EXPLANATION OF THE LEVEL OF LOCAL GOVERNMENT POLICE EXPENDITURES IS GENERALLY DRAWN FROM THE RESULTS OF EMPIRICAL ANALYSES. HOWEVER, THE ASSUMPTIONS OF THE UNDERLYING STATISTICAL MODEL ARE NOT REALISTIC FOR THE POLICE FUNCTION, AND AS A CONSEQUENCE THE REGRESSION RESULTS DO NOT PROVIDE USEFUL CLUES AS TO THE REASONS FOR EITHER THE RAPID GROWTH IN POLICE SPENDING AND EMPLOYMENT OR THE WIDE VARIATION IN PER CAPTIA EXPENDITURES AMONG CITIES. THE MODEL USED HERE IS DEVELOPED BY COMBINING AN EXISTING BODY OF PUBLIC FINANCE RESEARCH WITH A GROWING BODY OF LABOR ECONOMICS RESEARCH DEALING WITH LOCAL AND STATE GOVERNMENT, PARTICULARLY WITH ESTIMATING THE EMPLOYMENT DEMAND FUNCTIONS AND THE DETERMINANTS OF THE WAGE RATE. THE ESTIMATION INVOLVES USING THREE STRUCTURAL EQUATIONS--COMPENSATION, EMPLOYMENT, AND CRIME--IN ORDER TO IDENTIFY THE DETERMINANTS OF INTERCITY VARIATIONS IN POLICE COMPENSATION RATE AND EMPLOYMENT. RESULTS SUGGEST THAT IMPORTANT EXPLANATORY VARIABLES HAVE BEEN EXCLUDED FROM EARLIER EXPENDITURE DETERMINANTS STUDIES (E.G., EFFECT OF UNIONIZATION AND EDUCATION ON POLICE WAGES AND THE INFLUENCE OF NUMBERS OF POLICE ON CRIME). THE STRUCTURAL MODEL ALSO FACILITATES INTERPRETATION OF THE CONTRIBUTIONS OF SUPPLY AND DEMAND VARIABLES TO PUBLIC EXPENDITURE LEVELS AND ENABLES RESEARCHERS TO TRACE EXTERNAL SHOCKS THROUGH THE MODEL OF THEIR ULTIMATE EFFECTS ON EMPLOYMENT, COMPENSATION, AND EXPENDITURE LEVELS. STUDY DATA, THE MATHEMATICAL MODEL, AND REFERENCES ARE PROVIDED. (AUTHOR ABSTRACT MODIFIED--DAG)