ORDERS:
FINAL ORDER AND DECISION (voluntary loss cost)
STATEMENT OF THE CASE
This
matter is before the Administrative Law Court (“ALC” or “Court”) pursuant to
requests for a contested case hearing made by the Consumer Advocate for the
State of South Carolina (“Consumer Advocate”), the South Carolina Small
Business Chamber (“Small Business Chamber”), the South Carolina Chamber of
Commerce (“State Chamber”), various local Chambers of Commerce in South
Carolina and the Homebuilders Association of South Carolina, the National
Federation of Independent Business, the South Carolina Manufacturers Alliance,
and the South Carolina Trucking Association (“various chambers of commerce”),
the National Council on Compensation Insurance, Inc. (“NCCI”), and Companion
Property and Casualty Insurance (“Companion”) regarding the filing (“Filing”)
NCCI made with the Department on July 1, 2005 requesting a 32.9% overall
increase in the current loss cost level for the voluntary workers’ compensation
insurance program in South Carolina.[1]
Pursuant to notice to the parties, a hearing in this matter was held
before me on April 24, 2006 through April 27, 2006, at the ALC in Columbia, South Carolina. After a careful review of the file and all of the evidence
presented, I find and conclude that an increase of 18.4% in the loss cost level
in the voluntary workers’ compensation insurance program is necessary and will
not be excessive, inadequate, or unfairly discriminatory.
PROCEDURAL HISTORY
This matter
began on July 1, 2005 with the Filing made by NCCI recommending a 32.9% overall
increase in the current loss cost level for the voluntary workers’ compensation
insurance program in South Carolina. The Filing proposed an effective date of
November 1, 2005.
The Department
provided notice of the Filing (“notice”) to the Consumer Advocate and published
the notice in The Post and Courier, The Greenville News, The
News and Press, The Rock Hill Evening Herald, and The State newspapers as required by S.C. Code Ann. § 38-73-910(A)(Supp. 2005). The notice
stated that any insured or affected party could request in writing by August 1,
2005, a public hearing upon the proposed rate increase before this Court.
On July 14, 15 and
18, 2005, and on August 1, 2005, the Consumer Advocate, the State Chamber, the
Small Business Chamber, and various chambers of commerce, respectively, filed
requests for a contested case hearing with respect to the Filing pursuant to
the South Carolina Administrative Procedures Act (“APA”) and S.C. Code Ann. §
38-73-910 (Supp. 2005). These requests were consolidated by order dated August
30, 2005 and resulted in a contested case bearing Docket No. 05-ALJ-09-0277-CC.[2]
On August 24, 2005, the Court conducted a teleconference
with the parties. During the teleconference the Department questioned
this Court’s jurisdiction to conduct a hearing on this issue of the increase
because the Department’s Director (“Director”) had neither approved nor
disapproved the Filing. The Department then made an oral Motion to Dismiss for
lack of jurisdiction.[3]
Separate and apart
from this proceeding, on September 2, 2005 the Director disapproved the Filing;
in the letter of disapproval, the Director stated that the 32.9% increase was
excessive. Subsequently, on September 6, 2005, the Department filed a written
Motion to Dismiss Docket No. 05-ALJ-09-0277-CC on the ground that the Court
lacked jurisdiction because the Department had elected to extend the period for
review of the Filing for an additional sixty (60) days as authorized by S.C.
Code Ann. § 38-73-960 (2002). This Motion was denied on November 8, 2005.
Two additional
requests for a contested case hearing were then filed challenging the
disapproval of the Filing, one by NCCI filed on September 9, 2005, which was
assigned Docket No. 05-ALJ-09-0355-CC, and one by Companion filed on September
16, 2005, which was assigned Docket No. 05-ALJ-09-0364-CC. These cases were
also consolidated with Docket No. 05-ALJ-09-0277-CC for discovery and hearing,
and all motions to intervene filed in the Voluntary Cases were granted.
In addition, on
October 10, 2005, the Director issued a Corrective Action Order (“Corrective
Action Order”) that addressed the rates to be charged for workers’ compensation
insurance obtained in South Carolina’s assigned risk program. The voluntary
and assigned risk programs are related. The Consumer Advocate then requested a
contested case hearing on October 12, 2005, to challenge the Corrective Action
Order; this request resulted in a case bearing Docket No. 05-ALJ-09-0406-CC
(“Assigned Risk Case”). However, this case was not consolidated with the
Voluntary Cases for hearing and was heard separately on April 28, 2006.
MOTION TO REOPEN RECORD
On June 23,
2006, some two months after this Court had concluded the trial of these cases,
the Consumer Advocate wrote a letter to the Court, stating that on June 19,
2006 it had received information (based upon an article contained in the June
22, 2006 issue of Charleston’s The Post and Courier newspaper) that NCCI
had problems with data accuracy and reliability in all the states in which it
operated. It requested that the Court reopen the hearing in the Voluntary
Cases to consider after-discovered evidence and what it considered to be
previously undisclosed evidence by NCCI.
In its response
filed on June 29, 2006, NCCI argued that the information contained in the
newspaper article had been in the media for over one year and that the Consumer
Advocate should have made specific requests concerning such during discovery
and/or addressed the issue during trial. In support of its argument for denial
of the request, NCCI attached to its response letter an affidavit of Peter
Burton, NCCI Senior Division Executive for State Relations. In his affidavit,
Mr. Burton stated that in 2004 NCCI began investigating a rate-filing related
inquiry from the New Hampshire Department of Insurance relative to payroll data
for a particular classification code. He stated that during its investigation NCCI
found that some payroll data had been excluded from certain loss cost filings
in 2003 and 2004; however, he stated that the data did not impact the financial
information which provided the basis for the overall loss cost calculations for
New Hampshire. Also, he stated that this payroll-based data was used solely
to develop classification code relativities that determine how the overall loss
cost change is allocated among the different classification codes. He stated
that the calculation of classification code relativities in the Filing in South Carolina is based upon the appropriate payroll data.
The Court finds
that the information concerning this rate inquiry could have easily been found
by the Consumer Advocate in its research and preparation for discovery and in
its preparation for trial. Furthermore, the Court finds that even if it were
to reopen the Record to consider any such information, such data would not be
of any meaningful benefit to the Court in deciding the loss cost level in the
voluntary workers’ compensation program in South Carolina since it would not
impact the financial information necessary to determine the loss cost. Thus,
the request is denied.
DISCOVERY AND PREFILED TESTIMONY
The parties
conducted active discovery, including submission of prefiled testimony, through
April 21, 2006. NCCI, the Department, the Consumer Advocate and Companion
filed prefiled testimony, as briefly summarized below, of the following
individuals who were all qualified as experts in the field of actuarial science
and rate making:
By NCCI
NCCI prefiled
testimony and supplemental testimony of Jay A. Rosen, a Fellow of the Casualty
Actuarial Society (“FCAS”), and the Director and Actuary for NCCI; testimony
and supplemental testimony of Dennis Mealy, FCAS, and the Chief Actuary for
NCCI; testimony of Barry I. Llewellyn, an associate of the Casualty Actuarial
Society (“ACAS”), and the Senior Divisional Executive for Regulatory Services
for NCCI; and testimony of Robert F. Conger, FCAS, a principal of Towers
Perrin, a past president of the Casualty Actuarial Society, and a consulting
actuary.
In their prefiled testimony, Mr. Rosen and Mr. Conger opined that an
increase in loss costs in the voluntary market of 32.9% would not lead to loss
costs that are excessive, inadequate, or unfairly discriminatory.
By the Department
The Department
prefiled testimony and supplemental testimony of Matthew P. Merlino, FCAS, the
owner and an officer of Merlino & Associates, Inc., and a consulting
actuary.
Mr. Merlino
opined that the loss costs should be increased by 22.4%. However, after
reviewing materials not available when he prepared his prefiled testimony, he
changed his opinion and, in supplemental prefiled testimony, concluded that an
increase in loss costs in the range of 39 - 44% was needed.[4]
By the Consumer Advocate
The Consumer
Advocate prefiled testimony and supplemental testimony of Martin M. Simons,
ACAS, and a consulting actuary. He opined that the data upon which all
parties’ actuarial calculations were based had not been shown to be accurate.
He stated that he could not support an increase in the voluntary loss cost
levels based on that data. However, because he concluded that the South Carolina workers’ compensation system was in a “troubled state” he recommended a
12.7% increase. He opined that an increase of 12.7% would not be excessive and
that the voluntary market in South Carolina would further deteriorate if there
were no increase. Mr. Martin Simons also
recommended a reduction of 5.0% to the loss costs
applicable to "F" classes, which was lower than the 5.9% requested by
NCCI.
By Companion
Companion
prefiled testimony of Jerelyn S. Boysia, FCAS, and Director of Actuarial
Services for Companion. She opined that an increase of 32.9% was necessary.
STIPULATIONS OF
FACT
In making certain
determinations herein, the Court relies on the following stipulations of fact
entered into and filed by the parties on April 13, 2006, and finds them to be
established facts:
Background
1. This
is a contested case to establish loss costs for workers’ compensation insurance
to be written in the voluntary program in South Carolina. This case is the
consolidation of three cases, all dealing with that subject. This case was
also consolidated for the purposes of discovery and pre-trial matters, but not
for trial, with that certain case before this Court bearing docket number
05-ALJ-09-0406-CC.
2. All
employers with more than a minimal number of employees are required to provide
workers’ compensation benefits for their employees. This coverage can be
obtained from three sources: (i) self-insurance programs, whether individually
or through self-insurance groups, (ii) the voluntary program, or (iii) the
assigned risk program, also known as the involuntary program.
3. Self-insurance
programs are not before this Court in this case.
4. The voluntary program consists of all employers whose
applications for workers’ compensation insurance have been accepted by an agent
for a workers’ compensation carrier licensed in South Carolina. The employer
chooses the insurer in the voluntary program.
5. The assigned risk program consists of employers whose
applications for workers’ compensation insurance coverage in the voluntary
program have been declined by at least two carriers licensed in this State.
Since these employers are unable to secure workers’ compensation insurance in
the voluntary program, assigned risk insurance is the program of last resort
for those employers to obtain the required insurance.
6. Policies
issued in the assigned risk program are issued by servicing carriers or direct
assignment carriers. The employer cannot choose its servicing carrier or its
direct assignment carrier.
7. A
direct assignment carrier is required to insure in the aggregate, the
percentage of the risk covered in the assigned risk program that equals the
percentage of that carrier’s participation in the voluntary program. Otherwise
a direct assignment carrier cannot refuse to write a policy for an employee
that has been assigned to it.
8. A
direct assignment carrier processes, issues, and services the policies it
issues to assigned risk employers.
9. All
employers obtaining assigned risk coverage that is not written by direct
assignment carriers are insured under policies issued by servicing carriers.
The assigned servicing carrier cannot refuse to write a policy to any
applicant/employer who has been refused coverage in the voluntary program, who
completes the application process properly, and who pays the required premium.
10. In calendar year 2004, the last year for which the data is available,
the total workers’ compensation insurance premiums reported by the insurance
carriers with respect to South Carolina voluntary and assigned risk programs
approximated Five Hundred Fifty-Five Million Dollars ($555,000,000).
11. By Order
dated March 1, 1990, the then Chief Insurance Commissioner of the South
Carolina Department of Insurance (the “Department”), acting with authority
granted by S.C. Code Ann. § 38-73-1430, directed that workers’ compensation
rate filings be treated differently for the voluntary and the assigned risk
programs. Specifically, he ordered that rate filings for the assigned risk
program propose full rates while rate filings for the voluntary program propose
only “loss costs.”
12. The rate
filing herein complies with that Order.
Loss
Costs v. Full Rates v. Premiums
13. In the
workers’ compensation field, the term “losses” means medical benefits paid to
or for the benefit of persons injured in workplace accidents, and lost wages
and other compensation paid with respect to those accidents. These lost wages
and other compensation are sometimes referred to as indemnity benefits.
14. In South Carolina, workers’ compensation insurance “loss costs” are the sum of the medical and
indemnity benefits plus the cost of providing these benefits. This cost is
referred to as “loss adjustment expense” or “LAE”. Loss costs represent the
costs that all insurance carriers have in common.[5]
15. Each
carrier determines its own final rates in the voluntary program by combining
its own expenses with the loss costs. These costs include production expenses,
a provision for Second Injury Fund assessments, general administrative
expenses, and a profit and contingency factor. These carrier-specific expenses
are used to develop a loss cost multiplier, or “LCM”, which is applied to the
loss costs to determine the carrier’s final rate.[6]
16. A “rate”
is the charge per unit of exposure (usually $100 of payroll) levied by an
insurer to cover those costs associated with providing workers’ compensation
insurance coverage. The “premium” is the dollar figure resulting after
multiplying the rate by the number of exposure units associated with a
particular risk.[7]
17. Workers’
compensation insurance risks may be grouped by classifications with separate
rates, where those classifications can be demonstrated to have variations in
hazards, or expense provisions, or both, which have a probable effect on losses
or expenses. S.C. Code Ann. § 38-73-430(3). Pursuant to this statute, the South Carolina workers’ compensation insurance market is divided into approximately 550
classifications.
Parties
18. Petitioner,
National Council on Compensation Insurance, Inc. (“NCCI”) is the largest
corporation in the United States dealing with workers’ compensation data,
statistics and research. It was formed at the urging of the National
Association of Insurance Commissioners and is active as a nonprofit
organization in 39 states. As contemplated in S.C. Code Ann. § 38-73-1210(A),
it is the rating organization licensed to make workers’ compensation insurance
rate filings in South Carolina on behalf of its members. Every workers’
compensation insurer writing business in South Carolina must be a member. S.C.
Code Ann. § 38-73-510.
19. NCCI
collects workers’ compensation experience and data from carriers with respect
to their experience inside the State of South Carolina as well as throughout
the United States.
20. The
Director of the Department (the “Director”) or her designee is required by
statute to approve the rate for each classification under which workers’
compensation insurance is written in South Carolina. S.C. Code Ann. §
38-73-490. The Director or her designee is authorized to disapprove any
workers’ compensation rate that is not fair, reasonable, adequate, and
nondiscriminatory. S.C. Code Ann. § 38-73-990.
21. The
Consumer Advocate for the State of South Carolina (the “Consumer Advocate”) was
admitted as an Intervenor in this case as an “affected party” pursuant to S.C.
Code Ann. §§ 37-6-607 and 38-73-910(A). The Consumer Advocate has a
discretionary and statutory duty to represent consumers in matters of rates.
S.C. Code Ann. §§ 37-6-604(A)(1) and 37-6-609.
22. Companion
Property & Casualty Insurance Company, Inc. (“Companion”) was admitted as
an Intervenor in this case. Companion is a South Carolina insurance company
that writes workers’ compensation insurance in South Carolina and will be
affected by the loss costs that are the subject of this case.
23. The South Carolina Chamber of Commerce (the “State Chamber”) was admitted as an Intervenor in
this case. The State Chamber has more than 2300 member companies and their
employees in South Carolina, a number of which will pay workers’ compensation
insurance premiums based on the loss costs that are the subject of this case.
24. The SC
Small Business Chamber of Commerce (the “Small Business Chamber”) was admitted
as an Intervenor in this case. The Small Business Chamber is a statewide
advocacy organization representing small businesses, a number of which will pay
workers’ compensation insurance premiums based on the loss costs that are the
subject of this case.
Rate Filing and Related Procedures
25. NCCI
made the filing, which is the subject of this case, on behalf of all licensed
workers’ compensation insurance carriers doing business in South Carolina and
as authorized by S.C. Code Ann. § 38-73-1210.
26. In
reviewing rate filings, the Director or her designee may take into account
recently passed legislation which will have an effect on insurance rates. S.C.
Code Ann. § 38-73-915(A).
27. A
workers’ compensation rate filing is subject to a waiting period of sixty (60)
days before it becomes effective. S.C. Code Ann. § 38-73-960. The period may
be extended by the Director or her designee for an additional period not to
exceed sixty days if the Director gives appropriate notice within the waiting
period. Id.
28. The
filing herein was made on July 1, 2005 and sought an effective date of November
1, 2005.
29. The
filing herein seeks an overall average increase of 32.9% to the current
voluntary loss cost level.
30. An
increase in workers’ compensation insurance rates may not be effective until at
least twelve months has lapsed since the last increase in such rates. S.C.
Code Ann. § 38-73-920.
31. The last
increase in workers’ compensation rates in South Carolina was effective July 1,
2004, which is more than twelve months prior to the date the changes proposed
in the filing herein would be effective.
32. S.C.
Code Ann. § 38-73-910(A) requires a public hearing before the Administrative
Law Court, upon a request received by an insured or affected party within
fifteen (15) days of public notice of such filing, as a condition for approval
of an increase in workers’ compensation insurance rates.
33. Proper
notice of this filing seeking an increase in workers’ compensation insurance
rates was released by the Department on or about July 11, 2005, and published
in newspapers of general, statewide circulation at least 30 days in advance of
this hearing and more than 30 days before the proposed rate change would take
effect.
34. The
Director disapproved this filing on September 2, 2005.
Positions
of the Parties
35. All
parties who submitted actuarial evidence in this case agreed that an increase
in the average voluntary loss costs was warranted. The parties’ prefiled
testimony represent that the following increases are reasonable and in
accordance with statutory requirements:
NCCI +
32.9%
Companion + 32.9%
Consumer Advocate +12.7%
Department +
39-44%. However, not considering the 2003 Second Injury
Fund law change, +22.4%
36. One of
the components of NCCI’s 32.9% increase is a .4% increase in the loss
adjustment expenses. Such increase is warranted and appropriate.
37. The
class code sensitivities currently in effect were established pursuant to this
Court’s February 26, 2003 Consent Order and Order of Dismissal entered in the
matter styled National Council on Compensation Insurance, Inc. v. South
Carolina Department of Insurance, et al., bearing docket no.
02-ALJ-09-0537-CC.
38. The
current class code relativities need to be updated to reflect current class
relativities.
39. The following persons, all of whom submitted prefiled
testimony, are qualified as expert witnesses in the areas of actuarial science
and rate-making:
Jay
A. Rosen
Robert
F. Conger
Dennis
Mealy
Barry
I. Llewellyn
Matthew
P. Merlino
Martin
M. Simons
Jerelyn
S. Boysia
ADDITIONAL FINDINGS OF FACT AND DISCUSSION
Having carefully
considered all of the testimony and evidence presented at the hearing in this
matter, having closely passed upon the credibility of the witnesses and having
considered the burden of persuasion by the parties, by a preponderance of the
evidence I make the following Findings of Fact, which are in addition to the
foregoing Stipulations:
General
40. No member of the public appeared at the contested
case hearing to present a protest or other position with respect to the
Filing. An opportunity was offered to the public at the hearing for comment;
none was offered.
41. The Filing seeks:
(a) an increase of thirty-two
and nine-tenths percent (32.9%) in the overall average loss cost level for
coverage written in the voluntary program for other than “F” classifications;
and
(b) a decrease of five and
nine-tenths percent (5.9%) in the overall average loss cost level for coverage
written in the voluntary program for “F” classifications; and
(c) a revision in the existing workers’ compensation classification code
relativities.
42. None of
the parties contested the revised workers’ compensation classification code
relativities proposed by NCCI in the Filing.
43. All
parties agreed to a decrease of five and nine-tenths (5.9%) percent in the loss
cost level for the “F” classifications except the Consumer Advocate which recommended
a decrease of five (5.0%) percent.
Accuracy and Reliability of the Data
44. South Carolina had the lowest workers’ compensation rates in the nation in 1998 but will
become 23rd in the United States if a 32.9% increase in the current
loss cost level is approved. Before approving an increase, the Court must be
satisfied by the preponderance of the evidence that the data relied upon by the
actuaries was accurate and reliable. The Court inquired during the hearing
how previous loss cost level increases of 17.5% (approved on April 1, 2003) and
11.3% (approved on July 1, 2004) could have been based on accurate and reliable
data, especially since NCCI is now recommending an additional 32.9% increase
slightly more than a year after the approval of the last increase.
Mr. Rosen
testified the problem with prior filings was not with the data but that NCCI
had underestimated the projected ultimate cost of claims used in the Filing.
Further, he testified that he would not be surprised if NCCI’s current estimates
of loss cost will increase again when future data is reviewed. Mr. Llewellyn
opined that the data used was actuarially appropriate and credible. Mr.
Merlino testified he had no concerns with the quality of the data, only with
NCCI’s methodology. However, Mr. Simons opined that he did not believe any
actuary could attest that any increase in loss costs would not be excessive,
inadequate, or unfairly discriminatory based upon this data.
45. NCCI
used data it received from carriers in making the calculations in the
Filing. Generally, this data is reported to NCCI in two forms. The first type
of data is financial call data that is supplied by carriers at NCCI’s
request. The second type of data is statistical unit data that is
periodically supplied by carriers. These types are commonly used in all NCCI
filings.[8]
46. The
Consumer Advocate and Small Business Chamber argued that the data used by NCCI
in the Filing is flawed and, by extension, the use of the allegedly flawed data
casts doubt on the results it yielded. Their argument was essentially
three-fold: (1) the financial call data did not include data from 16.3% of the workers’ compensation insurance carriers that
write in South Carolina; (2) NCCI is unable to ensure that Second Injury
Fund and subrogation recoveries are properly accounted for (deducted from
losses) and reflected in the data submitted by carriers to NCCI; and (3) the
error rates contained in the South Carolina test audit results cast doubt upon
the accuracy of the data reported by the carriers and the resulting data used
by NCCI in the Filing. NCCI, Companion, and the Department disagreed with the Consumer
Advocate and Small Business Chamber’s contention that the data used in the
Filing was flawed.
Missing
Data; not submitted or excluded.
47. Financial
data from nine insurance companies was not submitted or was rejected by NCCI
because it did not meet its quality standards.
48. NCCI
uses three tests to assess the quality and reliability of financial call data:
(1) an arithmetic check is performed, (2) followed by a reasonableness check
(to ensure all unusual fluctuations in the carrier’s data was sufficiently
explained) and (3) later the data is reconciled to the annual statement data
submitted by carriers to the Department.
49. The
financial data of AIG Insurance Company (“AIG”) was excluded because it did not
meet NCCI quality standards.[9] AIG wrote approximately
11%-12% of the policy year 2003 premium volume in the voluntary market and it is
the largest workers' compensation insurer in that market in South Carolina.
However, NCCI discovered “accounting anomalies” and “a couple of significant
business units” that AIG was not “able to” or “had not been” reporting to NCCI
and excluded its data from consideration in its Filing.[10]
All parties agreed that the AIG data and that from several other small
carriers, given their irregularities, should not be included in the Filing and
their absence was not a reason to disapprove the Filing.
Even though the data
from AIG (the largest workers’ compensation insurer in the voluntary market in
South Carolina) and several other carriers was not used in the Filing, I find
that even without such, the data in the financial call database used to develop
the proposed increase in loss cost level was sufficient to produce a filing
that was actuarially sound and appropriate. Although Mr. Simons questioned the
overall reliability of the data used in the Filing, he made his recommendation
of an increase of 12.7% by excluding the AIG data. Therefore, I find that
NCCI’s exclusion of the AIG data and that from several other insurance
companies was appropriate and the exclusion does not raise questions about the
quality of NCCI’s data. Further, I find that the data used by NCCI is sufficient
in quantity to base an actuarial judgment and the exclusion of the data
does not materially affect the accuracy or reliability of the Filing.
Second
Injury Fund; deduction of fund reimbursements and subrogation recoveries from
losses.
50. The Second
Injury Fund (“SIF” or “fund”) reimburses employers and/or their insurers
for at least a portion of claims for injured employees whose injuries were made
worse by a prior injury. It operates on an annual
cash flow basis and is funded continuously through equitable assessments to
each member (insurers, self-insurers and the State Accident Fund). Currently,
fund reimbursements are in excess of $100,000,000 annually. Insurers are
required to reduce their overall loss cost data by these reimbursements.[11]
51. The
fund is not authorized to reimburse an employer on a claim unless the injured
employee has suffered a previous injury. The second injury (upon which a claim
is made to the fund) must combine with or aggravate the impairment from the
prior injury sufficient to cause a liability or impairment that is
substantially greater than would have occurred from the second injury alone. Generally,
fund reimbursements occur a few years after a policy effective date and
possibly years after an injury occurs. The fund normally pays a claim two
weeks after it is accepted; however, a claim that was accepted in the year 2000
could have a genesis of five to ten years earlier. In fact, 26% of the claims
accepted by the fund in 2006 were more than four years old. A total of 10,543
claims have been accepted since 2000.
52. NCCI offered
several responses in support of its recommendation that the data it used accurately
reflected recoveries and supported its recommended increase in the loss cost.
First, it acknowledged that it does not audit data
used in a Filing and does not audit carriers (except in a limited number
of assigned risk cases) to determine whether they are properly applying and
reporting fund reimbursements and subrogation recoveries. However, it argued that
unaudited data may be used for actuarial purposes and that Actuarial Standard
of Practice (ASOP) 23 (that relates to the quality of data) does not require
data to be audited by an actuary.[12]
Secondly, it
noted that insurance companies in South Carolina must certify that they have
reduced their medical and indemnity reserves to the threshold limits of
reimbursements as a condition for reimbursement from the fund. When the fund
subsequently accepts a claim, it notifies both the carrier and the employer
that the claim has been accepted. However, the Court notes that the fund does
not notify the insurance agent, NCCI, or the Department of the acceptance of a
claim or when actual reimbursement is or will be made.
Thirdly, NCCI asserted
that licensed insurance companies in South Carolina are required to file annual
statements and that the overall loss cost indication in this matter was based
on the reported financial data reconciled to the carriers' annual statement
data. It noted that the Department periodically conducts financial examinations
of licensed insurance companies to ensure their financial systems and records
are in compliance with its insurance laws and accounting requirements. It
asserted that if a carrier underreported fund reimbursements or subrogation
recoveries, it would have been discovered in these financial examinations. The
Department does not examine annual statements filed by carriers; it relies on
certified public accounting firms (“CPA’s”) retained by carriers to examine
their financial statements and to provide opinions both to the carriers and the
Department. The Department was unable to confirm that the CPA’s retained by
carriers (whose data was used in the Filing) verified in their opinions that
fund reimbursements and subrogation recoveries were accounted for.[13]
Fourth, NCCI
argued that its statistical plan requires a reduction of the gross incurred
cost of a claim by the amount of any paid or anticipated recovery from a fund
when the fund determines a carrier is eligible for reimbursement.[14]
It noted that carriers must file a correction report identifying any recovery
from a third party or a fund, and that upon receipt of the correction report,
NCCI recalculates the experience rating of the carrier. The recalculation
causes a decrease in the premium charged to the employer. However, the failure
of carriers to reduce the incurred loss files in even a small or moderate
percentage of cases will substantially overstate the loss costs and result in
substantial cost to employers. Because of this incentive, it opined that
carriers, their agents and employers attempt to monitor all claims and notify
NCCI promptly of appropriate data to ensure they pay the appropriate premium.[15]
53. John
Gardner, an insurance agent who sells workers' compensation insurance,
testified that he often encounters situations where the experience rating for
the past two rating periods is not corrected following a recovery. He
explained that a loss goes into an employer’s experience mod file for three
consecutive years. Given the fact that it may take several years from the time
of injury to the processing of a claim by the SIF, the customer's experience
mod might have already been impacted by the loss for several years. In such
cases, Mr. Gardner stated that he has had to request a correction to the
experience mod and request a refund. Also, he stated that his 70 customers
have cases involving the SIF reimbursements and subrogation recoveries ten to
twenty times a year.
54. Notwithstanding, NCCI was unable to verify, based upon the financial
call data for the 2002- 2003 period, the procedure and process that carriers in
this State use, if any, to reduce their reported experience in recognition of
fund recoveries or subrogation recoveries. The rules of its statistical
plan only require carriers to identify claims where there has been a fund
reimbursement or a subrogation recovery. NCCI could not identify the amount of
the recoveries, the date of their receipt, or whether the carriers had sought a
reduction in their experience.
Based
upon the preponderance of the evidence, the Court is unable to determine when
or how fund reimbursements or subrogation recoveries, received or anticipated
by carriers, reduced the losses utilized by NCCI in the Filing.[16]
Apparently, there was no auditing or examination of the data either by the
Department or by NCCI. Because of concerns over the sufficiency of the
quantity and reliability of the data available and when and how reimbursements
were deducted from the losses submitted in the Filing, this data should not
have been considered in the Filing (if it was) and will not be considered by
the Court in its decision. Since the Department has, subsequent to this
Filing, instituted a new policy requiring its outside actuary in its financial
audits to investigate and review for information
concerning fund reimbursements or subrogation recoveries, the quality of data
in future filings should have a much sounder basis for actuarial calculations
for loss cost increases or decreases.
The 2003 amendment of the Second Injury Fund statute; Quantity and Reliability of the data since its effective
date.
55. Prior to the amendment to the Second
Injury Fund Statutes in 2003, the majority of claims were accepted by the fund
under the “unknown claim” provision. The 2003 statutory change deleted the
“unknown condition” claims. See S.C. Code Ann. § 42-9-400. After the
passage of the amendment, the SIF commissioned Mr. Simons to prepare a study of
its claims, liabilities, future liabilities, and assessments. Mr. Simons
stated in his report that the amendment would cause substantial reductions in
further liabilities of the fund and he expected that the fund’s accident year
reimbursements would decrease by approximately 40% to 50%.
56. When
NCCI learned of the passage of the amendment, it requested the fund to provide
it with data relating to the distribution of “unknown condition” claims and all
other claims upon which it could base an analysis of the amendment’s impact.
The fund notified NCCI it could not provide the data. NCCI then contacted the
carriers to obtain their opinions and expectations of the impact of the
amendment on the workers’ compensation costs. Based on the information it
collected, NCCI considered the effects of the amendment on the overall workers’
compensation costs to be negligible for the 2002- 2003 period and therefore did
not address it in calculating loss costs in the Filing. None of the expert witnesses who prefiled testimonies on
January 27, 2006 addressed the possible extent of the statutory change, either.[17]
57. During the hearing, the Department was especially
concerned with the effect that the amendment (concerning SIF claims and
reimbursements) could have on calculating voluntary loss costs, especially
since fund reimbursements make up approximately 83% of total reimbursements.
Originally, Mr. Merlino assumed that the amendment would have a
negligible impact on voluntary loss costs. However, after receipt and review of
Mr. Simons’ report (in which Mr. Simons concluded that fund reimbursements to
carriers would be significantly reduced as the result of the amendment), Mr.
Merlino submitted supplemental pre-filed testimony opining that the effect of
the amendment would cause a significant reduction in carriers’ recoveries from
the fund; further, he opined that the loss cost level needed to be increased in
the range of 39 – 44%.
At the hearing,
Mr. Merlino testified that he had reviewed Mr. Simons’ supplemental pre-filed
testimony and that an additional adjustment to his proposed loss cost level (in
his amended prefiled testimony) was necessary to account for the reduction in
expenses incurred by carriers to administer, or to contract with third-party
administrators who administer claims with the SIF. This adjustment resulted in
his final loss cost indication of a 37 – 42% increase and supported his
conclusion that NCCI’s proposed loss cost increase of 32.9% was not excessive.
Mr. Merlino agreed that if he had used the evaluation date and data used by Mr.
Simons in his 2005 SIF report, his indication would clearly be different. However,
he felt that the data pertaining to claims prior to 2003 was “very mature” while
data postdating the passage of the legislation was immature.[18]
58. Mr.
Rosen reviewed the same information considered by Mr. Merlino; however, he felt
the available data concerning the effect of the amendment was too immature on
which to base an actuarial opinion. Mr. Llewellyn opined that the amendment
would affect loss costs but that he had not been able to calculate its impact.
Mr. Conger opined that the effect of the amendment would probably be an
increase in loss costs in the short term with a decrease later on. Finally, Ms. Boysia anticipated the effect to be a 60% reduction in
fund recoveries. However, she testified that Companion was presently reducing
its level of recoveries by only 35%. Notwithstanding, she stated that the data
was immature.
59. Mr. Simons
stated that the purpose of the report he prepared for SIF in 2005 was to
estimate future assessments for members of the fund, not to estimate the impact
of the amendment on future loss costs. Further, he testified it had only been
two or three years since the passage of the amendment, and that since very few
claims had been made, the data was too green and immature to be considered in
the Filing. NCCI, the Consumer Advocate, and the Small Business Chamber were
also concerned whether the data available since the amendment’s passage was of
sufficient quantity and reliability to be considered in determining the future
loss cost level.
Notwithstanding
the immaturity and unreliability of the data, I find by a preponderance of the
evidence that the amendment will most probably at some point in time result in
a reduction of fund reimbursements to carriers, will reduce monies received by
carriers to offset their losses, and will increase the losses of carriers.
Further, I find that the amendment contributes to the troubled state of
workers’ compensation in this state and indicates that an increase in the loss
cost level is needed and required. This Court cannot discount the effect the
demise of this source of reimbursement will have on the loss cost level.
However, since the data is so immature and green, the Court cannot ascribe any
percentage factor for usage in determining the amount of the loss cost increase
that is presently needed.
Test Audits
60. NCCI uses a voluntary test audit program. The
program involves NCCI reviewing selective policies and thereafter preparing
quarterly reports that it sends to the carriers whose data is reviewed as part
of the test program. This program is conducted in Oregon, Florida and Alaska, also. The test audits in this matter covered policy years 2002 and 2003
and addressed a sampling of policies that had annual premiums between $4,000.00
and $250,000.00.[19]
61. There
were 98,620 worker’s compensation policies in effect for employers for the six
quarters ending June 30, 2003. Of these policies, 18,510 were within the
premium size subject to selection for the test audits. Only 690 policies (0.6%
of the total policies written in this state) were selected. The selection of
employers for audit is largely a random process; however, if an employer is
audited in one year, he is excluded from the test audit the following year.
62. For the
six quarters ending June 30, 2003 (for which financial call data was
requested), 54.24% of the policies audited by NCCI
had errors; further, the error ratio in South Carolina was trending upward and
was more than 100% higher than the error rates in other states. Despite these
high error ratios, NCCI did not increase the sample size to see if the results
were an anomaly. NCCI reported that about 51% of the errors involved incorrect
classifications with the remaining errors involved premium or payroll
differences.[20]
Because classification errors impact the premium paid by an employer to its
carrier for the coverage, if an employee is classified in a position that lists
functions less risky than the duties the employee performs, the carrier will
collect less premium than it should.[21]
63. NCCI also uses
the financial call data to determine classification relativities, loss costs,
assigned risk rates, and individual employer experience rating calculations.
After it completes the test audit, it notifies the carriers of any errors it
found. After sixty days, a carrier may submit a report to NCCI disputing the
report or listing corrections to the errors reported by NCCI. NCCI performs no
follow-up with the audited carriers after it sent its reports to them. Thus,
it has no data to show that any errors it found in the instant test audit and
which were subsequently reported to the carriers were or were not corrected.[22]
With respect to the argument that error rates
contained in the South Carolina test audit results cast doubt upon the accuracy
of the data reported by the carriers and the resulting data used by NCCI in the
Filing, I find that, based upon the evidence submitted in this case, this
argument is entitled to little consideration. While
errors of the type and magnitude described in the test audits are great cause
for concern (especially the failure of NCCI to check to see if the errors had
been corrected by the carriers), the Court is unable to find, based upon the
preponderance of the evidence, a quantifiable impact such errors had on the
accuracy or reliability of the data used in the Filing. In future years, audits
by NCCI should reflect more accurate data collections and error corrections. Accordingly,
I find the data used in the Filing is sufficient in quality, notwithstanding
the Court’s concerns with the data used in the test filings, to make actuarial
determinations.
Methods, Assumptions, and Trends
used by NCCI [23]
Paid loss method vs. paid
loss plus case reserves method.
64. In
calculating a proposed loss cost level increase, actuaries most commonly use
the following two methods: paid loss or paid loss plus case reserves. In the paid
loss method, an actuary estimates future loss costs using benefits previously
paid by insurers on reported claims during a selective experience period. In
the paid loss plus case reserves method or paid plus development projection, an
actuary estimates future loss costs by including in the data base previous
losses paid on reported claims during a selective experience period plus
amounts set aside (reserves) to cover future payments on claims during the same
selective experience period.[24] NCCI used an average of both
methods in this Filing and normally uses both methods when preparing an
analysis.
65. In
the Filing, NCCI recommended a proposed increase of 32.9% based upon an equal
weighing of the data, using both methods, for policy years 2002 and 2003.[25]
It believed that the usage solely of the paid loss method would produce an
understated loss cost indication. On the other hand, NCCI offered testimony
that loss cost indications could be overstated through use of only the paid
loss plus case reserves method because of recent indications of changes in the
level of carriers’ case reserves. Therefore, NCCI concluded that an average of
both methodologies, each of which is actuarially accepted, would produce a loss
cost indication that was most appropriate. Further, it felt that because of
the increased time it takes to close claims in this state along with the
increasing payments on those claims, the two methods were needed.
Both Mr. Conger and Mr. Rosen stated that when reserves have been
strengthened, the paid loss method is the more stable and preferable one to use
to ensure that loss costs are not overstated. Mr. Conger did not believe
there had been a change in the level of case reserves in South Carolina. However,
Mr. Merlino testified that the reserves, particularly medical reserves, had been
strengthened by 40% and had been strengthened over the last four evaluations.
Mr. Simons also testified that medical reserves have strengthened by almost 50%
and indemnity reserves have strengthened by 20%, for a combined strengthening
of close to 30%.
66. Mr. Simons noted that NCCI had reported the industry in South Carolina was making progress on reserve deficiency and that workers’ compensation
reserves had strengthened by $9 billion between 2001 and 2004. He testified
that the paid plus case reserves method is directly influenced by insurance
industry reserve practices and that changes in reserves directly impact data
and cause significant changes in the underlying development of workers’
compensation insurance claims and in loss cost trends over time. Further, he
testified that when reserves are substantially strengthened during the
selective period used in a filing’s calculations, the paid plus case reserves
data and the resulting indicated loss costs are distorted.
67. Mr. Merlino initially recommended an overall loss
cost increase of 22.4%. Unlike NCCI, he relied strictly on the paid loss
methodology and gave no weight to the paid plus case reserves methodology
because he concluded that the paid plus case reserves methodology was producing
a highly biased projection due to the recent case reserve strengthening. He
also used five policy years of data experience instead of the two years used by
NCCI.[26]
He noted that the use of five years of experience increased his indications for
paid projections. He opined that the latest two years of experience indicates
an increase of 19.1% as compared to five years of experience that indicates a
22.4% increase.[27]
Commenting on his use of five years of experience, Mr.
Merlino stated that generally two
years of
experience would be reasonable for developing loss cost indications. However,
in view of his concerns that NCCI was historically underestimating its
projected loss costs, he opted for the experience period that raised the
indications.
68. According to Mr. Conger, the paid plus case reserves method
takes advantage of the knowledge of claims examiners. Therefore, in his
opinion, NCCI’s use of the average of the paid loss and paid plus case reserves
methods likely produces a loss cost indication on the low side of the best
estimate. In addition, Mr. Conger believed that if NCCI had used the
methodology advanced by Mr. Merlino, it would have indicated an increase of
26.9%.
The Court finds, based
on the preponderance of the evidence presented, that there has been a
significant strengthening in the case reserves, particularly in the medical
case reserves, and, therefore, the use of the paid plus case reserves method
overstates the amount of loss costs. The use of the paid loss method, as
suggested for use by both Mr. Simons and Mr. Merlino, is the more appropriate
method to use based upon the strengthening of reserves by the carriers that
write the voluntary workers’ compensation insurance in this state. Further,
the Court finds that the usage of more than two years of experience was needed
because of the strengthening of reserves over a number of years and the need to
use more experience in determining future projections.
Voluntary Experience vs. Voluntary and Assigned Risk
Experience.
69. NCCI
opined that its loss cost analysis considered voluntary and assigned risk
exposures and losses to: (1) utilize the largest volume of data; (2)
make benchmarks consistent year to year; and (3) to encourage competition as a
matter of good public policy.[28] Since 2004, it has calculated voluntary loss costs by
using a combination of voluntary and assigned risk business.
70. The
difference between NCCI’s and Mr. Merlino’s proposed loss cost increase is
attributable, in part, to NCCI’s use of the combined experience from both the
voluntary and assigned risk programs to calculate the voluntary program loss
costs. NCCI cited several justifications for using the combined data from both
programs. First, NCCI opined
that the use of both the assigned risk and voluntary program data utilizes all
of the credible data available in this state to calculate the overall loss cost
level change. Second, it opined that the usage allows it to determine the
individual loss costs for each of the approximately 550 separate job
classifications in this state and to set those loss costs at levels equal to
the average employer’s experience in each of those classifications. Third, it
argued that this combination results in loss costs that will be consistently
calculated for each filing and would not be dependent on the relative sizes of
the voluntary and assigned risk programs that vary from year to year. Fourth, it
argued that since the voluntary program had better loss experience actuarially
than the assigned risk program, if it calculated loss costs by using only
voluntary program data, the result would be a loss cost indication in the
voluntary market that would be better than the average experience in this
state. Fifth, NCCI argued that the use of this method fostered good public
policy by encouraging depopulation of the assigned risk program, provided
incentives for carriers to voluntarily write as much business as possible, and
would result in an actively competitive voluntary program. Sixth and last, it opined
that the use of both the voluntary and assigned risk program data is
appropriate if, as in this instance, the objective is to establish a loss cost
level that is the best available indication of the average cost for claims and
claims handling to insure employers in this state. Both Mr. Conger and Mr.
Merlino agreed that it is important to consider the purpose of the Filing when
calculating loss cost levels.
71. The
market share adjustment factor used by NCCI considers the distribution of
voluntary versus assigned risk exposures for the experience period and also
includes an assumed distribution of business for the prospective period. NCCI
assumed the distribution of business for policy years 2002 and 2003 on the
actual distribution underlying the data. The distribution of assigned risk
assumed, after adjusting to a current level basis, was 15.8% for policy year
2002 and 20.3% for policy year 2003. For the
prospective period, NCCI assumed that all risks would be written in the
voluntary market and assumed that the distribution of business for the
prospective period would be 0% for the assigned risk market and 100% for the
voluntary market. According to Mr. Merlino, if one assumed that the
prospective period had an assumed distribution of business between the
voluntary and assigned risk equal to the 2003 policy year distribution, the
voluntary rate distribution would be reduced from+ 32.9% to 23.7%. This was
the basis of his original recommendation.
72. According
to Mr. Merlino, NCCI has proposed loss costs for the voluntary market that, on
average, are appropriate for the business included in NCCI’s combined voluntary
and assigned risk data base. However, since the assigned risk market has a
higher average loss cost than the voluntary market, he felt that NCCI’s
approach would result in a level of loss costs that would be too high for the
voluntary market. Further, Mr. Merlino stated that even though the Filing
includes voluntary and assigned risk business, NCCI's assumptions are
inconsistently applied to the exposures and losses. It was also his opinion
that the voluntary data is more representative of risks that will be written in
the voluntary market than the usage of the combined voluntary and assigned risk
data. Mr. Merlino separated the two markets and used only the voluntary market
data for his calculation of voluntary market loss costs.
73. Mr. Simons opined that although it is appropriate
to look at loss costs for the voluntary and assigned risk markets together, it
was not appropriate to look at it in the manner NCCI used in the Filing. He
believed that an appropriate analysis should include the effect of changes
taking place in the assigned risk as well as changes in the voluntary market. He
stated that there are interrelationships between the voluntary and assigned
risks, and noted that as employers are shifted between the voluntary and
assigned risk markets, changes in assigned risk trends will often result in
offsetting changes in voluntary trends. Therefore, in his opinion, it is
essential that the voluntary and assigned risk markets are trended separately.
This
Court is cognizant of the order dated March 1, 1990 by the then Chief
Commissioner of the Department, acting under the authority granted by S.C. Code
Ann. § 38-73-1430 (2002), wherein the Commissioner directed that workers’
compensation rate filings must be treated differently for the voluntary and the
assigned risk programs. Further, this Court has previously found that
employers in the assigned risk market usually have greater losses than other
employers, have greater risks, and the carriers refuse to insure them. National
Council on Compensation Insurance Inc. v. S.C. Department of Insurance,
Docket No.: 00-ALJ-09-0687-CC (2001). Therefore, I find, based on the
preponderance of the evidence presented, that calculating
voluntary loss costs by using a combination of voluntary and assigned risk
business data is inappropriate and unreasonably overstates the voluntary
loss costs. The Court further finds that it is appropriate to look at both
loss costs for the voluntary and assigned risk markets at the same time,
understanding that they have a close interrelationship, that employers often
shift between them, and that changes in the assigned risk trends often result
in offsetting changes in voluntary trends. However, both markets must be trended
separately.
Trending methodology - experience
period/number of years used.
74. Trends adjust the historical data to account for the impact of inflation on
losses and premiums between time periods. These trends reflect the relative
difference in the rate of changes in indemnity and
medical costs to the rate of change in payroll.
The
trend component of the loss cost formula is used to actuarially estimate future
changes in the underlying historical claim costs from the time period of the
historical data underlying the filing to account for the period the proposed
rates will be in effect. Whereas loss development estimates the adequacy of
current reserve estimates, trend estimates the change in the loss ratio from
the historical period to a future period. Trend factors are calculated
separately for indemnity and medical loss ratios.
75. The trends used
by NCCI in the Filing were calculated using an 8-year exponential trend
procedure.
76. The final
annual trend factors were analyzed separately for indemnity and medical losses,
including frequency and severity trend analyses and indicated countrywide trend
factors. Once the indemnity and medical cost ratios were developed and brought
to the current benefit level and trended, NCCI indicated that the current loss
cost levels should be increased by 32.4%. Further, NCCI added the proposed
change in the loss adjustment expense of 0.4% to arrive at its recommended increase
of 32.9%.
77. Both Mr.
Merlino and Mr. Simons used different periods of historical data to determine
trends. Mr. Simons used data beginning in 2000 because it was substantially
more favorable than the eight years of data NCCI used. He opined that the
trends from 2000 were much lower and were even better in 2003 than in 2002.
Also, he testified that the use of the latest indicators is more responsive to
positive change and that the use of more recent data was actuarially sound
given the conditions of the South Carolina marketplace. Finally, Mr. Simons opined that it is illogical that trends would be improving
so rapidly if experience was deteriorating at such a pace that an increase of
32.9% would be needed from the loss cost approved only a year before the Filing.
Subject to his reservations as to the reliability of the data used in the Filing,
Mr. Simons presented an indication for a loss cost increase of 12.7%. His
indicated change was calculated from an equal weight of the most recent policy
year and accident year of South Carolina paid loss indications based upon
8-year exponential trends. Lastly, Mr. Simons recommended that loss costs
applicable to “F” classes (governed by Federal laws) should be reduced by 5.0%.
78. Mr.
Merlino (as did Mr. Simons) used a five year experience base instead of the
eight year base used by NCCI. However, he did use NCCI's
trend factors. He did not adjust his projections for the missing data or the
data accuracy concerns (as expressed by Mr. Simons). The later increase in his
indications (subsequent to the review of the 2005 SIF report by Mr. Simons) was
due solely to his analysis of the 2003 law change regarding the Second Injury
Fund, which has already been addressed.
79. Other
factors that were considered in the trending was the previous underestimating
of future loss cost claims in earlier filings by NCCI and the deteriorating
state of the voluntary market in this state.
80. Also, the
Small Business Chamber contended that NCCI did not consider data from the South
Carolina Workers’ Compensation Commission (“Commission”). It noted that in the
executive summary (contained in the Filing), NCCI listed the following factors as
affecting loss costs in this state:
(1) workers’ compensation insurance
provides benefits to covered employees in this state that include unlimited
medical care, wage replacement benefits, vocational rehabilitation services,
and death benefits;
(2) since the workers’ compensation
program provides unlimited medical care, it does not contain many of the cost
containment mechanisms associated with traditional health insurance;
(3) the increase in the
projected number of claims and the expected amount for which those claims will
ultimately be settled has an impact on the loss costs;
(4) indemnity (wage
replacement) benefits are tied to wages (wage levels that are ever increasing
by statute, i.e., average weekly wage) and this ultimately affects
benefit costs;
(5) attorney involvement
affects loss costs;
(6) there is a trend
towards claims remaining open longer.
As to the assertion
that injured workers are provided unlimited medical care in this state, this
Court takes judicial notice of the law in this state that in the workers’
compensation program the carrier chooses the treating physician and has the
right to decide which treatment will be provided. Further, the Court takes
judicial notice that physician’s fees are regulated by a fee schedule set by
the Commission. The Court also takes judicial notice that medical treatment is
governed by S.C. Code Ann. § 42-16-60 (2002) and is not unlimited under the law
of this state, except in cases of permanent and total disability and, even in
those cases, the carrier continues to choose the doctor and continues to authorize
any treatment as provided under the fee schedule, unless otherwise ordered by
the Commission. Accordingly, the Court finds, based on the preponderance of
the evidence presented and applicable law, that unlimited medical care is not a
valid reason to increase the voluntary loss cost level by 32.9%.
Concerning the impact
of claim frequency and claim amounts on future loss costs levels, Mr. Rosen
stated that lost time claim frequency has decreased according to its data.
Further, he admitted that the medical severity has gone up only 2.5% in South Carolina and had declined when compared to the rest of country in the 2002 – 2003 time
frames. All the witnesses testified that medical costs continue to increase
above the growth in wages. The Court finds, based on the preponderance of the
evidence, that claim frequency and claim amounts do not support an increase of
32.9% in loss costs.
The executive summary
also asserts that wage levels are tied to indemnity benefits and, therefore,
affect workers’ compensation costs. The data reflects that wages in South Carolina increased by 3% in 2002 - 2003 and remains 10% below the national average.
The parties stipulated that workers’ compensation premiums are based on rate
times the units of exposure, i.e., payroll, and this Court notes that
premiums automatically increase (by statute) when wages increase. According to
NCCI’s data, the cost of indemnity payments over wage growth decreased in 2002
– 2003. The Court finds, based on the preponderance of the evidence, that the
decrease in indemnity payments over wage growth does not support a 32.9% loss
cost level increase.
The executive summary
further asserts that the high level of attorney involvement in South Carolina workers’ compensation claims affects costs. NCCI did not review or consider
any data available from the Commission concerning such prior to submission of
the Filing. Data from the Commission reflects that attorney involvement in
workers’ compensation cases decreased during the applicable period (2002 – 2003).
Further, no evidence was adduced that showed any impact on loss cost where
there was attorney involvement. The Court finds, based on the preponderance of
the evidence presented, that attorney involvement does not support an increase
of 32.9% in the loss costs.
Also, the executive
summary states that there is a trend in South Carolina towards claims remaining
open longer and that this increases the costs to carriers in this state.
However, a chart contained in the executive summary shows that the percentage
of lost-time claims closed within the 2002 – 2003 period improved. Further,
data from the Commission shows that claims have been closing more quickly in
2003 than in 2001 and are closing even much more quickly under the leadership
of its new chairman and executive director.
Finally, NCCI did not
consider any effect on future loss costs resulting from improvements in the
Commission’s funding, staffing, and technical support.[29] The Commission’s budget was cut $740,000 by the General Assembly in 2001 - 2002
and $1,200,000 in 2002 – 2003. By statute the Commission is composed of seven
commissioners who serve as hearing officers. However, during part of the 2002
– 2003 time period, the composition of the Commission consisted of only five
(5) commissioners, thirty percent of its staff, and little or no travel money
or per diem to compensate Commissioners to travel to their assigned hearing
locations throughout the state to conduct hearings. Further, the computer
system at the Commission was old and during the 2002 – 2003 periods it did not
function properly. With additional funding in the last several years, a new computer
system has been installed and new software programs are being installed over a
three year time period; further, the Commission has been increased to its full
complement of seven members, additional staff has been hired, and travel money
for the commissioners has been restored.[30] Hearings on claims are being heard more timely now.[31]
Both Mr. Rosen and Mr. Conger testified that these improvements would be
positive changes.
The Court finds that
NCCI should have considered available Commission data showing a trend towards
claims closing faster because of recent improvements in Commission funding,
staffing, and technical support. This trend, which contradicts NCCI’s
assertion that claims are remaining open longer, does not support an increase
in the loss costs of 32.9%.
Accordingly, I find,
based on the preponderance of the evidence presented, that including the evidence
of improvements in Commission funding, staffing, technical support, and claims
processing, the usage of five years of historical data (versus eight years) for
trending was responsive to positive changes in the marketplace.
CONCLUSIONS OF LAW
The Court must
make two determinations. The first is whether the data used by the various
actuaries is sufficient to support an actuarially sound adjustment to the
existing loss cost levels. If the answer to the first question is yes, then
the Court must determine the amount of the increase. In this case, all the
actuarial testimony reflected that an increase was justified; therefore, the primary
issue before the Court is the amount of the increase.
Based upon the
foregoing Stipulations and Findings of Fact, as well as the discussion, I
conclude, as a matter of law, the following:
1. The South Carolina Administrative Law Court is
empowered to hear this case pursuant to S.C. Code Ann. § 38-73-910(A) (Supp.
2005) and Chapter 23, Title 1 of the South Carolina Code of Laws (2005), as
amended.
2. Generally,
a request for insurance loss cost revisions is governed by S.C. Code Ann. §§
38-73-10, et seq. (2002).
3. Before
an increase in premium rates may be granted, notice of the filing must be
published in all newspapers of general, statewide circulation at least thirty
days in advance of the insurer’s proposed effective date of the increase. S.C.
Code Ann. § 38-73-910(A) (Supp. 2005). Further, a copy of the notice must be
sent to the Consumer Advocate. Id. An affected party may then request
a public hearing on the proposed filing with the Administrative Law Court
within fifteen (15) days of the date of the notice. Notice was timely published
and the Consumer Advocate acknowledged receipt of this notice. I find and
conclude that all requirements of these statutory provisions have been met.
4. The
Administrative Procedures Act requires that notice of the hearing must be
provided to all parties at least thirty days prior to the hearing. S.C. Code
Ann. § 1-23-320(a)(2005). Notice of the place, time, date and subject matter
of the hearing was timely given to all parties.
5. S.C.
Code Ann. § 38-73-960 (2002) requires that a proposed loss cost increase must
be on file with the Department at least sixty days before it becomes
effective. I find and conclude that this statutory requirement has been met.
6. S.C.
Code Ann. §§ 38-73-10(a)(1) (2002) and 38-73-430(4) (2002) require that
insurance rates must not be excessive, inadequate, or unfairly discriminatory.
S.C. Code Ann. § 38-73-430 addresses certain provisions that must be
complied with in the making of rates.
7. S.C. Code
Ann. § 1-23-350 (2005) requires that a final decision in a contested case shall
be in writing and shall include findings of fact and conclusions of law. The
findings of fact in a contested case must be based upon the evidence and
matters officially noted during the course of a hearing. S.C. Code Ann. §
1-23-320(g)(i) (2005). The decision of an Administrative Law Judge who
conducts and hears a contested case is a “final decision” as defined in the
Administrative Procedures Act. S.C. Code Ann. § 1-23-610 (2005).
8. NCCI has the
burden of proof to show, by a preponderance of the evidence, that the proposed
loss costs will not be excessive, inadequate or unfairly discriminatory. S.C.
Code Ann. § 38-73-10(a)(1) (2002). In civil cases,
generally, the burden of proof rests upon the party who asserts the affirmative
of an issue. 29 Am. Jur 2d, Evidence § 127 (2d
ed. 1994). The preponderance of the evidence “is evidence which is of
the greater weight or more convincing than the evidence which is offered in
opposition to it.” Black’s Law Dictionary 1220 (8th ed.
2004).
9. By
Order dated March 1, 1990, the then Chief Insurance Commissioner of the
Department, acting with authority granted by S.C. Code Ann. § 38-73-1430
(2002), directed that workers’ compensation rate filings be treated differently
for the voluntary and the assigned risk programs. Assigned risk market program
filings should propose complete rates, and voluntary program filings should
propose only the projected losses and the projected loss adjustment expenses (i.e.,
the loss costs per S.C. Code Ann. § 38-73-10 et seq. (2002)). I conclude that the Filing complies with that Order.
10. NCCI
is the nonpartisan rating bureau required by statute for workers’ compensation
insurance. S.C. Code Ann. § 38-73-510 (2002). NCCI collects workers’
compensation experience and data from carriers with respect to their experience
in this State as well as throughout the United States. NCCI made the Filing on
behalf of all licensed workers’ compensation insurance carriers doing business
in South Carolina and as authorized by S.C. Code Ann. § 38-73-1210 (2002). The
Filing was made on July 1, 2005 and it sought an effective date of November 1,
2005 as well as an overall average increase of 32.9% to the current voluntary
loss cost level. An increase in workers’ compensation insurance rates may not
be effective until at least twelve months has lapsed since the last increase in
such rates. S.C. Code Ann. § 38-73-920 (Supp. 2005).
11. The
Director of the Department or her designee is required by statute to approve
the rate for each classification under which workers’ compensation insurance is
written in South Carolina. S.C. Code Ann. § 38-73-490 (2002). The Director or
her designee is authorized to disapprove any workers’ compensation rate that is
not fair, reasonable, adequate, and nondiscriminatory. S.C. Code Ann. §
38-73-990. In reviewing rate filings, the Director or her designee may take
into account recently passed legislation that will have an effect on insurance
rates. S.C. Code Ann. § 38-73-915(A) (2002).
12. A
workers’ compensation rate filing is subject to a waiting period of sixty (60)
days before it becomes effective. S.C. Code Ann. § 38-73-960. The Director or
her designee may extend the period for an additional period not to exceed sixty
days if the Director gives appropriate notice within the waiting period. Id.
13. The
Consumer Advocate for the State of South Carolina was admitted as an Intervenor
in this case as an “affected party” pursuant to S.C. Code Ann. §§ 37-6-607 and 38-73-910(A)
(Supp. 2005). The Consumer Advocate has a discretionary and statutory duty to
represent consumers in matters of rates. S.C. Code Ann. §§ 37-6-604(A)(1) and
37-6-609 (Supp. 2005).
14. The
purpose of this hearing was to determine whether NCCI’s request for a 32.9% increase
in the current voluntary loss cost level is “fair, reasonable, adequate, and
nondiscriminatory” as required by S.C. Code Ann. §§ 38-73-10, 38-73-490, and
38-73-910 (2002 and Supp. 2005). All employers with more than a minimal number
of employees are required to provide workers’ compensation benefits for their
employees. This coverage can be obtained from three sources: (i)
self-insurance programs, whether individually or through self-insurance groups,
(ii) the voluntary program, or (iii) the assigned risk program, also known as
the involuntary program. Self-insurance programs are not before this Court in
this case.
15. In
the workers’ compensation field, the term “losses” means medical benefits paid
to or for the benefit of persons injured in workplace accidents, and lost wages
and other compensation paid with respect to those accidents. These lost wages
and other compensation are sometimes referred to as indemnity benefits. In South Carolina, workers’ compensation insurance “loss costs” are the sum of the medical and
indemnity benefits plus the cost of providing these benefits. This cost is
referred to as “loss adjustment expense” or “LAE.” Loss costs represent the
costs that all insurance carriers have in common.
Each carrier
determines its own final rates in the voluntary program by combining its own
expenses with the loss costs. These costs include production expenses, a
provision for Second Injury Fund assessments, general administrative expenses,
and a profit and contingency factor. These carrier-specific expenses are used
to develop a loss cost multiplier, or “LCM,” which is applied to the loss costs
to determine the carrier’s final rate.
16. Workers’
compensation insurance risks may be grouped by classifications with separate
rates, where those classifications can be demonstrated to have variations in
hazards, or expense provisions, or both, which have a probable effect on losses
or expenses. S.C. Code Ann. § 38-73-430(3) (2002). Pursuant to this statute,
the South Carolina workers’ compensation insurance market is divided into
approximately 550 classifications.
The class code
sensitivities currently in effect were established pursuant to this Court’s
February 26, 2003 Consent Order and Order of Dismissal entered in the matter
styled National Council on Compensation Insurance, Inc. v. South Carolina
Department of Insurance, et al., bearing docket no.
02-ALJ-09-0537-CC. The current class code relativities need to be updated to
reflect current class relativities. Since none of the parties contested the
revised workers’ compensation classification code relativities proposed by NCCI
in the Filing, the Court finds that the relativities as proposed are
appropriate.
All parties,
except the Consumer Advocate, agreed to a decrease of five and nine-tenths
(5.9%) percent in the loss cost level for the “F” classifications; the Consumer
Advocate recommended a decrease of five (5.0%) percent but did not provide
sufficient evidence that the decrease recommended by the other parties was not
appropriate. Accordingly, I find that a decrease of five and nine-tenths
(5.9%) percent in the loss cost level for the “F” classifications is
appropriate and must be approved.
17. NCCI
used data it received from carriers in making the calculations in the
Filing. Generally, this data is reported to NCCI in two forms. The first type
is financial call data supplied by carriers at NCCI’s request. The second type
is statistical unit data supplied periodically by carriers. These types are
commonly used in all NCCI filings. The Consumer Advocate and Small Business
Chamber argued that the data used by NCCI in the Filing was flawed and, by
extension, the use of the allegedly flawed data casts doubt on the results it
yielded.
The Court is
concerned with the evaluation process and the quantity of data used by NCCI
before it makes a recommendation for a loss cost increase, especially based
upon testimony by Mr. Rosen that in the two immediate loss cost filings it
underestimated the projected ultimate cost of claims used in the filing and
testimony that NCCI failed in its evaluation to have a process that ensured
that the data it used in the Filing was accurate. However, I find and conclude
that NCCI’s exclusion of the AIG data and data from several other small
insurance companies was appropriate, was sufficient in quantity to base an
actuarial judgment, and their exclusion does not raise questions about the
quality of NCCI’s data. Thus, I find and conclude that the missing or excluded
data does not materially affect the accuracy or reliability of the data used in
the Filing.
18. NCCI uses a voluntary test audit program in four
states, including South Carolina. As part of this program, it reviews
selective policies and thereafter prepares quarterly reports that it sends to
the carriers it uses in the program. The test audits in this matter covered
policy years 2002 and 2003 and addressed a sampling of policies that had annual
premiums between $4,000.00 and $250,000.00. Of the policies audited by NCCI,
54.24% had errors; further, the error ratio in South Carolina was trending
upward and was more than 100% higher than the error rates in other states. NCCI
reported that about 51% of the errors involved incorrect classifications and
the remaining errors involved premium or payroll differences.
NCCI performed no follow-up
with the audited carriers after it audited the information and sent its reports
to them. Thus, it had no data to show that any errors it found in the audits
and subsequently reported to the carriers were corrected. Prior to the
beginning of this trial, NCCI recognized that this process needed a follow-up
with the carriers that were audited and has amended its process to ensure that
errors reported to carriers are corrected.
While errors of the type and magnitude described in
the test audits are cause for concern, the Court is unable to find, based upon
the preponderance of the evidence, a quantifiable impact such errors had on the
accuracy or reliability of the data used in the Filing. I find the data used
in the Filing is sufficient in quality, notwithstanding the Court’s concerns,
to make actuarial determinations.
19. The
Second Injury Fund was created by the General Assembly for the purpose of
making payments in accordance with the provisions of S.C. Code Ann. §§
42-9-400, 42-9-410, and 42-7-310 (Supp. 2005). In general, the fund reimburses
employers and/or their insurer for at least a portion of claims for injured
employees whose injuries were made worse by a prior injury. A major concern
was whether the financial call data correctly reported the reduction of gross
loss costs by the carriers in this state.
Given the fact that the dollar amount associated with
reimbursements from the Second Injury Fund and from subrogation claims was in
excess of $100 million per year for the period under review in this case (which
represents a significant portion of the total workers compensation market under
consideration in this case), I conclude that the verification of reimbursements
issue is relevant to the Court’s decision in this case.
The Court is
unable, based on a preponderance of the evidence, to determine with specificity
when or how fund reimbursements or subrogation recoveries, received or
anticipated by carriers, reduced the losses utilized by NCCI in the Filing.
There was no examination or auditing of the data either by the Department or by
NCCI. Because of its concerns over the sufficiency of the quantity and
reliability of the data available and when and how reimbursements were deducted
from the losses submitted in the Filing, this data should not have been
considered in the Filing (if it was) and is not considered in this decision.
Notwithstanding
the immaturity and unreliability of the data since the passage of the amendment
to the fund in 2003, I find and conclude that the effect of the statutory
change will most probably result in a reduction of SIF reimbursements to
carriers in the future and thus will increase the losses of carriers. Further,
I find that the amendment contributes to the troubled state of workers’
compensation in this state. This Court cannot discount the effect the demise
of this source of reimbursement will have on the loss cost level. However,
since the data is so immature and green since its passage, the Court cannot
ascribe any percentage factor to such for usage in determining the amount of
the loss cost increase that is needed at this time in this state.
20. The Consumer
Advocate, the Department, and NCCI disagreed on the appropriate methodologies
for calculating a proposed loss cost level increase. Actuaries
most commonly use two methods, paid loss and paid loss plus case reserves. In
the paid loss method, an actuary estimates future loss cost using benefits
previously paid by insurers on reported claims during a selective experience
period. In the paid loss plus case reserves method, an actuary estimates
future loss costs by including in the data base previous losses paid on
reported claims during a selective experience period plus amounts set aside
(reserves) to cover future payments on claims during the same selective
experience period. NCCI used both methods in this Filing and normally
uses both methods when preparing an analysis.
The Court finds and
concludes that there has been a significant strengthening in case reserves,
particularly in medical case reserves, and, therefore, the use of the paid plus
case reserves method overstates the amount of loss costs. The paid loss
method, as suggested and utilized by both Mr. Simons and Mr. Merlino, is the
appropriate method to use in this review based upon the strengthening of
reserves by the carriers that write the voluntary workers’ compensation
insurance in this state. Further, the Court finds that the usage of more than
two years of experience, as used by both Mr. Simons and Mr. Merlino, was needed
because of the strengthening of reserves over the last few years.
21. NCCI's
loss cost analysis considered voluntary and assigned risk exposures and losses
in making its calculations. Both the Department and the Consumer
Advocate used only voluntary data. The Court agrees with the methodology used
by both the Department and the Consumer Advocate, finding and concluding that calculating voluntary loss costs by using a combination of
voluntary and assigned risk business data is inappropriate and
unreasonably overstates the voluntary loss costs. The Court further concludes
that, although it is appropriate to look at both loss costs for the voluntary
and assigned risk markets together, understanding that they have a close
interrelationship, that employers often shift between the two markets, and that
changes in the assigned risk trends will often result in offsetting changes in
voluntary trends, it is essential that both markets are trended and evaluated
for loss costs changes separately.
22. The Court
does not find any evidence supporting NCCI’s assertion that there is a trend
toward cases remaining open longer at the Commission that would support an
increase in loss costs of 32.9%. Further, the Court concludes that NCCI should
have considered available Commission data as factors in its trending,
especially since the data shows that claims are closing much quicker because of
recent improvements in Commission funding, staffing, and technical support.
Also, the Court
concludes that the usage of more recent data for trending, such as
improvements in Commission funding, staffing, technical support, and claims
processing, is more responsive to positive changes in the marketplace and
should be used versus the eight years of data that NCCI used.
23. I find and
conclude that NCCI has met its burden of proof in showing that an increase in
the voluntary loss costs must be approved. I further find and conclude that an
increase in the current voluntary loss cost level of 18.4 % (22.4% as initially
determined by the Department without any consideration given for the effect of
the 2003 statutory change to the SIF, less a 4% discount for recent
improvements at the Commission and data quality concerns) would neither be
excessive, inadequate, nor unfairly discriminatory.
ORDER
Accordingly, based
upon the above Findings of Fact and Conclusions of Law, it is therefore:
ORDERED that the current overall workers’ compensation insurance voluntary program loss
cost level be increased, on average (other than for “F” classifications), by
18.4%; and it is further
ORDERED that the proposed change in the workers’ compensation insurance voluntary
program loss costs for “F” classifications is decreased, on average, by 5.9%;
and it is further
ORDERED that the workers’ compensation classification codes be updated in accordance
with the relativities established by NCCI in the Filing that is the subject of
this case; and it is further
ORDERED that all changes ordered above shall be effective for new and renewal policies
issued on or after December 1, 2006.
AND IT IS SO ORDERED.
Marvin
F. Kittrell
Chief
Administrative Law Judge
October 2, 2006
Columbia, South Carolina
[1] The requests were consolidated for discovery and
trial and are referred to as the “Voluntary Cases.”
[2] By Order dated August 30, 2005, Docket No.
05-ALJ-09-0294-CC, which was originally assigned to the case filed on August 1,
2005 by the various chambers of commerce, was consolidated into Docket No.
05-ALJ-09-0277-CC and then dismissed.
[3] After listening to arguments on the Motion, a
written order was issued on August 30, 2005 which required the parties to
submit a written statement of their interpretations of the appropriate procedure
to be followed in the South Carolina voluntary workers’ compensation rate
cases.
[9] AIG
management informed NCCI in May or June of 2005 that it had discovered some
accounting anomalies that affected its ability to report data to NCCI.
Thereafter, NCCI started a remediation program with AIG and began notifying
regulators that AIG data was not going to be used in any NCCI filings.
According to NCCI, the anomalies consisted primarily of AIG not sending certain
data. It appears that NCCI itself discovered some of the inaccuracies during
the data validation process. NCCI had not received the corrected and validated
data from AIG at the time of the hearing.
The problems associated with missing AIG data do not
appear to be limited to South Carolina. Testimony was introduced regarding the
handling of the issue by the Florida Office of Insurance Regulation, an
equivalent of the Department in South Carolina. In connection with the NCCI
2005 filing for an overall decrease in workers' compensation rates in Florida, and due to the fact that the AIG data was also excluded from that filing, Florida regulators and AIG executed a consent order on January 5, 2006. That order found
that since AIG was the second largest writer of workers' compensation in Florida, its data became an integral part of the overall data used to develop workers'
compensation rates in Florida. The order incorporated a remediation that
required AIG to transmit to NCCI all information requested by NCCI and otherwise
required by Florida statutes within certain timeframes. Pursuant to the
order, AIG also agreed to pay an administrative fine in the amount of
$500,000.
[10] NCCI found coding irregularities in the AIG Data. Further, AIG had also
notified NCCI of certain irregularities in its data submissions that resulted
in data from certain of AIG’s business units not being reported to NCCI. After
unsuccessful attempts at accurately addressing AIG’s data quality issues, NCCI
decided to exclude the AIG Data from the Filing.
[11] A major purpose of the fund is to provide an
incentive to employers to hire employees who have suffered work-related
injuries that resulted in impairment. It provides protections to these
employers.
[12] The Court notes that NCCI did not opine that the
actuarial standard provides that data should not be audited or that unaudited
data is accurate. Furthermore, although
its recommendation of an increase of 32.9% in loss
costs is substantial and there was a high level of errors in its test audit,
NCCI did not audit the data used in preparing the Filing. However, NCCI
expressed confidence in its validation process throughout the hearing.
[14] Subrogation recoveries reduce a carrier’s incurred
losses, also. It is important to provide these credits to the claim files to
ensure an employer's experience rating modifier is not penalized due to actions
for which someone else has been determined to be liable.
[15] Companion seeks information on fund reimbursements
and subrogation recoveries in all its claims. Recoveries from the fund reduce
its losses. It is advantageous for Companion and other carriers to obtain and
report information on these reimbursements and recoveries to NCCI because they
reduce their losses. Ms. Boysia testified that Companion accounted for all
monies it received from the fund in the NCCI financial data call herein.
[16]
The Court does not suggest that fund reimbursements or subrogation recoveries
are not properly reported in the certified annual financial statements
insurance companies filed with the Department or in the certificates the
insurance companies submitted to the SIF to receive reimbursements after claims
had been accepted.
[18] When asked where he got the information he
relied on to make the estimates about the impact of the amendment, Mr. Merlino
admitted some of the information was anecdotal information in a memo he was
shown.
[21] For example, the risks associated with the job
of a roofer are greater than the risks associated with the job of a clerical
worker and, thus, the classification for a roofer results in higher premiums.
[22] Prior
to the beginning of this trial, however, NCCI amended its process to ensure
that errors reported to carriers are corrected.
[23] The parties disagreed on the appropriate
methodologies for calculating a proposed loss cost level increase.
[24] There are two kinds of reserves. Case reserves
are established on a claim-by-claim basis by individual claims examiners. The
other is the total provision that an insurance company is required to report on
its financial statement for its total obligation to pay.
[25] Policy year means premium and loss data on business
for a 12-month period for policies with inception dates within the 12-month
period.
[28] NCCI uses the assigned risk data to set a benchmark loss cost that can be used
by a new carrier in setting its rates. However, NCCI opined that the assigned
risk market has a materially worse loss experience than the voluntary market
and if it had only used the voluntary program data in its Filing, the indicated
resulting loss costs would be eight to nine points less than the 32.9%
recommended.
[29] The Small Business Chamber
and the Consumer Advocate asserted that the justifications used by NCCI
for the 32.9% increase were not supported by the data nor by factors received
from the South Carolina Workers’ Compensation Commission that showed recent
improvements in its funding, staffing, and technical support would reduce
future costs through the closure of claims more efficiently. NCCI, Companion
and the Department took the position that these concerns were immaterial for
the development of loss cost.
|