Lauren Coughlin
November 5, 2003
ECO-043
The Uninsured Population
Introduction
In
recent years the
Model
The percentage of the
population that is uninsured is directly related to the unemployment rate, per
capita income, population, political party of the governor, and percentage of
nonwhites. This translates into the
economic model:
Percent
Uninsured = b 0 + b
1*Unemployment Rate + b2*Population + b3*Minority + b4*Per Capita Income + b5*Democrat + e.
There is a strong economic relationship between the percent
uninsured (the dependent variable) and two specific independent variables –
unemployment rate and per capita income.
[See graphs below]

Figure 1 Figure
2
The unemployment rate and percent uninsured produce a positive relationship, and per capita income and percent uninsured produce a negative relationship (both as expected).
Data
To
estimate this suggested relationship, I used data obtained from the “2001
County and City Data Book” and “The 2000 Census.” My dependent variable is the percent of Americans uninsured in
2000 (UNINS). This is measured as a
percent for each of the 50 states (numbered in alphabetical order 1-50). Figure 3 below shows the histogram of
the rates for all 50 states.
The five independent variables
included in the regression were state unemployment rate (UNEMPRATE), state
population (POPULATION), percent of nonwhites (MINORITY), income per capita (INCOMEPC),
and political party affiliation of the governor (DEMOCRAT).

Figure 3
Results
The estimated
regression for the percentage of the population without insurance was best
represented in linear form. The final equation is as follows:
Percent Uninsured = b 0 + b 1*UNEMPLOYMENT RATE + b2*POPULATION + b4*PER CAPITA INCOME + e.
This regression produces the
following results:
|
Variable |
Coefficient |
Std.
Error |
t-Statistic |
Prob. |
|
UNEMPRATE |
2.042990 |
0.502487 |
4.065760 |
0.0002 |
|
POPULATION |
1.11E-07 |
7.52E-08 |
1.475513 |
0.1472 |
|
DEMOCRAT |
-1.678600 |
0.958033 |
-1.752132 |
0.0867 |
|
C |
5.151214 |
1.986943 |
2.592532 |
0.0129 |
|
R-squared |
0.323356 |
Mean dependent var |
13.07083 |
|
|
Adjusted R-squared |
0.277221 |
S.D. dependent var |
3.789569 |
|
|
S.E. of regression |
3.221755 |
Akaike info criterion |
5.257385 |
|
|
Sum squared resid |
456.7071 |
Schwarz criterion |
5.413319 |
|
|
Log likelihood |
-122.1772 |
F-statistic |
7.008933 |
|
|
Durbin-Watson stat |
2.032506 |
Prob(F-statistic) |
0.000592 |
|
On average, states with higher unemployment rates have a
greater percentage of uninsured individuals.
A 1 percent increase in the
unemployment rate generates a 1.88 percent increase in the state’s uninsured
population. Insurance status is
also closely associated with level of income.
A $5,000 increase in per capita
income will lower the uninsured rate by –1.52 percent. States with high per capita incomes have
individuals who have more wealth and are thus better able to purchase
health insurance. Connecticut has the highest per
capita income at $40,702, and also one of the lowest unemployment rates at 8.2
percent.
Conclusion
My final
regression supports the hypothesis stated.
The percentage of the population that is uninsured can best be reduced
by a decrease in the unemployment rate, and an increase in per capita
income. States should therefore work to
lower the unemployment rate, which will in turn raise per capita income, and
have an even stronger effect on the level of insurance held by
individuals. To my surprise, DEMOCRAT
was found to be insignificant. Because
democrats are known for being more liberal, I would have predicted that a state
with a democratic governor would be more likely to establish programs to
provide insurance for the uninsured.