Skip to main content
. Author manuscript; available in PMC: 2023 Dec 5.
Published in final edited form as: Hous Policy Debate. 2021 Jan 8;31(3-5):670–695. doi: 10.1080/10511482.2020.1847163

Table A1.

Comparison of results from negative binomial and (nonspatial) OLS regression models, Ellis Act petitions, 2011–15.

Negative binomial
(outcome: count of petitions)
OLS
(outcome: logged petition rate)
Coeff. Sig. SE Coeff. Sig. SE

Household income in $1,000 − 0.0052 0.0077 − 0.0001 0.0035
% Vacant  0.0315 0.0200  0.0144 0.0088
Median gross rent in $100  0.0295 0.0306 − 0.0098 0.0148
Median home value in $1,000  0.0013 *** 0.0005  0.0006 *** 0.0002
% Renter occupied units − 0.0350 *** 0.0053 − 0.0090 *** 0.0026
% Households with children − 0.0043 0.0090 − 0.0056 0.0042
% Black  0.0499 *** 0.0078  0.0186 *** 0.0034
% Hispanic  0.0290 *** 0.0087  0.0151 *** 0.0038
% Asian  0.0207 *** 0.0074  0.0084 ** 0.0034
% with college degree or more  0.0591 *** 0.0113  0.0274 *** 0.0051
Poverty rate  0.0114 0.0099  0.0056 0.0047
Unemployment rate  0.0325 0.0357 − 0.0045 0.0163
% without a high school diploma  0.0837 *** 0.0196  0.0271 *** 0.0084
LIHTC units  0.0010 0.0007 − 0.0001 0.0004
Public housing units − 0.0026 0.0016 − 0.0005 0.0005
Housing Choice Vouchers − 0.0035 ** 0.0017  0.0000 0.0007
Newly constructed units  0.0096 * 0.0053  0.0041 0.0026
Constant − 4.2437 *** 0.9902 − 1.8947 *** 0.4240
No. observations 806 806

Note. OLS = ordinary least squares; LIHTC = Low-Income Housing Tax Credit; SE = standard error.

*

p < .1

**

p < .05

***

p < .01.