An Ordered Probit Model for Estimating Racial Discrimination through Fair Housing Audits. CANOPY ROYCHOUDHURY and ALLEN C. GOODMAN

1
An Ordered Probit Model for Estimating Racial
Discrimination through Fair Housing
Audits.CANOPY ROYCHOUDHURY and ALLEN C. GOODMAN

• Wayne State University
• 1993
• Presented by Mahetem Gessese

2

• Introduction
• Two persons (auditors) with matched
  characteristics such as age, sex, and income
• Equally eligible for the housing unitDirect
  methods of measuring of discrimination
• Except one is black and one is white.
• The paper examines the severity of differential
  treatment
• The audit was done in metropolitan Detroit.

3

• Auditing technique, average and marginal level of
  discrimination
• John Yinger (1986)
• Tai a bRi eai
• Where a, is the audit i, the individual auditor,
  T is the treatment variable, R is binary variable
  for minority status and e, is the random error.
• Yinger argues that the OLS estimator b for beta
  is unbiased
• However the standard error of b is biased
• He recommends a paired difference- of-means test
  that would remove the bias from the standard
  error of b
• Blacks in Boston were informed 30 fewer units
  

4

• The model
• Dj a0 aj Wj sumBij sumvikCiCk gYt
  eij
• i1,2,.k
• j1,1..10
• Y Bx e

5
(No Transcript)
6
Existing theories of Discrimination1) The agent
prejudice 2) The customer prejudice 3) The
perceived preference4) The rip-off hypothesis
7
Data569 observations317 collected randomly252
complaints-driven
8
Regression results

• Dj a0 aj Wj sumBij sumvikCiCk gYt
  eij
• plim(? j) ? j 1(r2/r2w)/(1- ?j)

9
Conclusion
10
Conclusion

• Widespread discriminatory 1980 1990
• Marginal discrimination higher than average
  discrimination