Keeping up with the Schmidts

Keeping up with the Schmidts

Gundi KniesDIW Berlin and University of Bristol

Do better off neighbours cause unhappiness?

March 29, 2007 G. Knies/ SOEP & CMPO

Structure of the talk

1. Problem Formulation

2. Methdology

3. Data

4. Empirical Results

5. Discussion

March 29, 2007 G. Knies/ SOEP & CMPO

1. Problem Formulation (1)- Neighbourhood Research -

Empirical studies: comparison with better-off neighbours

increases propensity to riot (Gurr 1970, Canache (Gurr 1970, Canache 1996)1996)

“a person’s sense of contentment depends not on objective conditions, but on the subjective perceptions and comparisons of self to others” Lopez Turley 2002, S. 672-673

Theory of relative Deprivation

March 29, 2007 G. Knies/ SOEP & CMPO

1. Problem Formulation (2)- Happiness Research -

Impact of Satisfaction with the Community and Neighbourhood on Life Satisfaction:

Neighbours as a Reference Group:

Michalos 1986: Multiple Discrepancy TheoryLuttmer 2005: relative consumption

Sirgy & Cornwell 2002; Shields & Wooden 2003

March 29, 2007 G. Knies/ SOEP & CMPO

2. Methodology (1)

Individuals living in neighbourhoods where they are worse off than their average neighbour are unhappier

Given one’s own income: How is happiness affected by one’s neighbour’s income?

Micro-economic happiness modell:

LSi = α + β´Xi + γ’Zi+ εi

xi = per capita household income zi = per capita neighbourhood income

LSi = β1 log x1i + γ1 log z1i+ εi

March 29, 2007 G. Knies/ SOEP & CMPO

2. Methodology (2)

Playing Devil‘s Advocate:

- lower life satisfaction = feeling deprived?

- interaction with the NB

- neighbourhood infrastructure effects

- unobs. heterogeneity controlled

March 29, 2007 G. Knies/ SOEP & CMPO

3. DataGerman Socio-Economic Panel Neighbourhood Indicators

Characteristics ofindividuals and households (all years)NB Infrastructure (94,99,04)

IDs: address, hhid, persid

Ø disposable pc income (Infas)

ID= PLZ93, 98

PLZ

Impact of NB on... Life-Satisfaction

Zip-code areas:9-63,000 inhabitantsØ 9,000 inhabitants

(SOEP: 17,000)

March 29, 2007 G. Knies/ SOEP & CMPO

4. Empirical Results: Structure

I. Household Income, Neighbourhood Income and Average Happiness in 1999

(very similar results in 1994)

II. Multivariate Prediction 1999 (very similar results in 1994)

III. Further Hypotheses/ Robustness tests- Measure of Relative Deprivation

- Effects of Neighbourhood Infrastructure- Interaction with NB- Unobserved Heterogeneity

March 29, 2007 G. Knies/ SOEP & CMPO

4. Empirical Results (1)Mean Life Satisfaction by Classes of Household and

Neighbourhood Income 1999

5

6

7

8

1 2 3 4 5

household income

Ø l

ife s

ati

sfact

ion

nb y 3

only hh y

5

6

7

8

1 2 3 4 5

household income

Ø life

sati

sfact

ion

nb y 1

nb y 2

nb y 3

nb y 4

nb y 5

March 29, 2007 G. Knies/ SOEP & CMPO

4. Multivariate Prediction of Life-Satisfaction

Control Variablesb-Coefficients

ALL

pc Neighbourhood Income (log)

0.21*

per capita Household Income (log)

0.47**

Number of Observations 12,251

R² 0.1

Notes: Model controls for marital status, number of children in the household, disability status basic characteristics, employment status and type of community.Source: SOEP 1999 and neighbourhood indicators on the zip-code level. Author‘s calculations.

NBY>HHY

HHY>NBY

0.36* ~ 0

0.53** 0.39**

6,596 5,671

0.1 0.1

March 29, 2007 G. Knies/ SOEP & CMPO

4. Measurement of Relative Deprivation

Compared to others I did not achieved what I deserve

AllHHY>NBY

NBY> HHY

totally agree 7.4 5.4 9.3

agree slightly 24.4 20.6 28.0

disagree slightly 43.2 46.1 40.4

totally disagree 25.0 27.9 22.3

Total 100 100 100

Ordered Probit Results

NBY 0.05

HHY 0.38**

N 12,145

Pseudo R²

0. 04

March 29, 2007 G. Knies/ SOEP & CMPO

4. Interactions 1999

Control VariablesHappiness Model

Young Kids

Dog Work Socials

Neighbourhood y (log)

0.24* 0.08 0.23 0.2

young kid * NBY -0.31

dog owner * NBY 0.37

work * NBY

-0.24

socials * NBY 0.07

Household y (log) 0.46** 0.47** 0.43** 0.47**

Number of Obs. 12,438 10,868 7,173 12,224

R² 0.1 0.1 0.06 0.1

March 29, 2007 G. Knies/ SOEP & CMPO

4. Neighbourhood Infrastructure

Reduction of β neighbourhood income to 0.04

Not statistically significant:β >0: kindergarten, primary school, cityβ <0: banks, shops, doctors, public transport, youth club, club for elderly

Statistically significant effects:β <0: parks, sports ground, gym, bars

March 29, 2007 G. Knies/ SOEP & CMPO

4. Unobserved Heterogeneity

Fixed Effects Models

^NBY ^HHY N R²

All 0.42 0.33** 8,491 0.03

Stayers 0.39 0.41** 6,966 0.03

Movers 0.37 0.11 2,592 0.04

Source: SOEP 21. Author‘s calculation.

March 29, 2007 G. Knies/ SOEP & CMPO

5. Discussion

If anything, people in Germany are happier the more income their neighbours have!

Is the neighbourhood scale inappropriate?

Are not all neighbours relevant?

Is the theory wrong?

March 29, 2007 G. Knies/ SOEP & CMPO

Definitions: Income MeasuresIncome components Federal

Statistical Office

GSOEP

(CNEF)

market incomes + +

income maintenance transfers/ soc. sec. + +

other regular monetary transfers + +

taxes on income and assets - -

NI contributions, ‘other regular payments’ - -

assumed income from living in owner-occupied housing

+ +

asset income flows + (assumed) +

sick payments + (assumed) /

income of non-profit organisations + /

refunds from health insurers + (assumed) /

March 29, 2007 G. Knies/ SOEP & CMPO

Definitions

Income Measures: Σ (SOEP HHY* HH Pop-Weight)= SOEP National Y

Σ (PLZ total HHY)= Infas National Y Infas National Y = SOEP National Y

Assumptions: distribution of NB Y unaltered through three

additional income componentsPer capita incomes:

HHY/HH size = Σ NB HHY/ NB population