Population Research and Policy Review 16: 513–530, 1997.
1997 Kluwer Academic Publishers. Printed in the Netherlands.
Period and cohort age patterns of US migration, 1948–1993:
Are American males migrating less?
ANDREI ROGERS & SAMEER RAJBHANDARY
Institute of Behavioral Science, University of Colorado, Boulder, Colorada, USA
Abstract. Has the national intercounty migration level of American males been experiencing
a decline over the past half-century? What differences in age patterns have period and cohort
migration schedules exhibited during this period? These fundamental questions are addressed
in this paper. We ﬁnd that both period and cohort levels of migration have indeed been
declining, with the principal consequence for age patterns being a parallel decline in the levels
of young-adult migration. American males are moving less, and their peak migration rate in
the young-adult years has steadily moved to a younger age over time.
Key words: Cohort, Migration, Model schedules, Period, US males
The decline in national mobility levels during the 1970s initially attracted
relatively little scholarly attention. More recently, a number of scholars have
addressed this particular topic (e.g., Rogerson 1987; Long 1988; Edmonston
1990; Plane & Rogerson 1991; Kulkarni & Pol 1994; Pandit 1996). Of these
studies, however, only Edmonston (1990) attempts to develop and examine
cohort data derived from consecutive period age patterns of migration. How-
ever, his examination of cohort migration is limited to a regression analysis
includes a set of dummy variables for the age patterns
, a linear
time trend variable, and a linear cohort trend variable’ (Edmonston 1990: 13).
Our approach is based on direct estimation and extrapolation of cohort
patterns from period data. Starting with the period intercounty migration
data for males that are presented in Edmonston’s paper, we ﬁrst smooth
and then graduate these migration schedules with a parameterized model
migration schedule, thereby, eliminating irregularities and creating single-
year-of-age migration rates from those referring to ﬁve-year age groups.
(Although Edmonston refers to single-year-of-age migration data, his paper
presents data only for ﬁve-year age groups.) Using these graduated period
migration data, we then identify the associated age-speciﬁc (incomplete)
cohort migration schedules and ﬁt parameterized model migration schedules