Workers in the Decennial Census and in the ACS

Census Data for Transportation Planning:Preparing for the FutureMay 11-13, 2005TRB

Alan E. Pisarski

Some things never change!

“Then there are the incompatibilities in the definitions of a worker in the census and in the standard transportation survey.”

AEP; SR 121, July 1970

BACKGROUND - A little history

Last confs: 1994 and again 1996First discussion: 1970 HRB SR 121First of these meetings Aug. 21-23, 1973 Second Dec. 9-12 1984“What happened in last census/what should happen in next?”

THIS HAS THE FLAVOR OF 1994 WHEN WE THOUGHT THE WORLD WOULD CHANGE BUT WERE UNSURE HOW TO ADDRESS IT

Why 2 conferences in 90’s –When ACS was called “continuous monitoring”

94 Plans for CM/ACS were very unclear–we needed time to clarify and to planCTPP heavily embedded in process Risk of losing major source of JTW We in transportation community took the subject very seriously “Bet your job” approach too dangerousCensus went parallel pathsWe are better prepared now! (?)

The Worker QuestionTo us the decennial and the ACS are surveys of workersYes of course demog is very valuable; but the specific transportation interest is all about the workerWorker count and attributesActivity variables associated with

the worker:Plus work location

The Worker Question

Census tabulations do not reflect thatNo reported workers per seFind workers at mode use tabs; (but that is workers who worked last week)Get workers from labor force

MALE FEMALE LABOR FORCE INCREASE BY DECADE

024

68

1012

1416

1960 1970 1980 1990 2000

(mill

ions

)

MALEFEMALE

Long term trend in Commuting

86.0 117.8 69.4 OVERALL CHANGE

13.2 11.513.2 128.3 20009.7 19.2 18.4 115.1 1990

11.4 23.0 18.1 96.7 198013.3 19.5 12.8 78.6 197018.5 11.7 6.9 65.8 1960

58.9 1950

POPULATIONINCREASE%

WORKERINCREASE

%

WORKERINCREASE

(millions)

TOTALWORKERS

(millions)

YEAR

An important pattern to watch –pop vs pop 16-65 vs Civ LF

0%

5%

10%

15%

20%

25%

30%

35%

1960 1970 1980 1990 2000

POP 16-65 CIV LAB F

Pop/Pop 16-65/LF trends

0 5 10 15 20 25 30 35

1960

1970

1980

1990

2000

INCREASE (millions)

TOT POP16-65CIV LF

Fewer Workers = Fewer Commuters

WORKERS ADDED PER DECADE

12.2

19.8 18.4

13.3

0

5

10

15

20

25

1960-70 1970-80 1980-90 1990-00

Mill

ions

of W

orke

rs

WE ARE A NATION OF IMMIGRANTS - AGAIN

02

468

1012

1416

1900 1910 1920 1930 1940 1950 1960 1970 1980 1990*

2000 ADJUSTIMMIG

Decennial / C2SS comparisonsC2SS & Census almost identical structuresNo look at weeks worked or hoursSmall differences re workers (Nat’l)C2SS had higher LF partic and Unemp

Census 65.2%C2SS 66.2%

Emp. Status item allocation big difference:Census 11.1%C2SS 6.0%

ACS study

Decennial-CPS comparisonsCivilian Labor Force

3.27%4.5142.2137.72000

1.05%1.3124.8123.51990

1.15%1.2105.6104.41980

2.37%1.982.080.11970

2.37%1.669.167.51960

5.67%3.361.558.21950

% DIFFDIFFCPS(Millions)

DECEN(Millions)

CIV LF

135.4 est.

142.2

136.9

5.2

APRIL 2000CPS

2.7 million(50% higher)

7.9UNEMPLOYED

7.1 million(5.2% lower)

128.3WORKERS AT WORK

4.5 million(3.2% lower)

137.7CIV LF

7.2 million(5% lower)

129.7EMPLOYED

DIFFERENCE2000 DECEN

Compared to CPS

Tendency to show Emp as UnempCensus 90% successful in placing CPS employed in employedCensus 86% successful in placing CPS not in LF in that categoryOnly 40% successful in placing CPS unemployed in unemployedINDEX OF INCONSISTENCY!

There were group quarters distortions

College campuses went wrong somewhereIf omit GQ: decennial unemployment rate drops from 5.8% to 5.2% - much closer to CPS Athens, Ga. 16-19 yr olds (10% of pop) with 35% unemployment rates

Group Distortions – Emp./Pop.

5.961.455.5Black

9.766.156.4Hispanic

2.765.162.4White

3.057.954.9Female

3.771.868.1Male

-.612.513.1Pop 65+

3.565.762.2Pop 25+

3.464.661.2All

Diff.April CPSCen 2000Variable

HISPANIC SHARE OF POPULATION BY WORKING AGE GROUPS

17%

11% 12%

5%

13%

0%2%4%6%8%

10%12%14%16%18%20%

% <16 %16+ %16-65 % 65+ ALL

Geographic DistortionsIn no state is Census higher than CPS in Emp/Pop ratio32 state’s differences not statistically signif. Some states saw major swings in Emp.

CA – 1.3 millionFL – .5 millionTX – .7 million

Estimates of impacts

7 million worker gap (5? 3?)State specific?Hispanics not reporting work?Older workers looks too good?Differences strong among young and less educated and some work classes = questions too hard?Is work a more fluid situation?

THE WORK/WORKER VARIABILITY ISSUE

MULTI-JOBS PER WORKERPART TIME/OCCASIONALJOB LOCATIONSTART TIMESMODE/ROUTENATURE OF WORK

Cannot be resolved now! (?)

Need to get at individual recordsNot available until microfichingdone in 2006!Consider:

State impacts?Metro impacts?Other area/functional impacts?

For the FutureACS and CPS benchmarked to censusAnd future benchmarks to same pop estimatesVariations will be in emp/pop ratiosWill census make linkages more transparent?Is ATUS part of this set?Will we have 3 integrated surveys?

QUESTIONS

Will we have an official undercount statement from census?Will additional research pin down the issues and problems further?What about state and metro undercounts/overcounts? How big an issue for the future?

Thank you !

Alan E. Pisarski

Recommendations

Some things never change!

“Then there are the incompatibilities in the definitions of a worker in the census and in the standard transportation survey.”

AEP; SR 121, July 1970

Some things never change!“Then there are the incompatibilities in the definitions of a worker in the census and in the standard transportation survey.”“Finally there is the problem of relating the work trip to other, more comprehensive transportation measures such as peak hour travel and total trips.”

Alan E. Pisarski, Use of Census Data for Transportation Planning, SR 121, July 1970