led data enhancements worksho… · commodity contracts & other financial investments ($52,356)...
TRANSCRIPT
LED Data Enhancements
Introducing New Quarterly Workforce Indicators by Worker Race, Ethnicity, and Education
LED Partnership WorkshopArlington, VAMarch 2011
Erika McEntarferLEHD Economic Research Group
Center for Economic Studies, US Census Bureau
2
Overview
• As part of the LED budget initiative, LEHD staff have been working on multiple enhancements to the microdata and public use data products – Enhancing core microdata with additional
information on workers and firms– Expanding data coverage (federal jobs, self-
employed)– New product development (job-to-job flows)
• Today will introduce the first deliverable from the budget initiative and discuss upcoming enhancements
3
New LED data w/race, ethnicity, and education now being released
• First set of deliverables from the LED budget initiative– New QWI tabulations by race & ethnicity,
education by sex– New race, ethnicity, & education data in
OnTheMap• Enormous effort over the last year by the LEHD staff
to get the new variables in the microdata and into the production code
4
New race & ethnicity variables in QWI and OnTheMap:
• Detailed employment indicators by six race variables– White alone– Black or African American Alone– Asian Alone– Native Hawaiian or Pacific Islander Alone– American Indian or Alaska Native Alone– Two or More Races
• And two ethnicity variables• Hispanic or Latino• Not Hispanic or Latino
• Full cross-tabulations for QWI, some categories collapsed for OTM
5
Adding race and ethnicity to the LEHD microdata
Individual Characteristics File (ICF)
PIK DecennialRace
DecennialEthnicity
SSARace
QWIRace
QWIEthnicity
Person1 White alone Not Hispanic White White alone Not HispPerson2 Black alone Not Hispanic Black Black alone Not HispPerson3 . . Hispanic White alone Hispanic
Use Decennial race & ethnicity when available
When no Decennial response, use Census Numident race as conditioning variable for OMB compliant variable impute
6
0.85
0.87
0.89
0.91
0.93
0.95
0.97
0.99
1.01
1.03
1.05
1=Av
erag
e 20
07 e
mpl
oym
ent
Impact of Recession on Employment in Ohio by Race & Ethnicity
White, Non-Hispanic
Black, Non-Hispanic
White Hispanic
QWI Example: Has the recession impacted different workers differently?
Seasonality dominates Hispanic trends
African-American employment 89% of 2007 level
7
Education in QWI and OnTheMap
• New release has detailed industry and geography employment indicators by 4 education classes– Less than a high school diploma– High school graduate, no college– Some college or associates degree– Bachelor’s degree or greater
• Education data are for workers age 25 and older only• Education status is dynamic, particularly before age 25• Needed a static education variable given limited source
data• Older cut-off for OnTheMap
8
Why only four education categories?
• No education registry or census captures educational attainment for the entire U.S. population– Limited survey data on educational attainment
(Long Form, ACS, SIPP, etc)– Use state of the art imputation methods to capture
small-area estimation properties– Need to keep set of education categories at a
higher level of aggregation to make estimation feasible
9
0.9
0.92
0.94
0.96
0.98
1
1.02
1=Av
erag
e 20
07 e
mpl
oym
ent l
evel
Impact of Recession on Employment in Ohio by Education Category
Less than HS educ
HS diploma
Some College
BA or greater
QWI Example: Recession impact on employment by education.
Employment of high-school graduates in 2009:3 93% of 2007 level
10
711 Performing Arts, Spectator
Sports ($64,020)
621 Ambulatory Health Care Services ($56,028)
523 Securities, Commodity Contracts &
Other Financial Investments ($52,356)
551 Management of Companies ($39,440)
425 Wholesale Electronic Markets
and Agents and Brokers ($34,952)
334 Computer and Electronic Product
Manufacturing ($32,232)
325 Chemical Manufacturing ($31,428)
541 Professional, Scientific, and Technical
Services ($30,468)
524 Insurance Carriers & Related Activities ($29,380)
324 Petroleum & Coal Products Manufacturing
($28,960)
Highest Wage Industries for College-Educated Men in Ohio, by Share of Employment in Sector
QWI Example: What are the highest paying industries for workers with college degrees in my state?
11
$0
$10,000
$20,000
$30,000
$40,000
$50,000
$60,000
$70,000
Comparison of Annual Wages in Selected Industries for College Educated Men in Ohio and New Jersey
Average annual wage (NJ)
Average Annual Wage (OH)
QWI Example: Comparison of average wages for skilled workers across states.
12
Imputing education when no survey response can be found
• Use information from workers with complete education data– Classify or assign each worker to a given cell (a particular
combination of a set of known characteristics such as sex, age, industry, firm size, etc.)
– Within each cell we estimate the distribution of education using the workers with complete information
• The resulting distribution is then used in our sampler to generate implicates
13
Checking the quality of the education impute
• Great care was taken in developing the imputation model and stratifying variables
• Will discuss in detail one test of quality of result:– Use sample with reported education. (‘D Sample’)– Compare key QWI statistics for this sample using reported
education and compare to those using imputed education– Next slides show a sample of key results
14
$-
$2,000
$4,000
$6,000
$8,000
$10,000
$12,000
$14,000
$16,000
$18,000
Less than High School
High School Graduate
Some College or Associates
Degree
College Graduate or
Greater
Illinois: Wages by Education
$-
$2,000
$4,000
$6,000
$8,000
$10,000
$12,000
$14,000
$16,000
Less than High School
High School Graduate
Some College or Associates
Degree
College Graduate or Greater
Delaware: Wages by Education
Average Full-Quarter Wage by reported education
Average Full-Quarter Wage by imputed education (average over implicates)
$-
$2,000
$4,000
$6,000
$8,000
$10,000
$12,000
$14,000
$16,000
$18,000
$20,000
Less than High School
High School Graduate
Some College or Associates
College Graduate or
Greater
New Jersey: Wages by Education
Average Full-Quarter Wage by reported education
Average Full-Quarter Wage by imputed education (average over implicates)
15
0.0%
5.0%
10.0%
15.0%
20.0%
25.0%
30.0%
35.0%
Less than High
School
High School
Graduate
Some College or Associates
Degree
College Graduate or Greater
Illinois: Distribution of Employment by Education
0.0%
5.0%
10.0%
15.0%
20.0%
25.0%
30.0%
35.0%
40.0%
45.0%
Less than High School
High School Graduate
Some College or Associates
Degree
College Graduate or
Greater
New Jersey: Distribution of Employment by Education
B using reported education
B using imputed education (average)
0.0%
5.0%
10.0%
15.0%
20.0%
25.0%
30.0%
35.0%
Delaware: Distribution of Employment by Education
B using reported education
B using imputed education (average)
16
$-
$2,000
$4,000
$6,000
$8,000
$10,000
$12,000
$14,000
$16,000
$18,000
$20,000
Less than High
School
High School
Graduate
Some College or Associates
Degree
College Graduate or Greater
Less than High
School
High School
Graduate
Some College or Associates
Degree
College Graduate or Greater
Female Male
Delaware: Wages by Sex and Education
Average wage by reported education
Average wage by imputed education (average over implicates)
17
$-
$2,000
$4,000
$6,000
$8,000
$10,000
$12,000
$14,000
$16,000
$18,000
Less than High School
High School Graduate
Some College or Associates
Degree
College Graduate or
Greater
Less than High School
High School Graduate
Some College or Associates
Degree
College Graduate or
Greater
Less than High School
High School Graduate
Some College or Associates
Degree
College Graduate or
Greater
1 2 3
Delaware: Wages by FIPS CountyAverage Full-Quarter Wage by reported education
Average Full-Quarter Wage by imputed education (average over implicates)
18
Summary of results• Comparisons generally show close correspondence between
estimates using reported education and imputed education• At the statewide level:
• The difference between average full-quarter wages ranges from -6.7% to +6.4% with the smallest difference being 0.2%.
• The share of beginning of quarter employment varies by a range of -4.9 to 6.4 percentage points with the smallest difference being -0.1 percentage points.
• Results for full-quarter employment generally mirror those for beginning of quarter employment
• Differences in wages and employment by gender and geography are captured in estimates using the imputed variable.
19
Upcoming enhancements• Federal Workers in QWI and OTM
– Very soon– Integrating OPM data– Release of DC QWI and OTM
• New LED data on firm age and size– Age of firm (new firm, old firm) – Size of firm (small firm, large firm)– Due fall 2011.
20
Upcoming enhancements• Integration of self-employment data
– Add to OnTheMap to see where self-employed workers are
– Due to frequency and timing issues, planning only to release in OnTheMap
• New product: job-to-job flows– Identify flows of workers between firms,
industries, geographies– Better understand how labor markets adjust to
trade, demand shocks, where new workers come from etc.
21
Contact Us
Comments/Questions [email protected]
Center for Economic Studieshttp://www.ces.census.gov/
Longitudinal Employer-Household Dynamicshttp://lehd.did.census.gov