NEIGHBORHOOD EFFECTS

1
NEIGHBORHOOD EFFECTS

• BY Steven N. Durlauf
• Ania Bonarska Okafor Luke Emeka
• Development Workshop, 2007

2
Outline

• Introduction
• Review of theories of neighborhood effects
• Econometrics
• Empirical studies
• Evidence of neighborhood effects
• Critiques by the Author
• Critiques of the approaches in general
• Can we measure neighbourhood effects and does it
  matter?
• Conclusions

3
INTRODUCTION

• Social determinants of individual behaviour and
  aggregates outcomes from the angles of
• Role model effects
• Peer group influence
• Produce imitative behavior
• due to
• Psychological factors
• Interdependencies
• in constraints and information transmission

4
INTRODUCTION

• Poverty traps argument
• Reinforcement of changes in private incentives
  argument

5
REVIEW OF THEORIES OF NEIGHBORHOOD EFFECTS
6
Review of theories of neighborhood effects

• Theoretical models answer two basic questions
• Individual decision making
• Memberships in a neighborhood
• Choice within neighbourhoods
• Individual payoff function V(.)

(1)
7
Review of theories of neighborhood effects Contd.
endogenous neighborhood effects taken as
average then Setting 0 then individual
decisions will solve
(2)
(3)
8
Review of theories of neighborhood effects Contd.

• Neighborhoods models and complementarity between
  choices
• For the payoff function described in three
  complementarity implies that

9
Review of theories of neighborhood effects Contd.

• Propositions
• Number of equilibria in the Brock and Durlauf
  (2001a) model of binary choice with neighborhood
  effects
• Limiting behavior of average choice in the
  Glaeser, Sacerdote and Scheinkman (1996) model
• Average preferences in the Föllmer (1974) model
  of interdependent utility functions
• Stratification in the Bénabou (1996a) model of
  endogenous neighborhoods
• Neighborhood size and direct connection
  probabilities

10
ECONOMETRICS
11
Econometrics

• Estimation of neighborhood effects via
• Levels and changes in economic stratification and
  segregation
• Causal mechanisms
• Identification considering the cross

(4)
12
Econometrics contd

• When beliefs are rational then
• Substituting (5) into (4), the individual choices
  may be expressed in terms of observables via

(5)
(6)
13
Econometrics contd

• Collinearity problem because of the presence of
  Xn(i) and Yn(i) in equ. (6)
• Reflection problem as formulated by Manski (1993)
• Nonlinear-in-means neighborhood model of the
  form

(7)
14
Econometrics contd

• where is invertible and
  . .The
• self-consistent expected average choice in a
  neighborhood is determined by
• where
• 
  (8)

15
Econometrics contd

• Self-selection bias. The assumption that is
• 
  unappealing
• Unobservable and sibling data
• - consider pair of siblings s in a given family
  I. Ignoring endogenous effects, individual
  behaviour may be described by

(9)
16
Empirical studies

• Ethnography
• Experiments
• Econometric studies
• Correlation studies
• Quasi-experiments
• Aggregate studies

17
Evidence on neighborhood effects

• Classroom effects
• Peer groups

18
Evidence on neighborhood effects

• Social capital
• Segregation
• Social attitudes
• Home ownership
• Geography and social customs

19
Critiques by the Author

• Deep identification problems
• Neglected issues
• House market dynamics
• Capital gains
• Future neighborhood composition effects
• More attention on microeconomic foundations

20
Critiques of approaches in general

• Inadequate attention to physical and
  institutional characteristics
• bounded rationality limitation of knowledge and
  computational capacity
• Weak conception of neighborhood eg neighborhood
  is not fixed and cannot be studied in isolation
• Complex conceptualization of neighborhood could
  get lost

21
Critiques of approaches in general

• Selection bias eg residential sorting housing
  stock and economic base
• Inadequate account of the influence of
  self-identity and passion on decision making
  process
• Preferences are not always primitive and stable
  the influence of institutions
• Difficult of determining boundary for
  neighborhood study distance weighting rather
  than fixed boundary

22
Critiques of approaches in general

• Neighborhood might mean more to some people than
  others low income communities and advantaged
  areas
• No consensus emerged about which characteristics
  affected which outcomes
• Inconsistency of findings provide no clear basis
  for policy formulation and implementation

23
Neighborhood Effects

• Can we measure them
• and does it matter?
• Ruth Lupton
• London School of Economics, 2003

24
Whats wrong with Neighborhood Effects research?

• Quantitative research is not complementary with
  qualitative research on neighborhoods
• Area-based policies should not be dependent on
  existence of neighborhoods effect

25
Conclusions

• Well-developed theoretical literature
• Econometrics literature still in nascent stage
• Large empirical literature
• Suggestions for future research?
• Structural models
• More collaboration between quantitative and
  qualitative research

26
Thank you