Spending on Children by liaoqinmei


									Data and Construction
 of Economic Table

 Washington Child Support Group
        December 2007
           Session I

• Quick Review of Past Session

• Data -- Consumer Expenditure Survey

• Constructing the Economic Table

• Segue to Second Session
                   Quick Review
•   Income Shares Model -- Normative Standard is to continue the spending the
    children at a level consistent with would had be spent if the family remained

•   Implementing the Income Shares Model

     – Allocating total spending of the family to the children: USDA, Engel, and
       Rothbarth Methods
     – USDA and Engel provide very close estimates
     – Rothbarth and Engel provide lower and upper bound for the estimates of
       child rearing costs

•   Moving from Net Income to Spending on Children other than child care
    and excessive medical expenditures
    Implementing Rothbarth
•   Need to be able to estimate the        Adult Goods
    relationship between spending on
    Adult Goods, Total Spending and                                         No Children
    the number of children

•   Sample size and the variability of
                                                                                  1 child
    levels of total spending and number
    of children will be important

•   Definition of adult goods: Broad
    (clothing, alcohol, tobacco,
    entertainment, savings) or Narrow                    on Children
    (clothing) may influence our ability
    to estimate precisely the needed
    relationships                                                      TS              Total
Consumer Expenditure Survey
•   A nationally representative sample of consumer units (families)
    collected to understand how Americans consume. Its main statistical
    function is to provide information to update the budget share weights
    used in the Consumer Price Index.

•   Only US survey aimed at how Americans spend.

•   Each consumer unit is interviewed five times, once to collect
    background information then four times more to collect information on
    expenditures during the quarter, current family members, and income

•   My research examines annual expenditures and consequently I have
    restricted my sample to those units with at least three quarterly
                      2006 Report
•   Data was from the 1st quarter of 1998 through the 1st quarter of 2004
    (most recently available data is through the 4th quarter of 2005).

•   Sample Restrictions:

     •   Must contain a married couple between the ages of 18 and 60 years old
     •   The unit does not have any other adults (18+) living in the unit even if the
         member is an adult child of the married couple.

•   Distribution of Sample:

    Number of Children:       0          1          2         3          4        5+
                          3,338      1,778      2,611     1,116        311        91
     Distribution of Total Spending
           (July 2003 Dollars)

Number of Children                    0                1               2                3+

Average Expenditures           $44,728          $46,140          $49,834          $48,341

Expenditure at
  5th percentile               $15,553          $15,766          $18,493          $17,170
  25th percentile               27,270           27,916           31,275           28,440
  50th percentile               38,759           40,175           44,460           42,248
  75th percentile               54,854           57,837           61,934           59,600
  95th percentile               93,265           94,340           97,562          102,247

To put in $2007 (October), multiple the above numbers by 1.138. For example, the 95th
percentile for childless couples would be $106,136 in today’s dollars.
       Limitations of Data
• Can be used only to estimate the cost of children when there
  are 1, 2, and 3 children -- not enough sample for 4 or more
  children in the family to provide reliable estimates

• Estimates limited to families who spend roughly $110,000 or
  less. But remember that families who spend $110,000 a year
  will have net incomes in excess of $150,000 a year or $12,500
  per month.

• The public use data codes clothing spending on family members
  more than 15 years old as adult clothing. If the unit had O
  children over 15 years old, I recoded adult clothing to be equal
  to 2/(2+O) times the amount of adult clothing report in file.
    From Net Income to BCSO Table

                   Rothbarth, Engel, USDA Estimates


  Net                   Family
                                                       on Child
                         Use CEX to estimate the      Excessive
         Savings         relationship between Net      Medical
                             Income and Total
                                 Spending                (CEX)
      Previous PSI Report
• Used Engel and Rothbarth estimates from my 2000 report

• Estimated relationship between net income and total spending,
  percentage of spending devoted to child care and excessive
  medical spending using a sample from the 1996 to 1998 CEX

• Inflated all dollars amounts to reflect 2004 dollars

• The table would change if updated to reflect my estimates from
  the 2006 report and use more recent CEX data to update the
  needed relationships to move from net income to the BCSO
      Example: Net Income = $4,583 (month)
      Midpoint of $50,000 to $60,000 Interval

•   Based upon the 1996 to 1998 data, families with children in this income interval
    (adjusted for inflation) consumed 88.2% of their net income. Consequently at
    $4,583 of net income, the family is expected to spend $4,042.

•   Based upon the Rothbarth estimates from the 2000 report, the family is
    expected to spend 25.5% of the family’s total spending on one child. Hence
    total spending on the child is $1,031.

•   As percentage of total spending, child care was 1.7% and excessive medical
    spending was 3.5% per child (computed using the 1996-8 CEX data)

          Subtract: .017(4,042) + .035(1,031) = $105

•   BCSO amount = $1,031 - $105 = $926 or 20.2% of net income (this amount
    differs from the PSI report because of rounding of the entries in Exhibit I-1.
Low Income “Adjustment”



                   Net Income (NI)
        Smoothing the Table
• Construct net income intervals

• At the midpoint of each interval, compute as
  previously done the BCSO entry

• We could assume the spending was the
  same within the interval but this would lead
  to discontinuities as well as other problems

• The table reflects an linearization of the
  midpoints. In Exhibit I-2 of the PSI report,
  they report at the midpoint two numbers:
  the BCSO entry divided by the midpoint and
  the slope of the green line (marginal
•   Child Support Models: Collection of normative judgments whose
    purpose is to determine a level of child support to be paid by one
    parent to the other parent for the purpose of providing financial
    resources to raise the children.

•   Disagreement over whether the child support award can arise for two

     – Disagreements over the model -- whether the normative judgments that
       have been made are truly appropriate
     – Disagreements over whether the model has been appropriately
       implemented -- for example, one can disagree over whether the Economic
       Table truly represents the spending of two parents in an intact family on the
       children without disagreeing whether Income Shares is the appropriate
 Income Shares versus POI
• Income Shares: The financial obligation of the parents should
  equal what they would have spent on the children had the family
  remained intact

• Percentage of Income (POI)

   – Parents should continue to spend the same proportion of their
     income on their children as they would have had the family
     remained intact
   – AND the percentage the poor pay should not be more than the
     more wealthy pay -- the percentage should be the same for all
     families with the same number of children
    Normative Judgments versus

• Normative Judgments can’t be verified on the basis of objective
  observation -- Data can’t resolve the differences between
  alternative value judgments

• Assumptions are often made to implement a model, however, in
  theory assumptions should be able to be verified by examining
  data or used to decide which one of the competing assumptions
  is ‘best’. However, it is not always possible to acquire
  appropriate data to test an assumption.

•   As total spending increases, adults   Adult Goods
    spend more on themselves                                               No Children

•   Holding total spending constant,
    adults without children spend more
    on adult goods                                                               1 child

•   Adult goods are a reasonable
    indicator of the standard of living
    (preferences of adults for adult                     Spending
    goods are separable from other                      on Children

    consumption groups: shared
    consumption and child goods)
                                                                      TS              Total
           Next Session
       Alternative Child Support Models

–   Income Shares
–   Percentage of Income (POI)
–   Melson
–   Cost Shares
–   ALI (American Legal Institute)

To top