|
|
Giving Hope and Support to America's Children
A Guidebook For A Common
Methodology For Determining NDNH-Attributable Child Support
Collections
Department of Health and Human Services
Administration for Children and Families
Federal Office of Child Support Enforcement
Federal Parent Locator Service
The Guidebook for a Common Methodology for Determining
NDNH-Attributable Collections is designed to provide a set of
detailed, step-by-step instructions for estimating annual child
support collections attributable to new hire (W-4) or quarterly
wage (QW) proactive matches returned to states from the National
Directory of New Hires (NDNH). These proactive matches occur after
a submission of employment information to the NDNH, which results
in an automatic comparison of the data with child support case data
in the Federal Case Registry (FCR). Upon receipt of a proactive
match, states can issue an income withholding order (IWO).
Currently, all 50 States, D.C., and Puerto Rico, as well as 146
federal agencies, are regularly submitting employment information
to the NDNH.
Why a Common
Methodology?
The purpose of calculating an estimate of annual
NDNH-attributable child support collections is to quantify the
tangible benefits of using proactive matches returned to states by
the NDNH. Since many states have expressed an interest in
quantifying the results at the state level, it is beneficial to
have a consistent methodology employed so that the broader,
national benefits of the system can be demonstrated as well.
Information about the results of these data matches over time can
be useful for policymakers and managers at both the state and
federal levels.
To Conduct the Study You
Will Need
-
A random sample of NDNH W-4 or QW proactive matches for
noncustodial parents (NCPs) for a particular time period, usually a
federal fiscal year (FY). The Federal Office of Child Support
Enforcement (OCSE) uses samples of 1000 for its W-4 collection
studies and at least 1500 for its QW studies to ensure enough
matches with collections for robust results. This sample can be
supplied by OCSE upon request.
-
Staff to research the child support case event and collection
histories for the NCPs in the random sample and to conduct the
final calculations.
Overview of the
Guidebook
This Guidebook presents:
-
A discussion of the characteristics of the random sample;
-
Step-by-step guidelines for reviewing cases and recording
attributable collections; and
-
An explanation of the Formula used for calculating
NDNH-attributable collections.
Although the Guidebook provides detailed information for
calculating the benefits of NDNH data, technical support is only an
email or phone call away. For questions about the Common
Methodology, please contact Linda Deimeke of the OCSE at
202-401-5439 or via email at ldeimeke@acf.hhs.gov.
OCSE can provide you with the sample of proactive matches
necessary to analyze results and estimate collections. However, if
you choose to undertake the programming required to draw the sample
from your system, there are some things to keep in mind.
The Characteristics of the
Sample
The sample contains individual matches returned on NCPs based on
an employer report (W-4 or QW submission). Table 1 below shows what
OCSE includes in and excludes from the sample.[1]
Table 1. The Random Sample
Includes:
|
Excludes:
|
NCP matches
|
Custodial parents (CPs) or putative fathers (PFs) matches
|
NDNH to FCR proactive matches
|
FCR to NDNH proactive matches
Locate request responses
|
Interstate matches
|
Intrastate matches
|
|
For QW studies, OCSE also excludes duplicate information being
provided on a NCP from the same employer across multiple
quarters.
|
The following is a description of the reason OCSE excludes from
its sample the kinds of matches listed:
Matches on CPs and PFs are not useful for a study focusing on
collections, since cases involving those types of participants
would not result in an IWO upon receipt of the NDNH data. However,
OCSE does include matches sent whether or not they pertain to cases
that have a child support order in place, since NDNH to FCR matches
do not include an order indicator in the match response. In
addition, a case participant may have multiple cases, some with
orders established and others with no order in place.
FCR to NDNH Matches and
Locate Request Responses
OCSE excludes FCR to NDNH matches, since they may duplicate the
employment information that would have been returned to the state
earlier, upon the original submission of the employment
information. Locate request responses, like the FCR to NDNH
matches, duplicate data already returned. Given this duplication,
it is best to use the matches returned to the state at the point
closest to when the employer first reported the employment
information.
Most states get matches from their own State Directories of New
Hires (SDNH) before the original employment information is even
forwarded to the NDNH for interstate matching. OCSE excludes these
intrastate matches from its sample, since collections from such
matches would be captured in an SDNH-attributable collections
figure.
Please note that, in order to ensure the randomness of the
sample, these types of matches must be excluded from the universe
(the population of records eligible for the sample) before the
sample is drawn, rather than from the sample itself after it has
been drawn.
If OCSE provides you with the sample, it will provide the
Employer address or the Employer Optional address that was returned
with the proactive match. In addition, other information on the NCP
will be provided, including the:
-
First, Middle, and Last Name;
-
State Member ID associated with the participant;
-
Social Security Number (SSN);
-
Date of Hire (if provided with W-4 submission);
-
Reporting Quarter (QW only); and
-
Wage amount (QW only).
If you request that OCSE provide you with a random sample, we
will also provide the quantity of the universe from which the
sample is drawn, i.e., the total number of records eligible for the
sample.
Once you have the sample, you can research each sample match to
look for NDNH-attributable collections. Below is a description of
the recommended steps for reviewing the cases associated with each
NCP match, determining whether a collection should be attributed to
the NDNH, and recording the data.
-
Pull up the NCP's record by SSN or Member ID. Verify the name
and SSN match. The participant's name may have changed since the
match was sent, so the SSN will be the final determinant. An
SSN-name mismatch should be rare. If it occurs, see whether there
was a change to the name or SSN after the match was returned.
-
Since each participant can have more than one case and NDNH
collections can be spread over more than one case, it is important
to look for multiple cases. Look to see how many cases there are
for the obligor and verify that they are open. If a case is closed,
verify that it was closed or was pending closure before the NDNH
match arrived. If it was closed after the NDNH match arrived, there
may still be NDNH-attributable collections that took place before
the case closed. Also, remember that in some instances a
participant may be a NCP or PF in one case, and a CP or even a
child in another case. So if the NCP is associated with multiple
cases, be sure to only review those cases in which the matched
participant is a NCP.
-
Once you have identified the appropriate cases to examine, look
in the associated event histories for the NDNH match; verify that
it came in and the date it arrived. Remember that multiple NDNH
matches may arrive around the same time, especially for QW
matches.
-
If no IWO was sent, code the record as “No IWO
Sent.” Even if no IWO was sent, it is still valuable to find
the reason it was not issued to provide information on data usage.
There are multiple reasons an IWO may not be issued from an NDNH
match, including:
-
Case closed
-
No order
-
Address already known (from the CP or NCP, or from some other
source)
-
Pursuing other leads (IWOs submitted to other employers)
-
Medical Assistance Only (if no IWO would be sent in that
particular case)
-
Awaiting Uniform Interstate Family Support Act (UIFSA)
processing
-
Wages earned during quarter deemed insufficient (QW only)
-
Already paying from one of the following sources:
-
If an IWO was sent, record the date it was sent and the amount
of the IWO obligation, since this information may be useful for
later analysis. The IWO may have been issued automatically by the
system or initiated later by the caseworker upon review.
-
Then look at the payment history(ies) to see whether payments
came in on the case(s). It is important to distinguish between
payments that are from the specific NDNH IWO being analyzed and
payments from other IWOs or sources that are not attributable to
that sample address. Your state may use special codes to identify
child support collected through income withholdings and may also
store the identity of employers that are the sources of individual
payments. If not, other factors must be considered to ensure that a
collection is attributable to the particular NDNH IWO that resulted
from the sample match. The following are some suggestions for
determining whether payments are NDNH-attributable:
-
First, examine the case event history and confirm that the NDNH
match was received and that an IWO was issued to that employer.
Then look in the collections history to see whether payments from
that employer started coming in after the IWO was issued. In order
to determine whether a collection is from the particular IWO issued
due to the NDNH match, first look for a Federal Employer
Identification Number (FEIN) associated with each payment. If this
is not available, then look at the date the collections began and
ensure that they began at a reasonable time after the IWO was sent.
You are best able to determine what constitutes a
“reasonable” amount of time given your state's own
processes and the method used to issue the IWO - some are issued
automatically, while others may be issued by the caseworker after
reviewing cases in which the automated criteria were not met.
Be certain that the first collections you record are
attributable to the new NDNH IWO, as opposed to being the
continuation of collections from a previous IWO. Each state system
may differ in the kind of detailed information that can be used to
trace collections and ensure that the collections are attributable
to the NDNH match. Sometimes it may be necessary to examine
caseworker notes to determine whether a particular IWO was
successful. In addition to using dates to assess whether incoming
payments are attributable to the new IWO, some state systems store
check numbers associated with each payment. Alterations in amounts,
date cycles, or check numbers may be used to differentiate between
new, NDNH-attributable collections and pre-existing collections
from a previous source, as well as to help determine when the flow
of attributable collections stops. Since some NCPs have more than
one job, in cases in which the NCP was already making payments but
the new IWO was issued to obtain additional collections, look to
see whether any new payments started from this new source.
Collections coming from any source other than the IWO being studied
are not NDNH-attributable, even if the payments come from an NDNH
employer.
-
Sometimes an incoming NDNH match will have its employer address
overwritten by the address contained in a state's employer table.
Even if the IWO is sent to the address provided by your in-state
employer table, the collections are still NDNH-attributable, since
the IWO was triggered by the NDNH match.
-
Be cognizant of IWO collections that are coming from the
“parent company” of the employer to which the IWO was
issued. If an IWO is issued to an employer with a different name
than the one in the sample, or if there is an IWO collection coming
from an employer with a different name than the one to which the
IWO was issued, check to see whether the Employer FEIN is the
same.
If part or all of the NDNH employer address information has been
changed or overwritten by the in-state employer table before the
IWO was sent, it is possible that a different employer name is
shown as the IWO recipient or the source of IWO collections. If
you're not sure whether the NDNH employer and the IWO employer are
related, you may be able to ascertain this through Internet
searches on the employer name.
-
If there are NDNH-attributable collections, you may want to
print the appropriate payment histories to record the payments from
all the matches with collections at a later time. See the section
on “Recording NDNH-attributable Collections,” on page
15, for suggestions on entering the payments once review of the
matches is complete. The following are guidelines to remember when
examining payment histories:
-
For each NCP with a NDNH-attributable collection, record the
collections for each month. The first month of the collections
history (Month 1) is the first calendar month in which a
NDNH-attributable payment is received, even if the payment does not
reflect what a subsequent full month of payments would be. For
example, if the date of the first collection for a NCP is January
25, and subsequent months show weekly or bi-weekly payments,
January would be Month 1 even though it reflects only a partial
month's payments.
-
If there are multiple cases, the monthly payment is derived by
adding together all the payments for that month from all cases with
NDNH-attributable collections. Month 1 is determined by whichever
case receives the first payment. Although unusual, there may be
instances in which, for a NCP with two cases, Case #1 receives its
first attributable payment in January and Case #2 receives its
first attributable payment in February. In such an instance, Month
1 for all cases is January; Case #2 would simply not contribute any
Month 1 payments, and its first payments would simply be added to
the total for Month 2 (February).
-
Miscoding of incoming payments may occasionally occur.
Collections identical to the usual IWO amount may be coded under
another collection category, e.g., as a “personal
check,” but are coming in after the IWO has been issued and
are for the identical amount and in the same date cycle as other
IWO payments. For instances in which you believe a change in the
type of payment is a result of miscoding, consider the payments
attributable to the NDNH IWO. If your state system stores the check
number associated with a payment, you may be able to differentiate
between the payment sources.
-
Be aware of the potential for “payment reversals.”
For example, a payment history may show that a payment was
collected in Month 1, but part or all of a payment was subtracted
out at a later date. If this occurs, be sure to subtract it out of
the month in which the payment was originally entered. Also, some
payments may be backdated and need to be counted in the month the
payment or reversal should originally have been posted. For
instance, a payment received in March and backdated to January
should be counted as a January payment.
-
Be consistent in the dates used for data recording purposes. For
example, for each collection there may be both a receipt date and
an actual posting date. OCSE uses the receipt date to reflect the
true timeframe of the employer's response to the IWO, as there are
sometimes delays in posting payments to an account. Whichever is
used as the date of collection, OCSE recommends using that date
consistently as the collection date for all records reviewed.
-
Please note that, in the payment history report, there may be
several distribution categories, such as current obligations vs.
arrears. The payment entered should be the total payment,
regardless of its eventual distribution.
-
If one month passes with no collections, that month is
considered “skipped.” Not all NCPs have jobs with paid
vacation, and this or other absences may cause gaps in payment
records. In addition, NCPs working for temporary agencies may have
inconsistent payment records, even with an IWO in place. If a month
is skipped but the payments resume the following month, mark the
missed month “Skipped” and enter the subsequent
payments in the appropriate month, provided you are confident that
they result from the original IWO you are tracking.
-
If the payments from the IWO terminate for two or more months,
enter “Ceased” in the first month in which collections
from the IWO do not take place.
-
If collections are continuing to come in at the time you are
conducting your study, enter “Incomplete” for the month
following the last month for which data exists. OCSE has developed
survival rates that take the complete data from other cases into
account to project the probability that payments continued or
ceased during the incomplete months. These survival rates are
discussed in the section “The Formula,” which presents
the formula for calculating total NDNH-attributable collections
from the sample.
Recording NDNH-Attributable
Collections
The following are some suggestions for entering the
NDNH-attributable payments that you have found through your review
of the appropriate child support cases into a spreadsheet.
-
For each proactive match that resulted in NDNH-attributable
collections, create a separate record or row; enter each month's
collections for that record in a separate column; enter up to 12
months of NDNH-attributable collections. Each separate record is a
“hit,” i.e., a match with a successful collection; if
you end up with 100 collection records, you have 100 hits. Dividing
this number by the sample size results in the “hit
rate” discussed later. Using the spreadsheet's COUNT function
for Month 1 will confirm the number of records with a collection.
Table 2 below is an example of a collections spreadsheet and shows
just the first 2 NDNH matches that resulted in collections. Once
you have gone through the cases, you may want to hide columns
containing the NCP's Name, Member ID number, SSN, and the employer
information; you can keep that information in the event that you
want to revisit the case, but hiding the columns saves space when
performing the calculations.
-
It is suggested that when there are multiple payments within a
month, you enter each individual payment, separated by a
“+” sign, and let the spreadsheet do the summing of the
payments to get the month's total; this prevents any mistakes in
adding the payments on a hand calculator.
In Match #1 above, if you enter 150+40 and click on the
“=” sign in the formula box, the spreadsheet will total
the month's collections to $190. In addition, if there are frequent
payments of the same amount multiple times a month, you can use the
“ * ” (multiplication) sign, such as in Match #2.
Payments for Month Two are 20*4, which the spreadsheet will
calculate as $80 for the month when you click on the
“=”. Further, most spreadsheets have a built-in order
of operation such that it will always multiply before adding; in
Match #2 above, Month Three's payments of 20*3+15 will equal $75.
Note that in some spreadsheets you must click on the
“=” sign in the formula bar, even in months with only
one payment, for the spreadsheet to properly interpret the entry as
a month's collection. This is especially important in calculating
the number of hits, which, in turn, determines the hit rate.
-
If there are multiple cases with NDNH-attributable collections
for the same NCP, enter the collections from all the cases into the
same NCP record. (It is important to combine a NCP's
NDNH-attributable collections into the same record; creating a new
record would create an extra hit when there was, in fact, only
one—which would artificially inflate the hit rate.)
-
If collections are still being made at the time of your study,
and you have less than 12 months of payment history, enter
“Incomplete” for the month following the last month for
which data exists, as was done in Match #2 above. In such a case,
you would use the survival rates provided by OCSE in Table 3 and
discussed below.
This section describes the statistical formula used to calculate
the total collections attributable to the NDNH from the sample.
This formula was developed through a collaborative effort of OCSE
and state child support staff and is shown in Figure 1.
The Components of the
Formula
This section provides a description of the various components of
the formula and how to calculate them once all NDNH-attributable
payments have been entered.
Component 1 - The Number of
W-4 or QW Employer Addresses Returned
Component 1 is the total number of W-4 or QW proactive NDNH-FCR
employer address matches that were returned to the state during the
FY studied. This is the universe of records eligible for the random
sample. As mentioned earlier, if you request that OCSE draw your
sample, the exact number of the universe will be supplied with the
sample itself.
Component 2 - The Hit
Rate
The hit rate is the percentage of sample addresses that result
in at least one NDNH-attributable child support collection. A
“collection” is defined as any dollar amount paid by a
NCP that is greater than zero. The hit rate is obtained by dividing
the number of hits by the sample size.
Component 3 - Monthly Median
Dollar Amount From Collections
Using all the collections entered into the table, as described
in the previous section, you can use a spreadsheet's formula
function to calculate the median monthly payment. Be sure the cell
references include all the months and records with collections. It
is important to use the median, because monthly child support
payments can vary widely between cases and even within cases, and
the median is less skewed than the “average” by small
or large outliers that are unrepresentative of the payment
histories as a whole. A medianis the middle observation in a series
of numbers arranged from smallest to largest. For example, in the
series of numbers below…
5, 70, 180, 250, 400, 600, 1500
the median is “250,” because it is the middle number
in the series. A median is not the same thing as an
“average” or “mean” and is used to make
more precise measures of highly skewed or varied data such as
income. For instance, in the numbers above, the average is over
429, while the median is 250, a figure less affected by the
smallest and largest amounts that are often unrepresentative of the
data as a whole.
Component 4 - The Survival
Rate
OCSE has found that, over time, a portion of IWOs stop achieving
collections. The survival rate is the percentage of matches with
collections continuing to pay after Month 1. There are two methods
of calculating survival rates. If you have complete payment
histories for all the sample matches that achieved a collection,
take the number of collection records in each subsequent month as a
percentage of the number of Month 1 collection records. For
example, if there were 150 collections in Month 1, and in Month 6
there were only 75 collections, the Month 6 survival rate is 50%
(75 Month 6 collections divided by 150 Month 1 collections).
In some cases, collections may have started shortly before you
conduct your study. In these situations you may have some cases
with incomplete collection histories (i.e., at the time of the
study you don't know yet if the collections will cease or continue
in the future). If that is the case, you must use the estimated
survival rates provided in this Guidebook. These estimates have
been calculated based on several state studies conducted by OCSE
and are based on regression analysis designed to forecast the
likelihood of payments over time in situations in which some data
points are “censored” (i.e., incomplete). Table 3shows
the survival rates for NDNH-attributable W-4 collections in Column
2 and the survival rates for QW collections in Column 3. For both
W-4 and QW collections, 100% of the matches have collections in
Month 1. By Month 4, however, the likelihood that the
NDNH-attributable payments will continue is only 53.0% for W-4
collections. The reasons for the decrease over time may include a
change in the NCP's job status or a change to the obligation itself
(e.g., emancipation by the child or children in the case). The QW
collections tend to last longer, as evidenced by the Month 4
survival rate of 82.0%. This likely reflects the fact that NCPs
successfully located through QW matches are less prone to frequent
changes in employment.
Table 3. Survival Rates for W-4 and QW Collections
MONTH NUMBER |
W-4 SURVIVAL RATE |
QW SURVIVAL RATE |
1
|
100.0%
|
100.0%
|
2
|
80.5%
|
93.0%
|
3
|
65.6%
|
87.0%
|
4
|
53.0%
|
82.0%
|
5
|
44.4%
|
76.0%
|
6
|
38.6%
|
71.0%
|
7
|
34.5%
|
67.0%
|
8
|
30.5%
|
62.0%
|
9
|
26.5%
|
58.0%
|
10
|
25.3%
|
54.0%
|
11
|
23.1%
|
51.0%
|
12
|
21.3%
|
47.0%
|
The 12-month window for attributing collections to the NDNH-FCR
was established because it was estimated that, on average, it would
take that long before the NDNH information would have been
available from the next best source, e.g., obtaining QW data
directly from a State Employment Security Agency (SESA). In
addition, the window was designed to assess how NDNH data has
changed the enforcement of child support, and in that regard it is
important to note that:
-
Interstate W-4 data were never before available on a national
basis for child support purposes;
-
QW data were available to states prior to the NDNH but were
limited in several ways:
-
States had to specifically request access to the data for
specific obligors, rather than receiving the data automatically as
now happens;
-
States were limited in the number of QW records they could
request from SESAs;
-
The matching process took much longer than the process with the
NDNH, which matches immediately upon receipt of the QW data.
Because of this limitation, by the time obligors were found, they
had often changed jobs;
-
States did not have confidence in the QW data because of these
factors, and often the information was not used.
Calculating Total
NDNH-Attributable Collections
-
Once you have entered all the payments and have clicked on the
“=” sign in the formula box for each month's payment,
you can use the spreadsheet to calculate the hit rate and median
(ensuring that the cell references used in the formula function do
not exclude any months or collection records). You can then
calculate the amount of NDNH-attributable collections. An example
of a spreadsheet for calculating total collections is shown in
Table 4.
-
Enter the size of the proactive match universe from which the
sample was drawn, the hit rate and median calculated in the
collections spreadsheet, and the survival rates for the data type
you are studying from the appropriate table shown earlier.
-
Multiply the formula components across to calculate the
attributable collections for each individual month, then sum each
month's results at the bottom. The formula calculates the results
from an entire year's NDNH proactive matches for the particular
data type studied, using the sample results as estimators for the
universe of matches as a whole.
Table 4. Example of the Final Calculation Spreadsheet
Estimated NDNH Collections
|
Month
|
Number of Addresses Returned by the NDNH
During the FY Being Studied
|
Hit Rate
|
Median Monthly Collection
Amount
|
Survival Rate
|
Total Monthly NDNH
Collections
|
1
|
20,000
|
11%
|
$180.00
|
100%
|
$396,000.00
|
2
|
20,000
|
11%
|
$180.00
|
85%
|
$336,600.00
|
3
|
20,000
|
11%
|
$180.00
|
72%
|
$285,120.00
|
4
|
20,000
|
11%
|
$180.00
|
61%
|
$241,560.00
|
5
|
20,000
|
11%
|
$180.00
|
52%
|
$205,920.00
|
6
|
20,000
|
11%
|
$180.00
|
44%
|
$174,240.00
|
7
|
20,000
|
11%
|
$180.00
|
38%
|
$150,480.00
|
8
|
20,000
|
11%
|
$180.00
|
32%
|
$126,720.00
|
9
|
20,000
|
11%
|
$180.00
|
27%
|
$106,920.00
|
10
|
20,000
|
11%
|
$180.00
|
23%
|
$91,080.00
|
11
|
20,000
|
11%
|
$180.00
|
20%
|
$79,200.00
|
12
|
20,000
|
11%
|
$180.00
|
17%
|
$67,320.00
|
Total Annual NDNH-Attributable Child
Support Collections
|
$2,261,160.00
|
This approach to determining NDNH-attributable collections has
been designed using estimators derived from analyzing a sample,
since determining actual NDNH-attributable collections would
require that a state program their system to track
NDNH-attributable collections. Some states have done so, but others
have not had the programming resources available, so a sampling
approach using estimators was developed. The methods presented in
this guidebook could be used to study SDNH-attributable collections
in addition to NDNH-attributable collections, although OCSE has
never performed that type of study. If a state chooses to use this
guidebook to study SDNH collections, they should be aware that it
might be necessary for them to adjust the methods presented here in
conducting that type of study. In addition, a state studying
SDNH-attributable collections would have to draw the sample of SDNH
matches from their own system.
When using the approach discussed in this guidebook, OCSE
recommends that you incorporate quality assurance (QA) measures
into each stage of your analysis, such as revisiting a portion of
the child support cases that were reviewed for NDNH-attributable
collections, as well as ensuring that the data entry of the
payments is accurate and the final calculations are correct. As
part of OCSE's technical support, we would be willing to perform an
independent QA review of the results of your study. Therefore,
please feel free to send your final spreadsheets to OCSE. If you
have any questions about the Common Methodology or this Guidebook,
please contact Linda Deimeke of the OCSE at 202-401-5439 or via
email at ldeimeke@acf.hhs.gov. In addition, if you would like to
provide feedback about this Guidebook or suggest changes, please
feel free to contact us.
|