LEHD Longitudinal Employer Household Dynamics U.S. Census Bureau and U.S. Dept of Labor/ETA program

Post on 16-Dec-2015

219 views 2 download

Transcript of LEHD Longitudinal Employer Household Dynamics U.S. Census Bureau and U.S. Dept of Labor/ETA program

LEHDLongitudinal Employer Household

Dynamics http://lehd.dsd.census.gov/led/

U.S. Census Bureau and U.S. Dept of Labor/ETA program

Elaine MurakamiElaine.murakami@fhwa.dot.gov

FHWA Office of PlanningMay 2007

Outline

• Importance of employment data and home-to-work flow

• Early findings

• Describe LEHD Process

• Caveat Emptor! Review and Evaluation of the Data BEFORE using it for transportation purposes.

Data Sources for Transportation Planning

Source Number of MPOs % of MPOs responding

Decennial Census 138 91%

CTPP 112 74%

CPS 46 30%

Travel Surveys 21 14%

ACS 18 12%

State employment data 17 11%

MPO employment data 13 9%

Commercial Employment Data

4 3%23%

Different universes between Decennial Census/ACS and LEHD

Private for-profit 68%

Private not-for-profit 7% Government

15%

Self-employed10%

Different timeframes between Decennial Census/ACS and LEHD

Decennial Census + ACS

LEHD

Point in time, from sample survey

Survey form allows only one answer.

Accumulation of quarterly records in a calendar year

Individuals can have many different employers.

Currently using Census 2000 for small area home-to-work flows

New addition of 2004 data. Test of 2002/2003 data.

Common sources of error that affect all home-to-work results

• Incomplete address information results in inability to geocode work location

• Incomplete geographic reference files (especially new commercial areas) also results in inability to geocode work locations.

• Interpretation of WHERE is work? – Respondents in surveys, e.g. construction– Payroll offices, headquarters, school district hdqtrs

LEHD Chart

First release of “On the Map” (2002/2003 data)

• School districts often linked to district headquarters office, rather than a specific school. The district headquarters’ address is incorrect. (Colorado)

• State employees most often linked to an office in the State Capitol. (Illinois)

• Too many “out of state” commutes– 40% of workers from “out of state” in the “On the Map” data,

compared to 2% from “out of state” using the CTPP2000. (Illinois)

• Average distance of home-to-work exceeding 25 miles where CTPP2000 found about 12 miles – CTPP more consistent with regional survey and model results.

(California)

Workers who work in Skokie, IL

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

Lessthan 3.0

3.0-5.9 6.0 - 8.9 9.0 -11.9

12.0 -14.9

15.0 -19.9

20.0 -24.9

25.0 +

Distance to Skokie (Miles)

CTPPLEHD

Workers who live in Skokie

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

Lessthan 3.0

3.0 to5.9

6.0 to8.9

9.0 to11.9

12.0 to14.9

15.0 to19.9

20.0 to24.9

Over25.0

Distance from Skokie (Miles)

CTPP

LEHD

Downtown Chicago

Journey to Work Distance

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0 10 20 30 40

One way "crow flies" commute distance (miles)

Cu

mu

lati

ve

Fre

qu

en

cy

Dis

trib

uti

on

HBP SSF SRPBish Emery stancup walnut

shore is like SRP

SJC is like SSF

SF Bay : Hacienda Business Park Compare LEHD to CTPP

CTPP LEHD

Within 5 miles 25% 10%

Within 10 miles 40% 21%

Within 20 miles 65% 38%

Source: Steve Raney, Cities 21

Are these isolated cases?

• We do not believe that these problems are isolated incidents, but instead may reflect systematic problems that have not yet been identified and resolved.

• Recommend thorough review of the data and processes before use by Transportation Planners.

StARS

QCEW (Covered Employment)

MN Model

Where are the Workplace Locations from?

• Step 1. Establish a list of potential workplace locations. – QCEW provides the workplace addresses. Some

addresses are payroll processing offices.

• Step 2. Assign workers to a specific worksite using a model based on MN data.

• Step 3. Add noise to protect confidentiality of firms

Issues with QCEW• Variability among states in level of effort used to get

Multiple Worksite Reports for large employers. – Variability in reporting from large employers for multiple worksite

reports.• Professional Employment Offices (PEOs) provide staffing

services to businesses to handle payroll, timecards, benefits. – This is a known problem in Florida, but may be increasing in other

states. A few states now require PEOs to report individual addresses.

• State and Local Government workers are OFTEN assigned to one location only (similar to business establishments)– The USDOT Bureau of Transportation Statistics (BTS) sponsored

research in IL and FL found that in IL, state government workers were nearly always assigned to Springfield IL.

• School employees are often assigned to a School District office, rather than to individual schools.

Model based on Minnesota data. • The MWR does not link individual

employees by SSN to a specific worksite. – Only in Minnesota (MN) does the

Unemployment Insurance (UI) file with SSN provide a link to a specific worksite.

– Therefore, to assign individual workers to a specific worksite in all other states, a model was built using data from MN.

• Step 3 (Disclosure proofing) only adds noise to the data.

Residence Data: Statistical Administrative Research Systems:

Issues• College/university students are most likely to

report a parent’s address for their W-2, • Temporary employment may also create

problems where a permanent address is used rather than an actual residence.

• Potential problems with incorrect SSN linkages due to undocumented workers using “borrowed” SSNs.

• Moving residences within a year. – Most residence information is taken from IRS 1040’s

(April 1), but primary job is selected by highest earnings.

Caveat Emptor !

• The On the Map project has promise in the long run, but for now,

• Transportation Planners should not use the LEHD On the Map data without thorough review and evaluation as the first step.

• Problems should be conveyed to the LEHD project team, so that research on the causes can be conducted.