CC BY
6
and instrumental
Section 2. Mathematical methods of economics
https://doi.org/10.29013/EJEMS-20-4-11-16
Lang Shen, California, United States E-mail: shenlang.china@outlook.com
Jinan Liu,
Supervisor: Dr Ph D, Director of Merck Company
FINANCIAL WORRIES OVER HOUSING COST AMONG ADULTS IN2017
Abstract
Aim: This study aims to:
1. Examine the predictors of adults' financial worries over Credit Card Payments in 2017;
2. Build a predictive model for adults' financial worries over housing costs among adults by using a logistic regression model.
Method: Data in NHIS, The National Health Interview Survey, was used in this study. The number of people who are worried about housing costs was calculated. We run a generalized linear model to examine all the predictors. We randomly selected all the participants and put them into two groups: training data and testing data. Then we run a logistic regression model by using the training data. Optional cutoff, misclassification, receiver operating characteristic, sensitivity, and specificity were calculated.
Results: Out of 26,025 participants, 20,856 of them (80.14%) worry about their housing costs and 5,169 of them (19.86%) do not worry about their housing costs.
The logistics regression shows the older population tends to worry less about housing costs. Older populations may have higher income or savings. Females are more likely to worry about housing costs than males, which indicates some extent of gender inequality. The non-Hispanic population is 58% less likely to worry than the Hispanic population. Compared with other races, the black population is 31.5% more likely to worry about housing costs. Compared with people in the South and the West, people in the Midwest are less likely to worry about housing costs. Compared with people who are not working, people who are working are 41.2% less likely to worry about the cost of housing.
The area under the ROC curve is 0.6285. The optional cutoff time is around 0.55. The misclassification error is 0.1953, the sensitivity is approximately 0.24%, and the specificity is almost 100%.
Conclusions: In this study, we determined that there are many predictors for the financial worries over housing costs among adults in 2017. This research can help find the features of the population who are worrying about the cost of housing.
Keywords: Financial worry, housing, cost, logistic regression, model, prediction.
1. Introduction
Nowadays, with the increase in the cost of living, more and more people start to worry about the housing cost among adults. Housing costs are mainly divided into three parts, rent, mortgage, and other housing costs. Other housing costs can include several things, such as internet, electricity, gas, trash, water, sewer, and parking. Renters have to pay the rent and the property owners have to pay the mortgage. Financial worries over housing costs are a universal topic whether you are a renter or a homeowner.
I, as an international student attending U.S. university, have to rent a house. Housing costs for me include an internet bill, electricity, gas, trash, water, sewer, and parking. I need to pay these fees monthly. My monthly rent is approximately $1,000. While my other monthly housing costs are around $250. In the winter months, I pay more for electricity and have a higher water bill in the summer months. As a student, I always feel worried about my expensive housing costs.
The constant financial worry over housing costs also caught the publics' attention because it was taking over large cities. "In cities such as San Francisco and New York, a consistent 2.5% annual appreciation above inflation in housing prices and rents has resulted in a quadrupling of housing costs since 1950 and homeless-ness rates not seen since the Great Depression" (Derek Fidler [1]). It shows the housing cost is increasing without faltering as time goes by. As a result, adults' worries over housing cost also increases. According to The Legislative Analyst's Office (LAO [2]), the housing supply being less than the demand is one of the reasons for high housing costs. Especially in California, considered a beautiful place to live, more and more people want to live here, but the land is limited, which results in high housing costs (Chas Alamo [2]).
One article in GALLUP states 25% of homeowners and 49% of renters are very or moderately worried about not being able to pay housing costs. This article also shows people with lower-income are more likely to worry more about housing costs than people with middle-wage income and upper-wage income are (Jeffery [3]).
This research aims to study the financial worries of people. It determines the predictors of financial worries over housing costs among adults in 2017 and creates a model for financial worries over housing costs among adults in 2017 by using a logistic regression model.
2. Data and Methods
Data:
The main source of the information comes from The National Health Interview Survey (NHIS). The NHIS started in 1957 and collect many types of data covering a wide range ofhealth topics. It collects data from individual household interviews. The NHIS is one of the programs of the NCHS, National Center for Health Statistics. And the NCHS is one part ofthe center for Disease Control and Prevention (CD C).
The National Health Interview Survey data in 2017 was used in this research paper. URL: https://www.cdc.gov/nchs/nhis/about_nhis.htm
Optimal Cutoff for Binary Classification maximizes a given criterion.
Misclassification Error is the incorrectly classified part of all events in a given probability cutoffscore.
Sensitivity is the proportion of positive results out of the number of true positive samples.
Specificity is the true negative divided by all the negative results.
Model:
Logistic regression models were used to calculate the predicted risk. A logistic regression model is one
of the generalized linear models. This model predicts linear model. Then we put the generalized linear
the results from many sets of variables. model to the test data to check the accuracy.
The logistic regression model can be expressed Variable: as the formula: The outcome variable is, "How worried are you
ln(P/P - 1) = + + p2*X2 + ......+ fi*Xn right now about not being able to pay your rent, mort-
We spilt the data into two parts, train data and gage, or other housing costs?" (ASIHCST) Table 1 lists
test data, and we use train data to run a generalized all the variables that this research takes into account.
Table 1.- Variables in this research
SEX 1: male 2: female
1 2
AGE_P Age>18
HISPAN_I 0: Multiple Hispanic 1: Puerto Rico 2: Mexican 3: Mexican-American 4: Cuban/Cuban American 5: Dominican (Republic) 6: Central or South American 7: Other Latin American, type not specified 8: Other Spanish 12: Not Hispanic/Spanish origin
R_MARITL 1: Married - spouse in household 2: Married - spouse not in household 3: Married - spouse in household unknown 4: Widowed 5: Divorced 6: Separated 7: Never married 8: Living with partner 9: Unknown marital status
MRACRPI2 1: White 2: Black/African American 3: Indian (American), Alaska Native 9: Asian Indian 10: Chinese 11: Filipino 15: Other Asian* 16: Primary race not releasable** 17: Multiple race, no primary race selected
1 2
REGION 1: Northeast
2: Midwest
3: South
4: West
DOINGLWA 1: Working for pay at a job or business
2: With a job or business but not at work
3: Looking for work
4: Working, but not for pay, at a family-owned job or business
5: Not working at a job or business and not looking for work
7: Refused
9: Don't know
ASIHCST 1: Very worried
2: Moderately worried
3: Not too worried
4: Not worried at all
7: Refused
8: Not ascertained
9: Don't know
3. Results:
Out of 26.025 participants, 20.856 of them
(80.14%) worry about their housing costs and 5.169
AGE P
of them (19.86%) do not worry about their housing
costs.
I ispah n<
MRRICD
W>«>e
MdA^sî
WorNL» lace warr
Figure 1. Financial Worries over hoursing cost. Matrix of correlations between variables
The (figure 1) shows these variables, age, male, hispan_no, married, white, are all negatively correlated with worries of housing costs. The variable, black, are positively correlated with housing wor-
ries, which indicates race inequality may still exist. Compared with people living in the Midwest, people in the South and the West are more worried about housing costs.
Table 2. - Logistic Regression for housing cost
Estimate Std. Error z value Pr(>|z|)
(Intercept) 0.6017805 0.1034199 5.819 5.93e-09
AGE_P -0.0052475 0.0009782 -5.364 8.13e-08
Male -0.1825180 0.0323142 -5.648 1.62e-08
HISSPAN_No -0.8575671 0.0433944 -19.762 <2e-16
MARRIED -0.2293469 0. 332243 -10.291 <2e-16
White -0.2293469 0.061452 -3.813 0.000137
Black 0.3145143 0.0723105 4.349 1.36e-0.5
Midwest -0.2177289 0.522893 -4.164 3.13e-05
South -0.0748782 0.470457 -1.592 0.111473
West -0.0475053 0.0511635 -0.928 0.353148
Working -0.5659221 0.0607217 -9.320 <2e-16
Nwor NLor -0.6384145 0.0649852 -9.824 <2e-16
Table 2. Shows all the variables are statistically significant except South and West
ROC Curve
0.00 0.25 0.50 0.75 1.00
1-Specificity (FPR) Figure 2. ROC in the testing sample for Logistic Regression
The area under the ROC curve is 0.6285. The tion error is 0.1953. the sensitivity is approximately optimal cutoff time is around 0.55. The misclassifica- 0.24% and the specificity is almost 100%.
Table 3.- Odds Ratios according to the Logistic Regression
OR Worry increase
Age P 0.693 -30.7%
Male 0.828 -17.2%
Hispan No 0.420 -58.0%
Married 0.664 -33.6%
White 0.820 -18.0%
Black 1.315 31.5%
Midwest 0.819 -18.1%
South 1.024 2.4%
West 1.050 5.0%
Working 0.588 -41.2%
NworNLor 0.553 -44.7%
Discussion
Out of 26,025 participants, 20,856 of them (80.14%) worry about their housing costs and 5,169 of them (19.86%) do not worry about their housing costs.
The logistics regression shows the older population tend to worry less about housing costs. Older population may have higher income and savings. Females are more likely to worry about housing costs than males, which indicates some extent of gender inequality. The non-Hispanic population is 58% less likely to worry than the Hispanic population. Compared with other races, the black population is 31.5% more likely to worry about housing cost. Compared with people in the South and the West, people in the Midwest are less likely to worry about housing costs. Compared with people are not working, peo-
ple who are working are 41.2% less likely to worry about housing costs.
There are still many more factors the can be considered in this research, such as family background, level of education, and job type. One article in GALLUP split people into home owners and renters. It states 25% of home owners and 49% of renters are very or moderately worried about not being able to pay housing costs (Jeffery [3]). With more factors being taken into account, the results will be more accurate.
All in all, we determined many predictors for financial worries over housing cost among adults in 2017. The gender inequality and racial inequality showed in the research should be called to attention. When people are less worried about housing costs, they can spend more time being happy.
References:
1. Derek Fidler, Hicham Sabir. (2019. Jan 09). The cost of housing is tearing out society apart. Retrieved September 09, 2020. From URL: https://www.weforum.org/agenda/2019/01/why-housing-appreci-ation-is-killing-housing
2. Chas Alamo, Brain Uhler. (2015. March 17). California's High Housing Costs: Causes and Consequences. Retrieved September 09, 2020. From URL: https://lao.ca.gov/reports/2015/finance/housing-costs/ housing-costs.aspx
3. Jeffrey M. Jones. (2016. April 27). U. S. Renters Worry More Than Homeowners About Housing Costs. Retrieved September 09, 2020. From URL: https://news.gallup.com/poll/191102/renters-worry-homeowners-housing-costs.aspx
4. NHIS - About the National Health Interview Survey. (2019. January 16). Retrieved September 09, 2020. From URL: https://www.cdc.gov/nchs/nhis/about_nhis.htm
