New York State Department of Social Services, DAB No. 1076 (1989)

DEPARTMENTAL APPEALS BOARD

Department of Health and Human Services

SUBJECT: New York State Department DATE: July 21, 1989
of Social Services Docket No. 88-113 Audit Control No.
02-50217 Decision No. 1076

DECISION

The New York State Department of Social Services (New York or State)
appealed a disallowance by the Deputy Assistant Secretary for Management
Analysis and Systems (Agency) of $5,272,583 in federal financial
participation (FFP) claimed under several federal-state programs for
developmental costs of New York's Welfare Management System (WMS) for
the period October 1, 1975 to March 31, 1982. The Agency disallowed:
1) $5,014,821 based on an audit finding that the State used incorrect
statistics (related to "data elements") to allocate WMS developmental
costs among benefitted programs; and 2) $257,762 based on an audit
finding that the State's source documentation did not support part of
its claim for costs of non-personal services (such as travel, supplies,
and materials) associated with WMS.

We reverse the $5,014,821 disallowance related to cost allocation
because the State presented persuasive evidence which supported the
general validity of the State's statistics, and which established that
the Agency's audit finding was based on a flawed assumption that certain
data element dictionaries contained all data elements involved in WMS
development. While we find some errors in the State's statistics, the
record does not show that the State's claims generally overstated FFP.
(As we discuss below, if further information shows that the claims were
overstated, the Agency may require an adjustment.) We uphold in
principle the $257,762 disallowance related to documentation because we
reject the State's assertions that the auditors simply failed to review
sufficient documentation and that we should presume that State records
could be shown to support the full amount of its claims if sufficient
effort were devoted to that task. We find, however, that the particular
facts here may warrant a downward adjustment of the disallowance and a
brief opportunity for the State to provide evidence related to that
adjustment.

The WMS Project

In New York, the State supervises and sets standards for public
assistance programs administered by local administering agencies. New
York's Welfare Management System is a unified computer-supported system
to assist these local agencies in validating, collecting, storing, and
processing information on applicants and recipients of several forms of
public assistance, including federal-state programs under Social
Security Act Titles IV-A (Aid to Families with Dependent Children or
AFDC), IV-D (Child Support Enforcement), XIX (Medicaid), XVI
(Supplemental Security Income or SSI) and XX (Social Services), as well
as New York's Home Relief Program and the federal-state Food Stamp
program.

Under the WMS, local agencies use a common application form and
eligibility workbook to collect eligibility data for all included
programs. Information from these documents is then entered into a
central State computer system which performs various eligibility
analysis functions and provides statistical reports to aid in program
management.

New York obtained federal approval for the advanced planning document
for the "upstate WMS system" (the system used by the entire state except
New York City) in 1976 and approval for supplemental planning documents
in 1979 and 1982.

I. Did the State allocate to federal programs an excessive share of
upstate WMS developmental costs?

Summary of the allocation issue

The State had an approved cost allocation plan (CAP) designed to
allocate costs of developing (not operating) the WMS system. The CAP
specified that allocation percentages were to be calculated using
numbers of components (called "data elements") in the system being
developed, which were assigned to benefitted programs and used as a
surrogate measure of the relative benefit to each program. Under the
CAP, the State was to submit periodically to federal authorities for
approval any added (or deleted) data elements in its "dictionary" of
data elements used in functions actually being developed for the WMS
system. For added elements, the State was to also submit a proposed
determination about what programs that data element benefitted. The
State had submitted revisions to its data element dictionary (including
additions and deletions) on a periodic basis until October 1979, and
calculated its allocation percentages based on its lists of approved
data elements.

The Agency disallowed $5,014,821 in FFP based on a federal audit which
questioned whether the approved data elements which the State used in
allocating its costs were used in functions actually being developed by
the State. The auditors gave no specific examples of particular
approved data elements which were not being used; rather, the auditors
inferred their finding from the undisputed absence of many approved
elements from several "operational" data element dictionaries listing
data elements used in WMS prototype systems. The auditors also found
that, for some quarters, the State did not retroactively amend its
allocation to reflect subsequent Agency approval of data elements added
or deleted during those quarters, and that the State had not submitted
new data elements for approval after October 1979. The auditors
reallocated the costs based on the "operational dictionaries."

In our analysis, we start from the proposition that the State generally
has a burden of documenting its costs and that, in this instance, that
burden clearly includes showing that its claim was in accordance with
its CAP methodology. Under the CAP, data elements used to allocate
developmental costs should be those used in functions the State was
developing for the WMS system. Based on the record before us, we find:

o The State presented persuasive evidence, such as contemporaneous
State submissions, federal reviews, and testimony of State officials
involved in system development, that the approved "data elements" were
actually used in the development of the WMS system.

o The auditors' findings to the contrary were flawed in light of
the unrebutted evidence that the "operational dictionaries" the auditors
relied on were only a partial listing of data elements used in system
development. The auditors did not consider other evidence that the
State was developing data elements not reflected in those dictionaries.

o The mere presence in the "operational dictionaries" of some data
elements disapproved for cost allocation purposes does not support the
disallowance. Even if federal disapprovals were intended to preclude
use both for cost allocation purposes and other purposes, which is not
clear from the record, the appropriate remedy would be to remove
associated costs from the WMS developmental cost pool altogether, not to
use disapproved elements to allocate costs.

o While the written record does not support the State's assertions
that it did not need to make retrospective adjustments or to submit new
elements for approval after October 1979, the appropriate remedy for the
State's omissions is to make adjustments consistent with the CAP.

The record here does not contain sufficient information to determine
whether any reallocation based on our findings (or a potential finding
that the State should not have continued to develop disapproved data
elements) would result in a determination that the State's claims for
FFP were overstated. Accordingly, we reverse the disallowance. Of
course, the Agency may, if it wishes, require the State to provide the
further information necessary to determine whether the State's claims
were overstated.

A. Background: The approved CAP methodology-- allocation based on
numbers of data elements

On March 28, 1977, New York submitted a cost allocation plan (CAP),
describing the method it would use to allocate WMS upstate development
costs to benefitting programs, which was approved retroactive to 1975 by
federal authorities. Agency Exhibit (Ex.) C; New York Ex. 8. Under
this approved CAP, developmental costs were distributed among
benefitting public assistance programs based on the number of approved
"data elements" either directly or indirectly assigned to each program.
New York Ex. 8.

The term "data element" was not precisely defined in the CAP; generally,
a data element is a basic unit of information stored or processed in a
computer system. Agency Ex. R; Transcript (Tr.), p. 84. Data elements
could include information received directly from an applicant (such as
family size), information generated by the system (such as a Medicaid
eligibility status code), and information used in processing the
information (such as a recipient identification code). See, e.g.,
Agency Ex. G (data element lists).

New York assigned data elements which could be specifically identified
with a particular program's requirements directly to that program. New
York assigned "common" data elements, which served more than one program
or served general system requirements, to each program using the percent
of common data elements equal to the percent of the total recipient
count (for all programs using WMS) from that particular program. All
data elements, and their assignments to a particular program or as
common elements, were subject to federal approval.

Developmental costs were then distributed based on the percent of total
data elements assigned to each program.

This method was an indirect method of identifying the costs to each
program, since the number of data elements acted as a surrogate measure
to estimate the amount of developmental costs benefitting each program.
The costs distributed may have included some cost for defining and
analyzing the data element itself, but primarily consisted of the costs
of developing WMS functions employing those elements.

In the CAP, both parties recognized that, as the system was developed,
data elements would be added and deleted as necessary. For this reason,
the State was to submit for federal approval any revisions to WMS data
element dictionaries (compilations of names and definitions of data
elements used by system developers). While the CAP did not specify the
time frame for submission of revisions, the testimony of those involved
in the CAP negotiation process indicated that the parties had understood
that revisions were to be submitted for each quarter. Tr., pp. 65-66,
156. This understanding is supported by the practice the State followed
in the initial years of WMS development. See Agency Ex. E.

It is undisputed that, over the course of WMS development, New York
followed the CAP process to receive approval for a list of data elements
which totalled 550 elements (with one exception, discussed in note 3
below) as of October 1, 1979. After that time, New York submitted no
more revisions to its data element dictionary. It is also undisputed
that New York allocated its costs in accordance with these approved
elements, using the correct program classifications and recipient
counts.

For the most part, the controversy in this case centers on whether the
approved elements, as the CAP contemplated, were also functional pieces
of information integral to the WMS system being developed. If not, it
was undisputed that those data elements were not appropriate for use in
allocating costs under the CAP. See Agency Brief at p. 10; Tr., pp.
70-74, 98.

B. The record contains persuasive evidence that the approved data
elements were actually used in developing the WMS system.

The Agency asserted that many of the approved data elements were not
actually used by State employees developing the WMS system (contrary to
the CAP requirement). Since the State has an affirmative burden to
document its costs, see New York State Dept. of Social Services, DAB No.
520 (1984), we examine first the affirmative evidence in the record that
the approved data elements were actually used in WMS system development.

New York asserted that it did not have the simplest form of
documentation for the later periods at issue here: paper copies of the
full data element dictionaries. New York presented a witness who
testified that the full developmental dictionary was not maintained on
paper, but was stored in the computer memory for use by system
programmers. This, of course, did not relieve New York of its burden
to prove its claim; the lack of a single, complete document for each
quarter made more difficult the job of both the auditors and New York.
But, as we discuss below, the record contains other evidence that New
York was using the approved elements in developing the WMS system.

As part of the CAP process, New York submitted periodic reports of the
content of its in-use data element dictionary for federal approval. The
federal approval letters show that these reports were not merely pro
forma lists which were approved without question; federal authorities
examined each proposed element to determine whether it was consistent
with program requirements and reflected necessary WMS system functions.
See, e.g., Agency Ex. E. Furthermore, New York presented evidence that
before submitting data elements for approval, it was necessary to
already begin developing the underlying system function (because the
first step of developing a system function was to define the information
required and the information to be created by the function). Tr., pp.
30-35. Thus, New York's submissions were essentially reports of
developmental work in progress at the time of the submission. Since
these submissions are contemporaneous evidence that the elements
contained in the submissions were being used by system programmers, we
give them substantial weight in making our findings here.

In addition to these submissions, the record contains evidence of at
least three detailed federal reviews (by different program agencies)
during the disputed time period which specifically examined the
then-current data element dictionary and did not note any discrepancy
with the data elements used for allocation purposes. New York Ex. 13
(review report dated March 1, 1978); New York Ex. 11 (review report
dated May 25, 1979); New York Ex. 14 (review report dated August 27,
1979). While the second report explicitly stated that the reviewers had
not raised allocation issues with the State, the reviewers would
certainly have noticed if the State was not developing parts of the WMS
system involving over 35% of the approved elements (as the Agency
alleged in this disallowance).

The State presented testimony by State employees involved in the
developmental process that the approved allocation elements were
actually used by State employees in developing the WMS system during the
disputed period. State witnesses testified that the list of allocation
elements was based on the full developmental dictionary (with some
variances, as we discuss below, because some elements may not have been
appropriate for allocation purposes). See, e.g., Tr. pp. 73-74; 94-95.

The State also attempted to recreate the full developmental dictionary
in use in March, 1982. The auditors rejected this attempt because they
found that the recreated dictionary had duplicative elements and some
disapproved elements. New York Ex. 1, pp. 50-51. As we explain below,
we find that the mere presence of duplicative or disapproved elements
does not support the present disallowance. Furthermore, the auditors
did not allege that any approved elements were missing from the
recreated dictionary.

In addition to the direct evidence listed above, there is an inference
which can be drawn from inconsistency in the auditors' findings: the
auditors found that the number of data elements used in system
development in October 1977 was 428, fell to 271 in June 1979, and did
not rise above 428 until March 1981. New York Ex. 1, p. 49. In the
absence of any explanation, this dramatic change strongly suggests that
one of the figures was erroneous; it is not logical that the size of the
system would shrink as development progressed. As we discuss below, New
York provided a explanation of why the larger figure (which was
consistent with the approved data elements) was correct.

In sum, the record contains strong, credible evidence that all of the
approved data elements were actually used in developing the WMS system.

C. The audit finding that many approved data elements were not in
the "operational dictionaries" does not refute evidence that the data
elements were used in WMS system development.

The disallowance was based on the audit finding that the approved data
elements were a faulty basis for allocation because many approved
elements were not actually used in WMS system development. The auditors
gave no affirmative evidence of spurious data elements; rather, the
auditors inferred their finding from the undisputed absence of the
elements from several data element dictionaries the auditors examined
for the period subsequent to October 1977. New York explained, and the
Agency did not dispute, that these dictionaries were "operational
dictionaries" which were used for on-line tests of system functions or
prototype operational systems. See Tr., pp. 97-98.

The auditors relied on these "operational dictionaries" basically for
two reasons. The first was that the State provided no similar, easily
accessible listing of overall developmental data elements in actual use
by programmers. As we noted above, the State had no paper copy of a
developmental dictionary showing actual use of those elements. While
it is understandable that the auditors sought paper records to support
the State's allocation, nothing in the CAP or any applicable regulation
or guideline required New York to have a paper copy of the developmental
dictionary. The second reason the auditors used these dictionaries is
that the auditors thought that the "operational dictionaries" contained
all elements actually in the system being developed. Essentially, the
auditors asserted that there was (or should be) no distinction between
the full developmental dictionary and an "operational dictionary" for a
prototype system because each should reflect the system being developed.

New York presented unrebutted evidence explaining why these "operational
dictionaries" did not contain all data elements involved in
developmental work. These witnesses indicated that the prototype
operational system was tested in phases, starting in July 1978, which
were concurrent with development work. Id., pp. 97-98, 126; see New
York Hearing Ex. 19. Thus, the early "operational dictionaries"
reflected only part of the total developmental effort because they did
not contain whole functions which had not been yet reached the testing
stage. Tr., p. 105. For example, the State said that some of the 550
approved data elements related to a budget calculation function which
was not fully implemented until December 1980. Id., p. 104. The number
of data elements in the "operational dictionaries" gradually approached
the number of elements in the cost allocation listing as new functions
were implemented.

The Agency itself presented testimony by a witness with considerable
experience in computer system design and operations which supported New
York's use, for cost allocation purposes, of more than just those data
elements integrated into a prototype operating system. He stated that
all elements which the State considered to be within the scope of the
final system should be included in the full data element dictionary
(used for cost allocation purposes). Tr., pp. 195-204. He conceded
that the dictionary might include elements which have not yet been fully
developed or implemented, or even which are eventually discarded as the
system design evolves. Id., pp. 221-22.

While this witness testified that, in his opinion, there should be no
distinction between the full data element dictionary and an
"operational" dictionary because all system users should be working from
the same common vocabulary, this testimony appeared based more in
abstract theory than the facts of this case. Tr., pp. 215-217. The
witness did not account for the unrebutted evidence New York presented
that the system was developed in stages, that development work was still
continuing during the period covered by the "operational dictionaries"
the auditors used, and that some system functions had not been included
in the operational tests. New York Ex. 16; Tr., pp. 97-106.

The Agency witness asserted that all data element dictionaries should be
oriented toward the "end user" of the system (and thus should be similar
to any "operational dictionaries" designed for end users), but this view
was neither reflected in the reference materials submitted by the Agency
(Agency Ex. R), nor was it consistent with the facts. New York
presented unrebutted evidence that the "operational dictionaries" were
directed at particular audiences in the counties where the prototype
systems were tested, such as data entry operators, and did not contain
data elements which were not needed by that audience. Tr., pp. 97-98.
Since the project was tested in stages, New York's witnesses asserted
that users of the prototype system did not need to know data elements
involved in functions not yet integrated into the prototype system.
Tr., pp. 105, 127. Moreover, New York asserted that the "operational
dictionaries" would not need to contain internal system elements which
did not relate to information the operators would need to put into the
system or which the system would compute. Tr., pp. 97-98, 105. Our
examination of the data elements submitted for federal approval
indicates that many of them were related to system functions that "end
users" would not need to know. See Agency Ex. E; Tr., pp. 97-98.

In relying on the "operational dictionaries," the auditors did not
consider other evidence that New York was indeed developing all approved
elements. The auditors apparently gave no weight to the submissions
made during the CAP approval process, and dismissed New York's attempt
to reconstruct the full March 1982 developmental dictionary without
considering whether, even if flawed in some respects, it provided some
support for New York's use of the approved data elements. Id., pp.
50-51. Additionally, one New York witness stated that he had provided
three full dictionaries (October 1979, January 1981 and February 1982),
but these dictionaries were not in the audit workpapers and the auditors
did not acknowledge their receipt. Tr., p. 124. We do not need to
resolve the issue of whether these dictionaries were, in fact, provided
to the auditors since we find other strong evidence in support of the
State's position. We note only that it appears that the auditors did
not fully consider all of the evidence the State provided.

In sum, we find flawed the Agency's finding that the approved elements
were not used in actual WMS developmental work and should not have been
the basis for allocating the disputed costs. For similar reasons, we
find flawed the alternative allocation method the Agency used to
calculate the disallowance. The alternative allocation was based on
data elements found in the "operational dictionaries" which, as we
discussed above, did not fully reflect developmental work. These
dictionaries did not include all data elements related to the
developmental work and included elements, as we discuss below, which had
been disapproved for cost allocation purposes. Nor did the reallocation
reflect the approval process set up in the CAP. The Agency did not deny
New York's allegation that, for example, the auditors' reallocation for
one quarter in 1980 included 14 data elements specifically disapproved.
New York Reply Brief, p. 2; New York Ex. 6. Thus, the alternative
allocation did not reflect either actual developmental work, or the
approval determinations required by the CAP.

D. The mere presence of disapproved elements in "operational
dictionaries" does not support this disallowance.

The auditors found that the "operational dictionaries" they examined
contained some elements which had been disapproved. The auditors used
that finding as a basis to reject the State's allocation method and
substitute their own. The auditors did not specify particular
disapproved elements; thus, the discussion below is necessarily somewhat
abstract.

To some degree, this finding was grounded in a fundamental disagreement
about the meaning of federal approval of individual elements. New
York's witnesses indicated that the State had understood federal review
to be limited to use for cost allocation purposes: in particular the
characterization of data elements as common or unique to a particular
function. See Tr., p. 53, 68. The Agency's witnesses were not
consistent: the Agency's computer systems expert stated that federal
review also went to the issue of whether the element was warranted at
all, id., p. 199, but the Agency's cost allocation expert agreed with
the State. Id., pp. 159-60, 167-171.

New York's witnesses explained the State's view that some data elements,
while disapproved for cost allocation purposes, were related to computer
programming needs and efficiencies, were within the scope of the WMS
project, and were permissible system features. The State asserted that
these elements were, for example, internal system elements used for
processing commands, or elements which were redundant but served to
reduce access times for information for internal computer processing.
See Tr., pp. 68-69.

The Agency's computer systems expert witness essentially agreed that New
York may have identified as "data elements" items which were permissible
but should not be called "data elements" in the sense of the CAP
methodology. He stated that his understanding was that "data elements"
should include only items related to federal program requirements. Tr.,
pp. 206-09. This testimony indicates that the State may have included
data elements in its data element dictionaries which were nonetheless
inappropriate for cost allocation purposes. See also New York Ex. 5
(internal DHHS memorandum suggesting that computer process elements may
be permissible but are not appropriate for cost allocation).

In view of the Agency's own evidence on this issue, we find that the
presence of disapproved elements in the "operational dictionaries" does
not necessarily mean that the State's allocation was wrong. Whether or
not New York followed accepted computer terminology by including these
elements in the dictionary, the Agency provided no reason why their
presence in that dictionary would automatically invalidate the elements
used for cost allocation purposes.

Although some of the reasons given for disapproval imply that the State
should have discontinued developing certain data elements, the approval
letters for the most part appeared to relate only to cost allocation and
did not specify that disapproved elements were outside of the scope of
the federally approved system. Moreover, it is not clear from the
record that these particular elements were the "disapproved" elements
the auditors found in the "operational dictionaries." We see no reason
why the Agency could not look into this issue further and, if
appropriate, disallow costs attributable to unapproved activities to the
extent they had been charged to federal programs. That is not, however,
what was done in this disallowance.

In sum, we find that the presence of disapproved or deleted elements in
"operational dictionaries" does not support this disallowance because
the Agency did not find that New York developed any particular elements
which were outside the scope of (not necessary and reasonable to) the
approved system. Of course, this does not preclude the Agency from
identifying specific elements, if any, which were outside of the scope
of the approved WMS system and disallowing particular costs associated
with unapproved activities, but the Agency must articulate its basis
before disallowing those costs and the disallowance calculation must
reflect this basis.

E. New York's failure to submit continuing updates after October
1979 is properly addressed by an adjustment other than this
disallowance.

It is undisputed that the CAP originally contemplated that New York
would submit updates of the data element dictionary on a quarterly basis
for federal approval, but that New York has not submitted updates since
one submitted in October 1979, covering the period through September 30,
1979. New York explained that, while it had initially submitted updates
on a quarterly basis, federal responses on its submissions took between
ten and eleven months. Tr., p. 146. So that it would not have to track
multiple updates, New York then informed federal authorities that it
would no longer submit a new update until it received a response on the
previous submission. Id. New York asserted (and the Agency did not
deny) that, as of the date of the hearing in this case, it still had not
received a necessary response from the U.S. Department of Agriculture
(USDA) on an October 1979 submission related to food stamp elements
(each federal program agency was supposed to approve elements related to
its program). Id., p. 147. Consistent with its practice, New York
stated that it did not submit any revisions pending a response.

The record supports New York's argument that the parties had not
contemplated the length of time required for federal approval and that
New York's actions were the result of the strain on its resources caused
by the unforeseen time lag for federal approval. The documentary record
indicates that approval was a lengthy process (although the delays were
caused partially by requests to New York for more information). A
witness involved in filing claims for New York indicated that New York
had limited capacity to make retroactive adjustments in claims once
filed, because of the resources required for such adjustments. Id., pp.
141-43. The Agency did not dispute that New York submitted updates
shortly after receiving a response on previous submissions. Id., p.
146.

New York argued at the hearing that the CAP had been effectively
modified by the actions of the parties, and noted that the Agency had
not objected to New York's policy not to submit updates until receiving
responses on previous submissions. New York also alleged that federal
authorities had reviewed its cost allocation data element listing in
1979, and had instructed New York that further revisions would not be
necessary but that New York should instead focus on developing the
operational cost allocation method.

We find no documentary evidence to support New York's allegations that
the underlying agreement had been modified. The mere fact that federal
reviewers told the State to focus on operations does not support a
finding that the State no longer had to submit updates to its data
element dictionary in accordance with the approved CAP. The review
report does not explicitly modify the developmental cost allocation
method; it merely requests that New York focus more attention on
operational cost methodologies. New York Ex. 11, pp. 7-8. Furthermore,
New York does not appear to have understood that the report modified the
cost allocation methodology; subsequent to the report, New York
submitted revisions to the data element dictionary pursuant to the CAP
methodology. New York Ex. 15.

Moreover, the fundamental premise of the CAP was that data element
statistics would be updated as appropriate; otherwise, they would not
reflect the benefit to each program from the developmental work actually
occurring. Thus, we conclude that the Agency may require New York to
conform to the CAP requirements and account for changes to the
developmental dictionary used in each quarter of WMS development.

On the other hand, we find that New York's failure to submit updates is
an insufficient basis to support this disallowance. The record does not
show that changes to the developmental data element dictionary after
September 30, 1979 would result in allocation percentages different from
those used by the State. New York provided evidence that developmental
efforts had not changed substantially after the revisions New York
submitted in 1979. From July 1978 through 1980, New York was in a pilot
stage involving three counties and the final shape of the system was
being settled. Tr., p. 55. New York presented testimony that only a
few elements appropriate for cost allocation were added after the last
revision was submitted. Id., pp. 82, 90.

The Agency made no findings about what changes in developmental work
occurred during this period and did not rebut New York's evidence that
any changes in the data element dictionary were minimal. The auditors'
statistics show that prototype "operational dictionaries" during this
time period grew in size but that growth moderated as the size
approached the number of approved elements. New York Ex. 1, p. 49.

Thus, we find that the Agency may require New York to submit for
approval any revisions to the developmental dictionary for all quarters
after September 30, 1979, and that New York must revise its allocation
percentages for those quarters to reflect changes in approved data
elements. We do not, however, find, in New York's failure to do so
previously, a basis to uphold the present disallowance.

F. New York's failure to adjust its allocation percentages
retroactively is also properly addressed by an adjustment other than
this disallowance.

The audit findings in part relied on the failure of New York to adjust
each quarter's claim retroactively for subsequent approvals related to
that quarter. The record indicates discrepancies between the actual
developmental dictionary, as revised on an ongoing basis, and the list
of data elements used for cost allocation. Although New York added
elements to its data element dictionary as soon as they were defined,
New York did not always use those elements for cost allocation for
quarters between the date the element was added and the date the State
received federal approval. Tr., pp. 58-59. This was not entirely New
York's fault; the time lag for federal approval made inevitable some
discrepancy between the actual developmental dictionary and the cost
allocation elements, assuming that elements were added to the actual
developmental dictionary as soon as the elements were defined. Id., pp.
211-12.

The CAP language, and the reasoning behind the CAP, supports the view
that approved elements should have been applied retroactively to the
first quarter after the elements were added to New York's own
developmental dictionary. Agency Ex. C. The CAP language states that
the allocation percentages will fluctuate quarterly based on
modifications to the data element dictionary. Id. The clear
implication is that the allocation percentage for each quarter would be
based on the data elements used throughout the quarter. New York's
early submissions of revised dictionaries were consistent with that
implication, stating that New York would use additional approved
elements for allocation purposes for all claims after the quarter the
elements were originally added to the developmental dictionary. Agency
Ex. E.

While the early submissions by New York and federal approval letters
indicate an agreement that New York should retroactively adjust its
claims based on approvals received later, the later approval letters do
not specifically require retroactive adjustment. See Tr., p. 68; New
York Exs. 2, 3; Agency Ex. E. These later approval letters stated that
the elements were approved for addition to the "current listing of data
elements for use in the calculation of allocation percentages for WMS
claims." Compare Agency Ex. E, Letter dated April 4, 1978 with New York
Ex. 2, Letter dated August 16, 1979. New York's witness asserted that
New York did not understand this language to permit retroactive
adjustments. Tr., p. 68. New York further presented testimony that it
had not made retroactive adjustments because of severe resource
constraints in its administrative offices. Id., pp. 137-39.

In the context of the CAP methodology, we find that the language in the
approval letters did not represent a modification of the concept that
allocation percentages for each quarter were to be based on data
elements used in that quarter. This concept is based on the theory
underlying use of the data elements for allocation purposes: that they
represent the developmental work occurring during any quarter. The same
principles would support adjustments for periods for which the State did
not submit revisions for approval but should have because changes
(however minimal) were made in the developmental effort.

On the other hand, the failure of New York to follow the approved CAP
methodology in this one respect does not support this disallowance
because it does not justify wholesale disregard of the CAP. The
auditors based the disallowance on allocation percentages not calculated
in accordance with the approved CAP, rather than an adjustment within
the framework of the approved CAP. In view of the fact that New York
indicated that the information was still available to make a retroactive
adjustment, which may or may not result in a disallowance, we see no
reason to alter the negotiated and approved allocation method. New
York's witness stated that it had information on the dates each element
was actually added to the developmental dictionary, in the form of
contemporaneous indexes to the developmental dictionaries. Tr., p. 124.
Additionally, New York stated that it had maintained some form of audit
trail to identify elements added between submissions to federal
authorities. Tr., pp. 83-84. The State should be able to provide this
information to the Agency, so that the Agency can ascertain whether
retroactive adjustments are warranted and fulfill its role of reviewing
program classification of the elements (and obtaining USDA approval
where necessary). Although New York indicated that it had limited
capacity to process retroactive adjustments to its claims, we find that
it would not be too great a burden to make a one-time adjustment in the
context of the Agency's further review of these costs. Agency Ex. C.

Thus, the Agency is not precluded from requiring retroactive
adjustments, if warranted, in the allocation for specific time periods.
We do not, however, find a basis, in New York's failure to retroactively
adjust its allocation percentages, to uphold this disallowance per se.

II. Did New York adequately document non-personal service costs
claimed?

Summary of the documentation issue

The Agency alleged that New York had failed to adequately document
claims for non-personal service costs (costs other than salary expenses,
such as heat, light, travel, materials, supplies, and car expenses)
allocated to upstate WMS development. This allegation was based on the
auditors' attempt to reconcile total claims for non-personal service
costs with supporting documentation. New York argued that the full
amount claimed could be reconciled with the documentation, but that the
cost of doing so would be excessive. New York argued essentially that
the process by which vouchers were verified and assigned to particular
projects provided sufficient safeguards so that it was unnecessary to
further scrutinize the costs or to be able to track individual
expenditures. New York asserted, in light of this process, that the
Agency should bear the burden of affirmatively showing that New York's
claims were inaccurate. New York contended that the Agency's failure to
reconcile the claim with supporting documentation indicated only that
the Agency had been unwilling to commit the resources to meet its burden
of proof.

As we discuss below, we uphold the disallowance in principle based on
the general obligation of a grantee to document its costs. We give
particular weight to the uncontested fact that the disputed claims were
not supported by New York's own formal accounting records, and to the
uncontested fact that New York itself, after considerable effort, did
not succeed in fully reconciling a sample quarterly claim with
supporting documentation. We find, however, that some reduction of the
amount disallowed is appropriate in light of the parties' examination of
a single quarterly claim which indicated that a significant proportion
of the disputed costs could be reconciled with supporting documentation.
We provide New York with a brief opportunity to submit factual evidence
related to that adjustment (since the dispute before the Board revolved
largely around the issue of the burden of proof).

A. Background: The calculation of the disputed amounts.

New York began work on the WMS system prior to obtaining approval of a
CAP. See Tr., pp. 323-24. All WMS costs were, at that time,
accumulated under the code 016. In 1983, a CAP was approved
retroactively for WMS central office costs. The CAP required that WMS
costs should be subcategorized into six separate projects, including
upstate WMS development. The CAP stated that non-personal service costs
which could be identified with a particular WMS project would be charged
directly to that project, and remaining non-personal service costs would
be distributed among projects using percentages derived from staff
counts. New York Ex. 20. To prepare its claims, New York
retroactively categorized costs accumulated under the code 016, when
practical, into the six projects, apparently based primarily on the
information concerning the organizational unit which had incurred the
cost. Id., pp. 322-24. New York then subtracted the total amount it
could readily identify with the six projects from the total costs
accumulated under the 016 code to obtain the amount which it had not
been able to specifically identify with a particular WMS project. New
York entered that amount as the "016 Indirect" amount for each quarter,
and allocated the amount to WMS projects based on staff counts. Id.,
pp. 349-50.

The auditors recommended disallowance of amounts identified by the State
as "016 Indirect" on cost allocation worksheets and allocated to upstate
development, after finding that they could not reconcile these amounts
with the supporting documentation. The auditors tested the supporting
documents for one sample quarter, the first quarter of 1982. Although
the auditors reconciled 98 percent of the total non-personal service
costs with supporting documentation (including part of the costs
originally included in 016 Indirect category), the auditors could not
find support for the remaining amounts. Tr., pp. 398-400. The auditors
provided New York with an opportunity to reconcile the remaining
amounts, but New York did not do so. (For the sample quarter, New York
had during the course of the audit reconciled roughly the same 98
percent of the total claim.) New York Ex. 1, pp. 22-24.

The auditors therefore concluded that the amount claimed was in excess
of the total amount which could be verified from the documentation. Id.
While the auditors' recommended disallowance for the sample quarter
included only the individual costs for which they found no
documentation, the auditors questioned the total amounts in the 016
Indirect category for each remaining quarter. Id.

B. Neither the CAP allocation method nor New York's accounting
procedures, by themselves, can excuse New York from the responsibility
to document its costs.

In analyzing this issue, we start from the proposition that a grantee
has the burden to document its costs. See, e.g., New York State Dept.
of Social Services, DAB No. 520 (1984). There is a basis in both law
and logic for this obligation. In the general grant regulations at 45
C.F.R. 74.61(b), grantees are required to maintain records which
"identify adequately the source and application of funds for grant- or
subgrant-supported activities." Furthermore, this result is logical
because supporting documentation is generally in the grantee's hands,
and the records are organized by the grantee's own accounting system and
procedures.

New York asserted that any disallowance would be inconsistent with the
indirect method of charging costs permitted by its approved CAP; New
York relied on evidence that an attachment to an approved CAP revision
showed "016 Indirect" costs being allocated using staff counts. New
York Ex. 20. A CAP provision permitting use of an indirect method of
charging costs does not excuse a grantee from the burden to document its
costs, however. While a particular allocation method may alter the
method of documenting costs which would be appropriate--for example,
when using a cost pool method, it is not necessary to individually
identify which costs are distributed to a particular program--the
fundamental obligation to document the amount of costs attributable to
the cost pool remains.

New York further argued that it had met its burden to document costs by
accounting procedures which ensured a correct result in claiming costs.
New York presented undisputed evidence concerning its accounting
practices. New York's witness explained that disbursements from the
State Treasury were regulated by a voucher accounting system. Whenever
a disbursement was made, it was initiated with a voucher, which was
verified by the State. Tr., p. 302. Vouchers contained a sub-purpose
code which indicated the organizational unit whose budget was involved.
Individual agencies submitting vouchers also had the opportunity to
encode a three digit project code onto its vouchers, to categorize the
expenditure. New York used that field as a "billing code" to indicate
whether the voucher was allowable under a federal-state program or
project. Id., p. 306. Since this field was not mandatory, vouchers
could be processed by the State Treasury without it. The witness stated
that vouchers for WMS costs were always coded by the sub-purpose code to
identify the organizational unit whose budget was involved but did not
always contain the 016 project code which indicated that the voucher
should be billed to the WMS project. Id., pp. 317-21.

New York's witness conceded that the computer reports generated by its
accounting system did not fully reflect the assignment of costs to the
WMS program. Id., p. 333. Typically, some vouchers were unassigned or
might have been inappropriately assigned. New York would manually
review the unassigned vouchers to assign them to a particular program or
project, as well as to reassign obviously inappropriately assigned
vouchers, and would enter these manual assignments into a journal. Id.,
p. 317. New York did not actually change the underlying records of the
vouchers or the computer reports reflecting those vouchers. Id., p. 333
("Because it wasn't mandatory and we didn't have a lot of resources,
what the department was doing was just correcting it between the books
and the claim.") Although New York's witness agreed that it would have
been better to correct the underlying accounting records, the
corrections were made off the books prior to submission of a federal
claim, due to a lack of manpower. Id.

Since New York's own formal accounting records thus failed to support
its claim, we find that its accounting procedures did not generate
adequate documentation to meet the State's burden to document claims.
The auditors' review showed that neither of the accounting reports
generated by New York (the R6 and R7 reports) provided a basis for the
full claim. While the State may be able to document its claims through
additional records and adjustments, we see no reason to place the burden
of challenging the State's own formal records on the federal government
in this instance. A state is free to use any accounting system which
will meet program requirements, but must bear the burden of ensuring
that its accounting system provides adequate information. See New York
State Dept. of Social Services, DAB No. 794 (1986).

C. The record does not support a finding that the audit was
deficient.

We reject New York's argument that the problem was not its records but
the auditors' failure to commit the necessary audit resources and to
examine the appropriate records. The Agency's witness detailed the
two-year audit process, and we find no support for the contention that
the audit was flawed or conclusory. A federal auditor testified that
the auditors had made a determined effort to find supporting
documentation, that they had examined all records which the State had
provided to them, and that a shortfall still remained. Id., p. 404.
While the Agency auditor stated that he did not specifically recall the
records of manual adjustments, the State's witness stated that these
records had been provided to the auditors. Id., pp. 404, 356.

We find it significant that New York itself, after considerable effort,
did not succeed in reconciling the amounts attributed to the 016
Indirect category with supporting documentation, even with additional
information on manual adjustments to the accounting records.
Furthermore, we note that the State did not provide even one example of
documentation of a cost which the auditors had not included in their
calculation (such as a manual reassignment to the WMS 016 account of a
voucher or a cost on a voucher). We see no basis to require the federal
government to organize and retrieve the State's documentation in a
meaningful form when the State itself does not do so.

D. The partial reconciliation in one quarter provides evidence that
some costs can be documented.

Ordinarily, New York's failure to provide meaningful documentation would
end the inquiry, but here there is strong evidence that at least some of
the costs could be documented, and New York offered to provide further
evidence if necessary. It appears questionable whether the full amount
of the costs should be disallowed when the federal audit itself
presented evidence that some of the disputed costs could, indeed, be
documented. In examining the overall claim, the auditors performed an
intensive review of the amounts claimed for one quarter, and stated that
a test of that one quarter could be representative of the entire claim
for the purpose of determining if there was an auditable base. New York
Ex. 1, pp. 21-22; Tr., p. 395. For that quarter, the auditors
reconciled 98 percent of total non-personal service costs, including
many costs originally in the 016 Indirect line item. Although the
auditors concluded that they could not render an opinion on the entire
016 Indirect category unless all non-personal service costs could be
reconciled with supporting documentation, the auditors did not explain
the basis for this position.

New York also reconciled approximately 98 percent of the costs in that
one quarter. New York stated that, while the remaining costs could be
reconciled, this would not be cost-effective. New York instead took the
position (which we rejected above) that there was sufficient evidence
that New York's accounting system, considered as a whole, was adequate.
It was undisputed that the cost and difficulty of individual
verification of costs are high in this circumstance. The costs at issue
were incurred over a period of years during which the allocation method
was still being negotiated. It is not clear that the State could have
known in what form to keep the records. The records are voluminous and
complicated, apparently encompassing many hundreds of vouchers from
various different State budget areas, including many vouchers that were
manually adjusted but never recoded in the State's computerized
accounting records.

Although we find above that New York may not avoid documentation
requirements merely by asserting the validity of its accounting
procedures, we find that some adjustment in the disallowance may be
warranted based on the evidence that some of the costs could be
documented and New York's statement that it could reconcile the costs if
necessary. First, since the auditors raised the possibility that their
review of one quarter could be representative of other quarters for some
purposes, we find that New York should have an opportunity to submit
evidence that the results in that one quarter did, indeed, reflect the
percentage of costs which could be documented in other quarters. The
Agency should consider whether the disallowance should be adjusted based
on this evidence. Second, since New York asserted that the entire
amount for the quarter reviewed could be reconciled with supporting
documentation with further effort (but asserted that it did not have the
burden of doing so, a position rejected above), we provide New York with
one final opportunity if it wishes to show that the entire amount
claimed as non-personal services can be reconciled with supporting
documentation for that quarter. In light of the complexity of the
accounting records at issue, we find that New York should have 45 days
after receiving this decision, or such longer period as the Agency
permits, to submit evidence relating to these adjustments.

We therefore uphold the disallowance of $257,762 in nonpersonal services
costs, subject to the adjustments discussed in the previous paragraph.

Conclusion

In sum, we reverse the disallowance of $5,014,821 but do not preclude
further review of issues not directly raised by this disallowance:
whether the State included in its claim any unallowable costs (see pages
12-15), and whether the State should adjust its allocation percentages
to reflect changes in data elements after October 1979 and to apply
approved data elements retroactive to the first full quarter in which
the elements were used (see pages 15-19). We uphold the disallowance of
$257,762, subject to adjustment consistent with the discussion above on
pages 23-24. With respect to this adjustment, we permit New York a
limited opportunity to submit further information within 45 days after
receiving this decision, or such longer period as the Agency permits.


_______________________________ Cecilia Sparks Ford

_______________________________ Norval D. (John) Settle

_______________________________ Judith A. Ballard Presiding
Board