivor option).  The reasons most people don't choose this option are twofold:  1) they don't want their spouse to lose access to their benefit should they die; 2) keeping the annuity rate the same as the assumed rate provides reasonably priced "insurance" for the surviving spouse.  Decouple the two rates, reducing the annuity rate relative to the assumed rate, will increase the cost of the "insurance" for adding the beneficiary.  This can already be seen as an effect of lowering the assumed rate twice in the past four years, with another reduction probable for January 1, 2018.  So, what are current retirees doing?  I'm encountering more and more retirees choosing the Option 1 benefit and purchasing a relatively inexpensive term insurance policy on themselves to cover themselves against early death so the spouse is taken care of through an insurance or annuity program that is cheaper to finance than taking out the "insurance" on an Option 2, 2a, 3, 3a, 4 settlement.  Here's the rub.  If you decouple the annuity rate from the assumed rate, I anticipate that more and more people will consider taking Option 1 instead of the actuarially-affected and tested versions of Option 2, 2a, 3, 3a, 4.  Thus, instead of saving employers money, every employee who selects a Full Formula Option 1 benefit is unaffected by changes in the annuity rate or the mortality tables.  But, to the employers, this is the most costly option for an employee to take because PERS will require the employer to deposit cash to cover the cost of making up that employee's Option 1 benefit but using a lower earnings rate on which to base the contribution.  Thus, instead of saving employers' money, this will cost them even more, as if no disconnection occurred at all.  If large numbers of members start taking this Option, it will show up quickly in PERS' experience data, and downstream, employer rates will start to rise in direct proportion to the degree of disconnect between the actuarial rate and the assumed rate.  This is, of course, the exact opposite of what this bill hopes to achieve. I expect that any savings this bill might create will be ephemeral, and be totally negated by employees discovering the potential value of the Option 1 benefit.

 

A corollary problem exists with the Lump Sum Settlement.  Again, a Full Formula lump sum consists of employee contributions to Tier 1 or Tier 2 accounts plus a an amount required to generate the present value of the monthly benefit taken over the expected life of the member.  But the low earnings rates that gets factored into the employee lump sum means that the expected contribution from the employer has to be considered to be what it would take to generate the monthly Option 1 benefit, but under decreased earnings assumptions.  Thus, the cost to employers for Full Formula Lump Sum settlements will be higher under the reduced annuity rate than they would if the annuity rate and the assumed interest remained coupled.  The logic here is exactly the same as why a lowering in the assumed interest rate requires employers to contribute more to the system.

 

I welcome your thoughts on this matter.  I think you will find that if you ask PERS or ask its actuary to consider a scenario where increasing numbers of members choose the Option 1 Full Formula benefit, or the Total Lump Sum Settlement under Full Formula, that they will concede my points.  Obviously, these scenarios would not follow the current pattern of PERS retirees, but we are no longer living in obvious times.  When stressed, humans have an extraordinary capacity to adapt, and one significant adaptation would be to choose a different payout method for receiving their PERS benefit.

 

There are certainly some flaws in my argument.  The most obvious one is that some members will not be able to get affordable term insurance to offset the beneficiary concern.  Second, a no-beneficiary option requires spousal consent, which may be uncomfortable for some.  Nevertheless, I think this bill will expose some real flaws in the logic and estimates behind the bill's formulation.

 

 Bottom line though is this is a very BAD bill that could be far more harmful to far more people than anything otherwise proposed so far.  To quote the full context of the late Warren Zevon's borrowed title: "...send lawyers, guns, and money, the shit has hit the fan."  Indeed!   (Oh yes, one more turd-blossom in this offensive piece of legislation.  It would take effect on passage, so anyone not yet retired on the date this bill were to be passed, assuming the Governor would sign it - a slim likelihood, would be trapped by it.  This bill has a long way to go, but if you were thinking about retiring, my advice would be sooner rather than later with this bill now in the pipeline.)

Friday, February 17, 2017

Stranded

Leave it to the Rs in the Oregon Legislature to come up with a do nothing, save nothing bill concerning PERS.  In a particularly mean mood, SB 791 was introduced yesterday.  The bill effectively ends the current 1039 hour per calendar year post-retirement work limit for PERS retirees.  Instead, the bill requires an employee to be fully retired from the employer for 6 months before he or she can be hired back in a part-time capacity.  While this is the SOP for private employer pension plans, it hasn’t been a feature of public employer plans.  I presume the reason for this bill is to “end” what some see as “double dipping” - a misnomer if there ever was one.  In fact, I can’t see employers happy with this bill because it may well cost them more money rather than less.  In common cases, an employee who is hired into a 1039-hour position after retirement possesses some unique skill that either isn’t easy to recruit for, or isn’t easy to train a new person into.  The role of the retiree is to help train a new employee assume the duties of a retiree who possesses a unique skill set.  Under the 1039-hour rule, no benefits are paid, no contributions made to PERS, and usually (though not always) no health care benefits. If employers are required to wait 6 months before they can hire a retiree back, how then are they to train someone is a hard-to-fill, hard-to-learn position that is essential.  Obviously this isn’t always the case; indeed, it may describe only half the cases.  Instead, what is going to happen is that employers are going to have to hire replacements before the essential employee leaves, and do the training simultaneously.  This results in double salary and benefits, certainly not a savings but an added expense.

To add insult to injury, this bill does not have a date certain for a starting date, instead stating that it covers any employee still active after the bill is signed into law.  So, while you’re being distracted by the effects of SB 559 and SB 560, which both would take effect on January 1, 2018, you can end up being stranded by this bill, which serves to prevent you from moving out of an agreed upon retirement plan into a brief period of part time employment with no interruption in payroll.  

The worrisome aspect of this bill is that it appears innocuous on the surface, will have broad public support as appearing to do something that it doesn’t really do, it has a possibility of slipping through the cracks and passing since it is targeted primarily at a relatively small sample of those on the cusp of retirement.   Since it obviously does not prevent reemployment of retirees, just puts a significant delay in their path, the bill will neither save money or will it do anything to address whatever ails  PERS.

One final, curious, note.  This bill completely removes the hour limitations on work for a public employer after retirement, provided that three conditions are met:  1) the employee must have spent a minimum of 6 months retired and off the payroll of any public employer; 2) if you are receiving Social Security prior to reaching normal Social Security age (i.e. between 62 and 66), the earnings limit is constrained by Social Security earnings rule (lesson:  if you plan to go back to work after retirement, don’t start drawing Social Security until you’ve fully stopped working); 3) if you are at least normal Social Security age (i.e. 66 or 67 depending on birth year), you can work as much as you want without any earnings limitation.  The only real change is with 1) as the other limits have been more or less in effect since the 80s.  What this bill does is to remove all of the exceptions now in the statutes that permit certain people to work more than 1039 hours if they live in places with certain demographic characteristics.  But then see 1) above for the important caveat.  The rule does not envision exceptions to 1), which happen to be the reason the original set of exceptions were introduced.

Thursday, February 09, 2017

Walking the DINOsaur

Leave it to the group Was/Was_Not to write my blog title for me today.    Today, my post is about Senator Betsy Johnson’s(DINO, Scappoose) fixation and preoccupation with “inactive” PERS members.  In several hearings before the Senate Workforce Committee, where Johnson has insinuated herself as a non-voting, but vocal, member she has asked both Steve Rodeman and attorneys Greg Hartman and Bill Gary about why “inactives” can’t be paid, what sounds like, zero interest on their “inactive” balances.  Bill Gary wrote an op-ed on something along these lines in the Eugene Register Guard about two years ago.  In Gary’s telling, he was flabbergasted that a 5-year UO Professor who moved on to another position, and then at retirement some 25 or so years later, ended up with a higher benefit than a public school teacher working for 30 years.  In effect, that’s our DINO question, but put more bluntly.  Why do we have to keep paying earnings on these people’s money when they aren’t doing anything to earn it?  Folks, the answer to this question revolves around the concepts of “vesting” and “accrued benefits”.  To understand what these mean, let us compare the circumstances of an “inactive” member with that of a member who didn’t work long enough to be vested.  Currently, and for as long as I can remember, one has to work for 5 years at more than 600 hours per year to be considered vested.  (I don’t know if years are prorated based on time worked during the year; for simplicity, we are going to assume 5 full time years).  An employee who terminates employment (or is terminated) before vesting has NO (zero, none nada, zilch) options about what to do with employee contributions and earnings to PERS.  They cannot keep the money in PERS and they are entitled to no benefits.  The money can be cashed out, subject to a significant tax hit, or rolled over into another qualified plan, including a rollover IRA.  The employee has neither the expectation of, or entitlement to, any employER contributions.  So, at the magic 5 year vesting point, an employee becomes a vested member in the Public Employees Retirement System (PERS).  That vesting entitles them to leave their employee account open, continue to draw earnings on it (because PERS is using the money, and there is a price for that), and to receive benefits that include the employer contribution matched in whatever way the vested Tier requires.  This is a really important concept to grasp.  In order to secure the “accrued benefit”, the member must have a PERS account at the time of retirement.  At retirement age or after, that member is entitled to receive a retirement benefit based on all the contributions in and earnings from his/her account, PLUS the employer contribution that produces the highest legal benefit for that employee.  PERS does not allow an inactive member to “cash out both employee and employer contribution", EXCEPT AT RETIREMENT.  Were this allowed, we wouldn’t be having this discussion.  But Senator Johnson MUST understand that “vesting”, “inactive” and “accrued benefits” are tied together in a neat little Gordian knot that can’t be untied without making some major changes to the plan.  And, the only change that could be made would be to give “inactives” the opportunity to cash out of the system at the FULL VALUE of their benefit at the time of withdrawal (that means the equivalent of a total lump sum settlement that can be rolled into another retirement vehicle and annuitized using whatever rate the individual can secure).  This would have to be optional, not mandatory.  This, folks, is not rocket science but the “accrued benefit” is fully defined in terms of the existing PERS Contract, “vesting” is defined, and the conditions required of an “inactive member” have all been defined in statute.  The only option is to change the statute to allow the full cash out for inactive members at a time of their choosing, or to allow them to continue to accrue earnings on their investment until they decide to retire.  

In 2003, the Legislature tried to incentivize “inactives” to withdraw from the system.  What was offered was a pittance - 150% of their individual account balance.  The offer was open for, at most, 18 months and very few people took advantage of it.  The reason should be obvious.  Why would you willingly sacrifice 50% of your employer match when you could leave the funds in the system and get 100% of the match earning, at that time, 8%?  This tactic was a failure.  Nothing short of a total lump sum settlement would ever satisfy a vested, “inactive”, Tier 1 member, nor is it likely to satisfy a vested, “inactive” Tier 2 member.  Moreover, even if Senator Johnson could suddenly figure out a legal way to implement a rate cut (to zero) for inactives, the savings to the system would be minuscule.  Why?  Because virtually all remaining Tier 1 inactives are probably at or very near retirement age, and they could simply pull the plug before implementation.

So, here’s my message, if it isn’t obvious.  Senator Johnson:  “there is no way to get there from here.”  Walk away from this idea before you look really silly.  Losing in court would be an expensive proposition for the State, and the savings absolutely trivial in the process.  Take this DINOsaur and walk it straight to bed.

Monday, February 06, 2017

The Show Must Go On (and on and on, Long Post)

Rumors of my demise, my death, my apathy have been greatly exaggerated since my last update in August 2016.  Fact is, nothing I said in August was ever superseded by later or more informative news, and so I’ve had little to say publicly about PERS (I’ve said lots on the private forum, Pers Oregon Discussion, see link on left).  Now that the political circus is back in town, the tents set up, and all the clowns are meeting with their clown faces on, we have something to discuss.

On Wednesday February 1, the 2017 Oregon Legislature convened for its long session in which thorny issues like the state budget, transportation, health, and, of course, PERS are on the agenda for their needs and for their contribution to the State’s apparent $1.8 billion budget shortfall.  This year, most of the action will take place in the Senate’s Workforce  Committee, chaired by freshman Senator Kathleen Taylor(D, Milwaukee), and vice-Chair, the estimably malign Senator Tim Knopp (R Bend), who is back for is second swing at the piƱata, after contributing to the 2003 wreckage.  The Committee is also ably “assisted” by Senator Betsy Johnson (DINO, St Helens), who is not even a member of the committee.  In the opening salvo, the Committee heard a very long presentation from Steve Rodeman, Executive Director of PERS, on the financing of PERS, as well as the demography of its current membership.  At the end of his presentation, Rodeman presciently noted that “…The PERS situation is driven by math; as an agency director, there’s little margin in having an opinion about math”.  Indeed!!!

Prior to convening the Legislature, Senators Tim Knopp and Betsy Johnson convened a “Working Group on PERS”.  Ostensibly it was convened to flesh out ways in which the existing $21 billion unfunded actuarial liability (UAL) might legally be reduced.  The committee, composed of experts and interested volunteers, had two meetings - one in September and one in December.  Members of the group thought that their input would be sought when legal issues and all of the corollary issues related to reforming PERS (“race to the door”, loss of institutional memory, effects on agency recruiting, etc, as well as the actual budget impact)) would be hashed out.  In fact, nothing of the sort occurred, and after the December meeting, Tim Knopp and most of the Republican Senate caucus dropped two bills on the Legislature to be introduced at the beginning of the session.  Those bills, SB 559 and SB 560, cover a fair bit of ground and relate to some, though not all, of the ideas I discussed in my previous post in August.  Let’s go through them seriatim.

SB 559 covers the period of time used to compute the Final Average Salary (FAS) that is the benchmark for Full Formula (FF) retirement.  The bill has an emergency clause* and is set to begin on 1/1/18.   FAS is also the measure against which the Money Match (MM) retirements are compared.  This is the metric used by those hysterical newspaper headlines shrieking about those relatively few members who were able to retire at more than 100% of their final salary.  Currently, FAS is based on the highest three years of a member’s final ten years covered employment.  Usually, but not always, those are the last three years in a member’s career.  SB 559 proposes to change the time period from the  three years to the FIVE years.   It is estimated that this would reduce the UAL by about $700 million and reduce employer rates by about 65 basis points in 2017-19.  This is a tricky proposal.   Its purpose is to dilute the FAS used to calculate the benefit under Full Formula (Tiers 1 and 2, OPSRP).  Recall that the formula involves total years of service, a multiplier for each year of service (1.67% of FAS for Tier 1 and 2; 1.5% for OPSRP), and FAS.  Option 1 (the highest benefit possible without a beneficiary) is the starting point for these calculations.  Thus, a 30 year, Tier 1 member can earn 50% of FAS.  So anything that reduces the FAS will have the attendant effect of reducing the benefit since FAS is the only variable in the equation - years of service being measurable and constant for any individual and the multiplier being set in statute.  Of course, FAS is also influenced by other variables besides how many years the average is computed over.  Adding to FAS for Tier 1 and Tier 2 is accrued sick leave (for participating employers) and the value of accrued vacation time.  Another factor that can drive up FAS is the acquisition of overtime pay for those eligible.  SB 559 ONLY deals with the time period for the multiplier; SB 560 has other interrelated effects.  The bottom line is that spreading the salary over five years has a tendency to lower the FAS since the actuary uses a 3.5% salary multiplier to calculate expected salary.    An example will illustrate.   Suppose a member is earning $50,000 in calendar 2014 and can retire with 30 years at the end of 2018.  Salary in 2015 is $51,750; 2016, $53561; 2017, $55436; 2018, $57376.  Leaving aside other additions to the totals, the basic FAS under the current rules would be based on the sum of the last three years: ( $53561+$55436+$57376)/3 = $55458, with a benefit of $27729 (with rounding).  Under SB 559, note the change.  FAS = ($50,000 + $51750 + $ 53561 + $55436 + $57376)/5 = $53625/2 = $26812.   So by taking the average out over 5 years, the simple FAS is reduced by almost $2000 and the benefit reduced by nearly $1000.   Since the average state and school district employee salary is $56,028 (Rodeman’s presentation on 2/1/17), our example isn’t very far off the mark.  Assuming the salary growth assumptions are correct, this gives a pretty good idea of how much of an impact this could have on all employees retiring under FF and Formula + Annuity (although the effect would be halved for these).  I once thought the salary assumption was way off until I calculated my own average rate of salary growth.  While it didn’t increase linearly with time, the difference between my starting salary and my retirement FAS followed an average 3.5% growth trajectory per year.  However, my final three years’ salary were nearly identical, which illustrates how off this set of assumptions can be if you focus only on a specific group of years.  Many employees reach salary plateaus near the end of their careers and the growth trajectory ceases to follow the normal pattern.  I’d be surprised if the savings from this change are as much as the actuary projects.

SB 560 is much deeper, more harmful, and worth more detail.  The essence of SB 560 is to redirect employee contributions (the 6% paid currently into the IAP) into a another fund (a second IAP-like fund?) dedicated to the pension costs for the employee (the FF, MM, or F+A) beginning January 1, 2018.  It also forbids employers from paying the “pick up” on or after 1/1/18.  The second piece of SB 560 is to place a cap on salary used to compute FAS at $100,000 beginning 1/1/2018 (see SB 559 also on how this impacts).  This bill also has an emergency clause* that takes effect upon passage.  Both bills are referred directly to the Oregon Supreme Court for adjudication.  

On the face of SB 560, the redirect appears to be “wage theft”, clearly illegal.  On closer inspection, however, the structure of the second “individual” account is such that it still belongs to the employee. If the employee ceases service for a PERS-covered employer before reaching retirement age, the member can go inactive until retirement age and the second “individual” account (the redirected 6% plus earnings and/or losses) will be used to offset the pension costs (i.e. FF or F+A, or conceivably MM in the case of Tier 1 or Tier 2).  The current IAP will be frozen as of 12/31/17 and will only accrue earnings from here on out.  If a member chooses to withdraw completely from the PERS system before retiring, they would be entitled to the balance in their Tier 1 or Tier 2 account, the IAP, and the second individual account.  No employer contribution is made in this case.   In the case of OPSRP members, there is no “member account” in the same sense of there being a Tier 1 or Tier 2 account.  The OPSRP member would have two individual accounts - the IAP (which is supplemental to the pension), and the post 1/1/18 individual account that would be applied toward the cost of the pension portion of the Defined Benefit of OPSRP (a formula-based pension).

The wild card in this portion of SB 560 is the prohibition, beginning 1/1/18, of employers “picking up” the member contribution.  While this certainly could be a negotiating tactic, the “pick-up” itself is a subject of collective bargaining and cannot simply be turned off by legislative fiat.  I presume that the intent of the bill, although this is nowhere clearly stated, is that this becomes the mandatory condition as collective bargaining contracts expire after 1/1/18.  Regardless of its interpretation, the only way this ends up saving employers any money is if all the money contributed is diverted to offset pension costs in the future, and that the employers do not incur offsetting expenses in exchange for having to discontinue the pickup.  To be completely revenue neutral to the member, the 6% member contribution currently paid for by employers would have to be added to the base salary of the member and then deducted, pre-tax, from the employee’s check.  That would be the only way this would not be “wage theft” as far as I can tell.  Of course, my legal opinion is worth what you pay for it - bupkis, nada, nothing - as I am not a lawyer.

The $100,000 cap on FAS beginning on 1/1/18 will end up saving money only for those employees who are slightly over the $100,000 FAS near retirement.  Those who are significantly over the $100,000 FAS after 1/1/18 still have either the 3 highest or 5 highest (see SB 559) years to use in computing their FAS.  The bill only says that the FAS will be limited to $100,000 for years beginning on or after 1/1/18, so members in the higher salary brackets will simply end up using other years for their FAS calculations.  Once out beyond 5 years or so from 1/1/2018, this bill will start to have a serious impact.  It will have an immediate impact on recruiting high-salaried professional into management positions, into Professorial and Administrative ranks in Higher Education, and in recruiting for positions at OHSU’s Medical School and Dental School.  Worse still, however, is that $100,000 is an unrealistically low threshold with neither an inflation adjustment, nor a recognition that the current average salary in the PERS system is approximately $56,000 per year.  Since the actuary uses a 3.5% per year salary multiplier, it would take the average member who receives no other adjustment other than the multiplier per year, less than 18 years for the average salary to be over $100,000.  This comports with Steve Rodeman’s testimony on 2/6/16 to the Senate Workforce Committee that a potential “unintended consequence” of this legislation, for example, would be to push the average salary over $100,000 in 20-25 years.  

Needless to say, both of these bills contain plenty to piss active members off.  Public testimony opens on these bills on February 13, 2017, and I expect there to be considerable argument both for and against them. I urge readers who can attend to do so.  Testimony on real impacts of these bills would also help. [Added 2/7].

One other feature of these pieces of legislation.  While both refer any legal challenges directly to the Oregon Supreme Court, which makes the resolution doable in about two years rather than four, both bills prohibit the Supreme Court from awarding legal fees to the winner.  This is a direct fiscal challenge to the PERS Coalition and any individuals seeking to challenge separately.  In the past, attorney fees have been awarded to the winners (i.e. PERS Coalition through Bennett, Hartman, Morris and Kaplan LLC).  This served to offset legal expenses of the PERS Coalition acting on behalf of all of its participant members.  Under these new wrinkles in the bill’s initial language, win or lose, the expenses will be borne by each party to the case, with no chance of recovering them in the event of a win.  This is both diabolical, and probably (at least in my opinion) illegal.

As more information about these and other proposals emerge, I will update this post, or post a new one.

 

*Emergency Clause does not affect the implementation date of either bill.  That is firmly established at 1/1/2018.  However, by making the bill effective on passage, the legal process or sorting out whether the bill(s) violate contracts, the Oregon Constitution, or the US Constitution, as well as collective bargaining agreements (???), can start immediately after the bill is effective, not on its implementation date.