BASIC ASSUMPTIONS OF THE WELFARE APPROACH TO THE ECONOMICS OF CRIME ARE THAT ANALYSIS IS BASED ON TRANSLATION OF POLICY GOALS INTO MEASURABLE QUANTITIES, THAT GOALS OF SPECIAL AND GENERAL CRIME PREVENTION ARE CONNECTED TO MINIMIZATION OF THE TOTAL COST OF CRIME, AND THAT THE ALLOCATION APPROACH APPLIES NOT JUST TO GENERAL PREVENTION BUT TO A WHOLE SYSTEM OF BALANCES. TO ATTAIN THE SOCIAL GOAL OF INTERNAL SECURITY, GENERAL AND SPECIAL CRIME PREVENTION SHOULD BE PROMOTED; THE COLLECTIVE PRODUCT 'INTERNAL SECURITY' MUST BE DIVIDED INTO LAW, POLICE, AND THE PENAL SYSTEM TO FACILITATE ANALYSIS; AND INTANGIBLE COSTS AND BENEFITS MUST BE WEIGHED. SPECIFIC BENEFIT INDICATORS ARE THE FOLLOWING: OUTPUT CRITERIA OF THE DIRECT TYPE (E.G., CRIME RATES) OR INDIRECT TYPE (E.G., MEASURES OF POLICE BEHAVIOR TOWARD THE PUBLIC); MEASURES OF THE ADMINISTRATION OF JUSTICE (E.G., CASE PROCESSING); AND MEASURES OF EXECUTION OF PENALTIES (E.G., RECIDIVISM). CROSS-SECTOR MEASUREMENT PROCEDURES INCLUDE THE ALTERNATIVE COST CONCEPT (CALCULATION OF THE LOWER COST ALTERNATIVE FROM TWO MEASURES WITH EQUAL BENEFITS) AND THE DAMAGE CONCEPT (CALCULATION OF BENEFITS ACCORDING TO DAMAGES AVERTED). MODELS FOR SPECIFIC AREAS DISCUSSED INCLUDE CHAPMAN'S MODEL FOR THE POLICE (WHICH CONSIDERS THE SUPPLY-OF-CRIME EQUATION, THE ARREST RATE, THE DEMAND FOR POLICE SERVICES); AND NEV'S MODEL FOR ESTIMATING TOTAL COST TO SOCIETY OF THE EXECUTION OF SENTENCES AS WELL AS BLUMSTEIN AND LARSON'S MODEL FOR MEASURING THE EFFECTS OF THE VARIOUS SUBSYSTEMS OF THE CRIMINAL JUSTICE SYSTEM ON ONE ANOTHER. IT IS CONCLUDED THAT THE PROBABILITY OF CERTAIN EFFECTS AND BENEFITS FROM A PROJECT DEPEND ON A SUBJECTIVE ASSESSMENT OF CRIMINAL BEHAVIORAL PATTERNS WHICH MAY OR MAY NOT BE ACCURATE, AND THAT FURTHER SYSTEMS ANALYSES WITH EMPIRICAL ASSUMPTIONS ABOUT HUMAN BEHAVIOR ARE NEEDED. AN EXTENSIVE BIBLIOGRAPHY IS PROVIDED. --IN GERMAN.