Economics regression analysis paper
For U.S. states (minimum 30 states) find cross sectional data on poverty rate, state transfer-payment spending per capita and state unemployment rate.
a. Present data in a table (relegated to an appendix);
See appendix
b. State expectations and set up hypotheses and speculate on type I and type II errors of the coefficients wherever possible,
Transfer payments are fees that the government or business pays to individuals for free to increase their income and purchasing power. Most of the government's transfer payments are characterized by welfare expenditures. So our first expectation is that the more state transfer payments, the lower the poverty rate in the state. However, we still make two mistakes. One is that our assumptions are wrong, but the statistical calculations show that we have no good reason to reject the null hypothesis that the more the state government transfers payment, the more poverty there is in the state. Another mistake is that our null hypothesis may be correct, but the statistical results show that we may need to reject the null hypothesis. For example, in the long run, the government's welfare policy can not completely change the state of poverty, but it is counterproductive, more and more People expect the help of the government and do not change the status quo from their own efforts.
If a person is unemployed, it means that he has no material source. It can be seen that the higher the unemployment rate, the higher the poverty rate. However, we still make two mistakes. One is that our null hypothesis is wrong, but the statistical calculations show that we have no good reason to reject the null hypothesis that the higher the unemployment rate, the lower the poverty rate. Another mistake is that our null hypothesis may be correct, but the statistical results show that we may need to reject the null hypothesis. For example, although a person is unemployed, he can still be self-sufficient to meet the individual's living needs.
c. Run two regressions:
1. One with poverty line being the dependent variable and State-unemployment rate being the independent variable; present the result as it is done conventionally including the supporting information such as statistics, R-square and the number of observations,
Show on a graph the estimated line and pick an observation to show the Total Error, Estimated Error and the Residual Error.
2. Another regression with poverty line being the dependent variable and the other two variables as the independent variables and present results conventionally.
d. Interpret your results (of the second equation) on its compliance with expectations, tests of robustness and the parameter statistics,
According to the EXCEL output, the regression coefficient of the variable ‘unemp’ is 1.7971, the value of the T statistic is 3.9096, and the corresponding p value is less than the required significance level of 0.05, indicating that the state unemployment rate has a significant positive impact on the state's poverty rate.That is to say, when the other variables are kept constant, the state's poverty rate increases by an average of 1.7971 units for each unit of unemployment rate increase. The regression coefficient of the variable ‘transfer’ is -0.0007, the value of the T statistic is 0.0002, and the corresponding p value is 0.0093.It shows that the state government transfer payment has a significant negative impact on the state's poverty rate. That is to say, when the other variables are kept constant, the state's poverty rate is reduced by an average of 0.0007 units for each unit of transfer payment. As can be seen from the above, the output of the regression model is in line with our expectations.
e. Speculate on the unexpected results and make recommendations for improving the model and policy within the context of results and theory,
The higher the health of the family members, the greater the probability of participating in the labor market, the higher the probability of obtaining employment opportunities, the higher the wage level, and the easier the family is in a virtuous circle. It can be seen that there may be a certain relationship between the physical health of residents and the poverty of urban residents. That is to say, there are many factors affecting state poverty, so we can add other factors that influence the poverty rate to optimize the model.
According to the context of results and theory, we make the following suggestions.The government needs to actively create employment opportunities to link social welfare to employment services. In order to reduce the state poverty rate, government can increase transfer payments, such as social security benefits, pensions, unemployment benefits, benefits, and various subsidies.
f. Include the computer output as an appendix.
Appendix:
|
State Name |
state transfer-payment($) |
poverty rate(%) |
Unemployee rate(%) |
|
Alabama AL |
4,654.0 |
19.3 |
6.1 |
|
Alaska AK |
14,865.5 |
11.2 |
6.5 |
|
Arizona AZ |
4,043.3 |
18.2 |
6.1 |
|
Arkansas AR |
5,428.6 |
18.9 |
5 |
|
California CA |
5,703.2 |
16.4 |
6.2 |
|
Colorado CO |
4,957.5 |
12 |
3.9 |
|
Connecticut CT |
7,110.2 |
10.8 |
5.7 |
|
Delaware DE |
7,986.3 |
12.5 |
6.9 |
|
Florida FL |
3,310.9 |
16.5 |
5.5 |
|
Georgia GA |
3,473.9 |
18.3 |
6 |
|
Hawaii HI |
8,516.3 |
11.4 |
3.6 |
|
Idaho ID |
4,071.9 |
14.8 |
4.2 |
|
Illinois IL |
4,955.7 |
14.4 |
6 |
|
Indiana IN |
4,226.1 |
15.2 |
4.8 |
|
Iowa IA |
5,624.4 |
12.2 |
3.8 |
|
Kansas KS |
4,501.6 |
13.6 |
4.2 |
|
Kentucky KY |
6,469.7 |
19.1 |
5.3 |
|
Louisiana LA |
5,407.5 |
19.8 |
6.3 |
|
Maine ME |
5,716.3 |
14.1 |
4.4 |
|
Maryland MD |
5,670.4 |
10.1 |
5.1 |
|
Massachusetts MA |
7,654.2 |
11.6 |
4.8 |
|
Michigan MI |
4,756.3 |
16.2 |
5.4 |
|
Minnesota MN |
5,545.4 |
11.5 |
3.7 |
|
Mississippi MS |
5,350.2 |
21.5 |
6.4 |
|
Missouri MO |
4,246.2 |
15.5 |
5 |
|
Montana MT |
5,885.7 |
15.4 |
4.2 |
|
Nebraska NE |
4,301.9 |
12.4 |
3 |
|
Nevada NV |
3,601.8 |
15.2 |
6.8 |
|
New Hampshire NH |
4,890.4 |
9.2 |
3.4 |
|
New Jersey NJ |
6,580.9 |
11.1 |
5.8 |
|
New Mexico NM |
7,286.8 |
21.3 |
6.5 |
Source: https://cdn.americanprogress.org/wp-content/uploads/2016/02/23080039/StateofStates-fullreport2.pdf
http://www.dlt.ri.gov/lmi/laus/us/annavg.htm
https://www.usgovernmentspending.com/year_spending_2015NMdn_19ds2n#usgs302
Computer output for the first regression:
Computer output for the second regression:
1
Yishu Huang
4127
0
31
7955
2
9096
3
0002
0
4596
0
0007
0
1
797
1
2837
9
ˆ
2
.
R
N
.
-
.
t
)
.
(
)
.
(
transfer
.
-
unemp
.
.
Y
i
i
i
=
=
=
+
=
2487
0
31
0992
3
5016
0
5546
1
7751
6
ˆ
2
.
R
N
.
t
)
.
(
unemp
.
.
Y
i
i
=
=
=
+
=