Because more equal societies work better for everyone

Notes on statistical sources and methods

How Income Inequality is Measured

There are several ways to measure income inequality.

One way (the 20:20 ratio)  is to compare how much richer the top 20% of people are, compared to the bottom 20%.  Among the rich developed countries the 20:20 ratio varies from as little as 3 or 4 to as much as 8 or 9.  For example, in Japan and Sweden the income gap is fairly small: the richest 20% are less than 4 times as rich as the poorest 20%; but in Britain the richest 20% are over 7 times as rich as the poorest 20%, and in the USA they are over 8 times as rich.

Instead of the top and bottom 20 percent, you can compare the top and bottom 10 percent (the 10:10 ratio) or any ratio.  Some people measure what proportion of income goes to the poorest half of the population (the median share). In many societies, the poorest half of the population get around 20-25 percent of all incomes and the richest half get the rest.

Another measure of inequality is called the Gini coefficient.  It measures inequality across the whole society rather than simply comparing groups.  If all  the income went to a single person (maximum inequality) and everyone else got nothing, the Gini coefficient would be equal to 1.  If income was shared equally, and everyone got exactly the same, the Gini would equal 0.  The lower its value, the more equal a society is.  The most common values tend to be between 0.3 and 0.5.

Another measure is called the “Robin Hood Index” because it measures what proportion of a society’s income would have to be taken from the rich and given to the poor to get complete equality.

In our book, for all of our international comparisons, we use the 20:20 ratio measure of income inequality from the United Nations Development Programme Human Development Indicators, 2003-6. As survey dates vary for different countries (from 1992 to 2001), and as the lag time for effects will vary for the different outcome we examine, we took the average across the reporting years 2003-6.  For the US comparisons we use the 1999 state-level Gini coefficient based on household income produced by the US Census Bureau. 
 
Sources

United Nations Development Program. Human development report. New York: Oxford University Press, 2003, 2004, 2005, 2006. 
  
US Census Bureau. Gini ratios by state, 1969, 1979, 1989, 1999.  Washington, DC: US Census Bureau, 1999 (table S4).

[top]

 

Sources of data for the Index of Health and Social problems

Component

International data

US state data

Trust

Percent of people who respond positively to the statement “most people can be trusted”

1999-2001

World Values Survey 1

Reverse-coded

Percent of people who respond positively to the statement “most people can be trusted”

1999

General Social Survey 2

Reverse-coded

Life expectancy

Life expectancy at birth for men and women

2004

United Nations Human Development Report 3

Reverse-coded

Life expectancy at birth for men and women

2000

US Census Bureau, Population Division 4

Reverse-coded

Infant mortality

Deaths in the first year of life per 1000 live births

2000

World Bank 5

Deaths in the first year of life per 1000 live births

2002

US National Center for Health Statistics 6

Obesity

Percentage of the population with BMI > 30, averaged for men and women

2002

International Obesity TaskForce 7 8

Percentage of the population with BMI > 30, averaged for men and women

1999-2002

Corrected estimates from Prof Majid Ezzati, Harvard University, based on NHANES and BRFSS surveys 9

Mental health

Prevalence of mental illness

2001-2003

WHO 10

Average number of days in past month when mental health was not good

1993-2001

BRFSS11

Education

Combined average of maths literacy and reading literacy scores of 15-year olds

2000

OECD PISA 12

Reverse-coded

Combined average of maths and reading scores for 8th graders

2003

US Department of Education, National Center for Education Statistics 13 14

Reverse coded

Teenage birth rate

Births per 1000 women aged 15-19 years

1998

UNICEF 15

Births per 1000 women aged 15-19 years

2000

US National Vital Statistics 16

Homicides

Homicide rate per 100,000

Period average for 1990-2000

United Nations 17

Homicide rate per 100,000

1999

FBI 18

Imprisonment

Log of prisoners per 100,000

United Nations17

Prisoners per 100,000

1997-8

US Department of Justice 19

Social mobility

Correlation between father and son’s income

30-year period data from 8 cohort studies

London School of Economics20

N/A

1. European Values Study Group and World Values Survey Association. European and World Values Survey Integrated Data File, 1999-2001, Release 1. Ann Arbor, MI: Inter-university Consortium for Political and Social Research, 2005.

2. National Opinion Research Center. General Social Survey. Chicago: NORC, 1999.

3. United Nations Development Program. Human Development Report. New York: Oxford University Press, 2004.

4. US Census Bureau. Population Division, Interim State Population Projections, Table 2. Internet release date: April 21, 2005, 2005.

5. World Bank. World Development Indicators (WDI) September 2006: ESDS International, (MIMAS) University of Manchester.

6. US National Center for Health Statistics. Table 105, Statistical abstract of the United States. Washington, DC: CDC, 2006.

7. International Obesity TaskForce. Obesity in Europe. London: International Obesity TaskForce in collaboration with the European Association for the Study of Obesity Task Forces, 2002.

8. International Obesity TaskForce. Overweight and obese. London: International Obesity Taskforce, 2002.

9. Ezzati M, Martin H, Skjold S, Vander Hoorn S, Murray CJ. Trends in national and state-level obesity in the USA after correction for self-report bias: analysis of health surveys. J R Soc Med 2006;99(5):250-7.

10. Demyttenaere K, Bruffaerts R, Posada-Villa J, Gasquet I, Kovess V, Lepine JP, et al. Prevalence, severity, and unmet need for treatment of mental disorders in the World Health Organization World Mental Health Surveys. Jama 2004;291(21):2581-90.

11. Zahran HS, Kobau R, Moriarty DG, Zack MM, Holt J, Donehoo R. Health-related quality of life surveillance--United States, 1993-2002. MMWR Surveill Summ 2005;54(4):1-35.

12. OECD. Education at a glance. OECD Indicators, 2003.

13. US Department of Education NCfES. The Nation’s Report Card: Reading Highlights 2003. Washington, DC, 2004.

14. US Department of Education NCfES. The Nation’s Report Card: Mathematics Highlights 2003. Washington, DC, 2004.

15. UNICEF Innocenti Research Centre. A league table of teenage births in rich nations. Florence: Innocenti Report Card, 2001.

16. U.S. Census Bureau. Statistical Abstract of the United States:2000 (120th Edition). Washington: Census Bureau, 2000.

17. United Nations Crime and Justice Information Network. Survey on Crime Trends and the Operations of Criminal Justice Systems (Fifth, Sixth, Seventh, Eighth): United Nations, 2000.

18. Federal Bureau of Investigation. Crime in the United States 1999. Washington, D.C.: U.S. Government Printing Office, 1999.

19. US Department of Justice BoJS. Incarceration rates for prisoners under State or Federal jurisdiction. File: corpop25.wk1.

20. Blanden J, Gregg P, Machin S. Intergenerational mobility in Europe and North America. London: Centre for Economic Performance, London School of Economics, 2005.

[top]

 

Developing the Index of Health and Social Problems for 23 rich countries:

The index has 10 components:

  • Life expectancy (reverse coded)
  • Teenage births
  • Obesity
  • Mental health
  • Homicides
  • Imprisonment (log transformed)
  • Mistrust
  • Social mobility
  • Education (reverse coded)
  • Infant mortality rate

16 countries had at least nine of the ten measures.  A further 5 countries had eight out of ten. only two measures missing.  Two countries (Israel and Singapore) with fewer measures were excluded from the index but included in analyses of individual measures.

  • Countries with data on all ten measures: Canada, Germany, USA
  • Countries with data on 9 out of 10 measures, but no data on social mobility: Australia, Belgium, France, Italy, Japan, Netherlands, New Zealand, Spain
  • Countries with data on 9 out of 10 measures, but no data on mental health: Denmark, Finland, Norway, Sweden
  • Countries with data on 9 out of 10 measures, but no data on education: UK
  • Countries with data on 8 out of 10 measures, but no data on social mobility or mental illness: Austria, Greece, Ireland, Portugal, Switzerland

The Index of Health and Social Problems was created by taking the mean of the z-scores for each measure (averaged over the number of measures available for that particular country).

[top]

 

Developing the Index of Health and Social Problems for the USA:

The index has 9 components:

  • Trust (reverse coded – which is the same as mistrust)
  • Life expectancy (reverse coded)
  • Teenage births
  • Obesity – corrected data based on measured heights/weights
  • Homicides
  • Imprisonment
  • Education (reverse coded)
  • Infant mortality rate
  • Mental health

Of the 50 states, 40 have data for all eight measures.

Nine states are missing data on trust from the General Social Survey:

  • Alaska
  • Delaware
  • Hawaii
  • Idaho
  • Maine
  • Nebraska
  • New Mexico
  • Nevada
  • South Dakota

Wyoming has data on trust, but not on homicides

The Index of Health and Social Problems for the USA was created by taking the mean of the z-scores for each measure (averaged over the number of measures available for that particular state)

[top]

 

Pearson Correlation Coefficients (r) and P-Values for Associations with Income Inequality reported in The Spirit Level.*

Indicator

International data

US data

 

r

p-value

r

p-value

Trust

-0.66

<0.01

-0.70

<0.01

Life expectancy

-0.44

0.04

-0.45

<0.01

Infant mortality

0.42

0.04

0.43

<0.01

Obesity

0.57

<0.01

0.47

<0.01

Mental illness

0.73

<0.01

0.18

0.12

Education score

-0.45

0.04

-0.47

0.01

Teenage birth rate

0.73

<0.01

0.46

<0.01

Homicides

0.47

0.02

0.42

<0.01

Imprisonment

0.75

<0.01

0.48

<0.01

Social mobility

0.93

<0.01

-

-

Index

0.87

<0.01

0.59

<0.01

         

Overweight children

0.59

0.01

0.57

<0.01

Drugs index

0.63

<0.01

   

Calorie intake

0.46

0.03

   

Public expenditure on health care

-0.69

<0.01

   

Child well-being

-0.63

<0.01

-0.51

<0.01

Triple education score

-0.44

0.04

   

Child conflict

0.62

<0.01

   

Spending on foreign aid

-0.61

<0.01

   

Recycling

-0.82

<0.01

   

Peace index

-0.45

0.03

   

Paid maternity leave

-0.55

0.01

   

Advertising

0.60

<0.01

   

Police

0.52

0.04

   

Social expenditure

-0.45

0.04

   

Women’s status

-0.50

0.02

-0.30

0.03

Juvenile homicides

   

0.29

<0.05

High school drop-outs

   

0.79

<0.01

Child mental illlness

   

0.36

0.01

Pugnacity

   

0.47

<0.01

* For details of measures, see Richard Wilkinson and Kate Pickett, The Spirit Level (Penguin).Buy the book from Amazon. [external link]

 

 

[top]

Erratum

Analyses of public expenditure on health care, the Global Peace Index, advertising and women's status previously published on this page inadvertently excluded Singapore. These have now been corrected.  When Singapore is added back in, statistical significance is unchanged and the correlations substantively similar

Measure

Correlation (r) without Singapore

p-value

Correlation (r) with Singapore

p-value

Public expenditure on health

-0.54

0.01

-0.63

<0.01

Global Peace Index

-0.51

0.01

-0.45

0.03

Spending on advertising

0.73

<0.01

0.60

<0.01

Women's status

-0.44

0.04

-0.50

0.02

The correlation coefficient and p-value for the UNICEF index of child well-being previously published on this page was for the index prior to removal of the child poverty component (r = -0.71, p<0.01), which we believe should be removed as it is close to a measure of inequality.  The correlation coefficient and p-value for the corrected index are: r = -0.63, p<0.01).  The graph and regression line in The Spirit Level use the corrected index, as does our paper on this topic in the British Medical Journal. [PDF]