Poverty Threshold
Alternatives / Choices
Thesia I. Garner
Senior Research Economist
Division of Price and Index Number Research
Brookings/Census Bureau Conference on Improved
Poverty Measurement
October 20, 2009 (revised October 21, 2009)
www.bls.gov
The materials presented are based on research and do not reflect an official BLS position or the opinions of other BLS employees.
This paper builds on research
since 2009 JSM: Choices

Threshold and Assumptions/Choices/Alternatives







Concepts
Basic needs
Multipliers
Distribution
Reference unit
Updating
Use of equivalence scales

JSM 2009 restricted to 2007 thresholds

Current research
 Trends over time
 Underlying reasons for movements
–
–
Prices
Impact of 2007 CE questionnaire and processing changes
National Academy of Sciences
Recommendations

2.1-2.4 A poverty threshold with which to initiate a
new series of official U.S. poverty statistics should be
derived from Consumer Expenditure Survey [CE] data
for a reference family of four persons (two adults and
two children).






Basic bundle
Percentage of median expenditures
Multiplier for other needs
Updating
Real growth in consumption
3.1 Adjustments
The NAS recommended a Procedure.
3
Concepts Used to Define Thresholds

Spending
How much does it costs, or does one need to
spend, to meet basic needs?
But … for some needs, such as housing, spending does not
equal consumption (e.g., subsidized, own)

Consumption
How much does it costs, or what is the
consumption value, to meet basic needs?
Thresholds Defined in Terms of…
Food, Clothing, Shelter, and Utilities
FCSU

Spending
 Official poverty threshold
 FCSU without principal payments
 FCSU with principal payments

Alternative: Spending and Consumption
 Expenditures with adjusted for select consumption (e.g.,
rental equivalence for owner-occupants, market value of
subsidized and rent-control housing)
Caution….

For a more complete consumption based
threshold, would also need data on …
Rent controlled or government subsidized housing
School breakfast, school lunch, WIC, energy
assistance, elder coupons
Medical consumption not financed out-of-pocket
Goods and services received as gifts (“net” value
of gifts)

And values would be added to resources for
consistency
Threshold Equations
Have used in the past
(1.15* PL * Median) + (1.25* PH * Median)
2
Used the midpoints of the percentages and multipliers
Recent proposal
1.20*33rd percentile
Three-parameter Equivalence
Scale
one and two adults
[adults]0.5
single parents
[adults + 0.8*1st child + (0.5*children-1)]0.7
all other families
[adults + 0.5*children]0.7
Basic Threshold and
Alternatives

Basic
 Median FCSU expenditures
– With principal payments
 Percentages of median
– 0.78 and 0.83
 Multipliers
– 1.15 and 1.25

Alternatives







Without principal payments
Consumption - owners’ shelter
Quarterly data independence
Census family versus Consumer unit
Median versus 33rd percentile
Annual versus quarterly price adjustment
Housing payment status
9
2007 FCSU Thresholds:
Basic Assumptions
$30,000
$25,804
$25,000
$23,569
$20,000
Difference=$2,235
$15,000
$10,000
$5,000
$0
CF-no pp
CF BASE pp
“CF” refers to Census family composed of 2 adults and 2 children
2007 FCSU Thresholds:
Spending vs. Consumption
$30,000
$27,512
$27,956
$25,804
$25,000
$20,000
Difference;$1,718
$15,000
$10,000
$5,000
$0
CF BASE pp
CF REQ fixed pc
CF REQ vary PC
2007 FCSU Thresholds: Independent
Quarters vs. 4 Complete Quarters
NOTE: difference due to summation
and price adjustment at micro level
$30,000
$26,183
$25,804
$26,878
$25,000
NOTE: difference due to summation
and price adjustment at micro level
$20,000
$15,000
$10,000
$5,000
$0
CF BASE pp
CF 4 comp ind
NOTE: difference due
to demographics
CF 4 comp not ind
2007 Thresholds: Census Families vs.
Consumer Units and Price Adjustment
$30,000
$25,804
$25,151
$25,865
$25,179
$25,000
$20,000
$15,000
$10,000
$5,000
$0
CF BASE pp
CU 2+2, Median,
ACPI
Difference: Census Family vs CU
CU 2+2, Median,
QCPI
CU 2+2, 33, QCPI
Difference: Median vs 33rd percentile
2007 Thresholds for Consumer Units
by Housing Payment Status
$30,000
$26,732
$25,179
$25,000
$20,000
$14,833
$15,000
$10,000
$5,000
$0
CU 2+2, 33, QCPI
with mort, rent pay
no mort, no rent pay
Percentage Increase in FCSU Expenditures
(with principal payments, quarterly price
adjusted) from Previous Year
8.00%
7.23%
7.11%
7.00%
6.03%
6.00%
5.10%
5.00%
4.00%
3.00%
2.00%
1.00%
0.00%
2007
2008
FCSU median
FCSU 33rd PC
NOTE: changes at the median less sensitive than changes
at the 33rd percentile
Trends Over Time
Trends 1996-2008
 Changes related to

Prices
Overall Spending
Income

Changes related to CE questionnaire
design and processing
16
CE Interview-Based FCSU Thresholds Using Medians with and
without Principal Payments
$35,000
$30,000
$27,043
$25,804
$24,069
$25,000
$20,000
$16,749
$15,000
$17,109
$15,742
$16,018
1996
1997
$17,697
$16,544
$18,196
$17,044
$19,097
$17,906
$19,935
$18,765
$20,757
$21,218
$21,895
$22,769
$24,755
$23,569
$21,902
$19,376
$19,782
$20,138
2003
2004
$20,776
$10,000
$5,000
$0
1998
1999
2000
2001
FCSU (w pp) threshold
NOTE: historically consistent series
2002
FCSU (wo pp) threshold
2005
2006
2007
2008
Percentage Changes in FCSU Median Expenditures
from Previous Years
8.00%
7.61%
7.21%
7.00%
6.00%
5.71%
5.42%
5.06%
4.96%
5.00%
5.03%
4.80%
4.80%
4.39%
4.12%
3.99%
4.00%
3.44%
3.29%
3.25%
3.00%
3.17%
2.22%
2.10%
2.15%
2.00%
3.19%
3.02%
2.82%
1.80%
1.75%
1.00%
0.00%
1997
1998
1999
2000
2001
2002
2003
2004
2005
% increase in FCSU median expenditures with principal payments
% increase FCSU median expenditures without principal payments
2006
2007
2008
Annual* CPIs
2.60
2.40
2.20
SHEL_CPI
2.00
CPI
1.80
FOOD_CPI
UTIL_CPI
1.60
APP_CPI
1.40
1.20
1.00
Census Family Mean Income, Expenditures, and FCSU
(without principal payments) Thresholds within the 47th to
53rd Range of the FCSU Expenditure Distribution
$80,000
$70,000
Before Tax Money Income
$60,000
Ztotal (range)
$50,000
FCSU Median (wo pp)
FCSU wo pp Threshold
$40,000
Official Poverty Threshold
$30,000
$20,000
$10,000
2004
2005
2006
2007
2008
Changes in CE Interview Questionnaire
in 2007Q2: Affect on Thresholds


Mortgage edits in processing
Change in question wording for food away
from home expenditures
2007 Improvements
to Mortgage Edits

Prior to 2007Q2
Before 2006, Field Representatives (FRs) could not
report a mortgage as an “interest only” mortgage
In 2006, FRs given the ability to report mortgage
as “interest only”, but the information not used in
processing

Beginning with 2007Q2
Questionnaire revised to better collect interest
only mortgages
Mortgage edit processing began utilizing “interest
only” designation
2007 Improvements to Food Away
From Home Data Collection

Prior to 2007Q2

Beginning with 2007Q2
CE Interview survey collected usual monthly
expenditures for food away from home
Cognitive testing indicated collecting usual weekly
amount results in more accurate data
Questionnaire revised to collect usual weekly
expenditures for food away from home
Interview data for food away for 2007 more
comparable with the Diary data for 2007 as
compared to past years
Census Family Median FCSU Expenditures (without principal
payments) and in 47th to 53rd Percentiles of FCSU Expenditure
Distribution (in 2007 dollars using 2008 qtr 1)
$30,000
$25,000
$20,000
$15,000
$10,000
$5,000
$0
2007 (2005Q2-2006Q1)
2007 (2006Q2-2007Q1)
2007 (2007Q2-2008Q1)
Median Total FCSU (wo)
Mean Total FCSU (wo)
Food Away from Home (wo)
Mortgage Interest Payments (wo)
Reference Family Median FCSU Expenditures (without principal
payments) and Mean of Expenditures in 47th to 53rd Percentiles of
FCSU Expenditure Distribution (in 2008 dollars using 2009 qtr 1)
$30,000
$25,000
$20,000
$15,000
$10,000
$5,000
$0
2008 (2006Q2-2007Q1)
2008 (2007Q2-2008Q1)
2008 (2008Q2-2009Q1)
Median Total FCSU (wo)
Mean Total FCSU (wo)
Food Away from Home (wo)
Mortgage Interest Payments (wo)
Distribution of Census Reference Families within the
47th to 53rd FCSU Expenditure Distribution (without Principal
Payments) by Threshold Year
80.00%
70.00%
60.00%
50.00%
40.00%
30.00%
20.00%
10.00%
0.00%
Renter-Public Housing
Renter-Govt Pays
Own with Mortgage
Own without Mortgage
NOTE: in the 2007Q2-2008Q1 period reference families had different
characteristics than in other periods: fewer owners with mortgages
and more non-subsidized renters
Rent
Reference Family FCSU Median Expenditures (with and without out principal
payments) and Official Poverty Thresholds
$30,000
$30,000
$25,000
$25,000
$20,000
$20,000
in 2006$
in 2007$
in 2008$
$15,000
$15,000
$10,000
$10,000
$5,000
$5,000
$0
$0
Median FCSU (w pp)
NOTE: microdata adjusted by CPI
Median FCSU (wo pp)
Official Poverty Threshold
Changes in FCSU Median Expenditures from Previous Years Using
Different Updating
8.00%
2006Q-2007Q1
data updated to threshold year
7.00%
6.00%
5.00%
CPI price
adjustment to 2006
threshold
4.00%
3.00%
2.00%
1.00%
0.00%
% increase in FCSU median expenditures with principal payments
% increase in FCSU median expenditures without principal payments
Note: expenditures were on the rise during the 2006 period even before
Changes in CE design and processing.
12 quarters
of data updated
to threshold year
Summary and Conclusions

Endogenous choices that affect thresholds

Exogenous factors affecting thresholds

What can be done
Choices and Impact on
Thresholds

Higher
 With mortgage principal
payments
 Consumption
 Median
 Annual updating of
quarterly data
 Census family
 Quarterly CE data not
independent
 Owners with mortgages
and rents with payments

Lower
 Without mortgage principal
payments
 Spending
 33rd percentile
 Quarterly updating of
quarterly data
 Consumer unit
 Quarterly CE data
independent
 Owners without mortgages
Exogenous Factors



Underlying spending patterns
Underlying prices
CE questionnaire design and processing

RESULT: using 3 years of data smoothes
impact of exogenous factors
30
Contact Information
Thesia I. Garner
Senior Research Economist
Division of Price and Index Number
Research/OPLC
202-691-6576
[email protected]
www.bls.gov
Extra Information
32
www.bls.gov
FCSU Threshold: Out-ofpocket Spending
Food (includes Food Stamps value)
Clothing
Utilities (includes telephone)
For renters, shelter expenditures
For homeowners, non-vacation shelter
expenditures that include
–
–
–
–
–
Mortgage interest payments
Mortgage principal repayments (“pp”)
Prepayment penalties
Property taxes
Maintenance, repairs, insurance and other related
expenditures
FCSU Threshold Accounting
for Consumption

Out-of-pocket spending on
–
–
–
–
Food
Clothing
Utilities (includes telephone)
Rent of renters
Food as pay
 Rent as pay
 Rental equivalence of owners

NOTE: from 2009 JSM
Census Family
Consumer unit
CU 2+2, 33, QCPI, 2007
CU 2+2, 33, QCPI, 2006
$28,000
CU 2+2, 33, QCPI, 2005
Based on Medians
CU 2+2, 33, QCPI
CU 2+2, QCPI
CU 2+2, ACPI
CF equiv
CF 4 comp indp
CF 4 comp not ind
CF REQ vary PC
CF REQ fixed pc
CF 2007
$29,000
CF 2006
CF 2005
CF 1.20
CF vary pc
CF BASE pp
CF-no pp
2007 FCSU Thresholds Based
on Alternative Assumptions
Based on
33rd percentile
$27,000
$26,000
$25,000
$24,000
$23,000
$22,000
$21,000
$20,000
Average Food Away from Home and Mortgage Interest
Payment Expenditures in the 47 to 53 Percentile Range of
FCSU Expenditures (without Principal Payments)
$7,000
$6,000
$5,000
$4,000
$3,000
2006
2007
2008
Mortgage
Interest
Payments (wo)
Food Away from
Home (wo)
$2,000
$1,000
$0
Average Food Away from Home, Mortgage Interest and
Principal Payment Expenditures in the 47th to 53 Percentiles of
FCSU Expenditures (with mortgage Principal Payments)
$7,000
$6,000
$5,000
$4,000
$3,000
$2,000
$1,000
$0
2006
2007
2008
Mortgage
Interest
Payments (wpp)
Food Away from
Home (wpp)
Mortgage
Principal
Payments
Weighted Percentage Distributions of Reference Families in the 47th to 53rd Percentile Distributions of FCSU
Expenditures by threshold Year (without principal payments) by Demographic Characteristics
Data collection
period
Own
Food Renter-Public Renter-Govt
Own with
without
Other
Housing
Pays
Mortgage Mortgage Own Other
Rent Tenure
Welfare Stamps
2006 (2004Q22005Q1)
0.00%
0.00%
3.34%
0.00%
76.94%
4.00%
0.00% 19.06% 0.00%
2006 (2005Q22006Q1)
0.00%
0.00%
3.83%
0.00%
79.80%
4.34%
0.00% 15.86% 0.00%
2006 (2006Q22007Q1)
0.00%
0.00%
0.00%
4.41%
80.56%
3.94%
0.00% 15.50% 0.00%
2007 (2005Q22006Q1)
2007 (2006Q22007Q1)
2007 (2007Q22008Q1)
2008 (2006Q22007Q1)
2008 (2007Q22008Q1)
2008 (2008Q22009Q1)
2006 (2004Q22007Q1)
2007 (2005Q22008Q1)
2008 (2006Q22009Q1)
0.00%
0.00%
3.83%
0.00%
79.80%
4.34%
0.00% 15.86%
0.00%
0.53%
2.34%
0.00%
3.85%
78.96%
3.19%
0.00% 17.86%
0.00%
0.63%
0.63%
0.00%
5.28%
66.45%
2.70%
0.00% 30.85%
0.00%
0.00%
0.00%
0.00%
3.85%
78.96%
3.19%
0.00% 17.86%
0.00%
0.00%
0.00%
0.00%
3.37%
66.63%
2.72%
0.00% 30.65%
0.00%
0.00%
0.00%
0.00%
0.00%
76.27%
2.41%
0.00% 21.31%
0.00%
0.00%
0.00%
2.33%
1.33%
77.62%
4.74%
0.00% 17.64%
0.00%
0.87%
2.50%
1.85%
1.24%
78.24%
2.71%
0.00% 19.05%
0.00%
0.00%
0.00%
0.00%
0.92%
77.51%
2.26%
0.00% 19.90%
0.33%