Massachusetts Department of Social Services, DAB No. 1308 (1992)

Department of Health and Human Services

DEPARTMENTAL APPEALS BOARD

Appellate Division

SUBJECT:  Massachusetts Department of Social Services

DATE:  March 2, 1992
Docket Nos. 90-116 and 90-133
Decision No. 1308

DECISION

The Massachusetts Department of Social Services (State) appealed two
determinations affecting the State's funding under Title IV-E (Foster
Care) of the Social Security Act (Act):

(1)     Docket No. 90-116:  The Administration for Children and Families
(ACF) disallowed $14,114,086 in federal financial participation (FFP)
claimed as administrative costs under Title IV-E for the quarters ending
September 30 and December 31, 1986.  This amount represented the
difference between the amount the State had originally claimed for
certain social worker costs and the State's revised claims submitted in
1988.  The State subsequently proposed an allocation method under which
it would be entitled to $1,299,830 FFP for the two quarters, in addition
to its original claims (thus effectively acknowledging that $12,814,256
of the revised claims were unallowable).  Transcript of Hearing held
12/5/90 - 12/6/90 (Tr. I) at 8-9.

(2)  Docket No. 90-133:  The Regional Director, Region I, affirmed a
decision by the Division of Cost Allocation (DCA).  DCA had denied the
State's request that the modified amendment to its cost allocation plan
(CAP), as approved by DCA on July 13, 1989 with an effective date of
July 1, 1988, be given a July 1, 1986 effective date.

ACF based its disallowance on two grounds:  (1) that the retroactive
claim for the quarter ended September 30, 1986 was not timely filed; and
(2) that the State's claims were not in accordance with the CAP in
effect during 1986 and were calculated using the results of time studies
conducted in later periods.  DCA's decision to.deny retroactive approval
of the State's amended CAP (which called for use of a time study) was
based on the fact that the State had no valid time study data from the
retroactive periods.

For the reasons stated below, we uphold DCA's decision.  Retroactive
amendment of a CAP is appropriate only if specified regulatory standards
are met, and DCA reasonably found that those standards were not met
here.  Moreover, even an allocation method proposed prospectively must
be valid.  The allocation method the State proposed here (which differs
from the time study method in the amended CAP) essentially involves
"backcasting" of data to the disallowance period.  Under prior Board
decisions, the State had the burden of showing that "backcasting" was
valid.  While the State here satisfactorily answered some of DCA's
concerns about "backcasting," the State's presentation was nonetheless
flawed in certain key respects.

In light of our decision denying retroactive amendment of the CAP, we
also uphold ACF's disallowance of the claims for the two 1986 quarters,
as inconsistent with the applicable CAP.  We also conclude, as ACF did,
that the revised claim for the first of those quarters was barred
because it was not timely filed.

Title IV-E

Title IV-E of the Act (enacted by the Adoption Assistance and Child
Welfare Act of 1980, Public Law 96-272) provides for maintenance
payments for children in foster care and adoption assistance for special
needs children.  Title IV-E funds are made available for states which
have submitted, and had approved by ACF, state plans meeting federal
requirements.  Section 471 of the Act.  FFP is available, under an
approved state plan, for foster care maintenance payments or adoption
assistance for children who meet specified eligibility requirements.
Sections 472, 473 of the Act.  In addition, states may receive FFP for
expenditures "found necessary . . . for the proper and efficient
administration of the State plan," including certain training costs (75%
FFP) and other administrative expenditures (50% FFP).  Section 474(a)(3)
of the Act.

ACF regulations implementing Title IV-E are codified at 45 C.F.R. Parts
1355 and 1356.  Section 1356.60(c), which governs administrative costs
other than training, states:  "The State's cost allocation plan shall
identify which costs are allocated and claimed under [Title IV-E]."
This section specifies certain costs directly related.only to the
administration of the Title IV-E foster care program (such as
determination of eligibility), which may not be claimed under any other
federal program.  The section also gives examples of activities which
will be considered necessary for the administration of the foster care
program (such as developing a case plan and placing a child) and other
activities which are not allowable under Title IV-E (such as counseling
or treatment services to remedy a child's behavior).

The Cost Allocation Plan Process

A state participating in public assistance programs under the Act,
including Title IV-E, is required to determine what part of commonly
incurred expenditures, such as staff salaries, is allocable to each
program the state administers.  A state is required to submit a plan for
cost allocation to the Director, DCA, in the appropriate HHS regional
office.  45 C.F.R. .95.507(a).  This cost allocation plan is defined as
"a narrative description of the procedures that the State agency will
use in identifying, measuring, and allocating all State agency costs
incurred in support of all programs administered by the State agency."
45 C.F.R. .95.505.  The CAP must contain sufficient information to
permit the DCA Director to make an informed judgment on the correctness
and fairness of the state's procedures for identifying, measuring, and
allocating all costs to each of the programs administered by the state
agency.  45 C.F.R. .95.507(a)(4).

A state may amend its CAP for various reasons, including the discovery
of a material defect in the CAP or a change which makes the allocation
basis or procedures in the approved CAP invalid.  45 C.F.R. .95.509(a).
The regulation at 45 C.F.R. .95.515 describes when a CAP amendment takes
effect:

 As a general rule, the effective date of a cost allocation plan
 amendment shall be the first date of the calendar quarter
 following the date of the event that required the amendment. . .
 .  However, the effective date of the amendment may be earlier
 or later under the following conditions:

 (a)  An earlier date is needed to avoid a significant inequity
 to either the State or Federal Government.

 (b)  The information provided by the State which was used to
 approve a previous plan or plan amendment is later found to be
 materially.incomplete or inaccurate, or the previously approved
 plan is later found to violate a Federal statute or regulation.
 In either situation, the effective date of any required
 modification to the plan will be the same as the effective date
 of the plan or plan amendment that contained the defect.

If a CAP amendment is disapproved, the state may appeal the disapproval
first to the Regional Director and then to this Board.  Costs not
claimed in accordance with an approved CAP (or a pending proposed
amendment) will be disallowed.  45 C.F.R. .95.519.

Factual Background

In 1986 Massachusetts was claiming its Title IV-E administrative costs
under a CAP that had been approved by DCA in 1983 with an effective date
of October 1, 1982, and modified in 1985 (CAP I).  The part of CAP I at
issue here sets out the allocation method for a pool of costs (cost
pool) consisting of salaries and related costs of certain social workers
(and their supervisors).  As described at the hearing in this case,
these social workers provide an array of services to protect children
whose welfare is at risk and to strengthen families, including
investigating child abuse, recruiting and training foster homes,
identifying family problems, developing service plans, and providing
counseling and therapy.  Tr. I at 71-75. 1/

Under CAP I, the State allocated these social worker costs to Title IV-E
through the use of a ratio, with the "# of IV-E Cases Claimed" as the
numerator of the ratio and the "Generic Caseload Count" as the
denominator.  This ratio, expressed as a percentage, was then multiplied
by an amount equal to 50% of the total cost pool to determine the amount
of FFP the State could claim for these costs under Title IV-E.  The
State said that the actual ratios used during the period modified CAP I
was in effect (January 1, 1985 until March 31, 1987) ranged from 4% to
8%.  State Ex. 15, Att. B.

In 1987, DCA performed a review of CAP I.  A State official testified
that his understanding was that DCA was concerned that Massachusetts had
the highest rate of .administrative to maintenance claims in the
country.  Tr. I at 19.  In the course of its review, DCA learned that
the State was interpreting CAP I differently from DCA.  DCA found that
the State was using a count of individuals in the numerator of the
allocation ratio, and a count of family groups in the denominator.  In
DCA's opinion, this was impermissible because it was not using
comparable data.  Tr. I at 203.  DCA advised the State that it must
propose a new method for allocating social worker costs, suggesting use
of a study of social worker time (as DCA had also suggested to the State
previously).  Tr. I at 25-34, 199-100.

The State rejected the use of a time study and instead negotiated with
DCA about various proposed methods, generally based on caseload
statistics.  Tr. I at 28-34, 60-61.  DCA rejected all of these methods
except one, and informed the State that unless it revised its CAP to
include this method, the State's claims would be disallowed.  Tr. I at
203-204.  The State then submitted a CAP including this method, which
DCA approved effective April 1, 1987 (CAP II).  CAP II used a ratio with
"# of IV-E Cases Claimed" in the numerator and "Generic Caseload Count x
(Individuals per family count - 1)" in the denominator.  The State said
that the actual ratios used "during" the period CAP II was in effect
ranged from 1.8% to 3.5%.  State Ex. 15, Att. B.

Meanwhile, the State was concerned that it was not identifying all
children in foster care who were eligible for maintenance payments under
Title IV-E.  Thus, it hired a consultant who advised the State on how to
increase its maintenance claims and who assisted the State in developing
a time study method for allocating social worker costs.

On June 29, 1988, Massachusetts submitted a revised CAP (CAP III).  CAP
III proposed use of a random moment time study (RMTS) which would survey
a sample of social worker time.  Social workers were to identify the
predominant activity in which they were engaged during the "survey
moment" according to eight activity codes.  The codes which were
identified in CAP III as representing activities claimable under Title
IV-E were Code 1 ("Child Welfare Training"); Code 2 ("Title IV-E and
Medicaid Eligibility Information Gathering and Documentation"); Code 5
("Child Welfare Case Management/Administration - DSS Placement"); and
Code 6 ("Child Welfare Case Management/Administration - No DSS
Placement").  State Ex. 2.  .DCA asked for clarification of some of the
definitions and instructions for the RMTS system and began discussions
with the State over whether the State needed to apply an "eligibility
factor" to the RMTS results since some of the coded activities related
to children who were neither eligible for IV-E nor potentially eligible.
As a result, CAP III was clarified and was modified to include an
eligibility factor (the number of IV-E eligible cases divided by the
total number of "Out of Home" cases).  State Exs. 4 and 5.  The
eligibility factor was to be applied to activity Codes 1, 5, and 6, but
not to Code 2.  On July 13, 1989, DCA approved modified CAP III, with an
effective date of July 1, 1988.  State Ex. 6.

On July 13, 1989, Massachusetts submitted revised Title IV-E
administrative cost claims for quarters ending September 30, and
December 31, 1986.  These revised claims were calculated based upon RMTS
results for the last two quarters of fiscal year 1988, without applying
an eligibility factor, and totaled $14,114,086 (after deduction of the
amounts the State had claimed for those quarters under CAP I).  ACF
disallowed these revised claims on the basis that the claim for the
quarter ending September 30, 1986 was not filed in a timely manner and
that both claims were not in accordance with the approved CAP in effect
in 1986 (CAP I).  The State appealed this disallowance to the Board
(Docket No. 89-219).

In October 1989, the State requested that DCA approve a retroactive
effective date of July 1, 1986, for the State's modified CAP III. 2/
(As we discuss below, however, the State did not implement the RMTS
until 1988 and based its claims for periods before the RMTS was
implemented on "backcasting" RMTS data.)  The parties agreed to dismiss
without prejudice the disallowance for the two 1986 quarters, pending a
decision on this request.  DCA ultimately denied the request on the
basis that it could not be assumed that the time study data reflected
worker effort for any period other than the quarter sampled.  The
Regional Director upheld the denial, and the State appealed that
decision to this.Board (Docket No. 90-133) and reinstituted its appeal
of ACF's disallowance (Docket No. 90-116).

The Issues

During the course of this proceeding, the State acknowledged that it had
calculated its revised claims for the two quarters of 1986 using the
original version of CAP III, and needed to recalculate the amount using
an eligibility factor.  The State presented new calculations using
activity code percentages based on mean RMTS data for the period July 1,
1988 to June 30, 1989 and eligibility factors based on data from the
relevant 1986 quarter. 3/  The State contended that these calculations
showed that the State was entitled to receive for social worker costs at
least $1,115,761 in FFP for the quarter ending September 30, 1986
($627,465 more than the State had received under CAP I) and at least
$1,220,565 for the quarter ending December 31, 1986 ($672,365 more than
the State had received under CAP I).

The State argued that retroactive approval of revised CAP III was
required either on the basis that use of CAP I and CAP II resulted in a
"significant inequity" to the State or that those CAPs each had a
"material defect."  In addition to comparing allowable percentages under
the various CAPs, the State alleged the following in support of its
position that retroactive approval was necessary to avoid a significant
inequity:  (1) that ACF guidance on what administrative costs could be
claimed was unclear and untimely; (2) that DCA did not provide the State
with .any examples of acceptable time studies; and (3) that DCA had
forced the State to accept the CAP II method, even though DCA was aware
that the State was underclaiming maintenance payments (and therefore
administrative costs).

In response to objections by DCA and ACF (the Agencies) that the State
was using data from later quarters, the State argued that such
"backcasting" was permissible so long as the State showed there were no
significant differences in the two periods.  The State conceded that
there were some changes occurring in certain characteristics of its
caseload between 1986 and 1990, but said it could demonstrate that
social worker time spent on eligible Title IV-E activities was stable
during this period.  The State presented evidence, including analyses by
a statistical sampling expert, to show that its basic administrative
structure was the same and that changes in caseload characteristics
identified by the Agencies had no statistically significant effect on
social worker time (except one characteristic, which favored the State).

After first indicating that it would seek a waiver of the timely filing
requirement for its revised claim for the quarter ending September 30,
1986, the State argued at the hearing that this was not a new claim, but
fell within an exception for "adjustments to prior year costs."  The
State also argued that the Board should order the Agencies to pay
revised claims for the period January 1, 1987 to June 30, 1988 (based on
the calculation method presented to the Board), even though the State
had not yet submitted such revisions.

Although the State's arguments were framed as a request for retroactive
effect of CAP III, we note that the allocation method proposed is not
that called for in CAP III as approved.  The RMTS system contemplates
collecting survey moments on an ongoing basis and calculating a claim
for any quarter based on the moments collected during that quarter; the
proposed method "backcasts" RMTS data from later periods.

Thus, the issues presented are:

I.  Whether the State's revised claim for the quarter ending September
30, 1986 fell under an exception to the timely filing requirements;

II.  Whether "backcasting" of data is ever permissible, and, if so,
under what circumstances; .III.  Whether the State showed that a
retroactive effective date for its proposed method was necessary to
avoid a "significant inequity" or because of a "material defect" in CAP
I and CAP II; and

IV.  Whether the State showed that its proposed method, even though it
"backcast" data, was a valid method for calculating the amount to which
the State was entitled for social worker costs.

Discussion

I.  The claim for the quarter ending September 30, 1986 was not an
"adjustment to prior year costs."

Section 1132 of the Act (added in 1980) generally prohibits FFP in any
claim for expenditures made under Title IV of the Act, unless the claim
is filed within a two-year period after the quarter in which the
expenditure is made.  Pub. L. 96-272, .306.  The filing limit must not
be applied, however, "so as to deny payment with respect to any
expenditure involving . . . adjustments to prior year costs."  Section
1132(a) of the Act. 4/

Applicable regulations define "adjustment to prior year costs" to mean
--

 an adjustment in the amount of a particular cost item that was
 previously claimed under an interim rate concept and for which
 it is later determined that the cost is greater or less than
 that originally claimed.

45 C.F.R. .95.4.

This definition was published in 1981, shortly after the enactment of
section 1132.  46 Fed. Reg. 3,527, at 3,529 (January 15, 1981).  The
preamble explained that subsequent adjustments of this type were
unforeseen and unavoidable and that a broader definition of
the.exception would render the statute a nullity.  46 Fed. Reg. at
3,528.  The purpose of section 1132 is to ensure that states submit
final reimbursement requests in a timely fashion so that HHS can plan
its budget.  Thus, we have held that the exception should not be applied
where a state was aware of costs, but simply failed to claim them.  We
distinguished this situation from adjustments where an interim rate did
not include costs of which a state was not aware at the time it filed
its claim based on the interim rate or where a delay in establishing a
final rate was due to circumstances beyond a state's control.  Courts
have upheld this approach.  See South Carolina Health and Human Services
Finance Commission v. Sullivan, 915 F.2d 129 (4th Cir. 1990).

Here, there is some question whether the State determined that the
amount of "a particular cost item" is greater than originally claimed.
The State is not adjusting the amount of any of the social workers'
salaries or other costs which are the expenditures in which FFP is
claimed.  The State's revised claim was based on an adjustment to the
percentage figure applied to the same cost pool used in the original
claim.  The State's position appears to be that the cost pool should be
considered "a particular cost item," but the State provided no support
for this interpretation.

The more critical question here is whether the "ratio" used in CAP I
constitutes an "interim rate concept" within the meaning of 45 C.F.R.
.95.4.  The term "interim" means "provisional" or "temporary."
Webster's Third New International Dictionary at 1179.  The ratio
established by CAP I, however, is not identified as an "interim,"
"provisional," or "temporary" method of determining what percentage of
the social worker cost pool would be allocated to Title IV-E.  Rather,
CAP I contemplates that, once the numerator and denominator are known,
the resulting percentage will be applied to the cost pool and that this
will determine the amount of the State's claims.  The mere fact that, in
certain specified circumstances, a CAP may be retroactively amended does
not bring a ratio determined under a CAP within the concept of an
"interim rate."  Here, the delay in claiming was not due to
circumstances beyond the State's control which prevented it from
adjusting from an interim to a final rate within the two-year filing
period.  If the State had chosen to implement an RMTS system at an
earlier date, the State could have done so.  Moreover, the State is not
simply adjusting a rate according to a pre-established formula for
adjustment, but is attempting to change the method for determining the
rate.  .Thus, we conclude that the claim for the quarter ending
September 30, 1986 is not an "adjustment to prior year costs" and
therefore is barred because it was not timely filed.

II.  "Backcasting" is permissible in some circumstances, but the State
has the burden of showing those circumstances exist.

The Board has previously addressed (in several different contexts) the
question whether sample results may be applied retroactively to an
earlier, unsampled period.  The Board has held that the proponent of
such "backcasting" of data, in any context, has the burden of showing
the validity of the method proposed.  Where a party seeks retroactive
approval of a CAP, the proponent of the method must also show that one
of the regulatory standards for giving retroactive effect to a CAP
applies.  Establishing the validity of a proposed method is a separate
step, which is necessary in order to establish the amount claimed and
would be required even for prospective approval.

In California Dept. of Health Services, DAB No. 666 (1985), the Board
permitted California to claim enhanced funding for abortions paid for
between 1972 and 1977 based on data acquired in 1977 and 1978.  The
Board noted that the parties in that appeal had concluded that data from
the 1977-1978 period would be the best available evidence for
identifying what services in the earlier period were for family planning
and had agreed that there were no significant differences between the
periods to make use of the later period inappropriate.  California at 2.

In Ohio Dept. of Human Services, DAB No. 900 (1987), the Board stated:

     While sampling in its purest form envisions samples from the same
     period in question, common sense would dictate that samples from
     another period may be used if it can be established that no
     substantial change has occurred so as to invalidate the procedure.

At 11.  The Board concluded that Ohio had failed to produce sufficient
evidence to establish that there were no significant differences between
the data from the audited period and other periods.  See also Washington
State Dept. of Social and Health Services, DAB No. 924 (1987).  .In
Missouri Dept. of Social Services, DAB No. 1021 (1989), the Board
stated:

 The Board's analysis therefore permits sample results from one
 period to be used to support claims from contiguous periods when
 no better documentation is available, provided that it can be
 shown that there are no significant differences between the
 periods.  The Board has recognized this approach as an expedient
 tool, particularly when the parties are in agreement on the need
 to establish a claim amount.  The party asserting the use of
 data from unsampled periods has the burden of showing that
 circumstances relating to the sampled and unsampled periods are
 such that the data can be used for the unsampled period.  We are
 not prepared to state what degree of similarity in circumstances
 is necessary to support the retroactive application of sampling
 results or other data; each case must be judged by its
 particular circumstances.

At 14.  Missouri was seeking to use data from the last six weeks of one
quarter to substantiate its claims for earlier quarters.  The Board
found that the six weeks of the quarter Missouri sought to use as
representative of earlier quarters was significantly different from the
earlier quarters.  The Board accordingly found that the data from those
six weeks could not be used to substantiate claims for other quarters.

In Maryland Dept. of Human Resources, DAB No. 1020 (1989), a Board
decision cited by DCA as authority for denying the State's request for
an earlier effective date for its CAP amendment, Maryland sought to have
a CAP amendment applied retroactively in order to recover $4.8 million
in unreimbursed Title IV-E administrative costs.  Since the beginning of
the Title IV-E program, Maryland had a CAP which allocated the costs of
social workers' activities based on a time study known as the Random
Moment Study (RMS).  Maryland claimed that a new time study system, the
Social Services Time Study (SSTS), determined that its Title IV-E
activities were being underclaimed.  Maryland sought to apply the
results of the SSTS for the quarter ended December 31, 1985 to the prior
eight quarters that had used the RMS.

The Board found that Maryland's methodology was flawed and that, in any
event, Maryland should have known it was underclaiming because its RMS
system did not allocate a particular category of costs to Title IV-E,
even though those costs were specifically identified in the IV-E
.regulations as allowable.  The Board's decision was overturned in
Colvin v. Sullivan, Civil No. H-89-1652 (D.Md. Feb. 21, 1990), but was
ultimately upheld.

The district court found that 45 C.F.R. .95.515 was inconsistent with
the timely claims provisions in section 1132 of the Act.  Massachusetts
relied on the district court decision in its initial briefs here, but
that decision was overturned in Colvin v. Sullivan, 939 F.2d 153 (4th
Cir. 1991), which reinstated the Board's decision.  The Fourth Circuit
found that no incompatibility existed between 45 C.F.R. .95.515 and the
timely claims provisions of section 1132 of the Act.  The court further
found that 45 C.F.R. .95.509(a) --

 gives states an incentive to carefully examine their CAPs at all
 times, to discover any defects that might exist as early as
 possible.  Maryland could not prove that the RMS method was
 materially defective until it began measuring the employees'
 time by the SSTS method, on October 1, 1985.  Increasing
 disbursements from the date that Maryland began measuring by the
 SSTS method, and no earlier, is entirely appropriate.

939 F.2d at 156.  The court concluded:

 [W]e find that the Secretary's approving of the SSTS method for
 use only in periods where SSTS data had actually been collected
 was both logical and an appropriate exercise of the discretion
 mandated by statute.  The statute of limitations provision of
 the Social Security Act did not require the Secretary to accept
 an application for reimbursement that otherwise lacked
 sufficient support, and does not preclude the Secretary from
 setting more stringent standards for the approval of retroactive
 adjustments to CAPs.

939 F.2d at 157.

In sum, backcasting may be permissible to establish the amount of a
claim, but the State has the burden of showing that it is appropriate
under the particular circumstances.  In the context of a retroactive
amendment of a CAP, the State must (as a separate requirement) meet the
stringent standards in 45 C.F.R. .95.515.  As we discuss next, the State
did not meet those standards, even assuming the permissibility of
backcasting here.

III.            The State did not establish that retroactive amendment
of the CAP was required.

As the Fourth Circuit recognized in Colvin, the regulatory requirements
for CAPs reasonably set stringent standards for the approval of
retroactive amendments to CAPs.  The regulations place the burden on
states to carefully examine their CAPs and to discover any defects as
soon as possible.  Thus, retroactive amendment is appropriate only where
"an earlier date is needed to avoid a significant inequity" or "the
information provided by the State which was used to approve a previous
plan or plan amendment is later found to be materially incomplete or
inaccurate . . . ."  45 C.F.R. .95.515.  As we discuss in this section,
the State did not establish that either of these conditions was met
here.

 A.              The State did not establish a significant
 inequity.

The State argued that we should find there was a significant inequity
here because (1) ACF guidance on what administrative costs could be
claimed was unclear and untimely; (2) DCA did not provide the State with
any examples of acceptable time studies; and (3) DCA had "forced" the
State to accept the CAP II method, even though DCA was aware that the
State was underclaiming maintenance payments (and therefore
administrative costs).  The State also relied on the alleged differences
in the amounts claimed under CAPs I and II and the amounts determined
using the State's proposed method.

The State did not point to any specific lack of clarity in ACF's
regulations on administrative costs, other than referring to this
Board's decision in Missouri Dept. of Social Services, DAB No. 844
(1987).  In that decision, the Board held that ACF could not reasonably
interpret the statute and regulations to limit allowable costs of
determining eligibility for Title IV-E (and providing certain required
preplacement services) only to costs associated with individuals
actually found eligible.  We agree with the State that there was some
lack of clarity in ACF policy in this regard prior to the Missouri
decision.  However, the State did not provide any evidence that it had
been influenced in choosing a CAP method by ACF's earlier
interpretation.  Instead, the evidence indicates that the State opted
for a very unsophisticated approach in its CAPs because it was aware
that its major problem was in failing to identify Title IV-E eligibles,
and it chose to devote its energies to .that effort.  Tr. I at 28-34,
60-61, 203.  Moreover, the State was aware of the Missouri decision at
least as of May 1987.  Tr. I at 34-35; 205-206.  Although a state could
not take advantage of that decision unless it had a time study method
for allocating costs, the State did not submit such a method until June
29, 1988.  Thus, the State did not establish any connection between a
lack of clarity in ACF policy and the State's failure to implement its
RMTS system at an earlier date. 5/

We also do not find any inequity in DCA's failure to provide the State
with examples of acceptable time studies.  While of course it would have
been helpful if DCA had done this, the regulation puts the burden on the
State to propose a CAP method.  The validity of any time study method
may depend on the particular circumstances of the individual state's
organization, so DCA may have been reluctant to suggest that a time
study method approved for another state was appropriate for
Massachusetts.  DCA did mention to the State that it might wish to use
time studies, and the State had its own reasons why it did not do so.
Tr. I at 25-34, 60-61, 199-203, 222-225.  Moreover, the State could have
itself obtained approved time studies from other states (which it
apparently did not even attempt to do) or from a consultant (as it
ultimately did).

The fact that the State has the burden to propose a CAP method and has
the information necessary to determine what method is appropriate also
undercuts the State's position that DCA unreasonably forced the State to
accept the CAP II method.  DCA reasonably found that the other two
methods proposed by the State at the time had no basis in logic or
reason.  State Ex. 15, Att.; Tr. I at 218-219.  The method DCA said was
acceptable recognized some greater effort for IV-E cases than for other
cases handled by the social workers.  Tr. I at 204.  The State had
failed to show any reasonable basis for recognizing an effort greater
than that recognized by the CAP II method.  The mere fact that the State
was not claiming as much FFP in Title IV-E maintenance payments as it
could have did not require DCA to accept an illogical and .unsupported
CAP method for claiming administrative costs.  Indeed, any underclaiming
of costs was due to the State's own failure to determine IV-E
eligibility.

We also note that some of the administrative costs at issue here are
ones which may be allocated to Title IV-E, but which the State is not
required to allocate to that title.  These are the types of costs which
the State may choose to claim under other programs.  Indeed, the State's
original allocation of part of the costs to Title XX (Social Services)
could have resulted in their reimbursement by that program except for
the fact that there is a ceiling on Title XX costs and the State chose
to claim other types of costs under that title rather than the social
worker costs at issue here.  See Tr. I at 160-161.  Because of the
potential overlap between programs, the Title IV-E regulations
specifically provide that the State's option of a program under which it
will claim overlapping costs be exercised in an approved CAP.  In other
words, the State's failure to be federally reimbursed for some of its
social worker costs under Title IV-E can in the circumstances of this
case be viewed as resulting both from the State's option of an
allocation method and from its choice not to claim reimbursement under
Title XX or some other program. 6/

Finally, as discussed below, the State's proposed method does not
establish with a reasonable degree of certainty the amount the State
could have claimed if it had implemented an RMTS sooner.  Even if we
would accept the State's proposed method as establishing reasonable
estimates to be compared to the amounts the State actually claimed,
however, this comparison does not establish a significant inequity.  In
Maryland, we specifically did not reach the issue whether a disparity in
claim amounts under two methods would be sufficient to establish a
significant inequity.  Maryland at 7.  Assuming that such a disparity
could establish a significant inequity if it showed that a state was not
receiving a substantial amount of federal funds to which it was
entitled, however, we find that DCA was not.required to determine that
such an inequity existed under the circumstances here.

We are aware that, from one point of view, there seems to be a
substantial disparity between the amounts the State originally claimed
and the amounts determined by the State's proposed method.  For the
quarter ending September 30, 1986, 50% of the total cost pool for social
worker costs was $12,167,517 of which the State originally claimed
$488,296, or 4.01%.  The State's proposed method would increase the
claim by $627,465, to $1,115,761 (9.17% of the social worker cost pool).
In context, however, this increase can reasonably be viewed as not so
substantial that it requires retroactively amending the CAP.  The
increase is only 5.16% of the total social worker cost pool, and that
cost pool represents only part of the State's total administrative
costs.  Other cost pools were charged directly to Title IV-E or
allocated using a different method.  See State Ex. 2.  Moreover, if the
State had implemented the RMTS method sooner, presumably it would have
had higher costs from administering that method than it had from the
simple case ratio method used under CAP I.

The same analysis applies to the quarter ending December 31, 1986.  The
State originally claimed $548,200, which is 3.90% of $14,045,624, which
is the federal share (50%) of the social worker cost pool.  The State's
proposed method would increase the claim by $672,365 to $1,220,565, or
8.69%.  This is an increase of only 4.79% of the total cost pool.

While the claim disparity could be greater if viewed over the entire
period from the start of the disallowance period through the CAP II
period, the State did not establish what the disparity would be during
the CAP II period (or, indeed, that such a disparity existed). 7/  The
State provided information on the percentages it said it used to
construct its claims for FFP "during" the period January 1, 1987 to June
30, 1988 (when CAP II was in effect).  State Ex. 15, Att. B.  The record
indicates, however, that the State in August 1989 submitted revised
claims for the period July 1, 1987 to June 30, 1988, based on
retroactive increases in the Title IV-E caseload identified during the
State's internal review of IV-E eligibility.  State Exs. 7 and 8.  The
State presented no.figures showing how it revised its claims for the CAP
II period, nor how the State's proposed method would alter those claims,
if at all.  Thus, we do not know whether the adjusted eligibility
figures used in the CAP II method for this period resulted in revised
claims at percentages higher or lower than what the State's proposed
method would give.

Moreover, the State did not submit claims for the CAP II period
(explaining that it thought its claims for the 1986 quarters had delayed
approval of CAP III).  Tr. I at 108-109.  The claims would be untimely
if submitted now (as we found the claim for the quarter ending September
30, 1986 was untimely).  The State suggested that we could require the
Agencies to approve and pay claims for the CAP II period, but this Board
has no authority to grant waivers of the timely filing requirements.  45
C.F.R. ..95.22-95.34; New Jersey Dept. of Human Services, DAB No. 1142
(1990).  Thus, since claims may be barred for all quarters involved here
(except the quarter ending December 31, 1986), it would be speculative
to say that retroactive approval of the State's proposed method is
necessary to avoid a significant inequity because the cumulative amounts
the State could otherwise receive would be substantial.  There is no
assurance that the State has not already forfeited any entitlement it
might have to additional funds because of its own failure to submit
timely claims.

In sum, the State failed to meet its burden to show that retroactive
approval was needed here to avoid a significant inequity to the State.

 B.           The State did not show that retroactive approval
 was necessary because of a material defect.

The State argued that CAPs I and II contained material defects,
justifying retroactive approval of the State's proposed method.  In our
view, this argument is based on a misunderstanding of the CAP
regulations and, in any event, has no merit.

The regulations contemplate that a state is responsible for proposing an
allocation method since the state has the best knowledge of its own
administrative structure and organization.  A state must promptly amend
its CAP if there are organizational or programmatic changes affecting
the validity of a CAP or if a "material defect is discovered in the cost
allocation plan by the Director, DCA, or the State."  45 C.F.R.
.95.509(a).  In this context, section 95.515 provides that the
effective.date will be "the first day of the calendar quarter following
the date of the event that required the amendment" but may be earlier if
specific conditions are met, including that --

     The information provided by the State which was used to approve a
     previous plan or plan amendment is later found to be materially
     incomplete or inaccurate . . . .

Here, there was no finding of a material defect by DCA.  Indeed, ACF
policy specifically permits a state to use a caseload count as an
allocation method.  ACYF-PA-87-05, State Ex. 14.

The mere fact that the State considers CAP I and CAP II to be materially
defective is not in our view sufficient to establish either that a
defect existed, or that it was material.  Using a caseload count is
reasonable, even if not entirely accurate, because it is simple and
gives roughly equitable results.  Moreover, we do not consider it a
defect that the State might have received more reimbursement under Title
IV-E if it had instituted an RMTS sooner.  Allocating more of the costs
to Title XX was primarily the State's decision.  Also, for reasons
explained above, we do not consider the disparity between what the State
did claim for each disallowance quarter and what it might have claimed
to be substantial when viewed in light of the State's total
administrative costs.

Finally, DCA did not find the State's information to be materially
incomplete or inaccurate.  While as indicated above the State provided
little to support any particular claiming method, the record does not
indicate that the State's information was inaccurate, and DCA accepted
it as sufficient to support the methods actually used.

In sum, we find that the State did not show that an earlier date for its
CAP was needed because of either a significant inequity or a material
defect in CAP I or CAP II.

IV.  The State's proposed method is flawed.

The Agencies took the position that the State's RMTS data could not be
backcast to the disallowance period.  They presented testimony and
evidence on the changes taking place in the State during the 1980's,
including data comparing caseload characteristics during 1986 and later
periods.  The State sought to meet its burden to establish that its
backcasting method was valid through.testimony by a program official and
consultant on how any changes affected social worker time and through
testimony by a statistical sampling expert analyzing the correlation
between social worker time and the caseload characteristics identified
by the Agencies.  The Agencies replied by presenting their own
statistical sampling expert.

In this section, we first explain what the statistical analyses were and
what remaining issues there are concerning the validity of those
analyses.  We then address more general questions regarding alleged
differences between the two periods and the State's presentation.

 A.  The statistical analyses

As discussed above, the State's method proposed here used RMTS data
obtained during the period July 1, 1988 to June 30, 1989 (the RMTS
period) and eligibility data the State said was from the last two
quarters of calendar year 1986 (the disallowance period).  In actually
calculating the amounts it said was due for each of the two 1986
quarters, the State determined the mean (or average) percent of total
social worker time allocated by the RMTS to Codes 1, 2, 5, and 6.  The
State then multiplied each of the mean percents for Codes 1, 5, and 6 by
the eligibility factor for each quarter.  The resulting percents were
then added to the mean percent for Code 2 for the quarter to obtain the
resulting percentage (9.17% for the quarter ending September 30, 1986
and 8.69% for the quarter ending December 31, 1986).  The resulting
percent for each quarter was then multiplied by 50% of the total cost
pool to get the total FFP to which the State said it was entitled for
each quarter.  State submission of 12/5/90, Revised Atts. 2(a) and 2(b).

The State's testimony to support its method focused not on the
percentages allocated to each of the RMTS Codes, but on the sum of the
percentages of time allocated to Codes 1, 2, 5, and 6.  The State
referred to this sum as "aggregate claimable time," but this is somewhat
of a misnomer.  These Codes represent the types of activities which may
be claimable under Title IV-E, but the total claimable time is
determined only after applying the eligibility factor to Codes 1, 5, and
6.  While we use the term "aggregate claimable time" in our discussion
of the State's evidence (since that is the term the State used), we also
explain below why we think the focus should have been on the individual
codes.  .The State's statistical expert testified generally that use of
a mean was a method used "fairly commonly in statistical analyses" when
there are missing data.  Tr. I at 176.  She noted that the "aggregate
claimable time" obtained by the RMTS during the RMTS period was a
relatively stable figure.  Tr. I at 183.  (It ranged from 57.2% to
60.3%.)  She testified that this showed that there were no significant
differences in how social workers were spending their time in 1988,
1989, and 1990, despite the fact that there were very substantial
differences in other characteristics of the agency.  She also testified
that she had looked at data from at least two other states and found a
very similar pattern, indicating that even if there are changes in the
nature of the caseload and other characteristics of the agency,
caseworker time does not vary as a function of these changes.  Tr. I at
167-171.

The State's statistical expert also presented results of a correlation
analysis and a regression analysis.  The correlation analysis was
performed to determine whether any of the caseload characteristics
identified by the Agencies had a statistically significant relationship
to "aggregate claimable time" during the RMTS period.  The results of
this analysis showed that none of the identified characteristics had a
statistically significant correlation with "aggregate claimable time,"
except "group home expenditures."  The analysis also showed that, as
group home expenditures increased, "aggregate claimable time" decreased.
8/  Since group home expenditures were lower in the disallowance period
than in the RMTS period, this factor (considered alone) would indicate
that "aggregate claimable time" was likely higher in the disallowance
period than in the RMTS period.

The regression analysis presented by the State's statistical expert was
essentially use of a formula to weight different characteristics in
different proportions and then to multiply those weights against the
actual data in order to estimate an "aggregate claimable time."  Tr. I
at 176-177; State submission of 12/5/90, Rev. Att. 5.  The analysis
projected estimated percentages of ."aggregate claimable time" for each
month of the period July 1986 through March 1988.  The estimated
percentages ranged from 58.6% to 67.1%.  State submission of 12/5/90,
Att. 9.  As support for the validity of this analysis, the statistical
expert also used her regression model to calculate "aggregate claimable
time" percentages for each month in the RMTS period.  The figures
obtained from this calculation were within 2.4 percentage points of the
actual percentages obtained from the RMTS.  The projected mean percent
was 58.3% compared to the actual mean of 58.4% for the RMTS period.
Id., Att. 10.  According to the to the State's expert, this shows that
the regression model was doing a "reasonably decent" job of prediction.
Tr. I at 183.

Finally, the State's expert testified that "if anything, the estimates
that have been used in the claim appear to be under estimates rather
than over estimates, based on the data that are available with regard to
the [State] agency characteristics."  Tr. I at 183.

After examining the underlying information used by the State's
statistical expert for her analyses, the Agencies' statistical expert
agreed with her in some important respects.  He withdrew objections he
had preliminarily raised regarding the validity of the RMTS samples
drawn by the State, and certain aspects of the statistical analyses
presented.  He agreed that, of the caseload characteristics identified
by the Agencies, the only one with a statistically significant
correlation to "aggregate claimable time" was group home expenditures.
He further agreed that her calculations were accurate.  Transcript of
Hearing held 2/19/91 (Tr. II) at 6-7, 16-21.  His testimony did,
however, raise certain remaining questions, which we address next.

 B.           The remaining issues concerning the statistical
 analyses

The Agencies' expert testified that he had four continuing objections to
the State's analyses.  The first objection was that the State had not
established that the conditions in the disallowance period were
identical to the conditions in the RMTS period.  He conceded, however,
that the State did not need to establish identical conditions, but only
that those characteristics affecting social worker time were the same.
As we discuss below, the State's presentation on the two periods had
certain flaws which do call into question the validity of backcasting
data here.  .The Agencies' expert's second objection was that the range
of values for group home expenditures during the disallowance period was
different from the range of values during the RMTS period.  He said that
this affected the validity of the State's regression analysis.  Tr. II
at 8-9.  The State's expert did not deny that the range of values has
some bearing on the use of a regression analysis, but pointed out that
here the regression analysis was not being used to establish an exact
figure for "aggregate claimable time" for the disallowance period, but
only to substantiate the reasonableness of using the mean from the RMTS
period.  Tr. II at 22-23.

The Agencies' expert also pointed out that the group home expenditures
variable explained only 17% of the total variation in "aggregate
claimable time" during the RMTS period.  This left 83% of the variation
unexplained.  Tr. II at 11-12.  The State's expert responded that, given
how slight the variation was to begin with, this was not a meaningful
objection.  She testified that, in her experience with statistics in
social services programs, it is unusual to be able to explain as much as
17% of such a small variation.  She also testified that the remaining
variation could be explained as simply random fluctuations.  Tr. II at
23-24.

The State's expert may be correct that the difference in range of values
of group home expenditures does not undercut use of the regression
analysis to support use of the mean.  However, her own testimony
indicates that a regression analysis (if shown to have predictive power)
would be a better alternative than use of a mean.  Tr. I at 176.
Moreover, the State's expert did not deny that an inability to fully
explain the variation lessened the predictive power of the regression
analysis.  Thus, while we do not think these two objections by
themselves would be a basis for finding that use of the mean is not
reasonable, they point up the fact that the mean is a rough estimate,
not as exact as a statistical analysis where the variables are known.

The final -- and in our view most substantial -- objection raised by the
Agencies' expert related to the State's focus on the "aggregate
claimable time," rather than on the individual RMTS Codes.  He pointed
out that, while the "aggregate claimable time" figure was relatively
stable over the RMTS period, the percentages obtained by the RMTS for
individual codes did not have the same stability.  The importance of
this is that (as .pointed out above) the State's claim for FFP is not
based on the "aggregate claimable time" obtained by the RMTS, but on a
percentage obtained by applying the eligibility factor to the results
from Codes 1, 5, and 6 and adding that to the results from Code 2
(determining eligibility). 9/  The eligibility factors used here were
roughly 10%.  Thus, in calculating the claiming percentage, the results
from Code 2 have about 10 times as much weight as the results from the
other three Codes.  Yet, the RMTS results from Code 2 ranged from 2.43%
to 4.17%.  In other words, evidence concerning the reasonableness of
using the mean "aggregate claimable time" to backcast data does not
establish the reasonableness of using the means for the individual
codes, as the State did.  Moreover, using an inaccurate measure for Code
2 may overstate the claiming percentage.

Even if the State's method could reasonably be adjusted to account for
this problem, 10/ however, use of the method would depend on whether the
State showed there were no relevant differences between the RMTS period
and the disallowance period.  For reasons we discuss next, we find that
the State's presentation on the similarities and differences was flawed
in certain respects.

 C.           Other questions regarding differences between the
 two periods

The evidence presented by the Agencies regarding differences between the
RMTS period and the earlier period was for the most part either general
and anecdotal (and sufficiently rebutted by the State's witnesses) or
based on the caseload characteristics analyzed by the State's expert.
Moreover, the Agencies were identifying changes in the State's programs
without regard to whether those changes were of a type which would
affect .allocation of social worker time using the RMTS Codes.  We
nonetheless find the State's evidence comparing the two periods to be
flawed in certain respects.

The State's two witnesses on program matters testified that there were
no organizational changes affecting the social workers.  However, the
State did not support this testimony with any organizational charts or
other documentation from the time period.  Moreover, one of these
witnesses was a high level official responsible for maximizing the
State's claims for federal funds, and the other witness was the
consultant hired by the State to increase its Title IV-E claims.  The
consultant admitted that his compensation depended on the State's FFP
recovery from his activities.  Tr. I at 54.  The State presented no
testimony from the social workers to substantiate that there were no
changes affecting the allocation of their time.

The consultant stated his opinion that, while the State did initiate
certain projects (such as efforts to recruit more foster parents), these
initiatives would not affect the State's basic mission.  Tr. I at
143-144.  We do not give as much weight to his opinion on this as we
would to the opinion of social workers directly affected by such
initiatives.  Moreover, as we discussed above, even if these initiatives
did not affect the State's basic mission, they might have affected what
portion of social worker time was allocated to claimable RMTS Codes, and
thus had an effect on the ultimate claiming percentage.  In particular,
the time allocated to Code 2 may have been affected by an initiative the
State admitted taking to increase the number of children determined
eligible for Title IV-E.  While the consultant testified that most of
this effort would have been done by his consulting firm under contract
with the State or by a separate eligibility determination unit, he also
acknowledged that the social workers played a role in gathering
documentation for eligibility.  Tr. I at 145-149.

We found an additional flaw in the State's presentation.  The State said
that the eligibility factors used to recalculate the claims were based
on data from the disallowance quarters.  The Board asked the State to
substantiate this and to show how it had calculated the eligibility
factors.  The document the State submitted is a rough handwritten
summary listing numbers of eligible Title IV-E families from various
State regional offices.  State's submission of 12/11/90, Att.  The State
provided.no explanation or documentation of the source of these figures.
11/

In sum, while the State sufficiently rebutted much of the Agencies'
evidence presented to show that backcasting was not appropriate here, we
still have substantial remaining questions on the State's method
(particularly the State's failure to focus on individual Code results
rather than "aggregate claimable time") and on the accuracy and
completeness of the information presented here.  Thus, we conclude that
the State did not meet its burden to show that its method could
appropriately be used to determine the amount the State would have been
able to claim if the State had implemented the RMTS in June 1986 rather
than in 1988.

Conclusion

For the reasons discussed above, we uphold DCA's decision not to approve
a retroactive effective date for CAP III (or for the State's method
proposed here).  We therefore uphold ACF's disallowance of costs claimed
in excess of the amount allowable under CAP I for the quarters ending
September 30, 1986 and December 31, 1986.  We also conclude that the
claim for the first of those quarters was untimely.

 

     ________________________ Donald F.
     Garrett

 


     ________________________ Norval D.
     (John) Settle

 


     ________________________ Judith A.
     Ballard Presiding Board Member

1.  These types of services might be reimbursable under one or more
federal or state programs, including Title IV-E, state foster care,
Title IV-B (Child Welfare Services), or Title XX (Social Services).

2.  The State also requested in the alternative that the CAP III version
originally submitted by the State be approved.  DCA and the Regional
Director both denied this request, explaining why application of an
eligibility factor was required under ACF policy (specifically,
ACYF-PA-87-05).  State Exs. 1 and 3; Regional Director's Letter of
6/8/90.  The State did not appeal this disapproval to the Board.

3.  DCA and ACF argued that this revised calculation constituted an
"untimely claim."  Generally, states have to document an allocation
method at the time they submit their claims.  We do not agree, however,
that adjustment of an allocation method in the context of a dispute over
the method always constitutes a new claim for purposes of the timely
filing provisions.  When the new allocation reduces the amount at issue,
it could be viewed as simply a settlement offer.  Treating the offer as
a new claim would not be necessary to meet the purposes of the timely
filing provisions.  The State actively sought here to settle the case,
and the Board delayed proceedings several times to permit the parties to
negotiate.  As we discuss below, however, we find the claim for the
quarter ending September 30, 1986, was untimely because it was not filed
within the two-year period required and did not meet an exception to
that requirement.

4.  Section 1132(b) provides that the Secretary "shall waive the filing
requirement if he determines (in accordance with regulations) that there
was good cause for the failure by the State to file such claim" within
the prescribed period.  As noted above, the State at first indicated it
was seeking a waiver for the claim at issue here.  The State later said,
however, that it had decided not to pursue a waiver until its appeals
were decided.

5.  The State also suggested that DCA unreasonably delayed approval of
CAP III.  The record shows, however, that most of the delay was due to
problems with the CAP, as proposed.  State Exs. 1-7; Tr. at 205.  DCA
ultimately approved modified CAP III retroactively to July 1, 1988, the
beginning of the first quarter in which the State had fully implemented
the RMTS.

6.  The State argued that ACF's policy announcement ACYF-PA-87-05
required that Title IV-E bear its "proportionate share" of the State's
administrative costs.  This policy simply requires a proportionate
allocation of services provided to IV-E eligibles and to other program's
recipients.  Caseload count is specifically identified as an equitable
basis for allocating costs of such activities.

7.  The State also did not address the first part of the period CAP I
was in effect, other than to give 4-8% of the cost pool as the range of
percentages used during the CAP I period.

8.  The State's consultant offered the following as a possible
explanation.  The caretakers in group homes are more experienced than
foster parents generally.  Thus, the social workers would need to be
less involved in monitoring children placed in group homes.  Tr. I at
195-196.

9.  The State's expert apparently misunderstood this, believing that it
was the "aggregate claimable time" which drove the State's revised
claims.  Tr. II at 31-32.

10.  For example, the State's method could be adjusted to use the low
end of the range obtained for Code 2 during the RMTS period (2.43%)
instead of the mean for Code 2 used by the State (3.42%).  (This would
reduce by .99% the claiming percentages calculated for the disallowance
period, yielding a claiming percentage of 8.18% for the quarter ending
September 30, 1986 and 7.70% for the quarter ending December 31, 1986.)

11.  In a pre-hearing conference, the Board had specifically asked the
parties to document their positions and, in particular, asked the State
to document that the eligibility factors used to recalculate the claims
were based on eligibility data from the 1986 quarters.  See Board's
confirmation of