Methodological Improvements for CBO’s Analysis of the Distribution of Household Income

Congressional Budget Office
Methodological Improvements for
CBO’s Analysis of the
Distribution of Household Income
Distributional Tax Analysis Conference
Washington, DC
September 13, 2017
Kevin Perese and Bilal Habib
Tax Analysis Division
As developmental work for analysis for the Congress, the information in this presentation is preliminary and is being
circulated to stimulate discussion and critical comment.

2CONGRESSIONAL BUDGET OFFICE
Two major methodology changes in the
forthcoming report on the distribution of
household income:
• New income measure for ranking households and
calculating average federal tax rates
• Correction for underreporting of means-tested
transfers in household survey data

Why?
Distributional Analyses Have
Historically Been Tax-Centric

• Everyone pays taxes
(either directly or indirectly).
• The tax system has explicit
progressive/redistributive properties.
• There are high-quality tax data.
• There is a lot of theoretical work on tax
incidence in the economics literature.
Distributional Analyses Have
Historically Been Tax-Centric

But there’s more to
government than just taxes.

Increasingly, the distinction between tax
and spending policies is more about the
legislative process and less about the
impact on households.

CBO will use a new framework to analyze
how means-tested transfers and federal
taxes jointly affect the distribution of
household income.

CBO’s Previous
Distributional Framework

Market
Income
Labor Income: Wages and salaries
Business Income: Income from businesses and farms
Capital Income: Capital gains, interest, and dividends
Other Income: Mainly retirement income
CBO’s Previous Distributional Framework

Market
Income
Cash and In-Kind
Government Transfers
+
Social Security
Medicare
Medicaid
Supplemental Nutrition Assistance Program
Unemployment Insurance
Housing Assistance

Market
Income
Before-Tax
Income
=+
Cash and In-Kind

Market
Income
Before-Tax
Income
‒
=+
Individual Income Taxes
Payroll Taxes
Corporate Income Taxes
Excise Taxes
Federal Taxes
Cash and In-Kind

Market
Income
Before-Tax
Income
After-Tax
Income
‒
=+
=
Federal Taxes
Cash and In-Kind

Market
Income
Before-Tax
Income
After-Tax
Income
Federal Taxes
‒
=+
=
Used as the basis for ranking
households and as the denominator
in average tax rate calculations.
Cash and In-Kind

Strengths Shortcomings
• Before-tax income, a broad
income measure, is a proxy
for both overall economic
well-being and ability to pay
tax liabilities.
• Before-tax income is
therefore an appropriate
denominator for calculating
average federal tax rates.
• The framework is tax-centric,
so it is less suitable for
distributional analysis of
government transfers.
• Therefore, the redistributive
properties of transfers and
taxes are not treated equally.

CBO’s New
Distributional Framework

Market
Income
CBO’s New Distributional Framework
Labor Income: Wages and salaries
Business Income: From businesses and farms
Capital Income: Capital gains, interest, and dividends
Other Income: Mainly retirement Income

Market
Income
Social Insurance
Benefits
+
Social Security
Medicare
Unemployment Insurance

Market
Income
Social Insurance
Benefits
=+
Mkt Income
Plus Social
Insurance

Market
Income
Social Insurance
Benefits
=
Means-Tested
Transfers
++
Medicaid
Supplemental Nutrition
Assistance Program
Housing Assistance, etc.
Mkt Income
Plus Social
Insurance

Market
Income
Social Insurance
Benefits
Mkt Income
Plus Social
Insurance
=
Federal Taxes
‒
Means-Tested
Transfers
++
Payroll Taxes
Excise Taxes

Market
Income
Social Insurance
Benefits
Mkt Income
Plus Social
Insurance
=
Federal Taxes
‒
Means-Tested
Transfers
++
Payroll Taxes
Excise Taxes
Taxes Before
Refundable Credits
Refundable Tax Credits
(significant portions
classified as budgetary
outlays)

Market
Income
After-Tax
Income
Social Insurance
Benefits
=
=
‒
Federal Taxes
Before Refundable
Credits
++
Mkt Income
Plus Social
Insurance
Means-Tested
Transfers
Refundable
Tax Credits

Market
Income
After-Tax
Income
Social Insurance
Benefits
=
=
‒
Federal Taxes
Before Refundable
Credits
++
Mkt Income
Plus Social
Insurance
Means-Tested
Transfers
Refundable
Tax Credits
Used as the basis for ranking households
and as the denominator in average
tax and transfer rate calculations.

Strengths Shortcomings
• The framework allows
analysts to calculate means-
tested transfer rates, tax
rates, and net tax and
transfer rates.
• It recognizes that life-cycle
income patterns make social
insurance benefits difficult to
analyze in a cross-sectional
framework.
• The measure of income—
market income plus social
insurance benefits—does not
fully represent one’s ability
to pay tax liabilities.
• Social insurance programs
have some redistributive
effects that the framework
does not capture.

Before-Tax
Income
Framework
Average Federal Tax Rates
Market Income
Plus
Social Insurance
Benefits
Framework
Versus

-5
0
5
10
15
20
25
30
1980 1985 1990 1995 2000 2005 2010
Lowest Quintile
Middle Quintile
Fourth Quintile
Highest Quintile
Second Quintile
Percent
Before-Tax Income
as Denominator and Household Ranking Measure
Average Federal Tax Rates by Income Group, 1979 to 2013

-5
0
5
10
15
20
25
30
1980 1985 1990 1995 2000 2005 2010
Lowest Quintile
Middle Quintile
Fourth Quintile
Highest Quintile
Second Quintile
Before-Tax Income
Market Income Plus Social Insurance Benefits
Percent
Average Federal Tax Rates by Income Group, 1979 to 2013

Differences between the frameworks are
primarily at the bottom of the income
distribution.
Differences can be decomposed into two,
partially offsetting, steps:
1. New denominator, and
2. New household rankings.

Average Federal Tax Rates, by Select Income Groups, 1979 to 2013
-5
0
5
10
15
20
1980 1985 1990 1995 2000 2005 2010
Lowest Quintile
Second Quintile
Percent
Before-Tax Income

-5
0
5
10
15
20
1980 1985 1990 1995 2000 2005 2010
Market Income Plus Social Insurance Benefits as Denominator and
Before-Tax Income as Household Ranking Measure
Lowest Quintile
Second Quintile
Before-Tax Income
Percent

-5
0
5
10
15
20
1980 1985 1990 1995 2000 2005 2010
Market Income Plus Social Insurance Benefits as Denominator and
Before-Tax Income as Household Ranking Measure
Lowest Quintile
Second Quintile
Percent

-5
0
5
10
15
20
1980 1985 1990 1995 2000 2005 2010
Lowest Quintile
Second Quintile
Before-Tax Income
Percent

Correcting for Underreporting of
A Regression-Based Approach
With Preliminary Results

Explicit analysis of government transfers
requires a more thorough accounting of
transfer income.
CBO’s tax model draws its transfer income
data from the Annual Social and Economic
Supplement of the Current Population
Survey (CPS). The analysis in this section is
conducted using only CPS data, before
merging with administrative tax data.

Underreporting of transfer income in the
CPS has increased over time, as is well
documented in Wheaton (2008), Meyer,
Mok, and Sullivan (2009), and Moffitt and
Scholz (2009).
As a result, CPS-based analyses are likely to
understate income growth at the bottom
of the distribution and the role of transfers
in reducing income inequality.

CBO’s goal is to obtain a more complete
(although partially imputed) accounting of
income from government transfers in the
CPS with enough precision for quintile-
level distributional analysis.

This analysis is focused on three of the
largest means-tested transfer programs—
Medicaid, the Supplemental Nutrition
Assistance Program (SNAP), and
Supplemental Security Income (SSI)—and
the two largest social insurance benefits—
Social Security and Medicare.

Means-Tested Transfers

Reporting Rates in the CPS for Means-Tested Transfer
Programs, 1979 to 2013
CPS Recipients as a Percentage of Administrative Recipients
0
20
40
60
80
100
1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013
73
66
53
Medicaid
SSI
SNAP

Researchers typically use one of three
methods to correct for underreporting:
• Administrative matching,
• Rules-based simulation, or
• Regression-based estimation.

Administrative matching offers near-perfect
accounting, but administrative microdata are not
widely available.
Examples: Davern et al. (2009); Meyer and Sullivan (2008).
Rules-based simulation offers precise estimates at
the micro level but requires a significant research
investment.
Example: The Urban Institute’s Transfer Income Model—see
Zedlewski and Giannarelli (2015).

CBO is opting for a regression-based estimation
approach, which is tractable for analyzing multiple
programs over many years but is less precise at the
micro level.
That approach allows CBO to perform quintile-level
distributional analyses of various transfer programs
from 1979 onward.
Example: Moffitt and Scholz (2009).

CBO’s preliminary regression-based
estimation has three steps:
1. Use reported data to estimate the
probability of receipt for all units.
2. Impute transfer receipt based on
estimated probabilities.
3. Assign transfer income to recipients.

Step 1. Predicted probabilities are
estimated using a probit model with CPS-
reported receipt as the dependent
variable.
Independent variables comprise individual
and household characteristics based on
program rules and other factors associated
with program participation.

Individual characteristics include age,
race, education, labor force status,
disability, marital status, and receipt of
other means-tested transfers.
Household/family characteristics include
income (as a percentage of the federal
poverty level), income composition,
household size and structure, and
geographic location.

Step 2. Transfer receipt is imputed to
nonreporters with the highest probability
of receipt until the administrative total is
reached. This process is repeated to match
the targets for each category (e.g., children,
seniors).

Percentage of Adults Receiving Benefits
0
10
20
30
40
50
60
0 25 50 75 100 125 150 175 200 225 250 275 300
Annual Family Income as a Percentage of the Federal Poverty Level
Medicaid Recipience Rates by Income, 2010
CPS (Reported Only)
CBO (Imputed Plus
Reported)
The Urban Institute’s
Transfer Income
Model

Percentage of Individuals Receiving Benefits
SSI Recipience Rates by Income, 2010
0
5
10
15
20
25
30
0 25 50 75 100 125 150 175 200 225 250 275 300
Annual Family Income as a Percentage of the Federal Poverty Level
CPS (Reported Only)
CBO (Imputed
Plus Reported)
The Urban
Institute’s
Transfer Income
Model

Distribution of SNAP Recipients by Annual Household Income
as a Percentage of the Federal Poverty Level, 1979 to 2013
Percentage of Recipients
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
1979 1984 1989 1994 1999 2004 2009
CPS (Reported Only)
1979 1984 1989 1994 1999 2004 2009
CBO (Imputed Plus Reported)
100
90
80
70
60
50
40
30
20
10
0
49% or Less
50%–99%
150%–199%
200% or More
100%–149%

Distribution of SNAP Recipients by Annual Household Income
as a Percentage of the Federal Poverty Level, 2005 to 2013
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2005 2008 2011
The Urban Institute’s
Transfer Income Model
Percentage of Recipients
100
90
80
70
60
50
40
30
20
10
0
2005 2008 2011
49% or Less
50%–99%
150%–199%
200% or More
100%–149%

Step 3. Transfer income is assigned to
recipients. The methodology underlying
that assignment varies by program.

For SNAP and SSI, CBO derives the average
benefit per recipient from reported values
(by household size and income-to-poverty
ratio for SNAP, and by eligibility category
for SSI).
Those averages are then assigned to newly
imputed recipients and are adjusted as
needed to match administrative totals.

0
500
1,000
1,500
2,000
2,500
3,000
3,500
1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013
Average Annual SNAP Benefits per Household,
1979 to 2013
Inflation-Adjusted 2013 Dollars
Administrative
Data
Reporting
Households
Newly Imputed
Households

For Medicaid, CBO derives the average
cost to the government per recipient from
administrative data (by eligibility category).
Those averages are then assigned to all
recipients (CPS “reported” values are
overwritten).

Average Annual Cost per Medicaid Recipient,
1979 to 2013
0
1,000
2,000
3,000
4,000
5,000
6,000
1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013
Inflation-Adjusted 2013 Dollars
CPS Averages
CBO Averages (From
Administrative Data)

CBO’s regression-based approach has both
strengths and limitations.

The approach is straightforward to
implement and easily scalable across
multiple programs. Its distributional results
are similar to those of rules-based
methods.

It does not, however, account for false
positives in the CPS, and it assumes that
nonreporters have the same characteristics
as reporters. It has a limited ability to
simulate different policy scenarios.

Social Insurance Benefits

CBO uses a different approach for
imputing social insurance benefits.
For its analysis of the distribution of
household income, CBO does not currently
perform any explicit distributional analysis
of social insurance benefits, since they are
included in the base income measure.

Receipt of social insurance benefits is
difficult to model with a regression using
the data available in the CPS. It is
dependent on life-cycle income/labor force
participation, it is not means tested, and
there are no income data for children in
the CPS (which are needed to impute
Social Security survivors’ benefits).

Social Security Reporting Rates in the CPS, 2001–2013
0
20
40
60
80
100
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
Disability Benefits
Survivors Benefits
Old Age Benefits
85
97
47

To impute Social Security benefits, CBO
creates a pool of eligible recipients for
each type of benefit and randomly assigns
receipt until the administrative counts are
matched.
The average benefit for each benefit type
is then assigned to new recipients and
aligned to administrative totals as needed.

0
20
40
60
80
100
1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013
Medicare Reporting Rates in the CPS, 1979 to 2013
93

To impute Medicare benefits, CBO makes
no change to reported recipients.
CBO assigns the average cost to the
government per participant to all
recipients. Benefits from the Low-Income
Subsidy for Medicare Prescription Drug
Coverage are allocated separately.

Preliminary Conclusions

Income Inequality, 1979 to 2013
Gini Index
.35
.40
.45
.50
Market Income Plus
Reported Social
Insurance Benefits
Market Income Plus
Imputed Social
Insurance Benefits
Plus Reported Means-
Tested Transfers
Market Income Plus
Imputed Social Insurance
Benefits Plus Imputed
Means-Tested Transfers
Market Income Plus
Imputed Social
Insurance Benefits
0.00
0.35
1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013
.35

-8%
-7%
-6%
-5%
-4%
-3%
-2%
-1%
0%
1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013
0
-1
-2
-3
-4
-5
-6
-7
-8
Change in Gini Index, 1979 to 2013
Percentage Change Relative to Market Income Plus Imputed Social Insurance Benefits
Reported Means-Tested Transfers
Imputed Means-Tested Transfers

Ratio of Means-Tested Transfers to Base Income Measure
Means-Tested Transfer Rates by Quintile, 1979 to 2013
.0
.1
.2
.3
.4
.5
.6
.7
Lowest Quintile Second Quintile Middle Quintile
.0
.1
.2
.3
.4
.5
.6
.7
1979 1984 1989 1994 1999 2004 2009
Fourth Quintile
1979 1984 1989 1994 1999 2004 2009
Highest Quintile
1979 1984 1989 1994 1999 2004 2009
All Quintiles
CPS (Reported Only)
CBO (Imputed
Plus Reported)

.0
.1
.2
.3
.4
.5
.6
.7
1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013
Means-Tested Transfer Rates, Lowest Quintile, 1979–2013
Ratio of Means-Tested Transfers to Base Income Measure
CPS (Reported Only)

CONGRESSIONAL BUDGET OFFICE
Notes to Figures
■ Slides 48 and 70: The reporting rate equals the weighted sum of
recipients in the CPS (including CPS imputations) divided by the
number of recipients in the administrative data, adjusted for
recipients outside the CPS sampling frame. Where administrative
totals are available on a monthly basis, they have been converted to
reflect the total number of program participants across the calendar
year.
■ Slide 55: Adults are defined as individuals age 18 to 64 who are not
disabled.
■ Slides 55–58: Individuals or households are considered recipients if
they participate in the program at any point during the calendar
year.

Notes to Figures (Continued)
■ Slides 55, 56, and 58: The Urban Institute’s Transfer Income Model
(TRIM) is a microsimulation model that uses CPS data as a basis to
simulate program rules for various transfer programs. It uses those
rules to determine program eligibility, participation, and benefits.
The current version of the model, TRIM3, provides publicly available
imputations for most major welfare programs going back to 1993.
For more details, see Zedlewski and Gianarelli (2015).
■ Slide 72: The reporting rate equals the weighted sum of recipients in
the CPS (including CPS imputations) divided by the number of
recipients in the administrative data.
■ Slides 75 and 76: Reported and imputed social insurance benefits
are from Social Security and Medicare. Reported and imputed
means-tested transfers are from Medicaid, SNAP, and SSI.

Notes to Figures (Continued)
■ Slides 77 and 78: The base income for the “CBO (Imputed Plus
Reported)” quintiles and means-tested transfer rates is market
income plus imputed social insurance benefits. The base income for
the “CPS (Reported Only)” quintiles and means-tested transfer rates
is market income plus reported social insurance benefits. Means-
tested transfers include Medicaid, SNAP, and SSI benefits.

Definitions
■ Market income consists of labor income, business income, capital
gains (profits realized from the sale of assets), capital income
excluding capital gains, income received in retirement for past
services, and other sources of income.
■ Government transfers are cash payments and in-kind benefits from
social insurance and other government assistance programs. Those
transfers include payments and benefits from federal, state, and
local governments.
■ Before-tax income is market income plus government transfers.
■ Social insurance benefits are payments from Social Security for
workers, spouses, survivors, and the disabled; Medicare; and
unemployment insurance.

Definitions (Continued)
■ Means-tested transfers consist of payments and benefits from
Medicaid, the Supplemental Nutrition Assistance Program (SNAP;
formerly known as the Food Stamp program), housing assistance
programs, and several smaller programs.
■ Federal taxes analyzed here consist of individual income taxes,
payroll taxes, corporate income taxes, and excise taxes.
■ After-tax income is before-tax income minus federal taxes.
■ Income groups are created by ranking households by various
income measures, adjusted for household size.
■ Quintiles (fifths) contain equal numbers of people.

References
■ Michael Davern and others, “A Partially Corrected Estimate of
Medicaid Enrollment and Uninsurance: Results from an Imputational
Model Developed off Linked Survey and Administrative Data,”
Journal of Economic and Social Measurement, vol. 34, no. 4 (2009),
pp. 219–240, http://dx.doi.org/10.3233/JEM-2009-0324.
■ Bruce D. Meyer, Wallace K. C. Mok, and James X. Sullivan, The
Under-Reporting of Transfers in Household Surveys: Its Nature and
Consequences, Working Paper 15181 (National Bureau of Economic
Research, July 2009), www.nber.org/papers/w15181.
■ Bruce D. Meyer and James X. Sullivan, Using Two-Sample Methods
to Correct for Reporting Bias in Surveys, Working Paper 0902
(University of Chicago, December 2008),
https://tinyurl.com/y8jjnqma (PDF; 176 KB).

References (Continued)
■ Robert A. Moffitt and John Karl Scholz, Trends in the Level and
Distribution of Income Support, Working Paper 15488 (National
Bureau of Economic Research, November 2009),
www.nber.org/papers/w15488.
■ Laura Wheaton, Underreporting of Means-Tested Transfer Programs
in the CPS and SIPP (Urban Institute, February 6, 2008),
https://tinyurl.com/yd4caq7n.
■ Sheila Zedlewski and Linda Giannarelli, TRIM: A Tool for Social Policy
Analysis (Urban Institute, May 2015), https://tinyurl.com/y7sbos8l.

Methodological Improvements for CBO’s Analysis of the Distribution of Household Income

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Similar to Methodological Improvements for CBO’s Analysis of the Distribution of Household Income

Similar to Methodological Improvements for CBO’s Analysis of the Distribution of Household Income (20)

More from Congressional Budget Office

More from Congressional Budget Office (20)

Recently uploaded

Recently uploaded (20)

Methodological Improvements for CBO’s Analysis of the Distribution of Household Income