1. Introduction

Overview

The share of black population in the county. This interests me because I am interested in seeing if ethnicity plays a role in housing prices. To see if there is an obvious show case of red lining.

Hypothesis Development

Does a relationship exist between the population of black people have an affect on housing prices at the county level? Null Hypothesis: average black population has NO effect on median housing prices at the county level. Alternative hypothesis: average black population has an effect on the median housing prices at the county level.

2. Empirical Framework

Data Description

This Data is a demographic data representation. The location for the collection of the all data collected is United States Counties. The years being observed are 2013-2017. Home affordability is the dependent variable. Main explanatory variable is pop_black, to get the total percentage of all races to each location by getting total and dividing by the total population giving the sum.

View Data

## # A tibble: 6 x 2
##   Home_Affordability below_poverty
##                <dbl>         <dbl>
## 1               2.59        0.0472
## 2               3.46        0.0325
## 3               2.68        0.0835
## 4               2.43        0.0395
## 5               2.58        0.0563
## 6               2.25        0.0956


The home affordability is the price divided by the household income. The below poverty data is the total data of the population divided by poverty.

3. Descriptive Results

5-point summary

Statistic N Mean St. Dev. Min Max
Home_Affordability 3,218 2.832 1.137 0.673 11.820
below_poverty 3,220 0.047 0.027 0.000 0.204
no_hs 3,142 0.094 0.043 0.008 0.355
hs 3,142 0.237 0.055 0.050 0.451
bach_degree 3,142 0.094 0.038 0.015 0.334
bachelors 3,142 0.225 0.057 0.081 0.644
grad_degree 3,142 0.050 0.029 0.000 0.295
pop_hispanic 3,220 0.113 0.193 0.000 1.000
pop_white 3,220 0.749 0.231 0.000 1.000
pop_black 3,220 0.090 0.145 0.000 0.869
speak_english 3,142 0.853 0.113 0.033 0.972
speak_spanish 3,142 0.060 0.097 0.000 0.865
married 3,142 0.421 0.059 0.139 0.627


As you can see the min, max, and mean of each the data points, as shown by the no pop_black in some of the counties. Which shows we have to add a small constant to the variable

Histogram


Both of the data points reflect a negative price for the pop_black skewed

Histogram after Log-Transformation


After running it through log transformation the house proce is now following a normal distribution and the pop_black still has a negative skewed represented

Correlation Plot


The Correlation Plot represent reflects the that pop_hispanic and speak_english appear to be multicollinear, as well as married and bachlors. Indicating that you might not wanted these variables included with the others.

4. Map Results


The results of the map represent that the home affordability in the west coast and parts of the Florida are signficanlty higher than others. But the west coast is generally higher in price verses the mid-west regions

5. Regression Model Results

Effect of Local Poverty on Housing Affordability
Dependent variable:
log(Home_Affordability + 1)
(1) (2) (3)
log(below_poverty + 1) -1.693*** -1.562*** -0.528*
(0.189) (0.215) (0.276)
log(no_hs + 1) -0.383*** -0.027
(0.124) (0.146)
log(hs + 1) -1.670*** -1.839***
(0.093) (0.110)
log(bachelors + 1) 1.155*** 2.302***
(0.092) (0.190)
log(married + 1) 1.676***
(0.240)
log(pop_black + 1) -0.041
(0.057)
log(pop_hispanic + 1) -0.168**
(0.069)
log(pop_white + 1) -0.039
(0.076)
Constant 1.367*** 1.516*** 0.694***
(0.009) (0.030) (0.135)
Observations 3,140 3,140 3,140
R2 0.025 0.233 0.248
Adjusted R2 0.025 0.232 0.246
Residual Std. Error 0.235 (df = 3138) 0.209 (df = 3135) 0.207 (df = 3131)
F Statistic 80.424*** (df = 1; 3138) 238.074*** (df = 4; 3135) 129.052*** (df = 8; 3131)
Note: p<0.1; p<0.05; p<0.01


Summary of Main Findings

So after reviewing the table you are able too to see small negative effects with a 90% confidence interval. But if raise the the population you are able to see a .4% drop median house prices

Conclusion

Summary of Findings and Policy Implications

After reviewing the findings you can see that we are able to accept the Null hypothesis of density of the black/African American population had NO effect on the median housing prices in the county. Which had no significant effect on the data to prove that information regarding housing prices correlating to the population increase of the black population. The data doesn’t provide us with with clear picture of the actual data points to either prove or disapprove that there are disparities in this population set more data is required to gain a sufficient to make a final determination. A more thorough investigation will be needed to determine if any other variations are at play to find why the root cause has the been leading to a negative skew on the graphs