The Prince and the Pauper:

Movement of Children Up and Down the Canadian Income Distribution

Peter Burton and Shelley PhippsDalhousie University

P. Burton and S. Phipps Dalhousie University

IntroductionUse Statistics Canada’s National

Longitudinal Survey of Children and Youth (NLSCY) to study family income for a cohort of Canadian children between 1994 and 2004

Children 0 to 7 in 1994; 10 to 17 in 2004Longest panel of data yet available in

Canada

P. Burton and S. Phipps Dalhousie University

Five questions:1. What happens to the level of family income

as children grow up?2. What happens to income inequality among

children?3. How much movement up and down the

distribution takes place?4. What are characteristics associated with

being ‘stuck at the bottom’5. What changes are correlated with moving

up or down the distribution?

P. Burton and S. Phipps Dalhousie University

DataNLSCY representative of Canadian

child populationInterviews every 2 years (6 cycles,

spanning 10 years)Use information provided by the

‘person most knowledgeable’ about the child

Select 7,163 children with complete income and family size data

P. Burton and S. Phipps Dalhousie University

Question 1. Trends in Income Levels?

Pre-tax annual income from all sources including government transfers

Adjust for differences in need for families of different size using Luxembourg Income study ‘equivalence scale’ (square root of family size)

Actual income of $80,000 for family of 4 means ‘equivalent income’ of $40,000

P. Burton and S. Phipps Dalhousie University

Mean equivalent family income, in 2004 dollars

1994 1996 1998 2000 2002 20040

50001000015000200002500030000350004000045000

29918 3070634373

37403 38082 38276

P. Burton and S. Phipps Dalhousie University

Is income growth the same at all points in the distribution?

Decile cut points defined using the NLSCY (i.e., families with children)

1994 1996 1998 2000 2002 20040

20,000

40,000

60,000

80,000

100,000

120,000

P. Burton and S. Phipps Dalhousie University

Question 2. Trends in Inequality?Compute measures of income

inequalityAnnual income and six-year average

income

P. Burton and S. Phipps Dalhousie University

Measures of Income Inequality

1994 1996 1998 2000 2002 2004

Long-run Average Income

Coefficient of variation 0.679 0.788 0.753 0.801 0.716 0.701 0.622

Gini0.334 0.345 0.328 0.339 0.325 0.321 0.293

90:10 Ratio 5.357 5.486 4.506 4.496 4.689 4.656 3.807

90:50 Ratio 2.017 2.040 1.938 2.011 1.902 1.913 1.899

Theil 0.188 0.214 0.196 0.214 0.189 0.186 0.150

P. Burton and S. Phipps Dalhousie University

Theil Decomposition

Theil index allows de-composition of total inequality into ‘within group’ (same child across six cycles) plus ‘between group’ (average income across different children)

De-composition suggests inequality of ‘permanent income’ about 75 percent of total

P. Burton and S. Phipps Dalhousie University

Question 3. Are the Same Children Always at the Bottom of the Income Distribution?What percent of children who start in

bottom quintile in 1994 are again in bottom quintile in 2004?

P. Burton and S. Phipps Dalhousie University

1994 to 2004 Transition Matrix

Bottom Quintile 2004

2nd Quintile 2004

3rd Quintile 2004

4th Quintile 2004

Top Quintile 2004

Bottom Quintile 1994

0.51 0.25 0.14 0.06 0.04

2nd Quintile 1994

0.26 0.28 0.25 0.14 0.07

3rd Quintile 1994

0.12 0.25 0.30 0.23 0.10

4th Quintile 1994

0.07 0.14 0.22 0.34 0.24

Top Quintile 1994

0.03 0.08 0.08 0.24 0.58

P. Burton and S. Phipps Dalhousie University

Transition Matrix for Children of Immigrants

Bottom Quintile 2004

2nd Quintile 2004

3rd Quintile 2004

4th Quintile 2004

Top Quintile 2004

Bottom Quintile 1994

0.67 0.15 0.07 0.07 0.05

2nd Quintile 1994

0.18 0.31 0.34 0.07 0.10

3rd Quintile 1994

0.12 0.39 0.22 0.16 0.10

4th Quintile 1994

0.14 0.08 0.23 0.35 0.19

Top Quintile 1994

0.02 0.06 0.11 0.23 0.58

P. Burton and S. Phipps Dalhousie University

‘Lenses’ What happens during intervening

years?How many children ever exposed to

a position of low income?How many children always (in all six

cycles) in a position of low income?

P. Burton and S. Phipps Dalhousie University

0 10 20 30 40 50 60 70 80 90 100

0

10

20

30

40

50

60

70

80

90

100

26.1

42.2

54.3

65.1

73.9

81.987.8

92.997.0

100.0

0.01.3

4.79.9

16.6

24.2

34.8

46.5

60.9

78.2

100.0

Figure 2. Relative Income "Lens"

Always Below

Ever Below

Equivalent Income Percentile

Percentof Children

Link

P. Burton and S. Phipps Dalhousie University

0 10 20 30 40 50 60 70 80 90 1000

10

20

30

40

50

60

70

80

90

100

0.0

40.0

63.6

80.9

92.5

100.0

0.03.9

15.5

33.5

60.3

100.0

0.0

54.5

73.5

86.9

95.1100.0

0.0

9.1

22.7

42.3

64.3

100.0

Relative Income Lenses, Immigrants and NonIm-migrants

Immigrant, Always BelowImmigrant, Ever BelowNonImmigrant, Always BelowNonImmigrant, Ever Below

Equivalent Income Percentile

Percentof Children

P. Burton and S. Phipps Dalhousie University

Question 4. Characteristics of Children ‘Stuck’ at BottomEstimate probit models of the

correlates of ‘always’ being in the bottom quintile

Dependent variable uses full six-cycle history

Explanatory variables are ‘starting point risks’ (1994 values)

P. Burton and S. Phipps Dalhousie University

Key results from probit regressions for ‘always’ in bottom quintile:In order of size of association, a child is at

greatest risk of ‘always’ being at the bottom of the distribution for his/her cohort if he/she:◦Lives in a lone-parent family◦Has a parent with no paid work◦Lives in the Atlantic region◦Has a parent who is non-white

P. Burton and S. Phipps Dalhousie University

Simulated Probability of Always Being in the Bottom Quintile

Base Lone Parent Pmk Unpaid Atlantic Non-white0

0.51

1.52

2.53

3.54

P. Burton and S. Phipps Dalhousie University

Question 5. Which changes are associated with movements up or down?Estimate fixed effects models for

change in percentile position Explanatory variables are now

‘changes’ (so ethnicity and immigrant status dropped)

Coefficient Atlantic -11.810*** Quebec -4.044 Manitoba/Saskatchewan -10.005*** Alberta -3.972 BC -2.642 Less than High School -1.094 Some Post-Secondary 0.788 University 2.724***Lone parent -22.360***Pmk student -3.571***Number of siblings -2.827***Pmk no paid hours -7.062***

P. Burton and S. Phipps Dalhousie University

ConclusionsUse longitudinal data tracking a cohort of

Canadian children from 1994 to 2004 (from ages 0 to 7 until ages 10 to 17)

Real growth at all points in income distribution; no trends in inequality as this cohort of children grows up

75 percent of inequality is attributable to ‘permanent income’

P. Burton and S. Phipps Dalhousie University

‘Stickiness’ of relative income position◦Beginnning to end of period ◦Always in the bottom

High level of ‘ever exposed’ to low income

P. Burton and S. Phipps Dalhousie University

Largest starting point ‘risks’: parental marital and employment status, region of residence and ethnicity

Largest movements up/down the distribution: changes in parental marital status and regional moves

P. Burton and S. Phipps Dalhousie University

Thanks!