Transcript Available of IRS Hearing on Pension Plan Mortality Regs

UNITED STATES DEPARTMENT OF THE TREASURY
INTERNAL REVENUE SERVICE

PUBLIC HEARING ON PROPOSED REGULATIONS

"MORTALITY TABLES FOR DETERMINING PRESENT VALUE UNDER DEFINED BENEFIT PENSION PLANS"

[REG-112324-15]

Washington, D.C.

Thursday, April 13, 2017

PARTICIPANTS:

For IRS:

VICTORIA A. JUDSON
Associate Chief Counsel
(Tax Exempt & Government Entities)

LINDA S.F. MARSHALL
Senior Counsel
Qualified Plans Branch 1
Office of the Associate Chief Counsel
(Tax Exempt & Government Entities)

QIAN MAGEE
Actuary
Rulings and Agreements
Employee Plans Division
(Tax Exempt & Government Entities)

For U.S. Department of Treasury:

HARLAN M. WELLER
Senior Actuary
Office of the Assistant Secretary
(Tax Policy)

Speakers:

KENT MASON
American Benefits Council

MICHAEL POLLACK
Willis, Towers, Watson

BRUCE CADENHEAD

ERIC
The ERISA Industry Committee

---

PROCEEDINGS

(10:00 a.m.)

MS. JUDSON: Good morning, and welcome to the Public Hearing on Proposed Regulations: "Mortality Tables for Determining Present Value Under Defined Benefit Pension Plans. REG-112324-15."

My name is Victoria Judson. I'm Associate Chief Counsel at Tax Exempt and Government Entities.

And with me from the IRS, I am pleased to have Linda Marshall, Senior Counsel, Qualified Plans Branch 1, Office of the Associate Chief Counsel, Tax Exempt and Government Entities; and Qian Magee, Actuary, Rulings and Agreements Employee Plans Division, Tax Exempt/Government Entities; and from the Department of the Treasury, Harlan Weller, Government Actuary, Office of the Assistant Secretary, Tax Policy.

We have three people scheduled to speak today: Lynn Dudley, American Benefits Council; Michael Pollack from Willis Towers Watson; and Bruce Cadenhead from ERIC The ERISA Industry Committee.

Is there anyone else in the audience who isn't listed who would like to speak today?

SPEAKER: (off mic).

MS. JUDSON: Okay. So Kent Mason will replace Lynn. Okay. And as many of you probably know, our rules are that each speaker has 10 minutes, we have this wonderful light system, and it will start green, when you have three minutes left it will turn to yellow, and then it turns to red. Should any of the panelists have questions while you are speaking, we can pause it, and then often after you are finished with your comments there will be additional questions that don't count against the time.

And with that, I call our first speaker, Kent Mason from American Benefits Council.

MR. MASON: This is good technique. Thank you very much. As you mentioned, my name is Kent Mason, I'm a partner at the law firm of Davidson & Harman. I'm here today for the American Benefits Council. We thank you for holding this hearing and for the opportunity to testify.

As proposed, the mortality regulations would have a significantly adverse effect on employers sponsoring defined benefit plans, diverting billions of dollars away from jobs and job-producing activities to the pension plan.

In this context we believe, I think the critical thing is that this Regulation be subject to the cost-benefit analyses and rules set down in applicable Executive Orders 12866 and 13771.

Now these orders only apply to regulations that are a "Significant regulatory action," and are legislative rules as opposed to interpretive rules.

And as I'm going to discuss we strongly believe that this regulation falls in both categories. A relevant part, a significant regulatory action is an action that is expected to have an effect on the economy of $100 million or more annually. This regulation —

MS. JUDSON: Kent, can I just stop you and just ask? What is the baseline you are using under your theory? I mean, doesn't the statute require us to modify the tables to reflect changes based on reasonable assumptions? So isn't that statutorily prescribed? So it's regulatory not legislative?

MR. MASON: I think that there are two answers to that: one answer is that the executive orders are completely clear that they apply regardless of whether a regulation is statutorily required. They are explicit on that point. So that even if — so in other words, but there is a second part to answer your question, but the first part is, whether this is statutorily required or not, these Executive Orders still require the application of the cost-benefit analysis including the two for one, the 13771, and they are explicit on that point.

MS. JUDSON: I mean, assuming --

MR. MASON: But I understand, yours is a baseline question.

MS. JUDSON: Yes. I mean, I'm not necessarily saying we would agree with that here, but assuming we did --

MR. MASON: It's explicit, yes.

MS. JUDSON: -- what would be your baseline? Because you seem to be saying -- you are talking about billions of dollars and --

MR. MASON: We are going to talk about that, I'm going to talk about that. And the answer there is that the baseline and essentially, if this work is sort of an interpretive regulation, where you simply follow the statute, then I think the argument from the government, but from the Treasury has always been that the results of the regulation flow from the statue and not from the regulation. And that is -- in most tax regulations, that's absolutely right, and here that's clearly not the case.

MS. JUDSON: But why not? I mean, the statute says we are supposed to prescribe mortality tables using reasonable assumptions, so why is it not prescribed by the statute?

MR. MASON: Well, because there's nothing in the statute that says that this particularly version, in other words, you have legislative authority and to say, okay, we can actually -- so there's the argument you are making, which sort of argued, you could do anything, you could sort of have people living to 150, and that would have no effect.

MS. JUDSON: No, because it says --

MS. MARSHALL: -- there's a 10-year requirement in the statute.

MS. JUDSON: There's a 10-year review and we need reasonable assumption, so under your theory if you wait -- if after 50 years the actually mortality based on actual experience changed from living to 65 years to 85 years, we couldn't do that. Wouldn't that be described by the statute?

MR. MASON: Oh, no. That's absolutely not right, absolutely not right. Because the law, the executive orders are clear, you are required to do this. In other words, we are not arguing -- we are not arguing that you aren't supposed to do this, we are simply saying that by clear black-letter law in the executive orders that you are, regardless of whether this is statutorily required.

That's specifically addressed, you are required to go through the economic analysis, and that is spelled out, and sort of repeatedly in the guidance; so that the fact that you are required does not affect the application of these executive orders. And in terms of the baseline, you really can't say there's a particular baseline, because you were given vast, vast discretion, and so to argue that the --

MS. JUDSON: But it's supposed to be based on actual experience, but let's go back to --

MR. MASON: But I mean, it's not even based on actual experience; it's not.

MS. JUDSON: -- you are also saying that it's two for one so you are saying when Congress tells us we have to issue, will we also have to remove two?

MR. MASON: I'm not -- I didn't issue these executive orders, that's what they say. That's exactly what they say.

MS. JUDSON: Okay. So, what do you think the baseline is? Because it sounds like you are saying the baseline would be whatever the prior table was. Is that your position?

MR. MASON: I think the baseline is that you have great discretion. In other words, for example, you are sort of saying, based on prior experience, you have a 1 percent projection assumption. That's not based on anything, that's based on 12 people in a room, making up a --

MS. JUDSON: Right. But there are a lot of other assumptions too. So, what I'm trying to ask you, what is your baseline going to be? Is it going to be what the results would be under the old tables? Or, are you saying you distinguish between certain assumptions that are more -- the estimates and others that are based on historical data? I'm just asking, what is your position on that baseline?

MR. MASON: I think our position would be that on the baseline, there is no identifiable baseline. In other words, because the Reg -- nothing happens without this Reg. In other words, that's --

MS. JUDSON: Are you saying change, you know, from years ago, everything would have to count?

MR. MASON: It's not 50, not 50 years.

MS. JUDSON: But for 10 -- for the last 10 years, everything would have to count since 2008 everything would --

MS. MARSHALL: It hasn't really -- because we used RP-2000 for the last table. Okay, all historic changes from 2000 to 2013 you would count when you give --

MR. MASON: I didn't write these executive orders, but the executive orders are 100 percent clear.

MS. JUDSON: Well, I'm sorry I've led you off your comments --

MR. MASON: But it was fun, it was fun.

MS. JUDSON: -- but this has been very helpful, so you now may, please, continue.

MR. MASON: Okay. You know, and somehow we sort of woven in some of the things we were going to talk about, but I mean the question is, is this an interpretive Reg or a legislative Reg, and the law is clear, that it is a legislative Reg if you could not enforce this without the regulation. In other words, could you enforce a new mortality table based on the statute? The answer is obviously no. That's the legal test. So, this is a legislative Reg.

MS. JUDSON: And where do you get that legal from test from?

MR. MASON: D.C. Circuit.

MS. JUDSON: Okay. What case?

MR. MASON: And I will send you the case.

MS. JUDSON: Okay. Thank you.

MR. MASON: So, that is sort of the distinction, is could you enforce the statute, and here there's nothing to enforce. And so the second question is, you know, and again this gets back to one of the things you were saying, is there discretion, you know. Or do you simply say, well, there's one study out there, we had to follow it. And that's simply inconsistent with the statute.

The statute as you to look for experience tables, not only the published experienced tables, but you also published projections, improvement Regs which is pure speculation; I mean, people just have opinions, there's no experience that tells us what's going to happen in the future. It is pure speculation. And so to say that this is sort of dictated by the statute and the using of discretion, there's no "there" there.

And even with respect with the actual base tables, they don't say, pick that table and follow it 100 percent, they simply say, take into account available studies, and to sort of just say, okay, well that means we have to follow the one thing that we normally follow. That's not what the statute says.

You have tremendous discretion on the projection, you also have tremendous discretion on sort of whether to take that yes away -- table, sort of in the whole or to adjust it and modify it in conformance with some concerns that were articulated with respect to that table.

MS. JUDSON: So, looking at the statute -- so I understand what you are saying on the baseline. It says that we are supposed to -- The table shall be based on the actual experience of pension plans and projected trends in such experience, but you are saying --

MR. MASON: To take into account, let's read the whole thing, it says, "Take into account."

MS. JUDSON: Right.

MR. MASON: It did not say "follow" it says "Taken into account."

MS. JUDSON: I understand that. So, you are taking the position that, as we take that into account it all would be above the baseline, even if it's based on historical data?

MR. MASON: Well, I think the question is, on the baseline, and is -- and I would sort of invite you to sort of look through executive orders on this point, but the only baseline -- in other words, there is no one way you can identify that says: here is how the tables should be improved, or sort of, should be modified. And to include the 1 percent projection, there is no basis for that, to sort of say, well that's the right answer. So, I find it very, very hard to say, okay, we picked this, and so therefore that becomes the baseline, that's the --

MS. JUDSON: Kent, every time I -- every time I ask you, you move to the one projected thing, what about the historic data?

MR. MASON: Okay. Well, it's more fun to talk about projected.

MS. JUDSON: I understand but so for your -- the point you are trying to make, but I'm trying to ask about the historical data piece.

MR. MASON: Yes. And I think there, you have tremendous discretion to take into account available studies, and so to me -- and there was a lot of criticism of that study. And so the argument that despite the fact that a lot of people expressed a lot of reservations about the study, that that by definition becomes your baseline, I don't see how you get there.

MS. JUDSON: But probably my --

MR. MASON: In other words, that's a legal conclusion that that becomes the baseline with no support. That would be bypassing the notice and comment process by saying: here is our baseline, we are not going to sort of consult with any stakeholders on what the new baseline is, we are simply going to establish this, and say, that by definition has to be the new baseline without input.

MS. MARSHALL: So, you are saying we need notice and comment about the baseline? Because it also seems that under your argument there would never be any Regs that are not legislative Regs --

MR. MASON: That's not true at all.

MS. MARSHALL: Because you are always modifying.

MR. MASON: That's not true at all, that's not true at all. In other words, take any of their sort of regular code sections, you can enforce those based on the statute, 40189, 40184, you can enforce those base on the statute. This, you cannot. And that's a huge night- and-day difference under the law, as to why one is legislative and one is interpreted.

MS. MARSHALL: So you are saying that every point in the 40189 Regs we could have enforced without publishing those Regs?

MR. MASON: I think you have the discretion, you would have had the discretion, and this has happened in countless situations. You know, because there are a lot of situations where there's a general statute, and your enforcement approaches are much more refined than the general statute. So, in other words, would you come to every conclusion in the Regs, well obviously that's unrealistic, but in many, many cases throughout the code you enforce the statute in a nuanced way that is not sort of reflected directly in the code, and that's the way the process works. I mean, there are thousands of examples of that.

So, the question here is: can you enforce it under this code generally for -- and generally, the answer is yes, with almost all Tax Regs, here it's literally impossible. You cannot enforce it without the Reg.

MS. JUDSON: So, you think it doesn't matter that it prescribes that we are supposed to do it, look at actual experience and projected trends. You think in most cases -- in all cases you'll have a zero baseline, and --

MR. MASON: I don't see how you could -- I don't see a mechanism that's respectful of the notice and comment process that could give you a different baseline.

MS. JUDSON: But we have notice and comment. You are commenting --

MR. MASON: Yes. But now you are saying to me, you are saying to me, that without -- Forget about notice and comment, the baseline is this new table over here, that's never been subject to notice and comment. In other words, it's never been approved. How in the world could --

MS. JUDSON: But don't you think each time -- well, you would have to infinitely have other things -- but you are saying after you approve it, then you have an additional notice and comment, and that you already approved?

MR. MASON: No, no, I'm just saying -- No, I'm just saying you can't assume that the proposal is right in establishing the baseline. You can't. That would be invalidating notice and comment.

MS. MARSHALL: And so therefore, the baseline is this base table we are using that has a --

MR. MASON: How could it not be?

MS. MARSHALL: -- I think it has a base here of 1996.

MR. MASON: How could it not be?

MS. JUDSON: Okay, so you've used the baseline as 1996. Okay. All right.

MR. MASON: It is the 2008 Regulation, I never --

MS. JUDSON: But that has tables based --

MS. MARSHALL: Based on the base table within, you know, mid- point of 1996.

MR. MASON: 2000 was my understanding.

MS. MARSHALL: But with RP-2000, the report was 2000, right --

MR. MASON: Right, you're right, you're right. So, in other words --

MS. JUDSON: Okay. We understand, so your position is, we have a 1996 database baseline. Okay.

MS. MARSHALL: I didn't write the law.

MS. JUDSON: Okay.

MR. MASON: You know, you are the one -- I think you are the one who is sort of taking sort of this way, and sort of saying, we are going to -- without notice and comment, establish a new baseline and impose that baseline without regard to input.

MS. JUDSON: I will say, with all due respect, without notice and comment, we are here at a public hearing, and we have --

MR. MASON: But you are actually telling me that I have to use your proposal as my baseline, that's sort of saying, the notice and comment is irrelevant. So, I'm saying, no, you have to -- you cannot assume that what you've proposed is the baseline, which is really what you are saying. You are saying the proposal is the baseline, and that's inconsistent with notice and comment.

MS. JUDSON: Well, that's simply a -- we have different plans here of whether this is legislative regulatory a lot of whether t price. But this has been very helpful that you clarified this position.

MR. WELLER: Oh yes, let me take a turn for a moment here.

MR. MASON: You get a turn?

MR. WELLER: Yes, I get a turn. I've been sitting here quietly. So, separating out for the moment, the issue of the base table, from the Mortality Improvement Scale, I understand that you have a comment that the long-term assumed rate of improvement, that we have used for the 2018 tables by the way, we don't necessarily commit to using what the Society of Actuaries has issued each year; that that 1 percent that the Society of Actuaries used in late 2016, which we proposed to apply for 2018 is too high. And you say, you are suggesting that it imposes additional cost on planned sponsors, and I think if we just focus on that question, I think Vicky is asking: cost by reference to what? I know that there's been a suggestion in the comments that the Social Security Actuaries used 0.75 percent. So, if the argument is the cost is the effect of that 0.25 faster percent per year, what makes the 0.75 percent to prepare --

MR. MASON: I am not -- that is not my position.

MR. WELLER: Okay. So if your position is cost is relative and there are no improvements, it's got to be relative to --

MR. MASON: No. It's relative to the current tables.

MS. JUDSON: He's saying the cost is relevant to the --

MS. MARSHALL: To the 2008 regulations.

MS. JUDSON: -- to the 2008 table based on 1996 data, that's what he's saying. But could you tell us what you would project the marginal increase in cost -- the increase in cost due to using a 1 percent mortality --

MR. MASON: Versus?

MS. JUDSON: -- versus 0.75?

MR. MASON: No. I could not. I mean I could ask my more learned colleagues, you know, two of whom you are going to be speaking to soon. But off the top of my head I could not tell you.

MS. JUDSON: It might be interesting to have that information.

MR. MASON: It would be interesting.

MS. MAGEE: As an actuary, usually, the requirement is we measure the liability reasonable assumptions. So, when you say we have -- or IRS Actuary has complete discretion, or tremendous discretion, in (inaudible) mortality table and projection scale. I wonder, do you feel that it can be anything?

MR. MASON: No. Absolutely not! But I mean, as the reasonable assumptions it's not like -- we wouldn't be having this meeting if there was one set of reasonable assumptions.

MS. MAGEE: Right. So, can you explain -- I mean elaborate on the range of assumptions that you have in mind?

MR. MASON: Well, I don't thing -- You know, I think the question is -- I guess the short answer is no. Okay. I would need to, and I'll be happy to follow up on this but, you know, if your question is, what other sort of regulatory approaches on the whole -- on the base table as well as on the projection rate, what's the range of reasonableness with respect to that? I mean, that's beyond my capacity here today.

MS. MAGEE: And also taking --

MR. MASON: But I mean, I would say that that is going to be a measure and of -- Well, I guess my sort of reaction is, there is sufficient leeway in your sort of -- in the reasonableness and which is a broad, broad concept, to sort of get the point that this has a huge effect on my benefit plan.

MS. JUDSON: Yeah, but I think what would be interesting is to hear what the difference would be between that range of reason --

MR. MASON: And I --

MS. JUDSON: -- just as opposed to compared to a table based on 1996 data.

MS. MARSHALL: And are there other, you know, broad based studies of pension plan participants that you think we should have used, say, instead of our P.E. 2014?

MR. MASON: I think -- see that's, in some ways I've heard that sort of argument. Well, gee, this was the only table we had. If that's your standard, okay, so we have to follow it 100 percent because it's the only table we have. Well, for the rest of time that's always gonna be true. Because this process, no one else has the time and resources to produce this table. No one.

So but that is -- so the fact that only one entity produces a comprehensive table doesn't mean you can't -- statute certainly doesn't say it, and nothing in the law would suggest. You kind of say, okay, one entity produced this. There was controversy about the -- what they did. Rubber stamp is right. That is not --

MS. JUDSON: But you're turning that in --

MR. MASON: -- the right answer.

MS. JUDSON: And --

MR. WELLER: Let me clarify for a moment.

MR. MASON: You got one. But, no, I'm sorry.

MR. WELLER: Let me clarify. That, in fact, the proposed regulations do not use the RP 2014 mortality table.

MR. MASON: Which --

MR. WELLER: They --

MR. MASON: Which they --

MR. WELLER: They reacted -- there was a process. You asked -- talked about notice of comment. In 2010 we issued a notice soliciting comments. Not on -- in general on what studies are out there, etcetera.

MR. MASON: That was '13. It was not '15.

MS. MARSHALL: Well, we did it in '13 and '15, I believe.

MR. WELLER: And we got no response that there's any other tables besides --

MR. MASON: We wrote you repeatedly on this stuff.

MR. WELLER: There was no other tables --

MR. MASON: Absolutely --

MR. WELLER: -- besides the --

MR. MASON: -- we submitted --

MR. WELLER: -- Society.

MR. MASON: -- other tables. We actually absolutely did. I actually, in my folder, I have letters we gave you on other tables, and I will go get them and show them to you, the letters and you were copied on them, so that's not true.

MR. WELLER: I would like to see the letters again.

MR. MASON: I actually printed --

MR. WELLER: That's fine.

MR. MASON: -- them out.

MR. WELLER: That's fine. My recollection is the primary comments about the appropriateness of that table had to do with how the society moved the actual data from the study which centered around 2006. It was the years 2004 to 2008, centered in 2006, forward from 2006 to 2014.

There were a lot of concerns raised. We reacted to those concerns and, in fact, our proposed regulation does not use the society procedure that they used in the RP-2014 study to go from 2006 to 2014. It does, in fact, take into account updated data that was -- the society had not used.

MR. MASON: And I agree with you that that was one of the key points. In fact, probably the central point of the letters was the projection forward from 2006 to 2014, but it was not the only point that was contained in a series of letters. There were a number of other points expressing concerns that were not responded to.

MR. WELLER: Certainly there are concern -- the continued comments of concern about the long term projection. We're --

MR. MASON: Right.

MR. WELLER: -- certainly aware of that.

MR. MASON: Right, yes. And that's one of them, but it wasn't all. It was also the base table concerns. Absolutely.

Am I back on the clock? This is kind of fun. You know? I mean, but all right. Okay. Where the heck am I?

All right. Well, I think we have talked about most of the things. I think the only thing that I would say as we're wrapping up this section is, you know, under the two for one, obviously, there would need to be two regs repealed with a net incremental cost of zero, and under the most recent iteration of the guidance under Executive Order 13:7-71 every effort should be made to use the cost savings for those other regs should -- the cost savings for those other regs should benefit defined benefit plan sponsors. Not, sort of, other industries.

In other words, the idea shouldn't be cost on party A benefits party B. The rules say that every effort should be made. That if it's a cost imposed on A the benefits of the cost savings should go to A.

MS. JUDSON: So let me just make sure I understand. To the extent you're now saying you have a baseline. So there's a benefit because of -- you're looking at 1996 data and how it's changed mortality for the current period, and all that cost should then go back to DB plans as opposed to other --

MR. MASON: Yes.

MS. JUDSON: -- entities or tax payers.

MS. MARSHALL: It goes to find out their legislative regulations that we can modify?

MR. MASON: If we get to that point I would be thrilled and I would absolutely identify two regulations for you to -- legislative regulations for you to modify. I would love the opportunity to do that if we're in that stage of the process.

And, yes. The answer to your question, Vicky, is yes.

Back on the clock. All right. Well, all right. So I know you're gonna be crushed, but I'm gonna move on to a different topic. And that is effective date.

We strongly believe that there needs to be an 18 month period between the publication of the final regulation and the effective date. And we believe it for sort of two reasons primarily. I mean, this is going to have an effect of hundreds of millions of dollars on some companies. And to expect people in a few months to sort of change their business plans, you know, based on sort of new hundreds of millions of dollars for liability is not realistic. People need time.

I mean, the argument I've heard is, well, you know, what about, you know, what about sort of -- I can just look at the proposed reg. Well, that just means, hey, forget notice and comment. You should just sort of accept the proposed reg. You've got to -- there's got to be a notice and comment, and after it's finalized they need a significant amount of time to plan for a huge new liability.

I think the second point is this simply doesn't work. The 2018 doesn't work at all with respect to the partial credibility tables. In other words, you're now saying we'll publish regs at some point. Then somebody's gonna study those partial credibility regs. They've got to prepare a study over several months. They send it to the IRS which looks at it for a few months. Then there's back and forth for a few months.

By the time -- if we finalize today you know when they get that back? They probably get it back in March of 2018. This has to be done -- these partial credibility tables need to be approved months and months and months in advance, not after the fact. So to have a 1/1/18 is simply making this unworkable and throwing people into sort of the -- into a place where they cannot plan ahead, and that's not appropriate.

Now, Linda mentioned earlier, and I wanted to come back to it, well, wait a second. You're under statutory requirement to do this every ten years. But let's be really clear about this. That is a statutory requirement that only is applicable, and I know you're gonna protest, but I'm gonna come back and explain why I disagree with your protest. So I should be off the clock, but --

MS. JUDSON: We'll give you time off the clock --

MR. MASON: Okay.

MS. JUDSON: -- okay?

MR. MASON: Okay. So the notion here is that if this was, for example, January of 2019 there is no one in this room who would contend this should be effective 1/1/18 because of the 10 year rule. So in other words, because it would be totally unworkable, and what I'm saying, it's totally unworkable right now to make this 1/1/18.

And to illustrate how the Treasury and the Service have already come to that conclusion in other situations, countless other situations, including one in this very reg. The partial credibility rules, by statute, are effective 1/1/16. Not '17, not '18, not '19, 1/1/16. And that statutory deadline was simply disregarded.

So this notion that something is sort of in violet about sort of a statutory deadline for when you have to put the regs into effect it's inconsistent with your own regulation. What your own regulation stands for is the proposition that you need to have a reasonable period ahead of time to make this -- before this thing can become effective. You're reflected that in a way you've approached partial credibility. There's no reason to take a different position on the base table.

MR. WELLER: Kent?

MR. MASON: My last two topics --

MR. WELLER: Just to note --

MR. MASON: I almost got through.

MR. WELLER: Almost got through that. Was the organization you're representing sent a letter in July of 2015 saying you needed more time? The 2014 society study was out there. People have been using the society projections for audited financial statements since 2014. Why does it need more time?

MR. MASON: Well, that's actually not right. That's absolutely not right. In the sense that under accounting principles they've had the ability, all along, to show a different sort of set of materials. So that -- there's nothing mandatory about those standards for accounting purposes.

So this notion of well, gee, they've been in effect and everybody's been tickled with them. That's not true.

MR. WELLER: I didn't say they've been tickled with them.

MR. MASON: But they're not --

MR. WELLER: I said they --

MR. MASON: They're --

MR. WELLER: They've been using them.

MR. MASON: A huge amount are not using them. In other words, that is not, sort of, the across the board. I mean, there's sort of a whole industry sort of saying, yes, we can make a reasonable showing to our actuaries. In fact, some of the -- I heard that some of the auditors were saying initially that if you actually used it it was sort of inappropriately tough. When the original 2014 came into place.

The fact that the argument that, gee, everyone's used to this. Everyone's using it. That's not true. I mean, for accounting purpose that's not -- it's not mandatory for accounting purposes and many people don't use it for accounting purposes.

MS. MAGEE: So you are about -- your more concerned about the budgetary purpose because of the cash contribution instead of the software issue of implementing this two dimensional projection scale? Is that --

MR. MASON: You know, in part because -- I mean, as a plan sponsor organization the input that we receive was focused primarily on sort of the effect on the plan sponsor and the plan sponsor liability. I don't wanna discount, by any means, the sort of software transition.

I'm not sure that would require 18 months though. So the 18 months came, you know, and we were originally at 12 months. And as Harlan would tell me, you know, we've written before on 12 months. But we got a very strong set of input from plan sponsors saying on the liability side 12 is not enough. We need much more planning than that. So that was something which we were surprised about, and by the passion with which plan sponsors felt.

So I am not sort of arguing for 18 based on the software issues. I would yield to my, sort of, learned colleagues coming up, sort of on that. But I'm not sure at all that we would need 18 for the software issues. But, I don't -- that's not my area of expertise.

To relatively quick final things. I'm actually gonna be under time, I think. One is, I think, and I'm gonna say something nice which is -- and that is I think the communities' reaction to the process for creating substitute mortality tables, the simplified process was very favorable. People liked it.

But it did beg the question of why not have an automatic approval process? If it's so simplified, you know, what's gonna happen? Are we gonna send this over and it's gonna sit around for two, three, four, five, six months. In which case, this is a disaster. Or is it gonna come back instantly. And if it's gonna come back instantly why the heck didn't we just do an automatic approval process. That's what this cries out for. Because of how well you did in terms of simplifying the process.

Lastly, we were disappointed that existing substitute mortality tables were not grandfathered. In other words, suppose I did my study two years ago. I submitted it. I've got my full credibility substitute mortality table. I've gotta throw that out and start over. That just seems like unnecessary. They should be grandfathered. They submitted their data. They should be grandfathered under the old system.

And with 40, 50 seconds or something. I'm done.

MS. JUDSON: No, you're over by that. Our alert system had a malfunction.

It's now fixed.

MR. MASON: I tend to have that effect on people.

MS. JUDSON: Do people have any other questions?

MS. MAGEE: One last question I have is, when you talk about the 18 months delay, you mentioned that the people would now go to do the experience study. It's always a partial credibility. There's not enough time. I'm not quite sure I understand because it actually simplifies the calculation of the mortality table, so the experience study itself should be the same.

So I'm not sure why you need extra time to -- MR. MASON: Well, I mean --

MS. MAGEE: -- just because the --

MR. MASON: And I apologize because I don't fully understand, so let me just try and understand. We talk through this. I mean, what I'm saying is, okay, I've got this plan with 50,000 participants. Okay? Too small to use full credibility.

So right now they look at their experience and they think, well, it's much better than this new table that's being proposed. So once this new table and the partial credibility rules get finalized, then they can go and sort of hire someone to do an experience study. And to do an experience study, sort of, that would sort of meet your criteria.

That will take a few months. Submit it to you guys. You guys look at it for a few months, and then maybe there's back and forth. That whole process, you know, seven, eight, ten months. And if you started now, you're already past 2018. So I don't even know what my mortality, sort of, table is until after the beginning of the year. That's pretty bad for planning. That's my position.

MS. MAGEE: Okay.

MS. JUDSON: Any -- oh, go ahead.

MS. MAGEE: I think -- go ahead.

MS. MARSHALL: Do you have any data about the number of plan sponsors you expect to request substitute tables under the new regulations?

MR. MASON: Well, I don't. You know, I guess I would -- I don't wanna keep putting pressure on my two smarter colleagues --

MS. MARSHALL: On your colleagues.

MR. MASON: -- coming up. But, you know, they might have -- they would have a better feel than I do.

MS. JUDSON: Do the panelists have any other questions? Thank you very much.

Next, Michael Pollack from Willis, Towers, Watson.

MR. POLLACK: Geez that's a tough act to follow. Good morning, I am Michael Pollack of Willis Towers Watson. I'd like to thank you for the opportunity to express our views regarding the proposed changes to better plan mortality regulations. Before getting into specific suggestions, we would like to commend the IRS for balancing the many competing objectives that went into this and for modifying the regulations to reflect more up-to-date mortality tables and enabling the use of partial credibility.

Overall, we think that it was a very reasonable proposal. That being said, we do have some suggestions. The first issue I want to talk about is something that Kent did get into and that's mortality improvement. We believe that it's the responsibility of the IRS and Treasury, on an ongoing basis to select appropriate mortality assumptions, including the improvement assumptions to be used across the whole defined benefit system and in fulfilling this responsibility, we think that you should not be overly reliant on any particular view or any source of information and we are concerned that the regulations suggest that the plan really is to just simply adopt the actuaries improvement skills as they are published each year. I know it doesn't explicitly say that but you know, I think a reader could infer that.

An assumption regarding a future mortality improvement rate, if any, is fundamentally not an actuarial assumption or not an actuarial issue. It's highly speculative as Ken talked about. Historically, it's been quite volatile, and a wide range of views exist on the issue, and all of them tend to have logical arguments to support them that can neither be proven or disproven.

Our particular view is that the SOA's one percent assumption about ultimate improvement is too high. Since the publication of the 2014 study, improvement rates have consistently been well below what was assumed in that study, resulting in the society issuing MP 2015 and 2016 both of which had lower improvement rates, but in issuing those new scales, they maintained the one percent assumption long term. While we disagree with the one percent assumption, we are not here to assert that our view is right or that the society's view is wrong because only time will tell what the ultimate rate is going to be, but our point is that there is a variety of reasonable views that exist such as that of social security and IRS and treasury shouldn't be overly reliant on one view and shouldn't automatically defer to what the SOA says on the issue.

In addition, we think --

MS. MARSHALL: I read your letter on that. I wasn't clear on what exactly the recommendation was aside from "don't automatically defer to SOA" and we've explicitly not committed to following that in the proposed regulations. Do you have a recommendation on what we should be doing for the ultimate assumption?

MR. POLLACK: I think the main point really is, don't just defer. It's your responsibility to make a thoughtful decision each year, and again, we are not saying the SOA is wrong. I don't know. My crystal ball is not better than anybody else's. I mean there is a large body of research, but fundamentally, society's approach was a very mathematical approach in building models and, you know, human mortality does not necessarily --

MS. JUDSON: But just to explore, you're not saying you think it should be zero? You think 1 percent is too high and that there are other views there.

MR. POLLACK: Yes.

MS. JUDSON: But you are not advocating it be zero.

MR. POLLACK: I'll tell you my personal view, not necessarily -- I don't think zero is an unreasonable assumption. I think you say that to some people and they think you're crazy, they don't consider that reasonable. I think there is a lot of arguments that can support zero, but I am not saying that's the right assumption. I don't think there is a right assumption, but I think that's within reasonable range and again, that's me personally, as opposed to my firm.

So we think the IRS and Treasury should also consider the effects of the assumption on the system as a whole. Overly conservative assumptions would result in higher contributions than necessary as Kent talked about, higher lump sums in traditional plans than warranted particularly for younger individuals and lower annuities from private plans than warranted and in our view, none of these would be good for the system as a whole.

We also believe that changing the improvement scale every year introduces an element of randomness into the process as improvement rates are going to vary, they are going to go up and down year by year, just history has shown that and that randomness could frustrate planned sponsors in their planning efforts and to a lesser extent, planned participants and their expectations for benefit amounts so we don't think that's good for the system as well and we suggest the IRS consider an approach whereby there is no change to the improvement scale unless and until the current table or your current view varies by more than a certain amount from what the current table induces.

I understand that could be a little difficult to measure, to boil things down to a single rate and to execute on, but we think it's at least worth considering. You know, it's kind of -- we think you should look for the middle ground. You don't want to wait 10 or 15 years and have a dramatic change that causes the billions of dollars of contributions that Kent talks about but on the other hand, having little wiggles and waggles every year is not necessarily the best thing either.

MS. JUDSON: If we were to follow that, what kind of percentage would you recommend?

MR. POLLACK: I haven't given that a lot of specific talk. I think the measure -- you look at the central, probably the age-adjusted central death rate.

MS. MAGEE: But on the flipside, if that impact, if that the change is not much, then the impact on liability will not be much either. Even that way it just popped -- you know, it could just -- it won't come to any harm.

MS. MARSHALL: Yeah, I think Qian's point is that the whole point of doing it this way is to avoid the cliffiness (sic) that we previously had by not reflecting changes in mortality improvement and there is a tension between the stability that you are recommending by keeping a set of mortality improvement rates for a few years before changing and then that gets cliffier (sic) because --

MR. POLLACK: Yeah, absolutely, and to the comments we sent in 2005, we did support more frequent adjustments. It's just every year might be a little bit too much. I mean, if you step back from it actuarily and you say, does one more year of data really alter what my "best estimate" of improvements are. It's -- the next issue I want to talk about --

MR. WELLER: Can I interrupt you for a moment? The changes flow through to the lump sum mortality, lump sum tables under 417(e), so are you suggesting that those not change from year to year? Are you suggesting that we keep just the mortality improvement scale constant but in fact the static mortality tables that are used for lump sums would still vary from year to year?

MR. POLLACK: No, I'd have them flow through. I'd have consistency.

MR. WELLER: So you're saying that the lump sum factors would not change from year to year?

MR. POLLACK: Yeah.

MR. WELLER: Even though you have potentially even more dramatic shifts based on interest rate shifts in lump sum factors, you would keep this piece of the mortality flat?

MR. POLLACK: I understand interest rates can have a much more dramatic effect. So the next thing I wanted to talk about was transition. As Kent talked about, some plant sponsors sought and obtained approval to use substitute tables under the current regulations. There weren't a large number of such sponsors, but they tended to be quite large and you know, the effort and cost required to do the study and obtain approval was significant. The tables that they obtained approval for were based on substantial amounts of data that were and remain relevant to the mortality experience of their plants and they underwent a rigorous review by the IRS and Treasury. The issuance of new standard tables for DB plans doesn't affect the validity and relevance of those studies, and these studies were -- and it shouldn't affect the appropriateness of the continued use of those studies which were approved for a period of years, so we believe that the existing approvals -- can't use the word grandfathered -- that they should remain in place as they were approved.

Requiring that sponsors discard all the work that they performed in essentially revoking what was a long-term approval seems unfair and could frustrate those sponsors and perhaps diminish their commitment to DB system.

At a minimum, we think that these sponsors should be refunded some or all of their filing fees or be given credit to them towards an application under the new rules, once they are finalized and permitted to use the same data and experience study period from their previously approved submissions to minimize the amount of revisions and duplication that they have to do.

Another transition element, we suggest that the experience requirements be relaxed a little bit for the initial applications due to the timing of the release of the RP 2014 study; many planned sponsors perform mortality studies with an experienced study period that ended in 2013 or 2014 and these periods would not seem to meet the requirements in the proposal, but we recommend that the proposal be modified to accommodate these periods in situations where the planned population hasn't changed substantially.

We don't believe that the additional cost and effort required to add an additional year or two of data and redo a study is warranted. In fact, we support a change to a more flexible experienced study period more broadly to allow sponsors to better coordinate and leverage studies for the various purposes they might need them for.

The next issue is automatic approval. Before the proposed regs, we thought there should be an automatic approval process, and we had one of our colleagues participate in a group with some others from the industry that made a proposal. After reviewing the proposed regulations -- they are really very prescriptive in certain areas so we think automatic approval should be more broadly available. There is flexibility in terms of grouping and subdividing populations and there is flexibility in the experience period and one time you use a study for but there is really no flexibility in the actual determination of the rates themselves so we feel like if there is none of that grouping and the criteria for the experience period -- there simply should be automatic approval and if the IRS wants to put more restrictive requirements on the study period or the distance between the study period and the U state or the length of time that you can use the table for, that -- to qualify for automatic approval, that would seem reasonable but the actual process of the calculations, there is no -- there is not much flexibility in there so we are not sure what the review really consists of. People can do the math.

The last issue we wanted to talk about was a requirement that all credible populations within a control group use a substitute table if any population does. We understand the reasoning for this requirement and we agree with it but we question its practicality and our concern that it's going to serve to really diminish the use of substitute tables and the better measurements that would result by it.

In our experience, there are a lot of large and complex control groups with multiple plans where there is little coordination with the members of the control group on pension issues so we suggest that the intent of the requirement could be met almost as effectively by applying the rule on an employer level or a planned sponsor level rather than a control group level and by doing this, we think you'd make the use of substitute tables more feasible to a larger number of plans and that would benefit the system overall.

MS. MARSHALL: Control group rule is statutory, is there some -- I mean do you see some way we can get around that?

MR. POLLACK: I can get back to you on that. The last thing is just on the technical aspects and it's not really a comment but approval, you know, the comments I've made have been on the process around the tables as opposed to the mechanics of the calculations. There were some elements we might have done differently. We might have permitted graduation, for example, as opposed to requiring a factor to be, versus the base table but by and large, we think the approach that the calculations was reasonable and sound and practical and simplified and hence the arguments for automatic approval, so we wouldn't suggest changes to the mechanics of the calculations, and that concludes my testimony, thank you very much for the opportunity, and I am happy to take any questions.

MR. WELLER: Mike, you referred to grandfathering of existing tables. To what extent does that run into the problem of other plans -- if you say the plan sponsor or the control group, as the case may be which did not have full credibility so we are permitted not to have the plan's specific mortality under the old standards which will have partial credibility and thereby would need to use plan-specific mortality under the new rules. Do you think that creates an impediment to grandfathering existing approvals for plan-specific mortality?

MR. POLLACK: I think that the requirement in that situation wouldn't apply if you grandfathered those. They established, they used the tables under a procedure that didn't, not only permit, but didn't allow it for the other group so I think that the rest of the group would have their own independent decision.

MR. WELLER: That kind of approach would possibly frustrate the -- if one plan uses it, they all use it policy which you said you understand why Congress put it in place.

MR. POLLACK: But at the time the initial approval was granted, that wasn't an issue. I mean they were under the rules. They didn't have other fully credible populations, they couldn't do it so they didn't -- I mean they did what they could do under the current rule.

MS. MARSHALL: And under the grandfathering, would they switch to the new mortality improvement scale or would they keep the old scale too?

MR. POLLACK: I think the old scale would be acceptable, I mean that's what they -- they file an application for a period of years and it was approved for its use. I meant they still are the standards -- you know, the requirements about it still representing, you know, an accurate estimate, you know, of the population, so to the extent that in light of different scales or up-to-date tables, it was felt that they no longer represented that and then they would have to change but --

MS. JUDSON: Do our panels have any other questions? Thank you very much for your comments.

MR. POLLACK: Thank you.

MS. JUDSON: Next we have Bruce Cadenhead from ERIC. Welcome.

MR. CADENHEAD: Thank you. Good morning. I'm Bruce Cadenhead. My employer is Mercer, but I'm here representing ERIC today. We both submitted letters. Some people have noticed some similarities between the two, but the views on expressing care are -- this is -- this is on behalf of ERIC.

ERIC is the only national association that advocates exclusively for large employers on health, retirement and compensation public policies at the federal, state and local levels. ERIC's members provide comprehensive retirement benefits to tens of millions of active and retired workers, and their families.

Thank you for the opportunity to present my views here -- or ERIC's views here.

First, I just want to commend you on the proposed approach for developing the standard tables. I know that there's a lot of different options and I'm staying away from the procedural issues. But given the different views, I think the approach chosen represents probably one around which, at least, the most consensus has evolved. There are certainly different views, especially around the level of future mortality improvement. Again, the MP 2016 scale, I think moves well in the direction addressing some of the concerns that were originally related to the MP 2014, so I'm glad to see that the latest guidance -- the latest release from the SOA was reflected in the proposed guidance.

Also, support the choice to continue offering static option to all employers, and the approach developing that static option, I think, was very cleverly chosen to balance a number of competing objectives. So kudos on that. Just a very nice job of replicating the generational immediate annuity factors across a range of ages and interest rates. I think it's particularly important to maintain the static option as the applicable mortality table for purposes of 417E3. And if this is not addressed in the proposed regulations and it's not going to be part of the final regulations based on past practice and comments, and language in the preamble, we anticipate it's going to continue to be the 50/50 blend of the static combined tables which, I think, is an appropriate choice.

One concern that we had with regard to the standard tables had to do with the potential timing of updates. After the adoption of PPA, the IRS was able to publish mortality tables for multiple years at once and this made it easy to do contribution budgeting and provide reasonable benefit estimates well in advance. And I would note to the comment earlier about changes in interest rates sometimes making it hard to predict -- there are a number of plans out there that for other purposes of other conversions, use 417E mortality tables, and having those available well in advance allows greater certainty around different benefit options.

The SOA's now expected to provide updates to mortality projection scales annually with late October being the likely timing, which raises two potential concerns. One is the frequent updates assumptions to that cannot be easily estimated in advance, and the possibility that those updates could be made effective with little advanced warning. Although we don't expect that would be the case -- the table comes out in October -- there's the possibility it could be imposed for the -- for the year following which on based on the current proposal, we would expect and hope for at least a year of advanced warning so that the -- say the MP 2017 table released on October 2017 would not be made effective in 2018, but instead would, at the very earliest, be made effective 2019. Our letter includes suggested language that could be added to the regulations that would -- that would commit to that timing.

Further, I note that ARPEC has adjusted its methodology somewhat since the MP 2014 scale was first released in part due to concerns about year-over-year volatility in results as projection scales are updated. The MP 2016 scale which is quite a bit different from MP 2014 reflects five additional years of data, as well as somewhat more stable methodology. So we anticipate that future updates to the projection scale are likely to result in much smaller changes to projected mortality rates and present values year- over-year. And therefore suggest that the IRS need not wait for every update to the projection scale, but could instead prescribe the use of the most recent -- or a recent -- scale for multiple years under the expectation that you wouldn't see large changes and then could evaluate, you know, every couple of years instead of every single year. That's what I have to say on the standard tables.

Switching to the topic of planned specific mortality adjustment. The proposed approach of developing a single credibility adjusted mortality ratio is a welcome simplification to the current process, extending the availability of this option to plans with as few as a hundred death over the study period is also welcome. However, I do have a few concerns about the proposed approach. One is that the number of deaths required for full credibility is now quite -- is now higher -- potentially quite a bit higher, depending on planned demographics in the current approach, which dilutes the effect of plan experience. To address that, we propose a number of different ways of combining experience to achieve a higher level of credibility.

One is that plans be permitted to aggregate male and female experience in developing a mortality -- single mortality -- ratio for the plan as a whole. Making a single adjustment to a standard table is already a simplification in that it assumes that the single adjustment is appropriate across different population subgroups: salaried hourly for example, and across different age groups. Despite the fact that these subgroups will exhibit different rates of mortality relative to the standard table, the proposed approach produces, nevertheless, reasonable results when measured in terms of the effect on overall liability because those factors tend to cancel out when the focus is simply on the liability measure, and not the particular year's mortality rate.

So we think that similarly, reasonable results could be derived from a single ratio applicable to both the male and female standard tables. The premise behind making a plan-wide adjustment to a standard table is that plan participants share certain characteristics that make it reasonable to presume that they exhibit consistently higher or lower mortality relative to the standard population experience. And these characteristics are likely to apply to both genders, just as they are to other subgroups within the population. Permitting the aggregation of male and female experience would increase the overall credibility assigned to plan experience. And should there be any concern about the ability of this exertion, a secondary test could be imposed to assume that at least the experience for both genders is directionally similar. That is, the mortality ratio separately generated both higher than one or both less than one as appropriate.

Similarly, we believe it would be appropriate to allow similar employers to group experience in order to achieve a higher level of credibility. Again, secondary tests could be applied to ensure that employers included in the grouping have directionally similar experience. That they're in a similar industry and so that they -- there's reason to believe that they have similarities and that this -- although perhaps employers that are too small to have even partially credible experience on their own might be exempted from that secondary test.

Also suggest that plans be allowed to request -- to approve -- to use blue and white collar tables based on workforce characteristics. If such a demonstration can be made to the satisfaction of the service, then collar-based tables could be used in place of the standard table, or at least as the table to which the credibility adjustment is applied.

There are also a large -- sorry, a number of large multiple employer plans that might have partially credible experience when looking at the plan as a whole and where no single employer represents a majority of the plan. It appears in reading the regs, although I'm not entirely clear, that as drafted, they could take advantage of what's in the proposed regs, but it would be helpful if this was clarified. Also the requirement that all employers within the controlled -- all plans within the control group look to see if they have partial credible experience and reflect that -- could be a little bit more problematic when applied to this type of a situation. It would be difficult for the multiple employer plan in question to impose such a requirement on its members or even to verify that such analysis has been done. Therefore, suggest that multiple employer plans where no single employer constitutes the majority of the plan be exempt from any requirement to do a similar study for other plans maintained by its members -- or by the control groups of its members.

Finally, point out a potential problem with the transition period for newly affiliated plans. It appears -- and I read this a bunch of times and am not quite sure -- but it appears from our reading it, that a plan acquired midyear would have to gather data for the period ending one year after the acquisition and if that data is partially credible, then apply to reflect that experience for the immediately following plan year. And in some situations, the experience period would end after the application deadline. This may be a misreading, in which case we just request that the regulations be clarified, but otherwise, suggest that the regulations be modified to allow for a reasonable period of time to gather data for newly affiliated plans.

And that is the end of my remarks; made it just in time.

MS. JUDSON: That's fine. Did panels --

MR. WELLER: We just have a couple questions for you.

MR. CADENHEAD: Yep.

MR. WELLER: On the multiple employer plan suggestion, the idea of that -- there is no employer who has represented the majority, is that what you -- is -- any particular reasoning for that particular line as opposed to -- for instance, we -- there's a line that's been drawn in the 10 or more employer plans under 419AF6 that's no more than 10 percent; sort of corresponding 10 or more employers. You have a sense of the utility of having -- whether you're looking for a really broad based multiple employer rule? Or are you looking for a narrower multiple employer rule?

MR. CADENHEAD: Right. Well, I think -- I'd say that there's two really distinct types of multiple employer plans. There's one where there's a large employer and then a joint venture is kind of tacked on it and smaller employer. So we're not talking about those situations. But where there's really a plan that covers a large number of employers and where there may not be any, or there may only be one substantial employer that meets the 10 percent test that you reference. In -- and so really to make this workable -- I think if you imposed a requirement just on the substantial employers and not all of the other employers, that would -- that would go a long way towards making it easy to do because the substantial employers at least are involved in the ongoing running of the plan, and you would be able to kind of work with them to make sure that if they just had any other plans, that the plans went through the necessary analysis. But extending it to all of the employers where you could have situations where you've got an employer that's just in for a few of their employees, but they also participate in some larger plan, that's what I see as the more problematic situation.

MR. WELLER: Okay, the other question. You referred to aggregating male and female lives for purposes of getting the mortality ratio and doing the -- getting the experience. My understanding is that isn't standard actual practice to do a mortality study that combines disparate mortalities. Can you cite to -- I recognize the fact that we don't require people to split out an hourly and a salaried, but can you cite to a -- something which actually shows it's reasonable to put together things which you know are inherently of different mortality experience?

MR. CADENHEAD: So first of all, I wouldn't say that it would be reasonable if you were coming up with population-wide tables, which is not what we're talking about here. Or if you were asking the employer to come up with their own tables, as was the case under the current regulations. What we're talking about here is an adjustment to a standard table, and in this case, the way I see it is, it's more a matter of how you define the standard table. Is it separate standard male and female tables? Or is -- or are the male and females together -- the standard table?

As far as analysis around this, I'm not aware of any specific studies. In preparing comments and working with, you know, a number of groups to try to assess whether or not this is reasonable, I did look at a few things. I did some analysis myself to say, well, what if -- for example, you had a population that exhibited the same underlying mortality, let's say, as -- you know, say it was a blue-collar -- follow the blue-collar tables, right. The adjustment relative to the standard no-collar table for male and female is not exactly the same. It's a little bit more an adjustment for males than for females, but they both move in the same direction. So the test that I did is if you looked at that separately and made a separate adjustment, versus aggregating together and making a single adjustment and measuring the liabilities, how much of an effect would it have. And I don't have the exact numbers in front of me, but the results were pretty close in terms of the overall effect on liabilities. And if that enables you to take, say, a population that has half credible -- or half -- has half enough deaths to be credible for males and half for females, and put them together and allows you to be fully credible, then I think that gets you closer to -- on an overall basis, the -- a reasonable outcome.

MS. JUDSON: So is that based on the assumption that males and females are equally distributed among the -- among job classifications, or you know --

MR. CADENHEAD: Well, it's not -- I think that's not even necessarily a premise of the current approach because when you're aggregating experience, say, within a gender, you're still -- you've got multiple job classifications. You've got multiple ages and the different ages could still exhibit different relative mortality to the baseline, and you would expect that would, in fact, be the case. So I'm really arguing more in terms of the overall effect. If you -- if you aggregate the male and female, and apply a single ratio developed on that basis, it will be weighted by the number of participants: Who are blue-collar, who are white-collar, who are different ages, who are male, who are female and by their benefits. So that you get an appropriate adjustment, kind of, reflecting the characteristics of the workforce.

MR. WELLER: Let me make sure that I'm following. Are you determining the mortality ratio by comparison to the gender separate tables, but you're just raising the credibility factor? Is that what you're -- or are you comparing the mortality ratio to some kind of blended table in the first place?

MR. CADENHEAD: There -- there are -- there's two ways you could look at it. One, is to just increase the credibility and apply then separate ratios to male and female. But the way that I looked at it was to do the analysis with the male and female experience together. So when you do that calculation, you are looking at --

MR. WELLER: I referenced to, what?

MR. CADENHEAD: I referenced to the -- to the separate male and female tables. So in other words, you're -- you are coming up with an age -- I'm sorry, a benefit-weighted number of deaths and you are not applying that to a single assumption, a single queue, at a single age, but rather to all of the different ages for male participants. You're doing the same thing for female participants. All I'm saying that is that instead of separately coming up with the ratio, add those results together in developing the ratio that is then applied -- the same ratio to the male and to the female tables. The -- mathematically, the analysis works just the same whether you -- or at least in terms of being able to apply it. If you apply it separately to male and female, or if you calculate the results on a male or female basis, and then add it up and come up with the single ratio.

MR. WELLER: You're saying the expected deaths would be based on separate genders?

MR. CADENHEAD: That's -- right. The expected deaths on the -- on the standard table would still be based on the separate gender tables, yes. And would be based on, you know, each participant's age so you would take the data that you know about that participant -- age and gender, apply the standard mortality rate, come up with the expected benefit weighted deaths. Do the same thing for females and combine the results.

MS. MAGEE: I have a question on the industry-wide mortality table. So are you -- I mean, for -- if a certain industry wants to create a table. If that employer will -- can participate voluntarily. Like, how does this going to work out?

MR. CADENHEAD: So I think practically to address the concerns about having the table relate to the experience of the employers who were using it, you would probably want to have some requirement that the study was based on the employers using it, as opposed to being -- you know, anybody could use a table based on someone else's experience. But it would be a -- instead of having employers come in on their own, if they can demonstrate that they have -- you know, a common industry and that the argument is that they are likely -- their employees are likely to be facing the same kind of effects, that they could come in together and pool their experience and come up with a table that allows for a greater degree of overall credibility.

MS. MAGEE: Okay.

MR. CADENHEAD: And again, you might want to -- you might want to look to see are they all trending in the same direction? You don't want to have one employer, you know, way over here offsetting the experience of another employer.

MS. MAGEE: What we say, 10 companies or 10 patient plans in that industry, 5 will come together, create a table. The other 5 could just do -- use the standard table. Is that --

MR. CADENHEAD: I think if you're -- if you are part of this and you are found to be a reasonable part of the group then, everybody that's part of that group would -- I would expect you would want them to all use it, or all not to use it at the same time. Now, that may raise transition issues. What if one of the employers acquires another group, you probably want to allow a reasonable period of time so that it doesn't throw all of the employers out of being able to -- to do that kind of -- to continue to use the table unpredictably.

MS. JUDSON: Thank you very much. Is there anyone who hasn't spoken who would like to speak? This ends the public hearing on proposed regulations. Thank you very much for coming, for your thoughtful comments, and your responses to questions. It's -- the hearing is on mortality tables for determining present value under defined benefit pension plans, Reg 11232415. Thanks again.

(Whereupon, at 11:20 a.m., the HEARING was adjourned.)