Document Sample
79 Powered By Docstoc



           Richard 3. Roth, Jr.
   California Department of Insurance


           Dean R. Anderson
       Tillinghast/Towers Perrin

              David Koegel
          Gill and Roeser, Inc.

MR. ROTH: My name is Richard Roth. I am the Assistant Commissioner of Insurance
for the State of California, and the chief property casualty actuary.

I was one of the people who was instrumental in changing Schedule P. I want to hasten to
say [ wasn't the only person. It was a team effort and the regulators could not have done
it without the cooperation of a number of actuaries in the insurance profession. They
deserve credit for that also.

Now, the heading is a little bit wrong. It should read) actually, Changes to Schedules F
and P since there is no longer a Schedule O. The format will be that we~-e going to have
a t5 minute presentation on the changes to Schedule F. Now, this actually ties in with P
because the changes to F -- reinsurance is now an important part of Schedule P, as you
will soon see.

To give that presentation will be David Koegel who is Assistant Vice President of Gill
and Roeser, Inc. David is an actuary with that firm. David.

MR. KOEGEL: Thanks very much) Dick. Good morning everybody. As Dick mentioned,
I'm going to make about a 15 minute presentation -- probably running to 17 minutes -- on
the changes to Schedule F. So bear with me as I'll be speaking fairly quickly. I'm
presuming some knowledge of Schedule F in the audience which may be a bad

Feel free to flip through the pages of the handout as [ go through the slides.

3ust out of curiosity, by a show of hands, how many people here work for reinsurance
companies? Don't be shy. Oh, okay. It's actually less than half the audience.


Basically what I'm going to be speaking about is the new provision for overdue authorized
reinsurance. I would like to start off with an old quote on the subject of debt: "if you
want time to pass quickly, just give your note for 90 days."

This thought has taken on new meaning for the reinsurance market. The NAICWs new
provision for overdue authorized reinsurance will affect so-called "slow-paying"
reinsurers in that they will be forced to alter their payment habits unless they are willing
to give up market share. Some ceding companies are going to be very frustrated with
their reinsurers who are sow payers as they~-e forced to take a hit to surplus as a result.


My talk will be broken down into three sections: first) the applicability and the intent of
the provision; second, the two-tiered nature of the hit to surplus and, finally, the
applicable changes to the 1989 annual statement blank.


The provision applies to -- simply stated -- reinsurance recoverables which are more than
90 days overdue, which are not in dispute -- I'll talk a little bit more about disputes in a
moment -- and, which are from authorized reinsurers. The NAIC's intent of this
provision is for companies to recognize, or write off, if you will, potential
uncollectibility of a portion of such overdue recoverables.

There's a lot on this next slide, and I'm going to go through it fairly quickly.


I look at the provision as if it's being broken down into two tiers. Tier one, which is the
less severe tier, is equal to 20 percent of paid loss recoverables more than 90 days
overdue according to the terms of the treaty. Sometimes though, the payment date may
not be specified in the reinsurance treaty. If that's the case, then the aging will begin 30
days after a bill is presented to the reinsurer or its intermediary requesting payment.

Tier two, which is more severe, is also 20 percent not only of the paid recoverables, but
of all recoverables including unearned premium, paid and unpaid losses, IBNR and paid
and unpaid allocated loss adjustment expenses.      But only to the extent that those
recoverables are not secured. Those of you familiar with Schedule F know that there is a
section for unauthorized reinsurers where a surplus penalty may be avoided if you have
adequate security in the form of funds held, a letter of credit, a trust agreement, or
some other form of security.

Whether a reinsurer is subject to tier one or tier two is determined in the new Schedule
F, Part 2B, Section 1, which I'll talk about further after the next two slides.

To quickly go over real the changes to the NAIC convention blank in 1989, there's three
areas of change. The major one is Schedule F. The next one would be the balance
sheet. The third is the capital and surplus account.

Schedule F, Part IA, Section 1 has been amended to separate paid loss recoverables into
four categories: current, 30 to 90 days overdue, 91 to 150 days, and over 180 days. Also,
there has been a column added to that exhibit, column 4 to show reinsurance premium
ceded. That total should reconcile with the underwriting and investment exhibit, Part
2B, Column 3, Line 32 on page 8.

There is an amendment in caption only of Part 2 to read Part 2A and that's for
unauthorized reinsurers. Then there is the new Part 2B, Sections I and 2 in which the
surplus penalty for overdue recoverables is determined.

A f t e r I talk about the balance sheet changes and the capital and surplus account, we'll go
back to Schedule F in a little more detail.

With respect to the balance sheet there are two changes. Add a new line 13(C) on page
3. It's going to say, "Provision for Overdue Authorized Reinsurance." This is referred to
on slide number 6 that is up on the screen now.


So, there is a new liability line for overdue authorized reinsurance. Line l~ which used
to read "Provision for Unauthorized Reinsurance" is now going to say *Provision for
Reinsurance R so as to include the unauthorized hits to surplus as well as the overdue
authorized hits to surplus.

Finally, with respect to the capital and surplus account on page 4, line 21 will now read
eChange in Liability for Reinsurance" to reflect the change during the year in the gain or
loss to surplus due to the provision as shown on page 3, line 14.

Now we'll slow down a little bit and talk about Schedule F, Part 2B because that's really
the crux of the matter.


There are going to be five columns of information for each applicable reinsurer. You're
going to have to list in Column l the amount of "90 days overdue" and that comes from
Part IA, Section l, Columns l(c)and l(d).

Column 2 is going to be the total recoverable on paid losses from each reinsurer from
Part IA, Section I, Column l(e). If you'll notice, there is a footnote on this slide which
says, "Amounts in dispute are excluded." Let's just digress a minute to the subject of

Under the old annual statement, there was a definition on the bottom of the page that
explained what constituted a dispute.            Either the ceding company or the assuming
company had to have initiated arbitration or other legal action concerning the
recoverable. From the regulators I've spoken to -- and I guess it would depend on what
state you're dealing in -it appears that the definition is now a bit broader and the
reinsurer's w r i t t e n denial of payment could be sufficient to establish a dispute.

For more on the subject of disputes, there is an interesting article in this month's Best
Review called "When Reinsurers Refuse to Pay." You may want to reference that if
you're interested in learning more about the subject.

Now, back to the schedule. Column 3 shows amounts received during the prior 90 days.
So, for 1989 that would be the amounts received in the fourth quarter of 1989.

In Column # -- [%e given this a name -- it doesn't technically have a name, but for the
purpose of illustration let's to call it the Overdue Ratio. The way the Overdue Ratio is
calculated, is Column l, 90 days overdue, divided by Column 2 plus Column 3, which is
total recoverable unpaid plus the amounts received during the prior 90 days.

Amounts received during the prior 90 days was an important addition to the denominator
of the Overdue Ratio. The reason was so that companies would not have incentive to
purposely delay the collection of on-time recoverables in the last quarter of the year to
inflate the denominator thereby lowering the overdue ratio.

The NAIC included amounts received during the last 90 days so as not to create a
disincentive to collect. They would have defeated the major purpose of the rule if they
omitted these amounts from the Overdue Ratio calculation.

Finally, in Column ~ of Section l you list the amount of 90 days overdue if the Overdue
Ratio is less than 20 percent. You can see on the bottom of this slide the footnote which
says, "the total of this column is carried to line (1) appearing beneath Section 2 for
calculation of the provision of overdue authorized reinsurance where 20 percent of these
amounts are a direct hit to surplus."

Now we move on to Section 2. You can refer to the old schedule for unauthorized
reinsurance because it's identical in format to this schedule but for unauthorized
companies. If the Overdue Ratio for a given reinsurer is 20 percent or greater --

meaning they didn't appear in that last column of Section 1 -- that reinsurer is designated
aslow payer -- and all recoverables from that reinsurer are listed in this section.

Columns I and 2, the sum of which is in Column 3, represent total recoverables, including
IBNR, unpaid loss expense, et cetera, from slow-paying reinsurers. Compare that to
Columns 4 and 5, the sum of which is in Column 6, representing the security that you
have against those recoverables.

The difference, or the excess, of the slow-pay recoverables over the security against
those recoverables is subject to a 20 percent penalty.

Finally, on the very bottom of Section 2 you will notice there are five lines under the
caption "Calculation of the Provision for Overdue Authorized Reinsurance." This, in a
capsule, is a review of what I~e just spoken about.


Basically the provision is equal to 20 percent of the quantity A plus B minus C. A
represents the overdue paid recoverables from non-slow pay reinsurers -- that's where
the Overdue Ratio is less than 20 percent -- plus the difference of B minus C -- which,
again, is the total recoverables from slow payers -- less the security against those
recoverables, 20 percent of which is a direct hit to surplus. As you can see, lines 4 and 5
at the bottom of Section 2, is where that calculation is precisely done.

Now, this provision is pulled from Schedule F and recorded as a liability on page 3, line
13(e), "provision for overdue authorized reinsurance." It's included in line 14 on the
balance sheet in the provision for reinsurance. The change in this liability during the
year is recorded in the capital and surplus account on page 4, line 21, as the "change in
liability for reinsurance." It represents a direct gain or loss to surplus.

I should point out here that the overdue authorized piece in 1989 is going to be a loss, it
can't be a gain. The reason the footnote reads "gain or (loss)" here is because in future
years line 21 could be a gain if the provision decreases during the year. So in future
years, a gain or loss could be realized depending on whether the provision decreases or
increases. Many companies will, however, be facing a very large surplus penalty in 1989.

I'd say in the next month or two there may be some preliminary estimates of the impact
these penalties are going to have. But right now companies are just scrambling around. I
don=t know how many of you here work on Schedule F, but traditionally it is the last
schedule to be prepared. I think it's incumbent upon companies to start a bit earlier,
perhaps at the beginning of the fourth quarter, instead of waiting until the end of the

That's basically it. I apologize if I rambled a bit. I guess we can take questions at the
end. Now I'll pass on to Dick. Thank you.

MR. ROTH" We hadarranged to have copies of the little Schedule P for 1989 available
to you. As of yet they haven't come in. When they come in, we,re going to put them
outside in the hallway.

I would also like to say that this session is being tape recorded and it will be transcribed
and put in with the final book. I will make a presentation going through Schedule P and

then Dean Anderson will follow with additional comments. Then~ at the end, we'll have
the session open for questions.

On Schedule P, first of all, let me start off by going through what was wrong with the old
Schedule P. What was wrong with the old Schedule P is that we in regulation thought it
was a pile of garbage. We almost declared it useless9 and it was even more serious than
that. We were finding that investment analysts on Wall Street were using Schedule P --
they'd put the data into their PCs and run off an analysis and say that such and such a
company was insolvent, sell. This was causing severe problems. Even our own examiners
were being misled by Schedule P.

But changing it was not easy. It took me over three years of trying before I was able to
work with the industry and work with the other insurance departments to put together a
package that we think is a step forward.

One of the problems~ first of all~ is the Schedule P combined lines of insurance that were
not homogeneous. An obvious example is private passenger auto and commercial auto.
Commercial auto is basically the long-haul truckers, and that is quite different from the
ordinary private passenger. Also, commercial peril and homeowners was combined

You can't really actuarially examine both together and when a company was under
examination what we did was have the company give us special runs of each one of the
individual lines so that we could perform an actuarial analysis on each one of the lines.
Well, this meant essentially that we were taking Schedule P and putting it aside and not
using it at all.

Another problem was that there was no separation of the allocated and the unallocated
loss adjustment expense reserves even though the paid amounts were broken out. Now,
this is a problem because the allocated loss adjustment expenses are quite different from
the unallocated~ and also we were finding that some companies did not even have an

So9 one of the main reasons that [ wanted the unallocated set out separately was because
I wanted to know if the company had one. Also~ having the unallocated in Parts 2 and 3
was not necessary.

Another problem9 and this was becoming more severe with loss portfolio transfers~ was
reinsurance. Reinsurance was paying a greater and greater role in Schedule P. So, the
regulators were concerned about Schedule P and also we were working on changes to
Schedule F. The problem of non-proportional reinsurance where you have a bulk transfer
in or out was causing difficulties.

Now9 what has happened is that all non-proportional reinsurance is now treated as if it
were a separate line.     So9 if you have assumed non-proportional reinsurance, that's
treated as a separate line and taken out. The only assumed reinsurance that's in Schedule
P is assumed proportional.

Another problem is that Part 2 and Part 3 with the incurred triangles -- the incurred was
only for six years and the paid was only four years. This isn't really quite enough for the
liability lines. We wanted to expand those lines.

Also, the Internal Revenue Service was going back ten years and the Securities and
Exchange Commission was getting more and more involved in this. For those reasons,
and out of actuarial necessity, we felt that we needed more years portrayed on Schedule

Another problem was the IBNR. Technically IBNR means strictly those losses which
were incurred but not reported. Some companies meticulously put down incurred but not
reported in Schedule P, Part I. Other companies included in that the bulk actuarial
reserve. In other words, the development of the case reserves and a reserve for reopened
cases. So we never really knew which was in there.

Another problem was that it only gave it for one year, namely the current year. To get
the IBNR for past years, like the past five years, we had to retrieve five annual
statements in order to put five years of IBNR in our computer in order to run an
actuarial analysis.  I'II get more involved in this later. Schedule P, Part l(f) was
unsatisfactory because of the ambiguity of the definition of IBNR.

I first started in the   insurance industry about 197t~. One of my goals in life has been to
destroy Schedule O.      I succeeded. Schedule O I always felt to be a complicated and not
very useful schedule.     You have to read the headings practically twice just to understand
it. From an actuarial     standpoint it really has no use.

If you want to make a test of any kind, the only thing you can do is test all of Schedule O
lines together. If you look in the 1988 blank, they have an IBNR and so forth for all of
the lines. Well, the lines are such a hodgepodge. You've got surety, youh~e got fidelity,
you've got international, and you've got reinsurance.

The reinsurance line is usually the dominant line as far as adverse runoff. Taking all
these diverse lines together, some of which are not necessarily property -- reinsurance
does not necessarily have to be property. It often is casualty.

From a regulatory standpoint you couldn't analyze it and come up with anything. It did
give the one year and two year development which was useful for the IRIS test, but that
was it. I wanted to be able to analyze things like automobile physical damage and even
f i d e l i t y insurety.

What was needed was to take the Schedule O lines and integrate them into a format for
Schedule P. Also, Schedule P was covered with ratios, percentages, and so forth, which
were not particularly useful and could also be recreated by a computer very easily. So,
in order to create additional space it was felt it was desirable to eliminate all these
ratios and just t r y to put in raw data.

Another problem we had in regulation was this. When we examined the company we
always tried to get direct data -- direct means that this was directly written -- and then
make a loss reserve analysis of the direct data to determine what the reserves should be
on a direct basis. Then subtract off the reinsurance and compare then the net with what
the company reported.

Schedule P was s t r i c t l y on a net basis and we didn't have the direct. One of the
interesting fallouts of analyzing the direct and then going to the net is that you find out
what the ceded loss ratio is. You have two loss ratios you ultimately get, the direct, the
net, and the ceded.

Well, the ceded is a very interesting regulation because that tells you what the
profitability is of the business thatts being ceded to the reinsurers. Surprisingly, very few
-- particularly small companies -- have any idea how profitable or whether they are
making money or whether the business being ceded is profitable or not profitable. Once
they pay that reinsurance premium they don~t care, it's just gone.

Well, w h a t happens is if t h e business that is being ceded is consistently unprofitable, we
know t h a t two or t h r e e years down the line they~.e not going to have any reinsurance.
Also, it says t h a t t h e business t h a t they,re writing is probably underpriced and t h a t they
will soon have problems.

Lastly, t h e r e was a growing towards c l a i m s - m a d e policies, namely policies w r i t t e n on a
r e p o r t year basis. The s t r i c t reading of the instructions require that Schedule P be only
on an a c c i d e n t year basis. Now, it's possible to report c l a i m s - m a d e policies on an
a c c i d e n t year basis but most companies just ignored the heading and wrote in on a r e p o r t
y e a r basis. Anyway, it was necessary to clean up that ambiguity so the reporting would
be on t h e basis on which t h e policy was written.

So, putting all these together, we were getting more and more horrified at what was
happening, particularly with respect to reinsurance. The increased utilization, the
reinsurance, was really destroying Schedule P. We felt, as regulators, that the reporting
information should have as much integrity as possible.

Substantial changes in the annual statement are not easy. It's easy to change headings
but it's very difficult to make a substantive change. One of the areas that underwent a
change are the reinsurers. In the rl0s and early '80s they were resistant to requiring new
information and reporting information. They felt that they were responsible in the
segment of the industry and that the main regulatory interest should be the primary

Then along came the Mission Insurance Company in California and we discovered a new
concept.    It's called uncollectable reinsurance.   When that occurred, suddenly the
reinsurance industry also became interested in financial reporting. I told the reinsurance
industry, you~e in the same business I am, you need to know what the primary companies
are doing, what they are writing, what the profitability is of their business. One of your
main tools is Schedule P. They agreed. With their cooperation, Schedule P was changed

Now, what I would like to do with that background is point out some of the things that
were done. I have some slides of Schedule P as I had put together a final draft for the
printer.   The printer then rearranged the tables to accommodate the problems of
printing. But the content is identical to what Schedule P will look like in the new annual

Unfortunately, it's a little bit difficult to read but ['11 try and go through it if [ can. My
problem is that if stray from the microphone, I'll also stray from the recording


A t the top here, instead of just net, we have premiums earned, direct assumed and
ceded. What we are starting here is a breakdown of the business into direct assumed,
ceded, and net.

Now, the t e r m d i r e c t and assumed has a special meaning. One of the roadblocks that we
had in t e r m s of splitting things out like this was what to do w i t h companies which
o p e r a t e on a pooling basis. The companies that operate on a pooling basis did not want to
r e p o r t as d i r e c t business the business that was w r i t t e n directly, but then ceded into the
company pool and then retroceded back.

If a company is operating on a pooling business -- incidentally, the pooling gets rather
c o m p l e x - - w h a t we are doing is that the NAIC w i l l be issuing special instructions and
these may have gone out already. If not, you should communicate w i t h the N A I C central
o f f i c e and ask for a set of supplementary instructions to Schedule P. The supplementary
i n s t r u c t i o n s answers all the question that we have received so far, plus it goes into great
detail on the pooling.

L e t me summarize, though, what the pooling concept is. Under the pooling concept, a
company does not report as direct business, business w r i t t e n directly. This Schedule P
assumes t h a t all of the business was w r i t t e n by the pool regardless of what subsidiary it
was w r i t t e n under. It was w r i t t e n by the pool, and then a percentage is ceded back to,
say, one of the pulp companies.

Well, the business ceded back to the pulp companies is treated as assumed business so
t h a t assumed means assumed business from a pool on a company that has a pooling
business. It also has the standard meaning of assumed reinsurance, assumed proportional
reinsurance regardless of whether there is a pool or not.

Ceded - - and the instructions go into this -- means in general ceded to a non-affiliate.              If
t h e r e is a pooling agreement, it does not mean ceded to the pool.

Now, you can have a pulp company that's ceded out to a non-affiliate. If a pulp company
cedes out to a n o n - a f f i l i a t e , that's included in ceded. If there is no pooling agreement
and you cede to an a f f i l i a t e , that's also included in ceded. But generally speaking ceded
means ceded to a n o n - a f f i l i a t e .

What we w a n t to look at here is the p r o f i t a b i l i t y of the business that was retained plus
the p r o f i t a b i l i t y of the business that was ceded outside the group or corporate entity.

Then, going across, many of the columns are repeated. You have loss payments. You
have a l l o c a t e d loss payments, and you have unallocated loss payments. One column that
is added is salvage and subrogation.

Salvage and subrogation previously was only a Schedule O item. It is now an item for all
lines, including Schedule P. So, we'd like to know for the liability lines what the amount
of subrogation is.

Under losses unpaid, we have it split into a case basis, direct and ceded. Then, we've
introduced a new concept called bulk plus IBNR. What l did here is I wanted to t r y to
remove all a m b i g u i t y in the t e r m IBNR. What we want under the term bulk and IBNR is
the a c t u a r i a l l y determined reserve, the large reserve that is not connected to any
specific case. The reserves in the file that are w i t h respect to a given case are called
case reserves. A n y t h i n g other is called under the term bulk and IBNR.

So, you no longer have to actually calculate this true IBNR, you know, the incurred but
not reported. You don't have to calculate that any more. A l l we want to know is what is
t h a t large reserve t h a t youh-e setting up that's other than an individual case reserve.

The same thing goes for allocated loss adjustment expenses unpaid. You have a case
basis or a bulk and [BNR. Under the case basis many companies actually have a reserve
for that in their claim files. Other companies may use a formula or whatever.

Then we have the unallocated loss adjustment expense reserve. Then, going down to the
bottom of the page what you do is you calculate the loss and expense ratios for the direct
assumed, ceded, and net.

Another item which is a new item here is called a discount. Now, Schedule P must be
reported on an undiscounted basis except for the tabular reserves in workers' comp.
Other than that, all of Schedule P must be on an undiscounted basis.

However, if the insurance department to which you are reporting has agreed to allow you
to report on a discounted basis, and you do so on page 3 or on other parts of the annual
statement, then there is a reconciliation. In other words, in the very last box you report
that discount so that you can reconcile between Schedule P and the other schedules. But
since this Schedule P is an actuarial content, we canWt use standard actuarial techniques
to a discounted loss reserve.

What you see here happens to be the summary. So, all lines have to be added up to the
summary. In each one of the lines you have to report all of this detailed information for
each one of the lines, and the number of lines defined has been greatly expanded.

Talking about the lines, homeowners and farm owners is one line. Private passenger auto
liability is a line.             Commercial liability is a line.     Workers w comp, commercial
m u l t i - p e r i l . Those are the main differences. International is a separate line. Then we
have, for the lines which we do not care too much about in regulation -- for instance,
ocean, marine and a i r c r a f t and boiler machinery -- those are combined into one line.

We also have a property line which is combined into one line. Then we have fire lines and
then marine, earthquake and glass. That's not particularly critical from a regulatory
standpoint. So, in fact, in that line we only have two years. You don't have to report ten
years~ only t w o years.

L e t me c l a r i f y another thing. Even though the original draft of this had ten years for all
lines and then the NAIC committee wanted a federal income tax -- you might call it like
an environmental impact study -- we wanted a federal tax impact study.

The Internal Revenue Service treats the old Schedule P lines differently. They have a
d i f f e r e n t formula and it turns out on the property lines that if I had gone back ten years,
according to the IRS rules, this would have triggered a different provision in the IRS and
the companies would have had to discount ten years. So, if I went back five, I had to
discount five. However many years I went back, thatts the number of years that you'd
have to discount.

Well, some companies made a calculation and found that even for property lines this
would have substantially increased their taxes because it would have increased the
discount. There are two reasons why the property lines are only two lines. One is not to
upset the IRS rules, and, two, I donWt need more than two years anyway for a property
line. Two years is fine for fire lines. Auto physical damage -- I don't need two years.
YouWd get a very good idea with two years, plus prior) on what the indications are.

Again, this is the summary line, and each individual line also has this. Part 2 is the
standard actuarial triangle of the incurred losses and allocated expenses only, no
unallocated in there.

What is done is that the top portion is filled in with numbers and what we have here is a
one and two year runoff. That one and two year runoff -- the regulators look at the
runoff to see how well the company is doing in reserving. Also, that will provide an input
for the IRIS test or the NAIC IRIS test.

Now, I've changed the line prior and this has caused quite a bit of confusion and a number
of questions. On the line prior, for 1980 -- if you can see a little asterisk here --the only
thing I want in that column is the outstanding reserves for accident years 1979 and
prior.   In the current annual statement what they've been doing is on the prior
accumulating all the incurred amounts since the beginning of the company.

Well, if you do that, the numbers in some companies are getting huge and they couldn't
even fit them in the box, and all of the paid amounts for prior years is of absolutely no
usefulness from an analytical standpoint. So I'm saying to the companies to take all the
paid amounts which you have made for 1979 and prior and just don't bother to tell us
about it. I don't need it and a l don't want it. 3ust give me the outstanding loss reserves.

Then I treat that as an incurred amount and what I want going horizontally across is the
development of those open claims or unreported claims going across so that I have a prior
line which then can be analyzed. So it's actually identical to the current prior line
except that it doesn't have all the old payments in it.

The middle box is Part 3 which are paid amounts. In this case the paid amounts have
been only four years, but now I'm going back ten years. The paid amounts, again, are
only for the losses and unallocated loss adjustment expenses. On the prior, you only
indicate the amounts paid in [981 and subsequent.

A number of analysts and people like to take the paid triangle and use that as a
projection.    They say they like to do it because in paid amounts companies can't
manipulate quite as much. I don't buy that argument. But, anyway, I have a paid triangle
for those who would like to see the development of the paid. Also, it gives you a table
upon which you can base a discount calculation. Discount is based on the payment
stream.     That will give you a ten-year payment stream upon which to calculate a
discount if you would like to do that.

The last two columns -- they happen to have x's in them in this case but for the individual
lines they don't have x's -- this gives claim count information. The New York Insurance
Department wanted a claim count information in there. Many people use claim count
information in order to calculate an average reserve. That's fine and the claim count
information will be available in there.

l will hasten to add that everybody is somewhat cautious about using claim count because
claim count can vary significantly by company to company. You can have one car with
five people in it and some companies call that a claim count of one; some people call it a
claim count of five. Not only that, some companies will open a claim file immediately~
some companies will wait six months to open a claim file, depending on just the way they
choose to handle the claims.

So, a claim count can vary dramatically from company to company. Hopefully, though,
within one company there should be some consistency.            In any event, it can and
sometimes is a useful item, a useful bit of information for the analyst.

Pat 6 is a new schedule which is a triangle of the bulk and !BNR reserve. Here, this gives
you in one table a history of the bulk and IBNR reserve so you don't have to pull past
annual statements.

We have completed our intent to have in one annual statement -- namely, the latest
annual statement -- all of the basic information needed to analyze the annual
statement. See, as you go through, all the slides represent a complete Schedule P. So
what you~e running through here are the various lines. You can see towards the bottom
there where you have only two accident years, those are property lines. All the rest
would be liability lines.

! rushed to get to the end here. If you haven't seen it before, this is what the new
Schedule P looks like. It just goes on with all these tables and tables of data for each
individual line. You can then read off from these tables directly into a PC. This is what
it is designed to do.

One thing that's added here is the Schedule P interrogatory. The footnotes in Schedule P
were becoming cumbersome. What I did is that all the footnotes are still operative but
they have been moved to an interrogatory.

Another thing that I eliminated was Schedules G and K having to do with fidelity
insurety. I moved some of that information over to this schedule. Up on top here is the
calculation of the excess statutory reserve. That calculation is still identically the same
from past years, subject to the additions of -- well, the excess statutory reserve is the

The second thing under the Schedule P interrogatory, if you do have a substantial amount
of claims-made policies, there is an interrogatory with respect to claims-made policies.
If you have a substantial amount of claims-made policies, you still have to put together
Part ~.

[ still have the definition of loss expense and the formula for distributing unallocated loss
adjustment expenses is still the same.

Then you have to report net premiums in force because of a requirement on the fidelity
insurety. I have one last question on there, and I'll read it because it's important. [t
says, "The information provided in Schedule P will be used by many persons to estimate
the adequacy of the current loss and expense reserves, among other things. Are there
any especially significant events, coverage, retention, or accounting changes, which have
occurred which must be considered when making such an analysis? H

Also, the individual who was responsible for the earthquake questionnaire sent out -- all
licensed companies in California must submit an earthquake questionnaire. In that
questionnaire I included a question similar to this. I've been getting back some extremely
useful and interesting commentary on the market and so forth on that questionnaire, so I
decided to add this.

Also, this question is designed to cut down the number of telephone calls. When they get
a Schedule P in and it looks strange, the analyst will call up the company and start asking
questions. The company say, "Oh, yeah, we know, mand they start listing off two or three

reasons, you know, "we merged a company," or, "we had a loss portfolio transfer," or "we
had a shortage in the reserves," or "we had a dramatic reserve strengthening," and so
forth. Sometimes they will follow-up with a letter.

So, in order to start putting into the statement some of these dramatic changes, I left an
opportunity for the company to say, okay, if you had a dramatic reserve strengthening or
you had a merger or you had some significant reinsurance transactions, or whatever, you
can put it in here. This will alert strangers who pick up your annual statement to make
an analysis, they are forewarned that there are significant changes. So, if they get
dramatic or unusual results, they won't run to Wall Street and issue a sellout or
something like that.

That's Schedule P. l'd like to introduce Dean Anderson of Tillinghast, who is also an
actuary. He will follow-up with additional comments.

If there is time and interest, I will go over a quick loss reserve methodology which can be
used. Basically it's an incurred triangle and a paid triangle. But I'd just as soon let Dean
Anderson speak right now. Thank you very much.


MR. ANDERSON: Thank you, Dick. My talk is going to be from the perspective of a
consulting actuary or outside analyst who has to evaluate a company's loss reserves using
only publicly available information, which is primarily the Schedules O and P from past
annual statements and the new Schedule P.

First, I will talk about what, in my opinion, were some of the limitations of the old
schedules and what are some of the benefits from the new schedules, as well as the types
of reserving techniques that can be used with the new schedules.

I would like to emphasize that my viewpoint will be from somebody outside the company
who is using these statements. Therefore, I see many benefits from the new Schedule
P. [ recognize that this viewpoint may not be shared by many people who have the
responsibility of compiling all this new data and information.

The old schedules had a number of limitations. One major limitation to both Schedules O
and P was the fact that they were only on a net basis. Over the last decade there have
been many reinsurance transactions dreamed up that have had very distorting effects on
individual company Schedules O and P.         We are no longer dealing with primarily
straightforward excess of loss and quota share reinsurance treaties.

Unfortunately, these programs have not been limited to the larger companies or specialty
companies. Many companies that in the past had "vanilla" type reinsurance programs
now have these distorting effects in their Schedules O and P.

For example. We were doing a reserve study for a workers' compensation company a few
years ago.     This company tended to reassume their first layer of excess of loss
reinsurance about four years after the beginning of the accident year. For a given
accident, year, the first three or four years of evaluations was at one retention level, say
$100,000, and then for all subsequent evaluation points the retention was $2~0,000. They
took back the case reserves and the [BNR that were related to that first reinsurance
layer at that evaluation point.      5o, they showed a big increase, or what would be
perceived as adverse development, in their loss triangle.

In the year that they reassumed this first layer reinsurance they also reassumed the
premium that was associated with those losses. But the premium was assigned to the
calendar year in which the transaction was performed. So, that year's premium was
overstated and, therefore, the loss ratio was understated. While the accident year three
or four years previous to that has an overstated loss ratio because have the premium
relating to the exposure they picked up when they reassumed the losses.

The old Schedules 0 and P had very limited data to perform actuarial analysis. This was
especially true if you only had one annual statement. You could perform some additional
techniques if you went back and picked up three, four or five annual statements. But if
you only had one statement, you were limited to doing a runoff review or using a
prospective test.

Part 2 of the old Schedule P provided a runoff review but it did not provide the data
necessary to perform an independent analysis of where the current reserves stood. Part
39 which showed the current evaluations and a limited amount of paid loss runoff, allowed
you to perform a prospective test by comparing the current paid-to-incurred ratios to
what actually happened to the older years.

In Schedule O, Parts I and 2 showed only a two-year runoff which, as Dick Roth
indicated, for many of the lines was insufficient. For the property lines it was probably
sufficient, but over the last number of years some additional lines have been added. The
reinsurance and international lines have gone into Schedule O. Those lines had a great
deal of casualty exposure and two years of runoff was totally insufficient.

If you went to the trouble and gathered up multiple years of annual statements, you could
do some limited actuarial analysis. It was possible, using Part l of Schedule P, to
develop paid loss triangles, incurred loss triangles, paid allocated loss adjustment expense
triangles, paid total loss adjustment expense triangles, and incurred total loss adjustment
expense triangles.

You had somewhat of a problem in backing out IBNR especially in the loss adjustment
expense triangles, in order to get a reported (paid plus case reserve estimate) triangle.
But you could do, what I would characterize as, dollar development techniques if you had
access to more than just the current annual statement.

In Schedule O you were limited to only a two-year runoff period. In Schedule P you were
limited to a five-year runoff period if you used only one annual statement. If you were
able to get multiple annual statements, you could get as much as an I I year runoff for
Schedule P, and you could get a similar runoff period for Schedule O if you were only
interested in reviewing all the Schedule O lines combined. On an individual line basis you
were limited to only a two-year runoff whether or not you used one annual statement or
a number of annual statements.

The definition of IBNR was a problem because there was not a consistent definition from
company to company. Some companies would define IBNR as being all of their statistical
actuarial reserves -- meaning pure IBNR as well as the bulk amount for case supplement
- - while other companies used the strict definition of IBNR and only included in that
reserve an amount for unreported claims.

Now, it would be good to have a consistent definition from company to company because
one of the things we like to do is combine annual statements from a number of
companies, which you can either gather yourself or you can gather from Best's or some
other outside source information, and develop industry benchmarks for paid development

and incurred development.     We use these to supplement the company's own data when we
do a reserve study.

This is especially appropriate for new companies, companies where the line of business is
new to t h e m even though they have been writing other lines for a number of years, and
for small companies where their own information is not 10096 credible.

Many of these issues have been addressed and in our opinion corrected, or at least
improved upon, with the new Schedule P. First of all, it is on both net and direct bases
so that you can do a reserve study on the direct business as well as a study on a net basis.

One of the interesting outcomes of this, in addition to the one that Dick Roth mentioned
regarding seeing the profitability of the ceded business, is to see how the ceded reserve
amount that you come up with compares to the amount the company is reporting in the
annual statement.    This is especially useful for a company that has a great deal of
reinsurance through companies that would normally be unauthorized except for the fact
that they are maintaining a fund balance with the company or a letter of credit.

If in your analysis of reserves on both direct and net bases you come up with a ceded
reserve amount significantly higher than what the company is showing in Schedule F, you
may come to the conclusion that the letters of credit or the fund balance amounts are
insufficient. Of course, it could work the other way too if your estimate is lower, then
you can feel much more comfortable that the company has sufficient backing for their

The new Schedule P for the liability lines will show a t e n - y e a r runoff period, which is a
w e l c o m e improvement for many of the lines of business, especially the unusual lines that
w e r e previously included in Schedule O for lack of a b e t t e r place to put them.

The data base includes a great deal of additional information. Claim count information
is now available, which was not available in the past except on an open count basis. On a
current evaluation basis you can get claims closed with payment and claims closed
w i t h o u t payment. You also can get the reported claims and the open claims. So, if you
do get multiple statements, you can develop count triangles to supplement your dollar

The ten years of data from one annual statement will allow you to develop dollar
triangles on a combined loss and allocated basis using paid losses or incurred losses~ the
incurred losses can both include or exclude IBNR. I prefer looking at it excluding IBNR,
but l recognize there are some people who do like to do an analysis including IBNR. Both
methods can be used with the current annual statement.

If you want to develop separate loss and allocated triangles, that is possible too. But you
have to use multiple annual statements and use Part 1. By using multiple annual
statements you can develop count triangles on a reported basis~ closed basis, both with
and w i t h o u t payment~ open counts, for loss and allocated individually or combined. You
can develop trianges for paid losses and incurred losses. By using the counts and the loss
dollars, you can derive severity amounts so many more types of actuarial techniques
available to you.

You can use straight loss development techniques, both paid and incurred. You can use
counts times averages techniques.      By having counts and dollar amounts, you can
investigate whether or not there has been a change in the case reserve adequacy level.

Most of these techniques, other than just the straight paid and incurred loss development
techniques, were not available under the old Schedules O and P even if you had multiple
annual statements.

On the claim counts I should probably point out that they are available for the casualty
lines only. They are not recorded for the property lines, reinsurance categories, or the
international category.

There is a new definition of IBNR. IBNR is defined to mean bulk reserves plus IBNR.
This is an improvement because it now creates more consistently from company to
company. So it is possible to develop better industry benchmarks for loss payment
patterns and loss improved patterns.

You still have a problem on the allocated side in this area in that many companies do not
set allocated reserves on a case basis but strictly on a statistical basis. So, all of their
allocated reserve would get thrown into the bulk reserve and there would be nothing on
the case side. Any benchmark you come up with on an industry standpoint would have to
take that into consideration. Perhaps you could come up with two benchmarks, one for
companies that do case basis allocated reserves, and one for those that do not set up case
allocated reserves,

Splitting up some of the lines of business has been a great improvement. As Dick Roth
mentioned earlier, there is significant difference m the patterns between private
passenger automobile and commercial automobile. The old multiple peril line of business
for Schedule P is now split three ways: commercial multiple peril~ homeowners and
farmownersl and the remainder which is primarily ocean marine, aircraft, and boiler and

The splitting of the reinsurance into several categories is especially important. Also, the
split between the claims made and the occurrence policies is important.

I would recommend perhaps an additional split that has not been accomplished yet. That
is to separate excess type policies from primary policies. I think the patterns, especially
for general liability umbrella and excess policies, are probably just as different from the
underlying policies as claims made and occurrence are different from each other. Itd like
to see data split between the primary policies and the excess policies.

The separation of allocated and unallocated loss adjustment expense has been very
important. Now you have basically the same information on allocated as you do on
losses. You can perform the same types of actuarial techniques to both categories. This
was not possible in the past. The old Schedule P gave you paid allocated, but the reserve
had allocated and unallocated combined and had case and bulk reserves combined. The
new detail on allocated loss adjustment expense is a big improvement.

Perhaps one of the major ramifications and, I think, benefits from this change to
Schedule P is that it has forced many companies to start compiling data that they
probably should have been compiling in the past. This type of detail really was necessary
to do a good actuarial analysis a company's reserves.

But, unfortunately, I have found in dealing with many companies that the reserving data
base, which I feel the Schedule P information is at the minimum level of needed detail, is
developed by considering Schedule P as being the maximum that they have to maintain.

What has happened by requiring this greater detail in Schedule P, is that it has required
many companies to now maintain the data that is necessary to perform an adequate
actuarial analysis.

Thank you.


MR. ROTH: I'm going to take a brief minute to go back to the slides, the last ones I had,
so I don't leave you hanging as to what their content is, and then I'm open for questions.


MR. ROTH" I need to explain this because I received some questions as to why there is a
Part 6 -- why do you have this bulk and IBNR in there and what use is it?

The intent is this. On Part 2 you have an incurred triangle. Incurred was the case plus
the bulk and IBNR plus the paid amounts. What normally we do is we take the
information from Part 6, which is the bulk and IBNR, and subtract it item by item, and
that leaves a triangle of case reserves plus the amounts paid to date.

What we%e done is eliminated, in a sense, the major judgmental factor from the analysis
and using only the triangle of the case plus paid to date, we get a gradually increasing
triangle. Then, you get numbers that dramatically increase. This is 338,000 up to 915 --
actually, it's up to 918,000. You get this dramatic increase. So the objective is to find
the pattern, the factors that will take you from 318,000 to 9lg,000 and say, okay, if this
is the pattern here, then this should be the pattern down here for 198g, and this number
here should grow at approximately the same rate.

So, these numbers down here are the ratios of successive numbers so that means for you
to project these amounts for 1988, 1987, 1986, on up to what their ultimate amounts
would be. Once you get the ultimate amounts, then you can subtract the ultimate
amount from the case plus paid and that will give you a test of what the II~NR and bulk
should be.

This is an internal way of testing that, and this is the most common method used by the
California Insurance Department.

The other way to do it -- the common way to do it -- is to take the paid table, which is
Table 3, and that has strictly the amounts paid to date. There are no reserves in that at
all. Looking at that pattern, you also go horizontally and project what the ultimate paid
amount would be. Once you have the ultimate paid amount, you can compare that
ultimate projection with the projection that you had from Table 2 and then you can also
compare that with what the company says is the ultimate. So that gives you the common

You can also just look at Table 2, which is the incurred amount, and look across to see
how well the development is. But the basis technique we use in regulation -- or, at least,
the California Insurance Department -- for testing is to take the incurred amounts,
subtract the bulk and IBNR, leaving the case in paid, and then project on that. We do
that for even property lines.

Now, the part I enjoy most is usually questions. If you have any questions for any one of
us, we'd be happy to entertain them. Yes?

QUESTION= Are you saying that the property line appeared (Inaudible.)
MR. ROTH: They will stay that way. No, it won't grow. It will stay at two years. Yes?

QUESTION: (Inaudible.) separate reserves allocated (Inaudible)       -- a company doesn't
have to reserve -- the Schedule P?

MR. ROTH: Are you talking about the allocated?

MR. ROTH= The company does not have to have, although I think it should, an allocated
loss adjustment expense reserve by a case. In that c a s e you're going to have to have a
bulk reserve. I won't give in on that. You do have to have a bulk reserve in that case for
the allocated loss adjustment expense, and that can be calculated in a number of ways.
But you at least have to have a bulk reserve. In that case, the case column would be

QUESTION= Your Schedule O lines which have your three-year development on the new
schedule - - on your Schedule P lines you have the ten. But in your summary you have a
ten-year. Are you not going to have to develop your Schedule P lines somewhere for ten
years so that when you summarize it you in fact pick up the ten year summary across the
board for all those years?

MR. ROTH: Yes. This is also a question that has been asked. All the summary tables
are for ten years. So the question is: on the old Schedule O lines if there are only two
years, what do they do with the summary?

The answer is that you have to have in your work papers ten years, even for the property
lines, the old Schedule 0 lines. You have to have ten years in your work papers so that
the summary is the sum of ten years for all lines, including the two-year.

I mentioned that one of the initial versions of it had ten years for the property lines and
that's why I did -- but this dispute about our interest in the Internal Revenue Service
came up and we decided to yield to that and report explicitly only two years even
thought the companies are required to have ten years within their internal work papers.

Does that answer your question?

QUESTION= Yes. Thank you.

MR. R O T H =   Yes?

QUESTION= I have a question for David. On the Schedule F if I understood correctly, it's
possible to avoid a tier two penalty in its entirety if you have offsets.

MR. KOEGEL= To the best of my knowledge that's correct, yes.

QUESTION: Is it possible to avoid a tier I penalty?

MR. KOEGEL: I don't believe so, because of the way the calculation works. Security
against recoverables from a reinsurer with an Overdue Ratio not greater than 20 percent
cannot be u~sed to offset the penalty.

I think what companies feel is that the tier one penalty will not be that serious because
you're taking one-fifth of a number which probably won't be that great. But when you
get to the tier two, it's a bit more of a concern.

I don't know if anybody here is aware of what Gill and Roeser does. We are reinsurance
i n t e r m e d i a r i e s specializing in financially-oriented or finite risk type reinsurance.
Companies have been inquiring as to whether there are bonafide ways to lessen the
surplus penalty.

I believe that the spirit of the rule is to guard against bad debt from companies which are
really in trouble, the reinsurers that in all likelihood are going to go insolvent. There
are, however, a number of very reputable and solvent reinsurers out there which for one
reason or another are slow in reimbursing their reinsureds. Let's face it, 90 days is really
not that long of a period of time. There is a lot of administration. Sometimes there are
disagreements that are not officially in dispute which undoubtedly cause delays.

Perhaps on an individual case-by-case basis, to answer your question, if it's a tier one
penalty relating to a quality reinsurer, I imagine that the particular jurisdiction may
listen to any arguments that you have in terms of security, and so forth. There is nothing
in the rule that evidences that you can get around a tier one penalty, but, again, it has to
be taken on an individual case-by-case basis.

MR. ROTH" Yes?

QUESTION: I'm not quite sure how you use the lines (Inaudible.)

MR. ROTH: Okay. I wanted to break it down as much as I could without defining a new
line. All of the lines, even though there are quite a few lines, are all previously defined
lines mainly on page 1#, the state page. So, no new lines are created.

The c r i t e r i a is -- well, for instance, the separating of private passenger auto from
commercial auto.           The loss development patterns are substantially different.   The
average loss is substantially different.         The number of death claims is substantially

When you use the development patterns and when you take a triangle like this and go to
project and you've got a mixture of these two, you sometimes will get distortion. The
same thing with homeowners and commercial multi-peril, they are substantially

What I'm trying to do is to get lines that are -- you know, break it down to try to get the
same character in loss pattern.

QUESTION: I have a question regarding the prior. For the new columns the salvage and
subrogation can be bulk and IBNR. (Inaudible.) -- to go back to the beginning of the

MR. ROTH: The salvage and sub?

QUESTION, Yes. Prior to the last? In other words, you must state it in that prior run.

MR. ROTH: Okay. The answer is this. On the prior on Part 1 there is a prior under
salvage and subrogation. However, for that role you only are asked to put the salvage
and subrogation r e c e i v e d i m m e d i a t e l y in the c u r r e n t calendar year, not historical. Like
t h e years 1980 through '89, you've got those ten years. You have to record the salvage
and subrogation that's relevant to each one of those absent years. Then, for prior you
only put in t h e c u r r e n t calendar year.

I don't know if it seems strange to you or not, but that's all the information I need. Then,
t h e reason I only want t h a t c u r r e n t calendar year is for reconciliation purposes, to
r e c o n c i l e t h e t o t a l with the other exhibits in the annual s t a t e m e n t .

In fact, the prior line -- there is a footnote which answers the question. For prior report
amounts paid or received in current year only, and then report cumulative amounts paid
or received in the specific years.

All loss p a y m e n t s are r e p o r t e d net or salvage and subrogation received. There is an
oddity here, but it's okay. In the triangles, you report the triangles net or salvage or
subrogation r e c e i v e d even though you cannot a n t i c i p a t e in loss reserving any salvage and
subrogation. That rule still holds. You cannot a n t i c i p a t e salvage and subrogation.
However, you can r e p o r t it. So, it would seem like there would be a distortion. There is,
but it turns out not to bother us.

QUESTION: David, have you tried to make any e s t i m a t e s on what the potential hit to
surplus will be for the industry on the Schedule F requirements?

MR. KOEGEL: No, not at this point. We don't regularly make such estimates, although
we've talked with several companies we believe will be affected in a major way by the
new provision. I think that toward the end of October or November, when companies will
be doing pro formas on their Schedule F we'll have a better feel as to the potential
impact on the industry.
                SEPTEMBER 18, 1989



               DAVID KOEGEL, ACAS
               A S S I S T A N T VICE PRESIDENT
               GILL AND ROESER, INC.
               NEW YORK, NEW YORK




                      - - R.B. T H O M A S
                          FARMER'S A L M A N A C



          --     APPLICABILITY         AND INTENT

          --     T W O T I E R S OF A S S E S S M E N T

          --     CONVENTION          BLANK CHANGES















          PART 2B -SECTION 1.





     Amend Part 1A - Section 1 to separate paid loss
     recoverables into four categories:

        (a)   Current;

        (b)   30-90 Days Overdue;

        (c)   91-180 Days Overdue;

        (d)   Over 180 Days Overdue.

     Add Reinsurance Premiums Ceded (Column 4) to
     Part 1A - Section 1.

     Amend caption of Part 2 to read Part 2A.

-    Add the new Part 2B - Sections 1 and 2.





-   Add a new Line 13e - " P r o v i s i o n for overdue
    authorized reinsurance as per Schedule F, Part 2B
    , Section 2".

-   Amend Line 14 to read "Provision for reinsurance",
    which    shall  include    provisions      for both
    unauthorized and overdue authorized reinsurance.


-   Amend Line 21 to read "Change in liability for
    r e i n s u r a n c e " , to reflect the change during the
    year in Page 3, Line 14.



           SCHEDULE F - PART 2B - SECTION 1


*   COLUMN 1: Amounts "gO Days Overdue";

*   COLUMN 2: Reinsurance Recoverable on Paid Losses;

    COLUMN 3: Amounts Received Prior 90 Days;

    COLUMN 4: C o 1 . 1 / (Col. 2 + C o l . 3)
                 [the Overdue Ratio];

#   COLUMN 5: Col. 1 amount if corresponding Col. 4 is less
              than 20%.

*   Amounts in dispute are excluded.

#   T o t a l of this column is carried to Line (1) appearing
    beneath Section 2 for calculation of provision for Overdue
    Authorized Reinsurance, where 20% of these amounts becomes
    a direct hit to surplus.



          SCHEDULE F - PART 2B - SECTION 2



  COLUMN 1:        Unearned Premiums;

  COLUMN 2:        (a) Paid and Unpaid Losses Recoverable;

                   (b) IBNR Losses Recoverable;

                   (c) Paid and Unpaid ALAE Recoverable;

  C O L U M N 3:   Total of I -I- 2a -t- 2b -I- 2c;

* COLUMN 4:        Deposits    by and     Funds Withheld    from

  COLUMN 5:        Miscellaneous Balances;

  COLUMN 6:        Sum of 4 -I- 5 but not in excess of 3.

* A m o u n ts should be identified separately as letters of
  credit (L), trust agreements (T), funds deposited by
  and withheld from reinsurer {F), or other (O).



          SCHEDULE F - PART 2B - SECTION 2


  Q U A N T I T Y (a) -I- (b) - (c) WHERE:

 (a)   equals overdue paid recoverables from non-"slow             payers"
       as set forth in the total of Column 5, Section 1;

 (b)   equals all recoverables from "slow        payers"   as set forth
       in in the total of Column 3, Section 2;

 (c)   equals all amounts which secure recoverables from            "slow
       payers" as set forth in the total of Column 6, Section 2.



Shared By: