Mom's rising income lowers Dad's rate, not just his dolla

1
Evaluating Support Guidelines

• What They Do, and How They Came to Do It
• New Jersey Legal Services Annual Meeting
• November 21, 2006
• IRA MARK ELLMAN,
• Sandra Day OConnor College of Law
• Arizona State University
• Tara OToole Ellman
• Private Consultant

2
Outline

• Arizonas Guidelines as Example
• How to Tell What They Do
• Analysis Why Do They Do It?
• How Does the Consultant Generate the Numbers?
• Why Does This Method Yield These Results?

3
CHART 1 ANIMATION
95TH PERCENTILE
POVERTY
80TH PERCENTILE HH
US MEDIAN HH
4
HOW MUCH MONEY..
HOW MUCH MONEY PEOPLE MAKE Some useful
reference points
Poverty level for 2-adult,
1-child household 1,207 Median income
for all US households
3,550 80th Percentile for all US households
7,001 95th percentile for all
US households 12,500 All
above in dollars per month in 2002
5
CHART 1 BASIC TABLE
6
CHART 2 SUPPORT RATES
Total support obligation divided by combined
income
The support rate Dad pays on his income depends
on the parents combined income.
7
CHART 2 CONTINUED
Rates high at low incomes, fall steeply as income
rises
Plummet at 4200
Low at high income, fall slowly above 8100
8
The Effects of Falling Rates

• Dads with the same income pay different rates.
• Moms rising income lowers Dads rate, not just
  his dollar amount.
• Is this reasonable? A policy choice on which
  people may differ.

9
SUPPORT RATE EXAMPLES
Examples Dad earns 3,000, mom earns
nothing Dad pays 20 or 600
Dad earns 3,000, mom also earns 3000
Dad pays 14 or 420
10
POVERTY LEVELS
Comparing Outcomes Across Households
Difficult A common method uses official poverty
level One adult and one child 1,033.
Household at 2,066 is at 200 poverty level An
imperfect measure, but often used Good measure
for lower income families Simple to
understand.
11
CHART 3 INTRO MIDDLE INCOME
Intact family was at 300 of poverty level
Situation Middle Income Household with One Child
Outcomes AFTER payments
Possible outcomes for Mom and Child BEFORE
payments.
Moms Income Share
12
CHART 5 LOW INCOME MOM
Payment helps, but not nearly enough to restore
old living standard.
After 468 payment, Mom and Child at 150 of
poverty
Mom earns 30 of combined income, or 1065, near
poverty level.
Situation Low income Mom from a middle-income
intact household.
13
INCOME SHARES
Outcome Comparisons Show Childs Living
Standard Depends Mainly On Custodial Parent
Income
14
LOW INCOME MOM
A LOW INCOME MOM EARNING 1000 IN THREE
SITUATIONS (ONE-CHILD)
15
HCART 4 TWO CHILDREN
How two middle- class children can have very
different outcomes
Falls near poverty with low earning Mom.
Keeps middle class life style with high earning
Mom .
Conclusion Where noncustodial parent earns the
majority of income, our guidelines do not protect
children from large declines in living standard
when their parents separate.
16
Conclusions

• Childs Living Standard Depends on CP
• If CP Earns Much Less Than NCP, Childs Living
  Standard Is Much Less than Marital
• Parents
• Similar Incomes, Similar Outcomes
• Big Income Difference, Big Outcome Difference

17
Part II How Did It Get This Way?

• How Does Consultant Generate These Numbers?
• Why Does His Method Yield These Results?

18
Method Initial Overview

• Try to Measure How Much
• The average intact family
• Of the same total size as our separated family
• And the same total parental income
• Spends on their children
• Obligor Pays Proportionate Share
• Q Why Does This Method Yield Such Problematic
  Results?

19
The Answer Depends Upon

• The definition of child expenditures
• what number does consultant seek?
• The Methods Concept
• How, so defined, are they measured
• how good is our measure?
• The Methods Implementation

20
HOW ARE CHILD OUTLAYS DEFINED?From PSI Report,
Pg. 6
21
What Are the Marginal Expenditures?
Illustrative (not actual data)
22
Use of Marginal Expenditure is a Child Support
Policy Choice

• Child Support Amounts Necessarily Involve a
  Tradeoff Between
• NCPs desire to avoid support of CP
• Childs claim on NCP Support

23
A Marginal Measure Yields

• Where Parents Are Equal Earners
• Both Have Base Expenditures
• Allocation of Marginal Expenditures Protects Both
  the Child and the NCP
• Where NCP Earns Much More
• CP and Child Both Lack the Base
• Where CP Earns Much More
• Child and CP Are Better Off than NCP Before
  receipt of support payments

24
Basic Policy Choice

• Children and Custodial Parent Are One Household
  and Share a Living Standard
• Compromise therefore necessary between two valid
  claims
• 1. Child Support must be adequate to provide
  child with an appropriate living standard
• 2. Child Support should benefit the child and
  not the custodial parent
• Marginal Measure Weights Claim 2

25
Expenditure Definition Summary

• Backward Looking, Not Forward
• Asks what families spent when intact
• Not what separated families need now
• Some expenditures excluded to define
  consumption
• Looks for measure of marginal outlay
• Works fine for Equal Earning Parents
• Favors Higher Earning Parent When Incomes are
  Disparate.
• Does not consider childs outcomes
• Next Do we measure it accurately?

26
Part II How Did It Get This Way?

• How Does Consultant Generate These Numbers?
• Why Does His Method Yield These Results?
• So far How the Definition Yields Poor Outcomes
  for Poor CPs, even with NCPs with Higher Incomes
• Now Impact of the Implementation Measuring
  Marginal Expenditures

27
Accuracy of Measure 2 Issues

• Is the Measurement Method Sensible?
• Do We Have the Data Needed to Do It?
• First, the method

28
From PSI Report, Pg. 6
29
Source of Equivalence Table

• Two living standard estimators used
• Engel families are equivalent when the same
  percent of their outlays go to groceries
• Rothbarth Adults equivalent when spending same
  amount on adult items

30
Points to Ponder

• Different estimators reach very different
  results see NJ PSI report pg II-6.
• No empirical basis for choice between them
• PSI chooses Rothbarth because its results match
  Betsons intuition that Engel too high. Pg II-7.
• Now to the data

31
CES Data

• Links income, expenditures, and family
  composition
• Panels interviewed every three months
• Do panel members accurately report their income
  and outlays?
• No
• Both Under-reporting and over-reporting

32
Income Underreporting

• Problem well-known among economists
• Lower half report expenditures in excess of
  income
• PSI solution
• ignore excess expenditures
• Lower half spends 100 of their income
• Likely effect increase child outlay estimates
  at lower income levels (because they are really
  higher income)

33
Expenditure Underreporting

• Particular problems in higher incomes
• Indicator implied savings rate implausible
• Households from 70 to 90,000 gross

Expenditures include pension plan contributions
34
Expenditure underreporting

• Likely Effect child expenditure estimate
  declines as income rises
• In NJ, for 2 Children Declines to 19 of net
  income, from 38, Pg II-11
• Conclusions Data Problems
• Yield regressive child support schedule
• Cast doubt on Rothbarth measures

35
Why the recent decline in Expenditure Estimates
at High Incomes?

• Change in parental values? Upper income parents
  spending less on their children?
• Cost of childrens goods (neighborhoods?) gone up
  less than the goods in general?
• Or an artifact of the data problems (E.g.,
  increase in high income underreporting?)

36
Conclusion 1

• Current approach misconceives task
• Real task guidelines that properly balance
  competing policy concerns
• Child well-being
• Fair allocation of Support Burden Between Parents
• Consultants Task estimate marginal expenditures
  on child in intact families

37
Conclusions 2 3

• Method used to pursue mistaken task is
  conceptually flawed
• Relies on arbitrarily chosen estimator
• Conceptually flawed method relies on flawed data
  that yields regressive support table

38
Summary Conclusion

• Consultant currently
• Employs bad data
• To implement a conceptually
• flawed method
• Of determining an irrelevant fact
• We should probably do something different