REGIONAL MIGRATION VERSUS REGIONAL COMMUTING: THE IDENTIFICATION OF HOUSING AND EMPLOYMENT FLOWS
| Author | Richard Jackman,Savvas Savouri |
| DOI | http://doi.org/10.1111/j.1467-9485.1992.tb00621.x |
| Date | 01 August 1992 |
| Published date | 01 August 1992 |
Scoiiish
Journal
of
Polilrral Economy.
VoI.
39.
No.
3.
August
1992
I992
Scoiiirh
Economic
Sociely
REGIONAL MIGRATION VERSUS REGIONAL
COMMUTING: THE IDENTIFICATION OF
HOUSING AND EMPLOYMENT FLOWS
RICHARD
JACKMAN
AND
SAVVAS SAVOURI*
Centre for Economic Performance, London
I
INTRODUCTION
A
comparison
of
regional unemployment rates across Great Britain indicates
a high degree
of
persistence; across the ten regions
of
Britain the simple correla-
tion coefficient for regional unemployment rates between
1974
and
1987
was
0.92
(Jackman
et
al.,
1991).
One explanation for such persistence is the lack
of
geographical labour mobility
or
migration, which is in turn often associated
with housing market rigidities (McCormick,
1991).
A
common view is that
there cannot be a net movement
of
workers into a region
of
high labour
demand because
of
housing supply inelasticities. In the local authority sector
there is preferential allocation to local residents, while planning and land use
controls hold back new housing development in the private sector.
But this view seems to neglect the possibilities that people can move their
home from one region
to
another without changing their job while others may
change their job without changing their home. In other words, if there are
rigidities in the housing market, regional commuting may take on
a
more
important role relative to migration as a means
of
alleviating regional in-
equalities. There is also, of course, the fact that some housing is occupied by
retired,
or
non-working households, who can migrate without regard
to
labour
market conditions.
Clearly in principle we may divide migrants into two categories, those who
also change their job, and those who move home without changing their job.
(Those who change job without moving home will not be classified as
migrants.) This point was first discussed by Gordon
(1975)
and later developed
by Molho
(1982).
Both authors argued that one could distinguish the two types
of
migrants by reference to the distance
of
the move.
It
was, they suggested,
to be expected that long-distance moves would consist
of
migrants who were
*The Centre for Economic Performance is financed by the Economic and Social
Research Council. Financial support is
also
acknowledged from the Esmee Fairbairn
Charitable Trust.
Date
of
receipt
of
final manuscript:
17th
February
1992.
212
REGIONAL MIGRATION
V.
REGIONAL COMMUTlNG
273
also changing their jobs, whom they termed jobmovers,' while it would only
be close
to
regional boundaries
or
along well-developed transport routes that
people might move home across regions without changing their job. Most
migration models assume that people move for job reasons; according
to
Gordon
(1975,
p.
161)
the failure
of
such models
to
allow for migrants moving
home but remaining in the same job had led them to under-predict flows
between adjacent regions, since they failed to capture pure housing moves, and
over-predict
flows
across non-contiguous regions.
A
related problem with recorded migration flows is that they do not identify
commuters,
ie. those workers who
fill
vacancies outside their region
of
resi-
dence without changing their region
of
domicile. If we assume that these
latent
inter-regional migrants are largely between adjacent regions, then the effect
of
vacancies on migration flows will tend to be under-recorded for flows between
adjacent regions. That
is,
some job-moves would not be accompanied by
housing moves and hence would not be recorded as migration, while some
housing moves, which would count as migration, would not be accompanied
by job-moves.
As
evidence
for
the contrasting motives for short against longer distance
moves, the General Household Survey for
19762
records that, when asked for
the main reason for moving, only
4%
of
those moving under
5
miles and
18%
of
those moving between
5
and
50
miles answered 'Job-Study reasons' while
of
those moving
50
miles or more
53%
gave this answer, the respective shares
who answered 'Housing reasons' were
33%, 19%
and
9%
(Table
1).
In attempting
to
overcome such identification problems, Molho
(1982)
argued that given an array
of
bilateral migration flows it was possible
to
identify certain cells as consisting exclusively
of
job-mover flows (ie. the flows
between non-contiguous regions) and that the model should be estimated over
these flows alone. Molho went
on
to argue that the same model could then be
estimated over flows between contiguous regions, and that
a
comparison
of
the
TABLE
1
Distance moved in
last
move
by
main reason
for
moving
Distance
of
Last Move (Miles)
Main reason
for
move Under
5
5-49
50
or
more
~~ ~ ~ ~~
Housing Reasons
33
19
9
Environmental Reasons
9
12
6
Job-Study Reasons 4
18
53
Personal and Other Reasons
54
51
31
~
Source: General
Household
Survey,
1976,
Table
5.50
'
There are
two
exceptions; firstly, retirement effects where households move home over
long
distance but where no employment effects are present and secondly long distance
commuters. In
our
empirical work we will assume that retirement flows are adequately
captured by regional fixed effects (in fact prime age adults-I5
to
44
year olds-form
60%
of
total migration flows) and that
long
distance commuting forms a relatively small share
of
overall commuter flows.
'Unfortunately, this question has not been asked more recently.
To continue reading
Request your trial