Inequality in Australia: Does region matter

1
Inequality in Australia Does region matter?

• Riyana Miranti, Rebecca Cassells, Yogi Vidyattama
• and Justine McNamara

PRESENTED AT THE 2ND GENERAL CONFERENCE OF THE
INTERNATIONAL MICROSIMULATION ASSOCIATION,
OTTAWA, CANADA, JUNE 8 10, 2009
2
Measuring Inequality - Background

• Why we chose this topic ?
• Objectives
• to provide valuable information about regional
  inequality at a small area level
• to explore another use of spatial microsimulation
  and demonstrate its benefits

3
What are we going to do

• Measuring inequality at small area using Gini
  coefficients
• Reasons for use of Gini coefficients
• Most common measure
• Validation purpose publicly available at the
  national and state level
• Expand previous research with improvements
• disposable household income
• smaller geographical unit than any that has been
  previously used
• Using spatial microsimulation, as direct data are
  not available

4
Data source

• Reweighting process uses three sources of data
• 2006 Census
• Survey - SIH 2003-04 and 2005-06
• Validation use 2006 Census data, ABS published
  data and SIH 2005-06
• Limit the scope of study to New South Wales (NSW)
  and Victoria (Vic)
• Unit of analysis small area (Statistical Local
  Area)

5
Spatial methodology

• Spatial microsimulation SpatialMSM/09C
• Small area weights for every SLA
• Benchmarks variables
• Complex process of spatial microsimulation

6
Gini coefficient

• Has a value between zero and one
• Zero means perfect equality, everyone has the
  same level of equivalised income
• One means perfect inequality, one person holds
  all the income
• Smaller Gini coefficient more equal
• Equivalised hh disposable income

7
Validation of our estimates

• To see whether our Gini coefficient estimates are
  reliable
• 197 SLAs in NSW, and 198 SLAs in VIC
• Small area validation equivalised gross
  household income data, see next slide
• Aggregate data validation, at capital city and
  balance of state level equivalised disposable
  household income data overall looks good.

8
Validation small area validation (NSW)
The Spearman rank correlation is 0.958
9
Australian map
10
Distribution of small area inequality estimates
New South Wales
11
Distribution of small area inequality estimates
Sydney
12
Distribution of small area inequality estimates
Victoria
13
Distribution of small area inequality estimates -
Melbourne
14
Inequality and small area characteristics

• Econometric analysis of determinants of
  inequality is beyond the scope of this paper.
  However
• Previous research in Australia discusses several
  factors associated with inequality
• We find some similarities but also differences in
  characteristics among high inequality areas no
  One story fits all
• Need to look further into particular SLAs, which
  ones underlying difference

15
Conclusion

• Application of spatial microsimulation
• The validation shows that weights give reasonable
  results
• Does region matter ? Yes. There are substantial
  variations in inequality at small area level
• May help the policy makers/service providers to
  understand differences in order to better develop
  programs/policy.
• Future work ? Econometric estimation, spatial
  microsimulation in order to model policy changes

16
www.natsem.canberra.edu.au