This post looks at what demographic factors impact credit card debt, percent of consumers with personal loans, and bankruptcy filings. This analysis occurs at the state level. I analyze the impact of credit scores, percent female, median age, median household income, number of people in each race category (e.g. Hispanic, White, Black, Asian, American Indian), educational attainment (e.g. percent high-school or higher, percent bachelor’s degree or higher, and percentage advanced degree), and marriage rate on credit card debt, percent of consumers with personal loans, and bankruptcy filings.
Turning first to credit card debt, I run a regression of credit card balance on the above mentioned variables and find several statistically significant factors. Specifically, a one point increase in credit scores leads to a $34.32 decrease in credit card balance. Furthermore, a one percent increase in percent female leads to a $29,043.45 decrease in outstanding credit card balance. When it comes to median income, a one dollar increase leads to a $0.06 increase in credit card balance. And, the impacts of numbers of Blacks and American Indians on credit card balance are statistically significant, but the magnitude of the impacts are effectively zero.
Figure 1, below, summarizes the results of the regression. The r-squared for this regression, which measures the proportion of the variance in credit card balance that is explained by the independent variables, is 0.75. This suggests that 75% of the variation in credit card balance is explained by the statistical model.
Figure 1. Regression Results of Average Credit Card Debt on Demographic Variables | |||
Independent Variables | Coefficient | P-Value | Signficiant? |
Average Fico Credit Score | -34.32 | 0.005 | Yes |
Percent Female | -29,043.45 | 0.039 | Yes |
Median Age | 24.23 | 0.500 | No |
Median Household Income | 0.06 | 0.000 | Yes |
Number of Hispanics | 0.00 | 0.538 | No |
Number of Whites | 0.00 | 0.152 | No |
Number of Blacks | 0.00 | 0.048 | Yes |
Number of Asians | 0.00 | 0.407 | No |
Number of American Indians | 0.00 | 0.062 | Yes |
Percent High School or Higher | 3,101.56 | 0.547 | No |
Percent Bachelor’s Degree or Higher | -3,530.94 | 0.596 | No |
Percent Advanced Degree | 10,332.22 | 0.293 | No |
Marriage Rate | -6.71 | 0.749 | No |
Turning next to percent of consumers with personal loans, Figure 2, below, shows that median age and median household income are both significant predictors, but the magnitude of the impact is effectively zero. The same goes for the number of Whites and the number of Blacks. What is noteworthy is that a one percent increase in the percent with a high school degree leads to an 88% increase in consumers with a personal loan. This impact is very large and is also statistically significant.
The r-squared for this regression, which measures the proportion of the variance in percent of consumers with personal loans that is explained by the independent variables, is 0.73. This suggests that 73% of the variation in percent of consumers with personal loans is explained by the statistical model.
Figure 2. Regression Results of Percentage of Consumers with a Personal Loan on Demographic Variables | |||
Independent Variables | Coefficient | P-Value | Signficiant? |
Average Fico Credit Score | 0.00 | 0.970 | No |
Percent Female | -0.70 | 0.597 | No |
Median Age | -0.01 | 0.043 | Yes |
Median Household Income | 0.00 | 0.049 | Yes |
Number of Hispanics | 0.00 | 0.174 | No |
Number of Whites | 0.00 | 0.001 | Yes |
Number of Blacks | 0.00 | 0.076 | Yes |
Number of Asians | 0.00 | 0.829 | No |
Number of American Indians | 0.00 | 0.670 | No |
Percent High School or Higher | 0.88 | 0.087 | Yes |
Percent Bachelor’s Degree or Higher | -0.67 | 0.308 | No |
Percent Advanced Degree | 0.27 | 0.775 | No |
Marriage Rate | 0.00 | 0.198 | No |
Finally, turning to the regression results of bankruptcy filings on demographic variables, I find that the number of Whites and the number of Asians are statistically significant, but the magnitude of the effects are basically zero. Nevertheless, the r-squared for this regression, which measures the proportion of the variance in bankruptcy filings that is explained by the independent variables, is 0.85. This suggests that 85% of the variation in bankruptcy filings is explained by the statistical model, which is high.
Figure 3. Regression Results of Bankruptcy Filings on Demographic Variables | |||
Independent Variables | Coefficient | P-Value | Signficiant? |
Average Fico Credit Score | -161.73 | 0.401 | No |
Percent Female | 15,792.67 | 0.944 | No |
Median Age | -153.04 | 0.796 | No |
Median Household Income | 0.03 | 0.885 | No |
Number of Hispanics | 0.00 | 0.113 | No |
Number of Whites | 0.00 | 0.003 | Yes |
Number of Blacks | 0.00 | 0.261 | No |
Number of Asians | 0.01 | 0.018 | Yes |
Number of American Indians | -0.01 | 0.538 | No |
Percent High School or Higher | -16,627.77 | 0.845 | No |
Percent Bachelor’s Degree or Higher | 45,698.33 | 0.678 | No |
Percent Advanced Degree | -97,323.31 | 0.547 | No |
Marriage Rate | -52.10 | 0.881 | No |
This short analysis examines what demographic variables explain credit card debt, percent of consumers with personal loans, and bankruptcy filings. I find that credit scores, gender, race, and educational attainment are all significant predictors, albeit in different statistical models. An analysis of debt, loans, and bankruptcy should include these independent demographic variables.
Source: https://www.experian.com/blogs/ask-experian/consumer-credit-review/, https://www.abi.org/newsroom/bankruptcy-statistics, https://www.chamberofcommerce.org/credit-card-debt-by-state, https://www.experian.com/blogs/ask-experian/research/personal-loan-balance-amount-by-state/, https://www.cdc.gov/nchs/data/dvs/state-marriage-rates-90-95-99-18.pdf, https://en.m.wikipedia.org/wiki/List_of_U.S._states_and_territories_by_median_age, https://www.governing.com/gov-data/census/state-minority-population-data-estimates.html, https://en.m.wikipedia.org/wiki/List_of_U.S._states_and_territories_by_educational_attainment