IT WAS HYPOTHESIZED THAT SEVERAL INDEPENDENT VARIABLES WOULD AFFECT POLICE SALARIES EITHER THROUGH THEIR DIRECT IMPACT UPON SUPPLY AND DEMAND OR INDIRECTLY THROUGH THEIR INFLUENCE AS MARKET IMPERFECTIONS. THESE INDEPENDENT VARIABLES ARE: CRIME RATE; MEDIAN FAMILY INCOME; NUMBER OF BUSINESS ESTABLISHMENTS; MEDIAN ANNUAL INCOME OF OTHER WORKERS WHOSE SALARIES MIGHT BE COMPARED WITH THOSE OF POLICE; MONOPSONY POWER OF THE COMMUNITY (I.E., ITS STATUS AS SOLE BUYER, AS OPPOSED TO ONE OF SEVERAL CITIES IN THE SAME AREA COMPETING IN THE MARKET FOR POLICE PERSONNEL); THE PRACTICE OF MAINTAINING PARITY BETWEEN POLICE AND FIRE FIGHTER SALARIES; POPULATION; AND UNIONIZATION OF POLICE. THREE DEPENDENT VARIABLES ARE USED TO MEASURE COMPENSATION: THE MINIMUM AND MAXIMUM ANNUAL SALARIES FOR A POLICE PRIVATE, AND THE AVERAGE ANNUAL SALARY FOR ALL POLICE DEPARTMENT PERSONNEL. THE ANALYSIS SHOWS THAT, WHILE POLICE SALARIES ARE INCREASED AS A RESULT OF FORMAL NEGOTIATIONS BY UNIONS, THE INCREASES ARE MODEST. POLICE UNIONS APPEAR TO BE MOST EFFECTIVE IN RAISING WAGES IN CITIES WITH MONOPSONY POWER. IN THE ABSENCE OF UNIONS, THESE CITIES PAY LOWER WAGES THAN DO CITIES THAT MUST COMPETE FOR POLICE PERSONNEL. THROUGH FORMAL NEGOTIATIONS, UNIONS ARE ABLE TO COUNTERACT THE EFFECT OF MONOPSONY POWER ALMOST EXACTLY. RELATIONSHIPS WERE ALSO FOUND BETWEEN HIGH CRIME RATES AND HIGH POLICE SALARIES AND BETWEEN LARGE NUMBERS OF BUSINESS ESTABLISHMENTS AND HIGH POLICE SALARIES. THE PRACTICE OF MAINTAINING PARITY BETWEEN POLICE AND FIRE FIGHTER PAY SCALES HAS A DEPRESSING EFFECT ON POLICE SALARIES. TABULAR DATA ARE INCLUDED.