Alternative Measures of Income Poverty and the Antipoverty

Alternative Measures of Income Poverty and the Anti-poverty Effects of Taxes and Transfers Presentation to University of Maryland-AEI Seminar on Poverty Measurement Daniel H. Weinberg, Chief Economist May 10, 2005

Request from Seminar Planning Group Besharov and Green (AEI) worked with seminar planning group to develop alternative measures of income poverty and methods to gauge the impact of taxes and transfers They asked the Census Bureau about which measures were feasible to compute with existing resources before the next seminar 2

Five Decisions Must be Made for Any Poverty Measure • Income concept • Unit of analysis • Equivalence scale • Inflation adjustment • Data source 3

1. Five Income Variants • Money income (used in official measure) • Pre-tax pre-transfer: money income excluding meanstested cash transfers • Pre-tax pre-transfer plus return to home equity • Post-tax, post-transfer: money income plus realized capital gains, plus non-cash transfers, plus EITC, minus income and payroll taxes • Post-tax, post-transfer plus return to home equity minus property taxes 4

Income: Additional information Non-means tested transfers like Social Security included in all measures. Work expenses are not subtracted from income, nor are medical out-of-pocket costs. John Coder (Sentier Research) computed other measures that include some unreported transfer income; these will be discussed at the end of the presentation. 5

2. Two Units of Analysis • Family • Household Tabulations presented here do not include unrelated children under 15 in families or in households 3. Two Thresholds (Equivalence Scales) • Official • Experimental threeparameter 3 -parameter scale 1 adult: 1. 00; 2 adults: 1. 41; Single parents: [1. 8+0. 5*(children-1)]0. 7 Other families: [Adults+0. 5*children]0. 7 6

4. Two Ways to 5. Data Source Adjust Thresholds for Inflation • Published CPI-U: 1963 -1982 CPI-U 1983 -1999 CPI-U-X 1 2000 -2002 CPI-U-RS • CPI-U-RS: 1963 -1977 CPI-U, 1978 -2002 CPI-U-RS 2003 Current Population Survey Annual Social and Economic Supplement (CPS ASEC) 7

Measurement Issues Unable to partition all income sources into means -tested and non-means tested parts. Improved methods for valuing noncash benefits like housing have not yet been implemented. 8

Warnings This is a work in progress. No significance tests have been performed. Because all these estimates are from the same dataset, conventional measures of sampling error do not apply. The Office of Management and Budget is responsible for any changes to the official measure of poverty. 9

Poverty Thresholds, 2002: Examples Note: The 3 -parameter (CPI) threshold for a 2 -adult 2 -child family is equal to the official threshold for that family by design; others are lower. 10

Percent in Poverty : 2002 Family Measures using Official Thresholds and CPI Threshold and Inflation Adjustment Money income Pre-transfer +HE* 12. 1 12. 8 11. 6 Family Measures Official-CPI 3 parameter-RS Household Measures Official-CPI 3 parameter-RS HE*=imputed return to home equity 11

Percent in Poverty : 2002 Money Income Measures for Families & Households Threshold and Inflation Adjustment Money income Pre-transfer, pre-tax +HE* Family Measures Official-CPI 3 parameter-RS 12. 1 12. 0 10. 0 Household Measures Official-CPI 3 parameter-CPI 10. 5 10. 6 3 parameter-RS 8. 6 HE*=imputed return to home equity 12

Percent in Poverty : 2002 Money Income and Pre-Transfer Measures Threshold and Inflation Adjustment Money income Pre-transfer, pre-tax +HE* 12. 1 12. 0 10. 0 12. 8 10. 9 11. 6 9. 9 3 parameter-CPI 10. 5 10. 6 11. 2 11. 4 10. 1 10. 2 3 parameter-RS 8. 6 9. 5 8. 6 Family Measures Official-CPI 3 parameter-RS Household Measures Official-CPI HE*=imputed return to home equity 13

Measuring the Distributional Effect: The “Poverty Shares Change Index” Example: People in poverty in California Official measure: 4. 605 of 34. 570 million (13. 32%) Pre-transfer pre-tax income including imputed return to home equity, for people in households, using the 3 parameter thresholds adjusted for inflation using the CPIU-RS [Pre. T+HE*-H-3 p-RS] measure: 3. 305 of 24. 418 million (13. 54%) Ratio=13. 54/13. 32=poverty shares change index of 101. 6 Interpretation: People in poverty in California are a larger percentage of all those in poverty when the Pre. T+HE*-H-3 p-RS measure is used than when the official poverty measure is used (their share is 1. 6% larger) 14

[Official=100] 15

[Official=100] 16

Pre-transfer Poverty Shares by State: Largest Changes (versus share of official poverty) Families-3 parameter-RS • share increased 10% or more: – DC, Connecticut, Hawaii (115. 4) • share decreased 10% or more: – Missouri, Idaho (88. 6) Households-3 parameter-RS • share increased 10% or more: – South Carolina, Alabama, Connecticut, Louisiana, Mississippi, DC, West Virginia (118. 8) • share decreased 10% or more: – Wisconsin, Nevada, Colorado, Delaware, Alaska, Idaho, Minnesota (75. 7) 17

[Official=100] HE*=imputed return to home equity 18

Percent in Poverty : 2002 Addition of Post-transfer, Post-tax Measures Threshold & Inflation Adj Money income Pre-tsfr pre-tax 12. 1 12. 8 Pre-transfer Post-tsfr Post-transfer pre-tax +HE* post-tax +HE Familes Official-CPI 11. 6 10. 2 9. 4 3 parameter-CPI 3 parameter-RS Households Official-CPI 3 parameter-RS 19 HE*=imputed return to home equity HE=imputed return to home equity minus property taxes

Percent in Poverty : 2002 Official and 29 alternative measures Threshold & Inflation Adj Money Pre-tsfr Pre-transfer Post-tsfr Post-transfer income pre-tax +HE* post-tax +HE Families Official-CPI 12. 1 12. 0 10. 0 12. 8 10. 9 11. 6 9. 9 10. 2 9. 9 7. 9 9. 4 9. 1 7. 3 3 parameter-CPI 10. 5 10. 6 11. 2 11. 4 10. 1 10. 2 8. 7 8. 6 7. 9 7. 8 3 parameter-RS 8. 6 9. 5 8. 6 6. 0 3 parameter-CPI 3 parameter-RS Households Official-CPI (Post. T and Post. T+HE household estimates from Sentier Research. ) 20 HE*=imputed return to home equity HE=imputed return to home equity minus property taxes

HE*=imputed return to home equity 21 HE=imputed return to home equity minus property taxes

HE*=imputed return to home equity 22 HE=imputed return to home equity minus property taxes

HE*=imputed return to home equity 23 HE=imputed return to home equity minus property taxes

Post-transfer Poverty Shares by State: Largest Changes (versus share of official poverty) Families-3 parameter-RS • share increased 10% or more: – Nebraska, Indiana, Illinois, Ohio, New Mexico (113. 2) • share decreased 10% or more: – Nevada, Mass. , Maine, Alaska, Kansas, Connecticut (83. 1) Households-3 parameter-RS • share increased 10% or more: – Ohio, N Carolina, Indiana, Tennessee, S Carolina, Illinois, New Mexico, Louisiana (122. 6) • share decreased 10% or more: – Kansas, Hawaii, Utah, Wash. , Nevada, Idaho, Mass. , Maine, Vermont, Delaware, Conn. , Minnesota (70. 1) 24

Extent of CPS Unreported Income (from October seminar) Ruser, Pilot, and Nelson (2004) study compared BEA State Personal Income (SPI) with CPS ASEC money income for 2001 BEA: $8. 670 trillion CPS: $6. 446 trillion Difference: $2. 233 trillion However, adjustments to BEA SPI are needed to derive a concept consistent with CPS ASEC. Remaining gap is $806 billion. 25

Key Areas of CPS Response Error 1. Wages and Salaries: 3 percent underreporting accounts for $158 billion of the gap 2. Self-Employment income: 48 percent underreporting accounts for $302 billion of the gap 3. Interest and Dividends: 32 percent underreporting accounts for $132 billion of the gap 4. Transfer Programs: 23 percent underreporting accounts for $199 billion of the gap 26

Transfer Income Imputation Census Bureau analysts are unfamiliar with the adjustment methods used in the Urban Institute’s TRIM model so these adjustments were implemented by Sentier Research using the CPS ASEC public use file. Imputations were done for underreported Temporary Assistance for Needy Families, Food Stamps, Supplemental Security Income only. We also have a concern that undercoverage of the lowincome population by the survey might affect the reliability of such an imputation. 27

Effects of Adjusting for Unreported TANF, SSI, FS Income on Household Poverty Rate Measures: 2002 [Official poverty rate= 12. 1 percent] All estimates from Sentier Research. TANF=Temporary Assistance for Needy Families; FS=Food Stamps, SSI=Supplemental Security Income 28

All estimates from Sentier Research. 29

All estimates from Sentier Research. 30