GAB Decision 524
Docket No. 83-183
March 29, 1984
California Department of Social Services;
Ballard, Judith; Settle, Norval Garrett, Donald
The California Department of Social Services (California, State)
appealed a disallowance of $581,202 federal financial participation
(FFP) in the cost of child care services for the period October 1, 1978
through March 31, 1981, claimed under the Work Incentive (WIN) program
authorized by Title IV-C of the Social Security Act (Act). /1/ The
costs were disallowed to the extent that the State's claim was based on
payments to individuals who allegedly were no longer eligible for child
care services or for whom the State was unable to document participation
in an active WIN component at the time services were provided.
The purpose of the WIN program is to furnish recipients of Aid to
Families with Dependent Children (AFDC) with incentives, opportunities,
and necessary services to prepare for and obtain employment in the
regular economy and in public service. 42 U.S.C. 630. Certain AFDC
recipients must register to participate in the WIN program and, where
the WIN component makes available necessary services such as child care,
accept employment which is offered as part of that program.
There were three issues in this appeal. The Agency disallowed FFP in
child care payments to WIN registrants who had been employed in
government subsidized employment more than 90 days. An audit showed 65
such payments. The applicable regulation limits child care payments to
the first 90 days only for unsubsidized employment, but a WIN Handbook
(2) limits child care payments to the first 90 days for both subsidized
and unsubsidized employment. The first, issue is whether the Handbook
is binding where the regulation addresses only unsubsidized employment
and the Agency did not promulgate the Handbook rule under the
notice-and-comment procedure set out under the Administrative Procedure
Act.
The Agency also disallowed FFP in payments to individuals 1) who had
been removed from active participation in WIN and determined to be
ineligible for further child care services by the WIN sponsor, or 2) for
whom the State could not document registration, certification, and
active participation in WIN at the time services were provided. The
State did not appeal these two findings.
The Agency estimated the dollar amount of unallowable payments using
statistical sampling techniques. It drew an original sample of 201
claims from six counties and later expanded the sample to 391 claims.
The expansion of 190 claims was drawn from three of the six counties.
The estimate based on the sample was applied to these three and one
other of the original six counties.
The use of sampling and the sampling methodology were the subjects of
the other two issues:
(2) Whether the Agency had the requisite authority to use statistical
sampling methodology to calculate the dollar amount of the disallowance.
(3) Whether the Agency expanded the sample in a valid manner.
We find for the Agency on all three issues and uphold the
disallowance as revised. /2/ Our decision is based on the parties'
submissions and the transcript of a hearing held in Washington, D.C. on
January 10, 1984.
Analysis
1. Background
To qualify for FFP in the cost of child, care services to needy
families, a state must have a plan, approved by the (3) Secretary, which
provides, for WIN registrants, "(through arrangements with others or
otherwise) . . . child care and other social and supportive services as
are necessary to enable such individuals to accept employment . . . ."
42 U.S.C. 602(a)(19)(G). The State plan must establish a separate
administrative unit (SAU) to administer the program.
Under a WIN regulation promulgated in final September 18, 1975, the
SAU is charged with "developing and supplying social services necessary
to enable a registrant . . . to accept employment . . ." These services
include child care and "shall continue for a period of 30 days after the
start of unsubsidized employment and may continue for a maximum of 90
days at the discretion of the SAU." 40 Fed. Reg. 43182, 43187; see also
45 CFR 224.30(b)(2)(1978). /3/
The WIN regulations authorize the NCC to issue a WIN Handbook, which,
like other NCC issuances, "shall be binding on all regional, State, and
local WIN operations." 45 CFR 224.10(c). The May 1976 edition of the
WIN Handbook required "necessary supportive services for the duration of
a registrant's participation in all WIN components except OJT and PSE .
. . ." /4/ Agency Exhibit, p. V-2. The Handbook also states:
3. When an individual enters PSE or OJT, necessary supportive
services shall continue for a period of 30 days after the start of
subsidized employment and may continue for an additional 60 days at the
discretion of the SAU. Such services may continue during this time even
after the AFDC grant has been discontinued due to income earned from
such employment. No additional WIN supportive services may be provided
to the individual when the OJT or PSE enrollments ends and the
employment becomes unsubsidized.
4. Except as noted in 3 above, when an individual has been
certified, necessary supportive services shall continue for a period of
30 days after the start of unsubsidized employment and may continue for
an additional 60 days at the discretion of the SAU. Such services may
continue even after the AFDC grant has been discontinued due to income
earned from such employment. The extended services are not to be
provided unless they are necessary to enable the registrant to accept
and retain employment, and need (4) not be provided for the full 30-day
period if other means of providing the services are available.
Generally, WIN services will cease when their cost is treated as a
work-related expense in the recalculation of the assistance payment by
IM, or when transition to Title XX services is complete.
Id. at V-2 and V-3. /5/
2. The 90 day limit on child care services for WIN registrants in
subsidized employment is valid and binding on the State.
The State argued 1) that the Handbook imposition of a 90-day limit
for payment of child care services to WIN registrants in subsidized
employment was a substantive rule subject to the notice-and-comment
requirements of the Administrative Procedure Act (APA); and 2) that the
Handbook provision was inconsistent with several WIN regulations which
the State contended precluded limits on child care services for WIN
registrants in subsidized employment.
The Agency contended that the provision in question was an
interpretative rule and thus not subject to the APA. /6/
(5) According to the Agency, the NCC was interpreting the statutory
mandate (also reflected in the regulations) that child care services be
furnished as necessary to enable a WIN registrant to accept employment.
42 U.S.C. 602(a)(19)(G).
In State of N.J. v. Department of Health and Human Serv., 670 F.2d
1262 (3rd Cir. 1981), the court defined an interpretative rule as "a
statement made by an agency to give guidance to its staff and interested
parties as to how the agency intends to administer a statute or
regulation." Id. at 1281-1282. /7/ The court defined a legislative rule
as one which implements the statute and, in so doing, creates new law
affecting individual rights and obligations. The court held that the
agency was merely attempting to interpret a discrete provision of the
statute by resolving technical questions having to do with the format
and timeliness of the required application. Additionally, in
distinguishing between interpretative and legislative rules there, it
advised taking the agency at its word -- "if an agency that has the
statutorily delegated power to issue legislative rules chooses instead
to issue an interpretative rule, the court accepts that characterization
of the rule." Id. at 1282.
We agree with the Agency that the 90-day limitation on services to
registrants in subsidized employment is an interpretation of the
statutory language that services be furnished "as necessary to enable a
WIN registrant to accept employment." The rule at issue here interprets
precisely which services are necessary to accept employment, i.e., those
furnished within 90 days of employment. The rule by implication
concludes that services furnished beyond 90 days exceed the meaning of
services necessary to enable "acceptance" of employment. While it
is
conceivable that the Agency could have interpreted the language in other
ways, that alone does not detract from the conclusion that the rule is
an interpretative one. A variety of options may be available to an
agency through interpretation of statutory language. Whether the
statutory reference is interpreted as those services furnished for 90
days or more or less is reasonably a matter of "what the administrative
officer thinks the statute or regulation means," and thus an
interpretative rule. Gibson Wine Co. v. Snyder, 194 F.2d 329 (D.C.
Cir. 1952). Child care services are merely one aspect of WIN, a program
to enable AFDC recipients to become self-sustaining. Thus, it is
consistent with program purposes to interpret the statute as providing
WIN funds for child care only for a short period.
(6) The Agency witness testified that when the WIN appropriations
were less than originally expected, the Agency had to decide how to best
use the limited amount available. Transcript of January 10 Hearing
(Tr.), p. 19. The Agency alleged that it found that registrants in
subsidized employment could pay for child care as well, if not better,
than those in unsubsidized employment and determined that the same
limits should therefore apply to both. Tr. 29-30; Agency Brief, p. 5.
/8/
We conclude that the decision not to distinguish the two types of
employment for purposes of determining what services were necessary to
enable a registrant to accept employment was an administrative one
consistent with the Act and a proper basis for an interpretative rule.
The State argued that the 90-day limit was inconsistent with certain
other WIN regulations and these regulations should prevail. /9/ The
Agency contended that these regulations mandate child care as needed to
enable a registrant to accept employment. It pointed out that where
child care was needed for a registrant to continue employment (after 90
days), the regulations allowed the WIN component to arrange (but not (7)
pay) for child care. /10/ The Agency's interpretation of those
regulations is reasonable. As the Agency noted, child care could be
made available through other sources. Moreover, we find that the
regulations do not preclude the Agency from limiting child care payments
under WIN. Thus, the Handbook limitation is not inconsistent with the
cited regulations.
The State argued in effect that the rule was legislative (and thus
subject to APA procedures) because the 90 day limit on child care in
unsubsidized employment had been included in a regulation promulgated in
1975, prior to the WIN Handbook provision discussed above. We are not
persuaded that this inclusion makes the 90-day limit in subsidized
employment a legislative rather than an interpretative rule. The 90 day
limit for subsidized employment could, in our view, be an interpretative
rule for the same reasons as a rule concerning unsubsidized employment.
Its inclusion in regulations is by no means a conclusive demonstration
that it might be legislative. The Agency is not precluded from mixing
provisions which are really interpretative rules in with legislative
rules in a regulation. See Davis, Administrative Law Treatise, Vol. 2,
Sec. 7.8, p. 42 (1979). Moreover, no 90-day or other limit appeared in
the Notice of Proposed Rulemaking, nor did the preamble to the 1975
final rule indicate that it was addressed as a result of public
comments. 39 Fed. Reg. 33699 (September 19, 1974); 40 Fed. Reg. 43182
(September 18, 1975). Thus, it appears that the 90-day limit on
unsubsidized employment also was not viewed by the Agency as a rule
which required APA notice-and-comment procedures.
The Agency also argued that even if the rule in question were a
legislative one, it would not have been required to follow APA
procedures because the regulations established its authority to issue a
Handbook with binding rules and the regulations were promulgated after
the 1971 voluntary submission to APA procedures in HHS (then HEW) grant
programs. We (8) find the 90 day limit on subsidized employment to be
an interpretative rule and do not reach this issue. /11/
2. The Agency was not precluded from estimating the dollar value of the
disallowance based on the sample of claims which it took.
The Agency found 65 claims which contained payments to registrants
who had been in subsidized employment for more than 90 days. On the
basis of these 65 plus eight other claims which the Agency found to be
unallowable (and which the State did not appeal), the Agency estimated
the dollar value of such claims (which were potentially unallowable for
the same reasons) in four counties. Originally 201 claims in six
counties were sampled. That estimate was applied to claims in only five
of the six counties. The Agency auditors expanded the sample to 391
claims, drawing the 190 additional sampled claims from three counties.
The four counties to which the results of the expanded sample were
applied included the three counties from which the expanded sample was
taken and one other county from the original six.
The State asserted that the estimate thus obtained could not be used
because 1) there was no established Agency policy permitting
disallowances to be based on a sampling of unallowable claims filed
under the WIN program; and 2) the Agency auditors did not use proper
statistical methodology when they expanded the sample.
a. The Agency may use statistical sampling to estimate disallowances
in the WIN program.
The State contended that the utilization of sampling was improper
because the WIN program was "very intertwined with Title IV-A" of
the
Act, citing regulations showing the relationship between WIN and Title
IV-A. Post-Hearing Brief, p. 3. The State also relied on California
Department of Social Services, Decision No. 319, June 30, 1982, where
the Board held that Agency policy for disallowances based on errors in
determining eligibility for assistance under the AFDC-FC (Foster Care)
program (Title IV-A of the Act) was to disallow only for individually
identified errors. Thus, in Decision No. 319, the Board reversed a
disallowance to the extent that it was based on extrapolation from a
sample.
(9) The Agency noted that the Quality Control provisions specifically
apply to eligibility errors in AFDC payments made under Title IV-A, but
do not mention disallowances for services under Title IV-C (WIN), citing
45 CFR 205.40. The Board called the attention of the parties to
Pennsylvania Department of Public Welfare, Decision No. 485, December
21, 1983, in which the Board reversed a disallowance of FFP claimed for
payments to AFDC recipients who were not registered for WIN, or exempt
from registration, to the extent that the disallowance was based on a
sample. The Agency argued that Decision Nos. 319 and 485 involved error
determinations in AFDC financial assistance payments whereas
WIN-purchased child care is characterized as services. /12/ It cited 45
CFR 220.16(e)(11). /13/ The Agency noted that while Decision 485
involved an aspect of the WIN program, it did so only to the extent that
the requirement to register with the WIN program related to eligibility
for AFDC financial assistance.
As the Board noted in Decision No. 485, we will normally accept
statistical sampling evidence and the results of extrapolation, where
valid, as reliable evidence of the amount of unallowable costs. Id. at
p. 9. We have previously upheld disallowances based on valid
statistical sampling methods, even in the absence of any policy
statement specifically authorizing the use of such methods. See
Decision No. 319 and the decisions cited at page 6.
In Decision No. 319, however, the Board concluded that the Agency had
a policy to disallow for eligibility determination errors involving Aid
to Families with Dependent Children -- Foster Care (AFDC - FC) only when
they were identified individually and thus not when they were based on
extrapolation from statistical samples. In reaching its decision, the
Board examined the complex background of the Quality Control (QC)
program which applied to AFDC eligibility determinations, the revocation
of the regulation providing for disallowances under the QC program, and
issuances by the Social Security Administration (SSA) explaining
disallowance policy after the revocation. In Decision No. 485, the
Board (10) concluded that Agency policy to disallow only individually
identified errors also applied to erroneous payments made to AFDC
recipients who were not registered for WIN and who were not exempt from
registering. The Board based its conclusion on the regulatory section
that defined the scope of the policy in terms of case errors in AFDC
assistance payments, 45 CFR 205.40(a) (1975). The critical difference
between these cases and the instant case is that the Agency policy
applied in the earlier cases does not apply to child care services
authorized under Title IV-C of the Act. 45 CFR 205.40(a) (1975); 45
CFR 220.61(e)(1). In the absence of any policy not to disallow based on
statistical sampling in the instant case, we see no reason why we should
not accept such evidence as we would normally do. University of
California -- General Purpose Equipment, Decision No. 118, September
30, 1980; Illinois Physicians Union v. Miller, 675 F.2d 151, 155 (7th
Cir. 1982) Rosado v. Wyman, 322 F. Supp. 1173, 1180-1183 (E.D.N.Y.
1970), 397 U.S. 397, 419 (1971); Georgia v. Califano, 446 F. Supp.
404, 409 (N.D. Ga. 1977). Furthermore, the State has not provided us
with any persuasive legal or policy reason why the normal rule should
not apply to these services.
b. The State did not show that the use of the expanded sample was
prejudicial and unfair.
The State alleged that the Agency's expansion of the sample by the
use of additional claims drawn from three counties made the
extrapolation of the results of the expanded sample to four counties
prejudicial and unfair. The State did not cite any authority of its own
for this assertion, but instead relied on the testimony of the Agency's
expert witness.
The State cited with approval the expert's testimony that the method
used by the Agency auditors "may not have been the most efficient"
--
i.e., it may not have resulted in the "lowest precision" (sampling
error) possible. Tr. 131, 114. The State inferred that this
necessarily meant that the sample would have been improved by selecting
from the fourth county for the expansion. We do not agree that this is
a correct inference. To the contrary, the expert testified that
sampling proportionately in accordance with the number in each county is
"usually . . . very inefficient." Tr. 109, 115.
The State had the opportunity to do so, but did not offer any support
for its contention that the method used by the auditors was unfair, nor
any for its contention that it was (11) invalid. /14/ It conceded that
from the viewpoint, of the expert witness "the method used was
satisfactory." /15/ State's Post-Hearing Brief, p. 3. Whether the
Agency had the burden of demonstrating the validity of the sample was
not made an issue by the parties, but the Agency did establish the
validity of the sample by the testimony of the expert witness. The
State did not meet either a burden of initial proof or a burden of going
forward.
Conclusion
For the reasons stated above, we affirm the disallowance to the
extent that it has not been withdrawn. /1/ The disallowance letter was
issued by the National Coordination Committee (NCC) of the WIN
program of the Department of Health and Human Services and the
Department of Labor. Both Departments have responsibilities for the WIN
program; DOL provides job placement and HHS provides financial
assistance and supportive social services. The Office of Human
Development Services (Agency) is the HHS operating component now
responsible for WIN, and is the respondent in this appeal. /2/
The Agency also disallowed FFP in child care payments after WIN
registrants had participated in institutional training for more than one
year. The State appealed, but the Agency withdrew this disallowance
when the State submitted documentation that the State had waived the
one-year limitation as authorized by the WIN regulations. /3/
References are to the October 1, 1978 edition of 45 CFR, unless
otherwise indicated. /4/ On-the-job-training (OJT) and public service
employment (PSE) are two types of state and federal subsidized
employment. /5/ The appeal file submitted by the State contains
a copy of the October 1979 WIN Handbook provisions limiting child care
services to "the first 30 days the individual is employed, although in
unusual circumstances such support may be continued up to 90 days."
Exhibit 3, p. VII-3. The Agency attached to its Post-Hearing Brief an
affidavit by a federal official who alleged that the WIN SAU Chief for
the State of California participated in a redrafting of the social
services part of the Handbook in the summer of 1979. The federal
official, who was then the regional representative to the State of
California, alleged that he had no recollection of the State SAU Chief
objecting to the 90-day limitation. Inasmuch as the period at issue in
this case commenced October 1, 1978, we considered only the 1976
Handbook in our analysis. In any event, the difference in wording is
not significant in this case. /6/ The APA exempts
"interpretative rules, general statements of policy, or rules of agency
organization, procedure, or practice." 5 U.S.C. 553. Even
interpretative rules may not be binding unless there is actual and
timely notice. The State did not dispute that it had such notice. As
both parties observed, the APA also exempts matters relating to grants.
5 U.S.C. 553(a)(2). However, in 1971, this Department voluntarily
subjected its grants to APA requirements. 36 Fed. Reg. 2532.
/7/ The court affirmed a Board decision upholding a disallowance based
on an Action Transmittal requiring new applications for assistance under
Title IV-D of the Social Security Act (enforcement of child support).
/8/ Counsel for the State asserted earlier in the hearing "I suspect
that if you are out in the open marketplace (i.e., unsubsidized
employment) you can, on your own, bargain for more . . . income . . ."
Tr. 12. He did not later challenge the Agency's witness' statement that
"(the) experience in WIN has been that for both . . . OJT . . . and
public service employment . . . the average wages are higher than the
average entry wage into unsubsidized employment . . ." Tr. 29.
/9/ The State cited 45 CFR 224.30(a)(4) and (a)(7), 45 CFR 224.34(c)(1),
and 45 CFR 224.44. Section 224.30(a)(4) states that when child care is
required, it "must be available during the hours the individual is
engaged in a WIN component." Section 224.30(a)(7) states that an
individual shall not be referred to work unless supportive services
necessary for participation are available and that the withdrawal of
such services shall constitute good cause for refusal to participate.
Section 224.30(c)(1) lists child care as one of those supportive
services and states that such services must be responsive to breakdowns
in prior arrangements in order to ease or avoid disruption of
employment. Section 224.44 requires that necessary supportive services
be provided to individuals who are temporarily suspended from WIN
because of employment in another program but who need the services to
continue the other employment. /10/ The Agency witness noted that in
case of an emergency, WIN could fund child care for an
additional 30 days. The witness also commented that child care funds
were also available under Title XX. The states are matched three federal
dollars for every one they spend under Title XX, and nine for every one
under WIN. Tr. 35, 33. The witness had since 1972 worked with the
Division of Program Design of the WIN program at the central office in
Washington, D.C. She participated in the preparation of the WIN
Handbook in 1976 and "basically wrote" the 1979 revision. Tr. 28.
/11/ The Agency noted also that the APA requirement of publication in
the Federal Register may apply even to interpretative rules. 5 U.S.C.
552. However, the requirement is waived where the party to whom the rule
is being applied had actual notice. Here the Agency alleged, and the
State did not dispute, that the State had actual notice. Tr. 21-22 and
affidavit of Deputy to the Director, Region IX WIN Program (attached to
Agency Post-Hearing Brief). /12/ The State did not comment on Decision
No. 485 in its Post-Hearing Brief. /13/ 45 CFR 220.61( e)(11)
states: Child care expenditures for WIN participants must be charged as
a service expenditure and separately identified since Federal funds for
this purpose come from a separate appropriation. The Agency also cited
45 CFR 222.18(d), but there is no such listing. /14/ Prior to
the hearing, the State contended that the Agency erred because two of
the three counties it used for the expanded sample had "the lowest error
ratio". Appeal Brief, p. 5. The expert testified that as far as he
knew the auditors did not use the error rate in deciding where to expand
the sample. Tr. 125. He pointed out that the auditors were sampling
for a variable which was the dollar value of erroneous payments and had
no way of knowing whether a county with fewer erroneous payments would
have a higher dollar value per error. Tr. 122-123. The State did not
mention this point in its Post-Hearing Brief. /15/ The expert
stated that there was no requirement that the expansion be
"proportionate" (i.e., drawn from all four counties) and that any
disproportionality was completely cured by the use of weighting such as
was done here. Tr. 115-116. He said that expanding the sample by
drawing cases from three counties and then extrapolating the result to
all four was "statistically . . . valid." Tr. 128-129.
NOVEMBER 14, 1984