Ask the Scholar
Page 44 of 122
I can add historical knowledge about this page.
Page image
OCR
JUL-03-2000 12:16
OMB DEP DIR MGMT
202 395 6974 P.13/37
A second issue with this type of estimation is that there has been a large change in the
wage distribution over time and that low wage workers have seen a decline in their wages.
In the NLSY panel, the average age of the sample increases with time so that PEᵢₜ will be
correlated with time. It is important to control for these time effects.
11
where tᵢ represents calendar time. In this regression if every individual in the NLSY was
born in the same year, time would be collinear with potential experience and we could
not measure the effect using the instrumental variables approach. However, since the
individuals in the NLSY were born from 1957-1964 there is some limited variation. Our
work below essentially uses the variation across cohorts to separately identify the effects of
potential experience and time.
One nice aspect of this instrumental variables strategy is that it can be performed with
repeated Cross section data rather than panel data. First notice that if the a, were known,
we would not need individual observations on actual experience to run the second stage
regression. We could use variation across time, groups, and potential experience to estimate
the ß, and form estimates of Yj- Second, we do not need measures of actual experience to
estimate the aj. Notice that
While we cannot construct actual experience for each individual in the sample, we can
construct an estimate of E (AEit I j,t, PEit). By definition when potential experience is
zero, actual experience is zero. Assuming we have a measure of the number of weeks
worked in year t for each individual, we can construct
E (AEit I j,t,PEᵢₜ = 1) = E (Weeksit I 1, PEit =
E (AEit I j,t,PEᵢₜ= p+1) = E (AEit I j, 1, PEit = p) + E (Weeksit I i, 1, PEit = p)
Following the process iteratively we can form measures of the expected value of experience
for each group and cohort. The biggest drawback of this approach is that the selection
problem may be particularly severe. The formulation above assumes that each individual
works at least one week in each year. If some individuals were working in some years,
9
Page data
- Page
- 44
- Source index
- 0
- Type
- photo
- Media ID
- b7d31b793bcc403a
- Size
- unknown
Document data
- ID
- 565374850
- Core
- doc
- Type
- document
DTO data
{
"id": "565374850",
"sourceUrl": "https://catalog.archives.gov/id/565374850",
"contentType": "document",
"title": "TANF [Temporary Assistance for Needy Families] - High Performance Bonus - Final Rule [binder] [1]",
"citationUrl": "https://catalog.archives.gov/id/565374850",
"collections": [
"Records of the Domestic Policy Council (Clinton Administration)",
"Margy Waller's Files"
],
"iiifBase": "https://s3.us-east-1.amazonaws.com/NARAprodstorage/lz/presidential-libraries/clinton/wjc-dpc/7367483/7367483-20200317S-023-008-2025/7367483-20200317S-023-008-2025-001.jpg",
"thumbnailUrl": "https://s3.us-east-1.amazonaws.com/NARAprodstorage/lz/presidential-libraries/clinton/wjc-dpc/7367483/7367483-20200317S-023-008-2025/7367483-20200317S-023-008-2025-001.jpg",
"largeImageUrl": "https://s3.us-east-1.amazonaws.com/NARAprodstorage/lz/presidential-libraries/clinton/wjc-dpc/7367483/7367483-20200317S-023-008-2025/7367483-20200317S-023-008-2025-001.jpg",
"imageCount": 122,
"hasImages": true,
"source": "import",
"hasTranscription": false
}
Context sent to Scholar
Document identity
{
"localId": "565374850",
"label": "TANF [Temporary Assistance for Needy Families] - High Performance Bonus - Final Rule [binder] [1]",
"core": "doc",
"dtoType": "document",
"citationUrl": "https://catalog.archives.gov/id/565374850"
}
Document source metadata
{
"id": "565374850",
"sourceUrl": "https://catalog.archives.gov/id/565374850",
"contentType": "document",
"title": "TANF [Temporary Assistance for Needy Families] - High Performance Bonus - Final Rule [binder] [1]",
"citationUrl": "https://catalog.archives.gov/id/565374850",
"collections": [
"Records of the Domestic Policy Council (Clinton Administration)",
"Margy Waller's Files"
],
"iiifBase": "https://s3.us-east-1.amazonaws.com/NARAprodstorage/lz/presidential-libraries/clinton/wjc-dpc/7367483/7367483-20200317S-023-008-2025/7367483-20200317S-023-008-2025-001.jpg",
"thumbnailUrl": "https://s3.us-east-1.amazonaws.com/NARAprodstorage/lz/presidential-libraries/clinton/wjc-dpc/7367483/7367483-20200317S-023-008-2025/7367483-20200317S-023-008-2025-001.jpg",
"largeImageUrl": "https://s3.us-east-1.amazonaws.com/NARAprodstorage/lz/presidential-libraries/clinton/wjc-dpc/7367483/7367483-20200317S-023-008-2025/7367483-20200317S-023-008-2025-001.jpg",
"imageCount": 122,
"hasImages": true,
"source": "import",
"hasTranscription": false
}
Document source extras
{
"url": "https://catalog.archives.gov/id/565374850",
"naId": 565374850,
"levelOfDescription": "fileUnit",
"otherTitles": [
"7367483-20200317S-023-008-2025"
],
"recordType": "description",
"ocrSource": "nara-archive"
}
Page context
{
"seq": 44,
"pageIndex": 0,
"type": "photo",
"url": "https://s3.us-east-1.amazonaws.com/NARAprodstorage/lz/presidential-libraries/clinton/wjc-dpc/7367483/7367483-20200317S-023-008-2025/7367483-20200317S-023-008-2025-044.jpg",
"mediaId": "b7d31b793bcc403a",
"ocrText": "JUL-03-2000 12:16\nOMB DEP DIR MGMT\n202 395 6974 P.13/37\nA second issue with this type of estimation is that there has been a large change in the\nwage distribution over time and that low wage workers have seen a decline in their wages.\nIn the NLSY panel, the average age of the sample increases with time so that PEᵢₜ will be\ncorrelated with time. It is important to control for these time effects.\n11\nwhere tᵢ represents calendar time. In this regression if every individual in the NLSY was\nborn in the same year, time would be collinear with potential experience and we could\nnot measure the effect using the instrumental variables approach. However, since the\nindividuals in the NLSY were born from 1957-1964 there is some limited variation. Our\nwork below essentially uses the variation across cohorts to separately identify the effects of\npotential experience and time.\nOne nice aspect of this instrumental variables strategy is that it can be performed with\nrepeated Cross section data rather than panel data. First notice that if the a, were known,\nwe would not need individual observations on actual experience to run the second stage\nregression. We could use variation across time, groups, and potential experience to estimate\nthe ß, and form estimates of Yj- Second, we do not need measures of actual experience to\nestimate the aj. Notice that\nWhile we cannot construct actual experience for each individual in the sample, we can\nconstruct an estimate of E (AEit I j,t, PEit). By definition when potential experience is\nzero, actual experience is zero. Assuming we have a measure of the number of weeks\nworked in year t for each individual, we can construct\nE (AEit I j,t,PEᵢₜ = 1) = E (Weeksit I 1, PEit =\nE (AEit I j,t,PEᵢₜ= p+1) = E (AEit I j, 1, PEit = p) + E (Weeksit I i, 1, PEit = p)\nFollowing the process iteratively we can form measures of the expected value of experience\nfor each group and cohort. The biggest drawback of this approach is that the selection\nproblem may be particularly severe. The formulation above assumes that each individual\nworks at least one week in each year. If some individuals were working in some years,\n9"
}