U S C E N S U S B U R E A U Helping You Make Informed Decisions •1902-2002 Issued September 2002 P60-219 Poverty in the United States: 2001 Demographic Programs By Bernadette D. Proctor and Joseph Dalaker A U.S. Department of Commerce Economics and Statistics Administration U.S. CENSUS BUREAU Current Population Reports Consumer Income
41
Embed
Poverty in the United States: 2001 - Census.gov · Poverty in the United States: 2001 Demographic Programs By ... ed with a recession that began in March 2001.3 The increase in the
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
U S C E N S U S B U R E A UHelping You Make Informed Decisions •1902-2002
Issued September 2002
P60-219
Poverty in the United States: 2001
Demographic Programs
ByBernadette D. ProctorandJoseph Dalaker
A
U.S.Department of CommerceEconomics and Statistics Administration
U.S. CENSUS BUREAU
Current Population Reports
Consumer Income
P60
-21
9U
S C E N
S U S B
U R
E A U
Poverty
in th
e Un
ited States: 2
00
1 Cu
rren
t Pop
ula
tion
Rep
orts C
onsu
mer In
come
U.S. Department of CommerceEconomics and Statistics AdministrationU.S. CENSUS BUREAUWashington, DC 20233
OFFICIAL BUSINESS
Penalty for Private Use $300
FIRST-CLASS MAILPOSTAGE & FEES PAIDU.S. Census Bureau
Permit No. G-58
This report was prepared under the direction of John Iceland,Chief, Poverty and Health Statistics Branch. Katrina Wengert andAmanda Kohler provided statistical assistance. Charles T.Nelson, Assistant Division Chief for Income, Poverty, and HealthStatistics, Housing and Household Economic Statistics Division, pro-vided overall direction.
David Nguyen, Demographic Surveys Division, processed theCurrent Population Survey 2001 and 2002 Annual DemographicSupplement files. Caroline Carbaugh, Chief of the SurveyProcessing Branch, Jaime Hasiuk, Stacy Lyons, Mary ThriftBush, and Kirk Davis programmed the detailed tables and pro-duced the publication tables. Special thanks to Donna Benton forher assistance in programming the experimental poverty measures.
Tim J. Marshall, under the supervision of Greg Weyland,Demographic Surveys Division, prepared specifications for thecomputer-assisted interviewing instrument used to conduct theAnnual Demographic Supplement. Andrew M. Stevenson,Technologies Management Office, programmed the instrument.
Additional people within the Census Bureau also made significantcontributions to the preparation of this report. Kathleen S. Short,David Waddington, Elizabeth Gifford, Erika Steinmetz, QiWang, Shailesh Bhandari, and Larry Long reviewed the contents.
Sampling review was conducted by Jana Shepherd and Alfred G.Meier of Demographic Statistical Methods Division.
U.S. Census Bureau field representatives and telephone interviewerscollected the data. Without their dedication, the preparation of thisreport or any other report from the Current Population Survey wouldbe impossible.
Greg Carroll, Barbara M. Abbott, Jan Sweeney, Gloria Davis,and Mary Stinson of the Administrative and Customer ServicesDivision, Walter C. Odom, Chief, provided publications and print-ing management, graphics design and composition, and editorialreview for print and electronic media. General direction and produc-tion management were provided by Gary J. Lauffer, Chief,Publications Services Branch.
Acknowledgments
U.S. Department of CommerceDonald L. Evans,
Secretary
Samuel W. Bodman,Deputy Secretary
Economics and Statistics AdministrationKathleen B. Cooper,
Under Secretary for Economic Affairs
U.S. CENSUS BUREAUCharles Louis Kincannon,
Director
Poverty in theUnited States: 2001
P60-219
Issued September 2002
Suggested Citation
Proctor, Bernadette D. and Joseph Dalaker, U.S. Census Bureau,
Current Population Reports, P60-219,Poverty in the United States: 2001,
U.S. Government Printing Office,Washington, DC,
2002.
ECONOMICS
AND STATISTICS
ADMINISTRATION
Economics and StatisticsAdministration
Kathleen B. Cooper,Under Secretary for Economic Affairs
U.S. CENSUS BUREAU
Charles Louis Kincannon,Director
William G. Barron, Jr.Deputy Director
Vacant,Principal Associate Director for Programs
Nancy M. Gordon,Associate Director for Demographic Programs
Daniel H. Weinberg,Chief, Housing and Household EconomicStatistics Division
For sale by the Superintendent of Documents, U.S. Government Printing Office,
Internet: bookstore.gpo.gov Phone: toll free 866-512-1800; DC area 202-512-1800
Fax: 202-512-2250 Mail: Stop SSOP, Washington, DC 20402-0001.
U.S. Census Bureau Poverty in the United States: 2001 iii
Poverty data offer an importantway to evaluate the nation’s eco-nomic well-being. Because poorpeople in the United States are toodiverse to be characterized alongany one dimension, this reportillustrates how poverty rates varyby selected characteristics—age,race and Hispanic origin, nativity,family composition, work experi-ence, and geography. These datareveal how many people were poorand how the poverty populationhas changed. A description of theofficial measure of poverty may befound on page 5.
Whether one is in poverty or notprovides but one perspective oneconomic well-being. This reportdiscusses as well the extent ofpoverty (page 9) and more compre-hensive experimental measures of
poverty that account for noncashbenefits (such as food stamps) andtaxes (such as the Earned IncomeCredit) in income (page 13).
The estimates in this report arebased on interviewing a sample ofthe population. Respondents pro-vide answers to the best of theirability, but as with all surveys, theestimates may differ from the actu-al values.
HIGHLIGHTS
• The poverty rate in 2001 was11.7 percent, up from 11.3 per-cent in 2000.
• In 2001, people below thepoverty thresholds numbered32.9 million, a figure 1.3 millionhigher than the 31.6 millionpoor in 2000.
• At 16.3 percent, the povertyrate for children remained
higher than that of other agegroups, but did not changebetween 2000 and 2001.
• For people 18 to 64 years old,the poverty rate rose to 10.1 percent in 2001, up from9.6 percent in 2000.
• In 2001, there were 6.8 millionpoor families (9.2 percent), upfrom 6.4 million (8.7 percent) in2000.
• For non-Hispanic Whites, thepoverty rate rose between 2000and 2001 (from 7.4 percent to7.8 percent), as did the numberwho were poor (from 14.4 mil-lion to 15.3 million). Povertyrates for Blacks, Hispanics, andAsians and Pacific Islanders didnot change between 2000 and2001. However, the number ofpoor Hispanics rose to
U.S. Census Bureau Poverty in the United States: 2001 1
Poverty in the United States: 2001
The estimates in this report are based on the 2000,2001, and 2002 Current Population Survey AnnualDemographic Supplements (CPS ADS) and provideinformation for calendar years 1999, 2000, and 2001,respectively. These estimates use population esti-mates based on Census 2000. Earlier reports present-ing data for calendar years 1993 through 2000 usedpopulation estimates based on the 1990 census.
In 2001, the Census Bureau tested a sample expan-sion of 28,000 households to the CPS ADS. Thesample expansion was officially implemented in theestimates presented here. It is primarily designed toimprove the reliability of state estimates of children’shealth insurance coverage, but the larger samplesize also improves the reliability of national esti-mates of other topics.
Because results presented in this report from the 2001survey have been recalculated based on the expandedsample and the Census 2000-based weights, they maydiffer slightly from earlier estimates that did not incor-porate the sample expansion and were based on the1990 census. Appendix B presents more detail on theintroduction of the sample expansion and new popula-tion controls based on Census 2000.
All statements in this report have undergone statisticaltesting, and all comparisons are significant at the 90-percent confidence level. Further information onthe source and accuracy of the estimates is at www.census.gov/hhes/poverty/poverty01/pov01src.pdf.
NEW POPULATION CONTROLS AND EXPANDED SAMPLE
8.0 million in 2001, up from 7.7million in 2000.1
• The poverty rate in the Southincreased from 12.8 percent in2000 to 13.5 percent in 2001.The poverty rates in theNortheast, Midwest, and Westdid not change.
• The poverty rate for people liv-ing in the suburbs rose from 7.8 percent in 2000 to 8.2 per-cent in 2001; the poverty rate
did not change in central citiesor in nonmetropolitan areas.2
• How poverty is measured affectsone’s perception of who is poor.Six experimental measuresshowed lower poverty rates forchildren, Blacks, and people infemale-householder families thanunder the official measure, whilepoverty rates for those 65 andover varied greatly according tohow medical expenses weretaken into account.
• Four of six experimental povertymeasures showed an increase inthe poverty rate from 2000 to2001, while two showed nochange.
POVERTY IN THE UNITED STATES
The poverty rate and the numberof poor both rose in 2001, to 11.7 percent and 32.9 million, upfrom 11.3 percent and 31.6 millionin 2000. These increases coincid-ed with a recession that began inMarch 2001.3 The increase in thepoverty rate in 2001 was the firstyear-to-year increase since 1991-92. Figure 1 displays povertyrates and the number of poor overtime, beginning with 1959, thefirst year for which poverty statis-tics are available. Table 1 presentsthe number of poor and povertyrates in 2000 and 2001 for manydemographic groups and showswhich groups had statistically sig-nificant changes.
2 Poverty in the United States: 2001 U.S. Census Bureau
Figure 1.Number of Poor and Poverty Rate: 1959 to 2001
Note: The data points represent the midpoints of the respective years. The latest recession began in March 2001. Source: U.S. Census Bureau, Current Population Survey, 1960-2002 Annual Demographic Supplements.
Numbers in millions, rates in percent
0
5
10
15
20
25
30
35
40
45
2001 19951990198519801975197019651959
Recession
32.9 million
11.7 percent
Number in poverty
Poverty rate
2 In this report, “suburbs” refers to metropol-itan areas outside central cities.
1 Because Hispanics may be of any race,data in this report for Hispanics overlap slightlywith data for the Black population and for theAsian and Pacific Islander population. Based onthe 2002 Current Population Survey AnnualDemographic Supplement, 3.7 percent of theBlack population and 2.4 percent of the Asianand Pacific Islander population were of Hispanicorigin. For the poor population, Hispanicsmade up 4.7 percent of Blacks and 2.5 percentof Asians and Pacific Islanders (a percentagesimilar to the total Asian and Pacific Islanderpopulation). Despite the sample expansion,single-year data for the American Indian andAlaska Native population are not shown in thisreport because of their small sample size in the2002 Current Population Survey AnnualDemographic Supplement.
3 According to the National Bureau ofEconomic Research, a recession started inMarch 2001. For more information about therecession, go to www.nber.org.
U.S. Census Bureau Poverty in the United States: 2001 3
Table 1.People and Families in Poverty by Selected Characteristics: 2000 and 2001(Numbers in thousands)
Characteristic
2001 below poverty 2000 below poverty1 Difference (2001 minus 2000)2
-Represents zero. *Statistically significant at the 90-percent confidence level.For explanation of confidence intervals (C.I.), see ‘‘Standard errors and their use’’ at www.census.gov/hhes/poverty/poverty01/pov01src.pdf.1Consistent with 2001 data through implementation of Census 2000-based population controls and a 28,000 household sample expansion.2As a result of rounding, some differences may appear to be slightly higher or lower than the differences of the reported rates.3Data for American Indians and Alaska Natives are not shown separately in this table because of the small sample of that population.4Hispanics may be of any race.
Source: U.S. Census Bureau, Current Population Survey, 2001 and 2002 Annual Demographic Supplements.
Age
In 2001, the poverty rate for allchildren under 18 years of age was16.3 percent, higher than the ratesfor people 18 to 64 years old and65 and over (10.1 percent for each).People 18 to 64 years old account-ed for most of the net changebetween 2000 and 2001; both theirnumber of poor and poverty rateincreased (17.8 million and 10.1 percent in 2001, up from 16.7 million and 9.6 percent in2000).
In 2001, 11.7 million children, or16.3 percent, were poor. Childrenrepresent a disproportionate shareof the poor (35.7 percent) as theyare only one-fourth (25.6 percent)of the total population. However,their poverty rate and number ofpoor were unchanged from 2000.
Children under 6 have been partic-ularly vulnerable to poverty. In2001, the poverty rate for related
children under 6 years of age was18.2 percent, unchanged from2000. Of children under 6 living infamilies with a female householder,no spouse present, 48.9 percentwere poor, over five times the rateof their counterparts in married-couple families (9.2 percent).
Race and Hispanic Origin
In 2001, the poverty rate was 7.8 percent for non-Hispanic Whites,22.7 percent for Blacks, and 10.2 percent for Asians and PacificIslanders. For Hispanics (who maybe of any race), the poverty rate was21.4 percent.4 Non-Hispanic Whitesshowed increases in both the
number of poor and poverty rate:15.3 million non-Hispanic Whites, or7.8 percent, were poor in 2001, upfrom 14.4 million and 7.4 percent in2000, respectively. Even though thepoverty rate for non-Hispanic Whiteswas lower than that for the otherracial and ethnic groups, 46.4 per-cent of the poor were non-HispanicWhite. Non-Hispanic Whites togetherwith Hispanic Whites made up abouttwo-thirds of the poor (69.1 percent).
The poverty rate for Blacks in 2001,22.7 percent, did not change from2000 and also was not differentfrom the lowest ever measured (in2000 and 1999); however, thepoverty rate for Blacks remainedhigher than the rates for people ofother racial and ethnic groups in2001.5
4 Poverty in the United States: 2001 U.S. Census Bureau
Figure 2.Poverty Rates by Age: 1959 to 2001
Note: The data points represent the midpoints of the respective years. The latest recession began in March 2001. Data for people 18 to 64 and 65 and older are not available from 1960 to 1965. Source: U.S. Census Bureau, Current Population Survey, 1960-2002 Annual Demographic Supplements.
Percent
0
5
10
15
20
25
30
35
40
45
2001 19951990198519801975197019651959
Recession
10.1 percent10.1 percent
16.3 percent
65 years and over
Under 18 years
18 to 64 years
4 Data users should exercise caution wheninterpreting aggregate results for both theHispanic population and Asian and PacificIslander (API) population because these popula-tions consist of many distinct groups that differin socio-economic characteristics, culture, andrecency of immigration. In addition, the CPSdoes not use separate population controls forweighting the API sample to national totals. Forfurther information, seewww.bls.census.gov/cps/ads/adsmain.htm.
5 Poverty data are available for Blacks begin-ning with 1959, for Hispanics in 1972, for non-Hispanic Whites in 1973, and Asians and PacificIslanders in 1987.
U.S. Census Bureau Poverty in the United States: 2001 5
The Official Measure of Poverty
Following the Office of Management and Budget’s (OMB) Statistical Policy Directive 14, the Census Bureau uses a set of money income thresholds that vary by family size and composition to determine who is poor (see the matrix below).
Poverty Thresholds in 2001 by Size of Family and Number of Related ChildrenUnder 18 Years
(Dollars)
Size of family unit
Related children under 18 years
None One Two Three Four Five Six SevenEight
or more
One person (unrelated individual):Under 65 years . . . . . . . . . . . . . . . . 9,21465 years and over . . . . . . . . . . . . . 8,494
Two people:Householder under 65 years. . . . . 11,859 12,207Householder 65 years and over . 10,705 12,161
Nine people or more . . . . . . . . . . . . . 39,223 39,413 38,889 38,449 37,726 36,732 35,833 35,610 34,238
Source: U.S. Census Bureau.
If a family’s total income is less thanthat family’s threshold, then thatfamily, and every individual in it, isconsidered poor. The official pover-ty thresholds do not vary geograph-ically, but they are updated annuallyfor inflation using the ConsumerPrice Index (CPI-U). The officialpoverty definition counts moneyincome before taxes and does notinclude capital gains and noncashbenefits (such as public housing,medicaid, and food stamps).
Example: Suppose Family A con-sists of five people: two children,their mother, father, and great-aunt. Family A’s poverty thresholdin 2001 was $21,665. Supposealso that each member had thefollowing income in 2001:
Since their total family income,$25,000 was greater than theirthreshold ($21,665), the familywould not be considered “poor”according to the official povertymeasure.
While the thresholds in some senserepresent families’ needs, the offi-cial poverty measure should beinterpreted as a statistical yardstickrather than as a complete descrip-tion of what people and familiesneed to live. Moreover, many ofthe government’s aid programs usedifferent dollar amounts as eligibil-ity criteria.
Poverty rates and the number ofpoor are one important way of ex-amining people’s well-being. Othermore detailed measures of povertyare considered in the sections“Depth of Poverty Measures” and“Experimental Poverty Measures.”
For a history of the official pover-ty measure, see “The
Development of the OrshanskyThresholds and Their SubsequentHistory as the Official U.S. PovertyMeasure,” by Gordon Fisher, atwww.census.gov/hhes/poverty/povmeas/papers/orshansky.html.
Weighted average thresholds:Some data users want a summaryof the 48 thresholds to get a gen-eral sense of the “poverty line.”These average thresholds providethat summary, but they are notused to compute poverty data.
One person $9,039Two people 11,569Three people 14,128Four people 18,104Five people 21,405Six people 24,195Seven people 27,517Eight people 30,627Nine people or more 36,286
Despite its recent expansion to78,000 households nationwide, theAnnual Demographic Supplement tothe Current Population Survey is notlarge enough to produce reliableannual estimates for AmericanIndians and Alaska Natives. How-ever, Table 2 displays 3-year aver-ages of the number of poor Amer-ican Indians and Alaska Natives andtheir 3-year-average poverty rateand provides 3-year-average povertystatistics for the other groups forcomparison.6 For 1999-2001, thepoverty rate for American Indiansand Alaska Natives was 24.5 per-cent, with 0.8 million poor. Thispoverty rate was higher than thepoverty rates for non-Hispanic
Whites, Asians and Pacific Islanders,and Hispanics, but not differentfrom the rate for Blacks. AmericanIndians and Alaska Natives were theonly group to show a decline whenthe 2000-2001 average was com-pared with 1999-2000.
Nativity
The foreign born experienced nochange in their poverty rate or num-ber of poor between 2000 and2001 (16.1 percent, or 5.2 million,were poor in 2001). Of the foreign-born population, three in eight (36.9percent) were naturalized citizens,and the rest were non-citizens.7
While the poverty rate for the for-eign-born population as a wholewas higher than the rate for thenative population (11.1 percent), therate for foreign-born naturalized citi-zens (9.9 percent) was lower thanthe rate for natives.
Among naturalized citizens, boththe poverty rate (9.9 percent) andnumber of poor in 2001 (1.2 mil-lion) were unchanged from 2000.Among noncitizens, 4.0 million, or19.7 percent, were poor in 2001,also unchanged from 2000. Thenative population, however, hadincreases from 2000 to 2001 inboth their poverty rate (from 10.8 percent to 11.1 percent) andnumber of poor (from 26.7 millionto 27.7 million).
Families and UnrelatedIndividuals
The poverty rate for familiesincreased to 9.2 percent in 2001,up from the 26-year-low measured
6 Poverty in the United States: 2001 U.S. Census Bureau
Figure 3.Poverty Rates by Race and Hispanic Origin: 1959 to 2001
Note: The data points represent the midpoints of the respective years. The latest recession began in March 2001. Data for Blacks are not available from 1960 to 1965. Data for the other race and Hispanic origin groups are shown from the first year available. Hispanics may be of any race. Source: U.S. Census Bureau, Current Population Survey, 1960-2002 Annual Demographic Supplements.
Percent
0
5
10
15
20
25
30
35
40
45
50
55
60
2001 19951990198519801975197019651959
Recession
10.2 percent9.9 percent7.8 percent
22.7 percent21.4 percent
Black
Hispanic
White
Asian and Pacific Islander
Non-Hispanic White
6 Data users should exercise caution wheninterpreting aggregate results American Indiansand Alaska Natives (AIAN) because the AIANpopulation consist of groups that differ in eco-nomic characteristics. The CPS does not useseparate population controls for weighting theAIAN sample to national totals. For furtherinformation, see www.bls.census.gov/cps/ads/adsmain.htm.
7 Natives are defined as people born in theUnited States, Puerto Rico, or an outlying areasof the United States, and those born in a foreigncountry but who had at least one parent whowas a U.S. citizen. All others are foreign-bornregardless of date of entry into the UnitedStates or citizenship status. The CurrentPopulation Survey, the source of these data,does not extend to Puerto Rico or to the outly-ing areas of the United States, and thus thoseliving there are excluded from the officialpoverty statistics.
in 2000 (8.7 percent). In 2001, 6.8 million families were poor, upfrom 6.4 million in 2000. The num-ber and percent of non-HispanicWhite families (3.1 million and 5.7 percent, respectively) and Blackfamilies (1.8 million and 20.7 per-cent, respectively) who were poor in2001 were above the corresponding2000 levels. For Hispanic families,the number of poor rose from 1.5 million in 2000 to 1.6 million in2001, but their poverty rate did notchange (19.4 percent).
The number of poor and the pover-ty rate of married-couple familiesincreased from 2.6 million and 4.7 percent in 2000 to 2.8 millionand 4.9 percent in 2001. Blackmarried-couple families were theonly racial and ethnic group tohave an increase in both theirnumber of poor and poverty ratefrom 2000 to 2001.
Families with a female householderand no husband present experi-enced an increase in the number ofpoor (from 3.3 million in 2000 to3.5 million in 2001), but not theirpoverty rate (26.4 percent in 2001).Of those female-householder fami-lies, the number of poor non-Hispanic White and Hispanicfamilies increased, whereas Asianand Pacific Islander families experi-enced a decrease in their number ofpoor as well as their poverty rate.
In 2001, families with a malehouseholder, no spouse presenthad an increase in both the num-ber of poor and poverty rate (from0.5 million and 11.3 percent in2000 to 0.6 million and 13.1 per-cent in 2001). Of those male-householder families, Hispanicfamilies saw an increase in theirpoverty rate; however, the numberof poor increased for both Blackand Hispanic families.
For unrelated individuals, or peo-ple not living with any relatives,the number of poor increased (to9.2 million in 2001) as did thepoverty rate (19.9 percent in2001). Most of the 0.9 percentagepoint overall net increase occurredamong male unrelated individuals,whose poverty rate was 17.3 per-cent in 2001, up from 15.7 percentin 2000. The poverty rate forfemale unrelated individuals wasunchanged in 2001 (22.3 percent).Even though men accounted formost of the net increase in povertyamong unrelated individuals, mostpoor unrelated individuals werewomen (58.5 percent).
Work Experience
People who worked at any time dur-ing the year had a lower povertyrate than nonworkers (5.6 percentcompared with 20.6 percent), butamong poor people, many workedeither part-time or part-year (see
U.S. Census Bureau Poverty in the United States: 2001 7
Table 2.Number of Poor and Poverty Rate by Race and Hispanic Origin: 1999, 2000, and 2001(Numbers in thousands)
-Represents zero. *Statistically significant at the 90-percent confidence level.For explanation of confidence intervals (C.I.), see ‘‘Standard errors and their use’’ at www.census.gov/hhes/poverty/poverty01/pov01src.pdf.1As a result of rounding, some differences may appear to be slightly higher or lower than the difference of the reported rates.2Hispanics may be of any race.Source: U.S. Census Bureau, Current Population Survey, 2000, 2001, and 2002 Annual Demographic Supplements.
Table 3). Of poor people 16 yearsof age and older, 38.3 percentworked, but only 11.5 percentworked full-time year-round. In contrast, 69.4 percent of all peo-ple 16 years old and over worked,and 46.1 percent worked full-time,year-round.
The poverty rates just discusseddescribed workers and nonworkersfrom the individual’s perspective.But the poverty status of a personwho lives with relatives is not justdetermined by personal income,but by the total income of all ofthe family members. Thus, onefamily member’s work experienceaffects the poverty status of all theother members of that family.Therefore, Figure 4 shows povertyrates of people by what type offamily they lived in and bywhether any of their family mem-bers worked.
Figure 4 shows lower poverty ratesfor family members living with atleast one worker than for familymembers living with no workers—7.6 percent compared with 30.5 percent. The same patternheld when families were classifiedby family type.
Region
The South had the highest povertyrate in 2001 (13.5 percent), whichwas higher than its rate of 12.8 percent in 2000. From 1995to 1999, poverty rates of the Southand West were statistically indistin-guishable from each other. Thepoverty rates for the Northeast,Midwest, and West remainedunchanged in 2001—9.4 percentfor the Midwest, 10.7 percent forthe Northeast, and 12.1 percent forthe West. The South also experi-enced an increase in the number ofpoor, to 13.5 million in 2001, upfrom 12.7 million in 2000. TheSouth had a disproportionatelylarge share of the nation’s poor:
41.1 percent of the poor lived inthe South in 2001, compared with35.7 percent of all people.
Residence
The poverty rate increased in metro-politan areas outside central cities(“suburbs”), from 7.8 percent in2000 to 8.2 percent in 2001. For
people living inside central cities,the poverty rate was 16.5 percent in2001, unchanged from 2000. A dis-proportionate share of poor peoplelived inside central cities: 40.7 per-cent compared with 28.9 percent ofall people. Taking suburbs and cen-tral cities together, the poverty ratefor people in metropolitan areas was
8 Poverty in the United States: 2001 U.S. Census Bureau
Table 3.Work Experience During the Year for People 16 Years Oldand Over: 2001(Numbers in thousands)
For explanation of confidence intervals (C.I.), see ‘‘Standard errors and their use’’ atwww.census.gov/hhes/poverty/poverty01/pov01src.pdf.
Source: U.S. Census Bureau, Current Population Survey, 2002 Annual Demographic Supplement.
Figure 4.Poverty Rates of People in Families by Family Type and Presence of Workers: 2001
Source: U.S. Census Bureau, Current Population Survey, 2002 Annual Demographic Supplement.
TotalWith no workersWith 1 or more workers
(Percent)
In malehouseholder
families
In femalehouseholder
families
In married-couplefamilies
In families
7.64.7
21.3
10.2
30.5
15.1
70.2
48.0
9.95.7
28.6
13.6
11.1 percent in 2001, up from 10.8 percent in 2000.
Among those living outside metro-politan areas, the number of poorrose to 7.5 million in 2001, upfrom 7.0 million in 2000. Thisincrease did not translate into ahigher poverty rate; 14.2 percentwere poor in 2001.
State Poverty Data
Table 4 contains poverty rates forthe 50 states, the District ofColumbia, and the United Statesusing 3-year averages covering1999 to 2001 to improve the sta-tistical reliability of the estimates.(See the text box “InterpretingState Poverty Data” on this page.)Readers should be aware thatalthough New Mexico appeared tohave the highest poverty rate inTable 4 (18.8 percent), it was notstatistically different from the ratesfor Arkansas, Mississippi, andLouisiana, though it was higherthan the rates for the other states.Similarly, the 3-year-average pover-ty rate for New Hampshire, eventhough it looked lowest (6.2 per-cent), was not statistically differentfrom that of four other states—Minnesota, Maryland, Connecticut,and Iowa though it was lower than
the rate in the other 45 states andthe District of Columbia.
To compare changes in povertyrates at the state level, the CensusBureau recommends using 2-yearmoving averages (2000-2001 and1999-2000). Based on thisapproach, Figure 5 shows that twostates—South Carolina and Utah—showed increases while fourstates—California, Delaware,Massachusetts, and Nevada—showed declines in their povertyrates.
DEPTH OF POVERTYMEASURES
While categorizing people as “poor”and “nonpoor” is one summary ofeconomic position, in reality eco-nomic situations fall into a muchbroader spectrum. Two “depth ofpoverty” measures more fullyreflect the distribution of people’seconomic well-being. The ratio ofincome to poverty compares afamily’s income with its povertythreshold, and expresses that com-parison as a fraction. The incomedeficit tells how many dollars afamily’s income is below its pover-ty threshold. These measuresillustrate how the composition ofthe low-income population variesby the severity of poverty.
Ratio of Income to Poverty Level
In 2001, the number of “severelypoor” people—defined as thosewith family incomes below one-half their poverty threshold—roseto 13.4 million (4.8 percent), from12.6 million (4.5 percent) in 2000(see Table 5). The severely poorrepresented 40.8 percent of thepoor population in 2001,unchanged from 2000.
The number and percent of “nearpoor” (people with incomes at orabove their threshold but below125 percent of their threshold)remained unchanged in 2001, at12.4 million and 4.4 percent,respectively.
While some demographic groupsmake up similar shares of the popu-lation at varying degrees of poverty,others are unevenly distributed.Table 5 presents the number of peo-ple and percentage below multiplesof their poverty threshold—thosebelow 50 percent of poverty (“Under0.50”), those in poverty (“Under1.00”) and those below 125 percentof poverty (“Under 1.25”). Amongpeople aged 65 and over, 2.2 per-cent were below 50 percent of theirpoverty threshold, compared with4.8 percent for all people. However,
U.S. Census Bureau Poverty in the United States: 2001 9
Poverty estimates for states are not as reliable asnational estimates. These state poverty rateestimates are intended to provide a sense of theranges within which the poverty rates probablyexist. We recommend using caution whencomparing poverty rate estimates across statesbecause their variability is high.
Why show averages? Why not show the latestyear alone?
Averaging poverty rates over several years improvesthe estimates’ reliability. An estimate’s reliability is
measured by a 90-percent confidence interval: thesmaller the confidence interval, the more reliable theestimate. For instance, using 2001 data alone,Alabama had a confidence interval of +/- 1.98 per-centage points around its poverty rate, but using a3-year average, the confidence interval decreased to+/- 1.55 percentage points. For more informationon confidence intervals, see the CPS Source andAccuracy Statement at www.census.gov/hhes/poverty/poverty01/pov01src.pdf.
Interpreting State Poverty Data
10 Poverty in the United States: 2001 U.S. Census Bureau
Table 4.Percent of People in Poverty by State: 1999, 2000, and 2001
-Represents zero.*Statistically significant at the 90-percent confidence level.For explanation of confidence intervals (C.I.), see ‘‘Standard errors and their use’’ at www.census.gov/hhes/poverty/poverty01/pov01src.pdf.Source: U.S. Census Bureau, Current Population Survey, 2000, 2001, and 2002 Annual Demographic Supplements.
among the elderly, the proportionbelow 125 percent of poverty (16.6 percent) was not different thanthe proportion for all people (16.1 percent). Since people 65 andover and the total population hadsimilar percentages below 125 per-cent of poverty, but the older popu-lation was more sparsely populatedamong the severely poor, the olderpopulation therefore was more high-ly concentrated near or above theirpoverty thresholds.
Income Deficit
The income deficit for families inpoverty (the difference in dollarsbetween a family’s income and itspoverty threshold) averaged$7,231 in 2001 (see Table 6),unchanged from 2000.8
The average income deficit wasgreater for poor families with afemale householder with no hus-band present ($7,692) than for
poor married-couple families($6,840), as was the per capitadeficit ($2,378 compared with$1,833). The income deficit percapita is computed by dividing theaverage deficit by the averagenumber of people per family.Because families with a femalehouseholder and no husband pres-ent were smaller than married-cou-ple families, the greater per capitadeficit for female-householder fam-ily reflects their smaller family sizeas well as their lower income.
U.S. Census Bureau Poverty in the United States: 2001 11
TX
NM
WA
NV-2.1
MI
IL
MO
MT
WY
ID
UT2.4UT2.4 CO
HI
NE
AK
KS
OK
WI
IA
LA
ME
VT
IN
KY
TN
AL GA
OH
WV
NC
SC1.7SC1.7
NJMD
CTRI
MA -1.4
DE -1.8
SD
VA
-0.7CA
AZ
NY
DC
OR
PA
ND
MN
FL
NH
MS
AR
DecreaseIncreaseNo change
Percentage Point Change
Figure 5. States With Significant Changes in 2-Year Average Poverty Rates: 1999-2000 to 2000-01
Source: U.S. Census Bureau, Current Population Survey, 2000, 2001, and 2002 Annual Demographic Supplements.
8 All 2000 figures are expressed in 2001dollars.
12 Poverty in the United States: 2001 U.S. Census Bureau
Table 5.Ratio of Family Income to Poverty Threshold for People by Selected Characteristics: 2001(Numbers in thousands)
1Data for American Indians and Alaska Natives are not shown separately because of the small sample of that population.2Hispanics may be of any race.
Source: U.S. Census Bureau, Current Population Survey, 2002 Annual Demographic Supplement.
Example: “Depth of Poverty” Measures
Suppose Family A has five people—two children and three adults—and has an income of $25,000.
Ratio of income = Family A’s income = $25,000 = 1.15 to poverty Family A’s poverty threshold $21,665
Since Family A’s income-to-poverty ratio was at least as great as one, Family A is not poor. However, sinceits ratio was also less than 1.25, it would be considered “near poor,” and its five members would be talliedin Table 5 as “Under 1.25.” All people in the same family have the same ratio.
Since Family A’s income was greater than its threshold, its income surplus—the number of dollars above itspoverty threshold—was $3,335 ($25,000 - $21,665). Family A would be tallied in the bottom half of Table 6, in the column, “$3,000 to $3,999.”
U.S. Census Bureau Poverty in the United States: 2001 13
For poor unrelated individuals(people who do not live with rela-tives), the average income deficitwas $4,550 in 2001; the $4,231figure for women was lower thanthe $4,997 deficit for men.Because there were more femalethan male unrelated individualsaged 65 and over, and becauseunrelated individuals aged 65 andover have lower poverty thresh-olds, the lower average deficit forwomen reflects differences in age,not just income.
In 2001, 368,000 poor familieshad incomes less than $500 belowtheir poverty thresholds, a similarnumber to the 326,000 familieswith incomes within $500 abovetheir respective poverty thresholds.Therefore, slight modifications toraise or lower the poverty thresh-olds would likely have similareffects on the overall poverty rate,in opposite directions.
EXPERIMENTAL POVERTYMEASURES
Using different methods to meas-ure poverty changes one’s percep-tion of who is poor. To measurepoverty, two important compo-nents must be considered:
1. How does one measure a fami-ly’s (or person’s) needs?
2. What resources should onecount as income for meetingthose needs?
In 1995, a panel of the NationalAcademy of Sciences (NAS) issueda report that recommended newways to measure income, families’needs, and other aspects related tomeasuring poverty.9 Because theofficial poverty measure does not
show how taxes, noncash benefits,and work-related and medicalexpenses affect people’s well-being, the NAS panel observed thatthe official measure does not showhow policy changes in those areasaffect who is considered poor. Inaddition, the panel noted that theofficial poverty measure does nottake into account how the cost ofbasic goods (such as food andhousing) has changed relative toother goods since the early 1960s,when the official poverty measurewas developed. Moreover, it doesnot reflect the fact that those costsvary by geography. Nor do theofficial thresholds, according to theNAS panel, accurately account forincreased expenses and economiesof scale that occur as family sizeincreases. Hence, the NAS panelsuggested a way to construct anew poverty measure that address-es these issues.
Table 6.Income Deficit or Surplus of Families and Unrelated Individuals by Poverty Status: 2001(Numbers of families and unrelated individuals in thousands, deficits and surpluses in dollars)
Source: U.S. Census Bureau, Current Population Survey, 2002 Annual Demographic Supplement.
9 Citro, Constance F. and Robert T. Michael,Measuring Poverty: A New Approach.Washington, DC, National Academy Press,1995.
14 Poverty in the United States: 2001 U.S. Census Bureau
For further details about poverty measurement, see: “The Development of the Orshansky PovertyThresholds and Their Subsequent History as the Official U.S. Poverty Measure” by Gordon Fisher, available at www.census.gov/hhes/poverty/povmeas/papers/orshansky.html; Experimental Poverty Measures: 1999 by Kathleen Short, available at www.census.gov/prod/2001pubs/p60-216.pdf; and the National Academy of Sciences’s report, MeasuringPoverty: A New Approach, a summary of which is available at www.nap.edu/readingroom/books/poverty/.
Official and Experimental Poverty Measures: How Do They Differ?
Question Official measure Experimental measures
What counts as income? Gross money income (that is, before taxes) ofall family members living in the same housingunit, not counting capital gains.
Like the official measure, the experimentalmeasures add together the incomes of all fam-ily members who live together, except that theexperimental measures:
Use after-tax incomeInclude noncash benefits as income(such as food stamps and housingsubsidies)Deduct some work-related expenses(such as transportation and childcare) from incomeTake into account medical out-of-pocket expenses (each measure hasa different method for doing so)
What is used as a benchmarkfor need?
First computed in 1963-64, the thresholdswere originally based on U.S. Department ofAgriculture (USDA) food budgets, designedfor families under economic stress. SocialSecurity Administration analysts used 1955USDA data to find out what portion of theirincome families spent on food, then multipliedthe food budgets by the inverse of that factorto get the thresholds (with some adjustmentsfor two-person families and single people).Except when federal interagency committeesmade minor revisions, these thresholds haveonly been updated for inflation annually withthe Consumer Price Index (CPI-U), asdirected by the Office of Management andBudget’s Statistical Policy Directive 14.
Unlike the official thresholds, which have nofixed relationship between thresholds fordifferent-sized families (because they werederived with food budgets and spending data),the experimental measures start with expendi-tures for food, clothing, shelter, and utilities(and for some measures, medical expenses)for a family of four—two adults and twochildren—plus a small additional amount forother expenses. These dollar amounts areadjusted for larger and smaller families, basedon their relative need. The adjustments aremade using three parameters: the first reflectsthat children tend to consume less than adults,the second reflects that a doubling of familysize does not mean that every expensebecomes twice as high, and the third reflectsthat the first child in a single-adult family repre-sents a greater increase in expenses than thefirst child in a two-adult family.
What adjustments are made forgeographic differences in thecost of living?
None. The same thresholds apply to all partsof the country.
Some measures (labeled NGA) make noadjustment; others are adjusted using cost indi-ces by state and metropolitan/nonmetropolitanresidence, based on housing costs.
The Census Bureau has been con-ducting research to refine some ofthe panel’s measurement methodsand to examine how the NAS panel’srecommendations would affect thenumber of poor and the povertyrate.10 Six experimental measuresare discussed below. These meas-ures each account for work-relatedexpenses, noncash benefits (such asfood stamps and housing subsi-dies), and adjust thresholds by fami-ly size in similar ways, but themeasures differ among one anotherby how they account for health carecosts and whether they considergeographic differences in the cost ofliving.11 The first three measures(labeled NGA for “no geographicadjustment”) do not adjust thethresholds to account for geograph-ic differences in housing costs. Thelast three (labeled GA) do, but are
otherwise identical to their counter-parts that do not account for geo-graphic cost differences. (See Short,2001, for a full discussion of themeasures.)
The first measure most closelyreflects the NAS panel’s approachfor taking into account how med-ical out-of-pocket expenses (MOOP)affect poverty, but does notaccount for geographic cost-of-liv-ing differences as noted above.This measure is called “MOOP sub-tracted from income” (MSI-NGA).These medical expenses includehealth insurance premiums, copay-ments made at a doctor’s officethat are not covered by insurance,and other expenses paid out of thepatient’s pocket, such as over-the-counter medications.12 The MSI-NGA measure subtracts theseexpenses from family incomebefore comparing the income withthe family’s threshold, which inthis case excludes medical carefrom the family’s “needs” (thethreshold).
The second measure, “MOOP in thethreshold” (MIT-NGA), increases thepoverty thresholds to take MOOPexpenses into account, instead ofsubtracting these expenses fromincome. Using data from the 1997-1999 Consumer Expenditure Surveyand the 1996 Medical ExpendituresPanel Survey, the Census Bureaucomputed a threshold to allow forfood, clothing, shelter, utilities, andMOOP. How much money wasallowed for MOOP depended on thefamily’s size, the presence of elderlyfamily members, the self-reportedhealth status of the family mem-bers, and differences in healthinsurance coverage across families.Thus, for the MIT-NGA measure, thethresholds’ allowances for MOOPreflect expected—that is, average—medical expenses along thosedimensions.
The third measure, CMB-NGA (for“combined” methods), combinesattributes of both the MSI and MITmeasures. Like the MIT-NGA meas-ure, the CMB-NGA includes expect-ed MOOP expenditures in thethresholds. However, like the MSI-NGA measure, the CMB-NGA takesinto account variations in medical
U.S. Census Bureau Poverty in the United States: 2001 15
10 Short, Kathleen, U.S. Census Bureau,Current Population Reports, P60-216,Experimental Poverty Measures: 1999, U.S.Government Printing Office, Washington, DC,2001. See also the Census Bureau’s povertymeasurement Web site for additional studies:www.census.gov/hhes/poverty/povmeas.html.
11 The NAS panel did not advocate one par-ticular dollar amount upon which to basethreshold revisions; rather, they offered a rangeof thresholds. The analysis here uses the mid-point of the NAS panel’s range.
12 Since MOOP is not reported in the CPSAnnual Demographic Supplement, these valueswere imputed using statistical techniques. SeeShort (2001) for details.
Table 7.Experimental Poverty Measures: 1999 to 2001(Percentage poor)
*Statistically significant change from the previous year.For an explanation of the measures, see text and Short (2001).All estimates use Census 2000-based weights and may differ from previously published estimates.All thresholds are updated for inflation with the Consumer Price Index (CPI-U).See www.census.gov/hhes/poverty/povmeas/report.htm for measures updated using growth in median expenditures.Source: U.S. Census Bureau, Current Population Survey, 2000, 2001, and 2002 Annual Demographic Supplements.
needs across families. The CMB-NGA measure calculates the differ-ence between the expected MOOPand the actual amounts each fami-ly spent out-of-pocket for medicalcare and subtracts the differencefrom family income. This way,families who had greater thanexpected medical expenses may beclassified as poor when they other-wise would not be, but those whowere unexpectedly healthy areclassified as better off than theywould be under both the officialand MIT-NGA measures.13
The last three measures (MSI-GA,MIT-GA, and CMB-GA) includeadjustments to the thresholds toaccount for geographic differencesin housing costs; hence, they arelabeled GA for “geographic adjust-ment.” They are otherwise identi-cal to their counterparts that donot account for geographic costdifferences.
All six measures use the ConsumerPrice Index for All UrbanConsumers to update the thresh-olds from 1999 for inflation. Sixadditional measures are presentedon the poverty Web site; these usegrowth in median expenditures forfood, clothing, shelter, and utilitiescalculated from the ConsumerExpenditure Survey, based on 1999
data, to update the measures (seewww.census.gov/hhes/poverty/povmeas/reports.htm).
Data for each of these measuresand the official measure are pre-sented in Table 7. All of the exper-imental measures yielded a higherpoverty rate in 2001 than the offi-cial measure. The MSI-GA measurehad the smallest difference fromthe official measure (12.3 percentcompared with 11.7 percent forthe official measure), followed byMSI-NGA (12.4 percent), MIT-GA(12.7 percent), and MIT-NGA (12.8 percent). The experimentalmeasures were more stable fromyear to year than the official meas-ure. While the official poverty ratechanged from 1999 to 2000 and
16 Poverty in the United States: 2001 U.S. Census Bureau
Table 8.Experimental Poverty Measures by Selected Characteristics: 2001
Characteristic Official MSI - NGA MIT - NGA CMB - NGA MSI - GA MIT - GA CMB - GA
Note: While the experimental measures differ among one another in their computation of medical expenses and geographic variations in costs, they aresimilar in their scaling of thresholds by family size and their treatment of noncash benefits and child care and work-related expenses. See text and footnotes foradditional information and references.
MSI = Medical out-of-pocket expenses (MOOP) subtracted from incomeMIT = MOOP included in the thresholdsCMB = Combined methodsNGA = No geographic adjustment for housing costsGA = Geographic adjustment for housing costs
Source: U.S. Census Bureau, Current Population Survey, 2002 Annual Demographic Supplement.
13 While conceptually similar, the povertystatus of an individual or family may differbetween the MSI and CMB measures because ofdifferences in the methods used to calculate thecorresponding poverty thresholds.
from 2000 to 2001, none of theexperimental poverty rateschanged from 1999 to 2000 andtwo measures’ poverty rates (forMIT-NGA and CMB-NGA) did notchange between 2000 to 2001.
All of the experimental measuresshowed a different profile of thepoor population than did the officialmeasure (see Table 8). Each of theexperimental measures yieldedlower poverty rates than the officialmeasure for people in families witha female householder and no hus-
band present, whereas the oppositewas true for people in married-couple families and male house-holder families (see Figure 6).
These patterns occurred becausethe official measure does not addnoncash benefits or deduct taxesand work-related expenses fromincome, while the experimentalmeasures do. Hence, the experi-mental measures show relativelylower poverty rates for femalehouseholder families becausethose families received more in-
kind benefits and paid less in taxesand work expenses than didmarried-couple families.
Similarly, poverty rates by age fromthe experimental measures differedfrom the official measure, as shownin Figure 7. People under 18 hadlower poverty rates than under theofficial measure, while those ages18 to 64 had higher rates thanunder the official measure, andthose 65 years and over had differ-ences that were higher still.Including medical expenses when
U.S. Census Bureau Poverty in the United States: 2001 17
Figure 6.Experimental Poverty Rates for Individuals by Type of Family: 2001
Note: While the experimental measures differ among one another in their computation of medical expenses and geographic variations in costs, they are similar in their scaling of thresholds by family size and their treatment of noncash benefits and child care and work-related expenses. See text and footnotes for additional information and references. MSI = Medical out-of-pocket expenses (MOOP) Subtracted from Income MIT = MOOP Included in the Thresholds CMB = "Combined" methods NGA = No Geographic Adjustment for housing costsGA = Geographic Adjustment for housing costs Source: U.S. Census Bureau, Current Population Survey, 2002 Annual Demographic Supplement.
Percent
CMB-GAMIT-GAMSI-GA
CMB - NGAMIT - NGAMSI - NGA
Official
CMB-GAMIT-GAMSI-GA
CMB - NGAMIT - NGAMSI - NGA
Official
0 5 10 15 20 25 30
CMB-GAMIT-GAMSI-GA
CMB - NGAMIT - NGAMSI - NGA
Official
People in married-couple families
People in families with a female householder, no husband present
People in families with a male householder, no wife present
measuring poverty strongly affectedpoverty rates by age because theelderly tend to have high out-of-pocket expenses for health care.The opposite was true for children,who tend to be healthier on aver-age. The method by which oneaccounts for medical care alsoaffects poverty rates by age. Thepoverty rate for people 65 and overaccording to the MIT-GA measure(12.7 percent), though higher thanthe official measure, was not nearlyas high as the measures that sub-tract MOOP from income.
Poverty rates by race and ethnicityalso differed between the experi-mental and official measures, asshown in Figure 8. The experi-mental measures yielded slightlyhigher poverty rates for non-Hispanic Whites and lower rates forBlacks than the official measure.Among Hispanics (who may be ofany race), the experimental meas-ures all showed higher povertyrates than the official measure, butthe geographically adjusted meas-ures produced higher rates forHispanics than those with no geo-graphic adjustment. The higher
Hispanic poverty rates for the GAmeasures reflect that Hispanicstend to live in areas with higherhousing costs, such as California.
Finally, regional poverty ratesbased on the experimental meas-ures differed distinctly from theofficial poverty rates. Among thegeographically adjusted measures,the Northeast and West had higherpoverty rates than the officialmeasure, while the Midwest andSouth had lower rates. These dif-ferences by region reflect thelower housing costs in the Midwest
18 Poverty in the United States: 2001 U.S. Census Bureau
Figure 7.Experimental Poverty Rates by Age: 2001
Note: While the experimental measures differ among one another in their computation of medical expenses and geographic variations in costs, they are similar in their scaling of thresholds by family size and their treatment of noncash benefits and child care and work-related expenses. See text and footnotes for additional information and references. MSI = Medical out-of-pocket expenses (MOOP) Subtracted from Income MIT = MOOP Included in the Thresholds CMB = "Combined" methods NGA = No Geographic Adjustment for housing costsGA = Geographic Adjustment for housing costs Source: U.S. Census Bureau, Current Population Survey, 2002 Annual Demographic Supplement.
Percent
CMB-GAMIT-GAMSI-GA
CMB - NGAMIT - NGAMSI - NGA
Official
CMB-GAMIT-GAMSI-GA
CMB - NGAMIT - NGAMSI - NGA
Official
0 5 10 15 20 25 30
CMB-GAMIT-GAMSI-GA
CMB - NGAMIT - NGAMSI - NGA
Official
Under 18 years
18 to 64 years
65 years and over
and South compared with theNortheast and West.
More information on experimentalpoverty measures can be found onthe Census Bureau’s PovertyMeasurement Research Web site atwww.census.gov/hhes/www/povmeas.html. The Census Bureauplans to continue to issue reportson experimental poverty measuresin order to help policy makers,researchers, and the publicimprove their understanding ofhow measurement issues affect theperception of who is poor.
NOTES, ADDITIONAL DATA,AND USERS’ COMMENTS
CPS Data Collection
The information in this report wascollected in the 50 states and theDistrict of Columbia and does notinclude residents of Puerto Ricoand outlying areas. The populationcontrols used to prepare the esti-mates are based on results ofCensus 2000. Specifically, the esti-mates in this report are controlledto national population estimates byage, race, sex, and Hispanic origin,
and to state population estimatesby age, and are based on the newCPS sample expansion to about78,000 households nationwide.For more information on the CPSexpansion, see Appendix B.
Because the CPS is primarily ahousehold survey, people withunconventional housing who arenot living in shelters are excludedfrom these poverty statistics. TheCPS also excludes armed forcespersonnel living on military basesand people living in institutions.For further documentation about
U.S. Census Bureau Poverty in the United States: 2001 19
Figure 8.Experimental Poverty Rates by Race and Hispanic Origin: 2001
1 Hispanics may be of any race. Note: While the experimental measures differ among one another in their computation of medical expenses and geographic variations in costs, they are similar in their scaling of thresholds by family size and their treatment of noncash benefits and child care and work-related expenses. See text and footnotes for additional information and references. MSI = Medical out-of-pocket expenses (MOOP) Subtracted from Income MIT = MOOP Included in the Thresholds CMB = "Combined" methods NGA = No Geographic Adjustment for housing costs GA = Geographic Adjustment for housing costs Source: U.S. Census Bureau, Current Population Survey, 2002 Annual Demographic Supplement.
Percent
CMB-GAMIT-GAMSI-GA
CMB - NGAMIT - NGAMSI - NGA
Official
CMB-GAMIT-GAMSI-GA
CMB - NGAMIT - NGAMSI - NGA
Official
0 5 10 15 20 25 30
CMB-GAMIT-GAMSI-GA
CMB - NGAMIT - NGAMSI - NGA
Official
Non-Hispanic White
Black
Hispanic1
the CPS Annual DemographicSupplement, seewww.bls.census.gov/cps/ads/adsmain.htm.
Small Area Income andPoverty Estimates
The Census Bureau also computesimproved (in the sense of havinglower standard errors) annualpoverty data by states, as well asbiennial estimates for counties,based on models using data fromthe CPS, the 1990 decennial census,and administrative records. State-level estimates for 1998 are avail-able on the Internet at: www.cen-sus.gov/hhes/www/saipe.html.Estimates for income year 1999 willbe available later this fall.
Additional Data and Contacts
Detailed tables, historical tables,press releases and briefings, andunpublished data are availableelectronically on the U.S. CensusBureau’s poverty Web site. The Website may be accessed through theCensus Bureau’s home page atwww.census.gov or directly atwww.census.gov/hhes/www/poverty.html. CPS microdata areavailable for down-loading throughthe FERRET system. FERRET maybe accessed by clicking on “AccessTools” on the Census Bureau homepage or by clicking the FERRET linkon the poverty Web site. Technicaldisclosure avoidance methodshave been applied to these CPSmicrodata to prevent disclosure ofindividuals’ identities.
If you have trouble finding povertydata or have questions about
them, you may contact theHousing and Household EconomicStatistics Division statistical infor-mation staff by e-mail at [email protected] or by telephone at 301-763-3242.
Comments
The Census Bureau welcomes thecomments and advice of data andreport users. If you have sugges-tions or comments, please writeto:
John IcelandChief, Poverty and Health Statistics
BranchHousing and Household Economic
Statistics DivisionU.S. Census BureauWashington, DC 20233-8500
20 Poverty in the United States: 2001 U.S. Census Bureau
U.S. Census Bureau Poverty in the United States: 2001 21
Appendix A: TIME SERIES POVERTY ESTIMATES
Table A-1.Poverty Status of People by Family Relationship, Race, and Hispanic Origin: 1959 to 2001[Numbers in thousands. People as of March of the following year]
Year and characteristic
All people People in families Unrelated individuals
Total
Below povertylevel
All familiesFamilies with female
householder, nohusband present
Total
Below povertylevel
Total
Below povertylevel
Total
Below povertylevel
Number Percent Number Percent Number Percent Number Percent
22 Poverty in the United States: 2001 U.S. Census Bureau
Table A-1.Poverty Status of People by Family Relationship, Race, and Hispanic Origin: 1959 to 2001—Con.[Numbers in thousands. People as of March of the following year]
Year and characteristic
All people People in families Unrelated individuals
Total
Below povertylevel
All familiesFamilies with female
householder, nohusband present
Total
Below povertylevel
Total
Below povertylevel
Total
Below povertylevel
Number Percent Number Percent Number Percent Number Percent
U.S. Census Bureau Poverty in the United States: 2001 23
Table A-1.Poverty Status of People by Family Relationship, Race, and Hispanic Origin: 1959 to 2001—Con.[Numbers in thousands. People as of March of the following year]
Year and characteristic
All people People in families Unrelated individuals
Total
Below povertylevel
All familiesFamilies with female
householder, nohusband present
Total
Below povertylevel
Total
Below povertylevel
Total
Below povertylevel
Number Percent Number Percent Number Percent Number Percent
24 Poverty in the United States: 2001 U.S. Census Bureau
Table A-1.Poverty Status of People by Family Relationship, Race, and Hispanic Origin: 1959 to 2001—Con.[Numbers in thousands. People as of March of the following year]
Year and characteristic
All people People in families Unrelated individuals
Total
Below povertylevel
All familiesFamilies with female
householder, nohusband present
Total
Below povertylevel
Total
Below povertylevel
Total
Below povertylevel
Number Percent Number Percent Number Percent Number Percent
U.S. Census Bureau Poverty in the United States: 2001 25
Table A-1.Poverty Status of People by Family Relationship, Race, and Hispanic Origin: 1959 to 2001—Con.[Numbers in thousands. People as of March of the following year]
Year and characteristic
All people People in families Unrelated individuals
Total
Below povertylevel
All familiesFamilies with female
householder, nohusband present
Total
Below povertylevel
Total
Below povertylevel
Total
Below povertylevel
Number Percent Number Percent Number Percent Number Percent
rFor 2000, figures are based on a November 2001 weighting correction. For 1992, figures are based on 1990 census population controls. For 1991, figures are revisedto correct for nine omitted weights from the original March 1992 CPS file. For 1988 and 1987, figures are based on new processing procedures and are also revised toreflect corrections to the files after publication of the 1988 advance report, Money Income and Poverty Status in the United States: 1988, P-60, No. 166.
NA Not available.1Consistent with 2001 data through implementation of Census 2000-based population controls and a 28,000 household sample expansion.2Hispanics may be of any race.
Note: Prior to 1979, people in unrelated subfamilies were included in people in families. Beginning in 1979, people in unrelated subfamilies are included in allpeople but are excluded from people in families.
26 Poverty in the United States: 2001 U.S. Census Bureau
Table A-2.Poverty Status of People by Age, Race, and Hispanic Origin: 1959 to 2001[Numbers in thousands. People as of March of the following year]
Year and characteristic
Under 18 years 18 to 64 years 65 years and over
All people Related children in families
Total
Below povertylevel
Total
Below povertylevel
Total
Below povertylevel
Total
Below povertylevel
Number Percent Number Percent Number Percent Number Percent
U.S. Census Bureau Poverty in the United States: 2001 27
Table A-2.Poverty Status of People by Age, Race, and Hispanic Origin: 1959 to 2001—Con.[Numbers in thousands. People as of March of the following year]
Year and characteristic
Under 18 years 18 to 64 years 65 years and over
All people Related children in families
Total
Below povertylevel
Total
Below povertylevel
Total
Below povertylevel
Total
Below povertylevel
Number Percent Number Percent Number Percent Number Percent
28 Poverty in the United States: 2001 U.S. Census Bureau
Table A-2.Poverty Status of People by Age, Race, and Hispanic Origin: 1959 to 2001—Con.[Numbers in thousands. People as of March of the following year]
Year and characteristic
Under 18 years 18 to 64 years 65 years and over
All people Related children in families
Total
Below povertylevel
Total
Below povertylevel
Total
Below povertylevel
Total
Below povertylevel
Number Percent Number Percent Number Percent Number Percent
U.S. Census Bureau Poverty in the United States: 2001 29
Table A-2.Poverty Status of People by Age, Race, and Hispanic Origin: 1959 to 2001—Con.[Numbers in thousands. People as of March of the following year]
Year and characteristic
Under 18 years 18 to 64 years 65 years and over
All people Related children in families
Total
Below povertylevel
Total
Below povertylevel
Total
Below povertylevel
Total
Below povertylevel
Number Percent Number Percent Number Percent Number Percent
30 Poverty in the United States: 2001 U.S. Census Bureau
Table A-2.Poverty Status of People by Age, Race, and Hispanic Origin: 1959 to 2001—Con.[Numbers in thousands. People as of March of the following year]
Year and characteristic
Under 18 years 18 to 64 years 65 years and over
All people Related children in families
Total
Below povertylevel
Total
Below povertylevel
Total
Below povertylevel
Total
Below povertylevel
Number Percent Number Percent Number Percent Number Percent
rFor 2000, figures are based on a November 2001 weighting correction. For 1992, figures are based on 1990 census population controls. For 1991, figures are revisedto correct for nine omitted weights from the original March 1992 CPS file. For 1988 and 1987, figures are based on new processing procedures and are also revised toreflect corrections to the files after publication of the 1988 advance report, Money Income and Poverty Status in the United States: 1988, P-60, No. 166.
NA Not available.1Consistent with 2001 data through implementation of Census 2000-based population controls and a 28,000 household sample expansion.2Hispanics may be of any race.Note: Prior to 1979, people in unrelated subfamilies were included in people in families. Beginning in 1979, people in unrelated subfamilies are included in all people
but are excluded from people in families.
U.S. Census Bureau Poverty in the United States: 2001 31
Table A-3.Poverty Status of Families by Type of Family: 1959 to 2001[Numbers in thousands. Families as of March of the following year]
Year and characteristic
All families Married-couple families Male householder,no wife present
Female householder,no husband present
Total
Below povertylevel
Total
Below povertylevel
Total
Below povertylevel
Total
Below povertylevel
Number Percent Number Percent Number Percent Number Percent
rFor 2000, figures are based on a November 2001 weighting correction. For 1992, figures are based on 1990 census population controls. For 1991, figures are revisedto correct for nine omitted weights from the original March 1992 CPS file. For 1988 and 1987, figures are based on new processing procedures and are also revised toreflect corrections to the files after publication of the 1988 advance report, Money Income and Poverty Status in the United States: 1988, P-60, No. 166.
NA Not available.1Consistent with 2001 data through implementation of Census 2000-based population controls and a 28,000 household sample expansion.Note: Before 1979, unrelated subfamilies were included in all families. Beginning in 1979, unrelated subfamilies are excluded from all families.
32 Poverty in the United States: 2001 U.S. Census Bureau
INTRODUCTION
The 2001 Current Population Survey(CPS) served as a tool for testing twomethodological changes: (1) a sampleexpansion of the Annual DemographicSupplement (informally called theMarch Supplement) and (2) the intro-duction of new Census 2000-basedpopulation controls. The following sec-tion first describes how the methodolo-gy changed, then discusses how thechanges affected poverty estimates.
Description of CPS SampleExpansion: The Census Bureau testeda 28,000 household expansion in theinterviewed sample for the CPS AnnualDemographic Supplement in 2001.The original sample size of approxi-mately 50,000 interviewed householdsfor the 2001 CPS Annual DemographicSupplement was increased to approxi-mately 78,000. The primary goal ofthe sample expansion was to producemore reliable estimates of low-incomechildren without health insurance forthe State Children’s Health InsuranceProgram (SCHIP) through reduced vari-ances. Although the SCHIP sampleexpansion was specifically targetedtoward producing better children’shealth insurance estimates at the statelevel, other state estimates, as well asnational estimates, improved.
Description of Change inPopulation Controls: In order toproduce estimates for the entire civil-ian noninstitutional population,Census Bureau analysts must weightthe CPS sample results to independ-ent estimates of the population by
sex, age, race, and Hispanic/non-Hispanic categories. These independ-ent estimates are developed by usingcivilian noninstitutional populationcounts from the decennial censusesand projecting them forward to cur-rent years using data on births,deaths, and net migration. Beginningwith the 2002 CPS AnnualDemographic Supplement, the inde-pendent estimates used as controltotals for the CPS are based on civil-ian noninstitutional population bench-marks established by Census 2000.
CPS SAMPLE EXPANSION:EFFECTS ON POVERTY DATA
Table B-1 displays national-levelpoverty data from the original andexpanded CPS samples, weightedwith 1990 census population con-trols. Both samples yielded the samepoverty rate for the United States in2000: 11.3 percent. Poverty rates fordemographic groups were raised orlowered by the expansion with aboutequal frequency, and usually theextent of the change was small inrelation to the change’s standarderror. Statistically significant differ-ences were few and were not concen-trated among any single demographicgroup.1 National-level poverty
estimates from the expanded sample,therefore, look reasonable. At thestate level, the sample expansionraised and lowered poverty rates withroughly equal frequency (see Table B-2). For further analysis aboutthe CPS sample expansion and itseffect on income and poverty esti-mates, see www.bls.census.gov/cps/ads/adsmain.htm.
CENSUS 2000 POPULATIONCONTROLS: EFFECTS ON POVERTY DATA
Weighting the estimates with Census2000 population controls, instead ofthe 1990 census controls used in pre-vious reports, affected poverty rateestimates only minimally—see TableB-3. The poverty rate for the UnitedStates remained at 11.3 percent in2000, after reweighting with the newpopulation controls. Among demo-graphic groups, however, thereweighting tended to raise povertyrates more often than it loweredthem. Most differences between thetwo sets of estimates were 0.2 per-centage points or less, and occurredapparently randomly across demo-graphic groups. For further informa-tion about CPS weighting procedures,see Technical Paper 63RV, available atwww.bls.census.gov/cps/tp/tp63.htm.
Appendix B: SAMPLE EXPANSION AND INTRODUCTION OF CENSUS 2000-BASED POPULATION CONTROLS
1 Usually when two estimates are “significantlydifferent” it means that the difference was largeenough, in relation to the difference’s own standarderror, for us to infer that the difference is “real,” ormore accurately, that there was a less than 10 per-cent chance that the difference merely came fromsampling variation. In Table B-1, however, both setsof data estimate the same populations in the sameperiod. Therefore, “significant” here means that wewould have inferred that the estimates came fromdifferent populations, if we did not already knowthey were the same.
U.S. Census Bureau Poverty in the United States: 2001 33
Table B-1.Effect of CPS Sample Expansion on Estimates of People and Familiesin Poverty: 2000(Numbers in thousands)
Characteristic
Expanded sample Original sample Difference (expanded sampleminus original sample)1
- Represents zero.* Statistically significant at the 90-percent confidence level.For an explanation of confidence intervals (C.I.), see ‘‘Standard errors and their use’’ at www.census.gov/hhes/poverty/poverty01/pov01src.pdf.1As a result of rounding, some differences may appear to be slightly higher or lower than the differences between the reported rates.2Data for American Indians and Alaska Natives are not shown separately in this table because of the small sample of that population.3Hispanics may be of any race.Source: U.S. Census Bureau, Current Population Survey, 2001 Annual Demographic Supplement.
34 Poverty in the United States: 2001 U.S. Census Bureau
Table B-2.Number of Poor and Poverty Rate by State: 2000(Numbers in thousands)
State
Expanded sample Original sample Difference (expanded sampleminus original sample)1
Number
90-percentC.I. (±) Percent
90-percentC.I. (±) Number
90-percentC.I. (±) Percent
90-percentC.I. (±) Number
90-percentC.I. (±) Percent
90-percentC.I. (±)
United States . . . 31,089 628 11.3 0.2 31,054 879 11.3 0.3 35 533 - 0.2
-Represents zero.*Statistically significant at the 90-percent confidence level. For explanation of confidence intervals (C.I.), see ‘‘Standard errors and their use’’ at
www.census.gov/hhes/poverty/poverty01/pov01src.pdf.1As a result of rounding, some differences may appear to be higher or lower than the differences between the reported rates.
Source: U.S. Census Bureau, Current Population Survey, 2001 Annual Demographic Supplement.
U.S. Census Bureau Poverty in the United States: 2001 35
Table B-3.People and Families in Poverty in 2000, Estimated With Census 2000-Based and1990-Based Population Controls(Numbers in thousands)
- Represents zero.1As a result of rounding, some differences may appear to be slightly higher or lower than the differences between the reported rates.2Data for American Indians and Alaska Natives are not shown separately in this table because of the small sample of that population.3Hispanics may be of any race.Source: U.S. Census Bureau, Current Population Survey, 2001 Annual Demographic Supplement.