Since the 1950s, the dominant pattern of partisan change in the American electorate has involved movements between party identification and independence rather than direct or indirect shifts between parties. This article employs switching regression analyses to investigate the long-term evolution and short-term dynamics of independence between 1953 and 1988. The analyses reveal that a new 'independence regime' developed rapidly in the mid-1960s, with the 'tipping point' in the transition occurring in the second quarter of 1967. Under the new - but not the old - regime, short-term changes in the size of the independent cohort have reflected economic conditions as well as political events. These findings argue that future research on the dynamics of public support for political parties in the United States and elsewhere will profit by developing dynamic models which assess processes of long- and short-term change in tandem.