Measuring the Effectiveness of Political Parties at the State Level in the U.S.

I wanted to determine if Democrats or Republicans are better at governing at the state level.  In order to test this, I run two cross-sectional regressions.  As an independent variable, I collect data from the U.S. Census website to create a percentage of Democrat power for each state over a 10 year period: 1999-2009.  If the Democrats held a majority of the seats in the house, then they would be awarded a 1 for that year.  If the Republicans held a majority, then that state would get a 0 for that year.  If the house was split evenly 50/50, then it would be assigned .5 for that year.  I then take the average for the house over that 10 year period.  I do the same process with the Senate.  I then take the average of the House number and the Senate number and I get my percentage of Democrat control at the state level for each state.  Nebraska does not have party affiliation at the state level so they are excluded from all regressions.

For the first regression, the dependent variable is the sum of total debt and unfunded liabilities per capita for that state.   The argument is that better managed states should look for the best long term fiscal health of the state and keep their debt down.  A poorly run state government will use the promises of debt and unfunded liabilities to buy votes to keep themselves in power and thus run up their debts.  The data for debt and unfunded liabilities comes from the Forbes website.  The debt numbers are as of January 2010.


The results of the regression are below.  The data shows coefficient, then p-value.

Debt plus Unfunded Liabilities	Intercept	Democrat Control
	9793.0 (2.01e-14) ***	2362.9 (0.1)

Adjusted R-squared: 0.03636

The regression image is below:

As the chart shows, there is a slight increase in the regression line as the democrats gain a larger percentage of state power.  However, the p-value for the regression is above 10% (barely).  Thus, no conclusions can be made concerning Democrat control over the state House and Senate and the state's debt burden.

For the second regression, I use internal migration as the dependent variable.  This is taken from the 2010 U.S. census.  The data shows whether more people moved into the state or left the state due to internal migrations between 2000 and 2010.  The argument is that better run states should experience an inflow of people while poorly run states should experience an outflow.  I believe this is one of the best ways to judge the quality of government because it reflects the voluntary movement of people.  The independent variable is the same.

The regression results are:

Internal Migration	Intercept	Democrat Control
	0.02689 (0.0131) *	-0.03412 (0.0420) *

Adjusted R-squared: 0.0656

The results show that a state entirely controlled by the Republicans will have an inflow of people of 2.69%, while a state entirely under the control of the Democrats will have an outflow of .7% (.02689 - .03412).  The result is statistically significant at the 5% level.

The chart for this regression is below:

References:

http://www.forbes.com/lists/2010/44/debt-10_Global-Debt-Crisis_StateName.html
http://www.census.gov/