Graduation Year


Document Type




Degree Name

Master of Arts (M.A.)

Degree Granting Department


Major Professor

Michael J. Lynch, Ph.D.

Committee Member

John Cochran, Ph.D.

Committee Member

George Burruss, Ph.D.


Criminology, Economics, Social Structure, Public Policy


The present study was designed to examine the variations in criminal justice expenditures across states in relation to crime, measures of political party membership, and several control variables that also attempt to explain both property and violent crime. The year, 2009, was chosen for the analysis. Data in the present study were collected by Olugbenga Ajilore (2016) for the year 2009 and supplemented with other state level data. The Ajilore dataset is one of the few datasets that has reliable criminal justice expenditure data across states, which is also disaggregated by type. Criminal justice expenditure data is actually quite difficult to collect across states and is not widely available across states particularly over consecutive years/time.

The dependent variables in the current study is the crime rate, which is separated into two categories; violent and non-violent crime. Both variables are important and essential in understanding the effects of police expenditures and political influences. The independent variables are correctional direct expenditures, judicial and legal direct expenditures, police protection expenditures, state legislative composition, state control, and governor’s party. Each of these variables either measures the level of expenditures on crime control, or measures factors that may influence the level of expenditures on crime control. The control variables are imprisonment, population age, unemployment rate, poverty rate, education, and foreign born. These six control variables are utilized to accurately account for the other possible factors leading to the effect of police expenditures on crime.

An OLS regression of each criminal justice expenditure on crime was conducted in three models: expenditure/threat Hypothesis Models; expenditure/political party model, and reduced form models. Three equations were estimated for each model to help assess the effects of the independent and control variables on property and violent crime independently. The criminal justice expenditures were used in separate models due to collinearity. Models for total criminal justice expenditures were also estimated to address collinearity between individual criminal justice expenditure measures.

The study found that though Republican states increase criminal justice expenditures, this does not deter or decrease crime. The minority threat is also lightly supported in relation to politics and crime which lacks evidence to support the claim of Blalock’s minority threat hypothesis. There is an economic threat that can be seen in the reduced crime models that may indicate that there is in fact a power threat with Republican states. These findings display evidence of social control through politicians, mostly Republicans as the reduced crime models show an increase in poverty and criminal justice expenditures as crime increases. Lastly, the deterrent theory was seen to fail in this study as this research revealed that there is a positive relationship between politics and crime through criminal justice expenditures, specifically police expenditures.