This extension of the Markov model uses transitions between categorical states in the criminal justice process as the units of analysis and the unobserved instantaneous rate of transition between the states as the dependent variable. Using California data for felony offenses, instantaneous rates of transition are modeled from arrest to three outcomes: release, prosecution denial of complaint, and lower or superior court. Models are developed for each time the individual is processed by the system and the effects of race, ethnicity, age, sex, and other variables on the transition rate are examined. The analysis reveals that Chicanos have significantly higher rates of transition from arrest to disposition than do whites in their first shift regardless of outcome. Similarly, blacks move from arrest to release by police or the prosecutor more rapidly than do whites in their first shift. Being young increases the rate of transition for all races or ethnic groups. An arrest for auto theft results in speedier transition to disposition than robbery, while arrests for homicide, forgery, and narcotics result in slower releases by police and prosecutor. Most effects of race/ethnicity are insignificant by the time a person has undergone processing six or more times. Thus, Chicanos and blacks are particularly likely to be arrested by police with little grounds for the arrest and consequently are quickly released. This model reconceptualizes recidivism as an aspect of criminal justice processing rather than of defendant status. Tables, 11 footnotes, and 27 references are included.