INDUSTRY CHARACTERISTICS AND INTER‐REGIONAL WAGE DIFFERENCES
| Date | 01 May 1991 |
| Author | P. D. Murphy,D. H. Blackaby |
| Published date | 01 May 1991 |
| DOI | http://doi.org/10.1111/j.1467-9485.1991.tb00307.x |
Scortrsh
Joitrnal
o/
Polrrrcni
Econoin.v.
Vol.
38.
No.
2,
hlay
1991
1991
Scuiiirh
Economic
Sociei)
INDUSTRY CHARACTERISTICS AND
INTER-REGIONAL WAGE DIFFERENCES
D.
H.
BLACKABY*
University College
of
Swansea
AND
P.
D.
MURPHY*
University
of
Aberdeen
I
INTRODUCTION
One
of
the accepted characteristics
of
the wage structure is that there are
significant differences in average earnings at both regional and industry levels.
For
example, Tarling and Wilkinson (1982) find an inter-quartile range
of
20%
in average hourly earnings for
82
UK
industries in 1980, while Harris (1989)
has shown that full time hourly earnings differentials varied by approximately
30%
between the highest and lowest paid regions in both 1970 and 1982.' Of
course differences in observed industry and regional earnings are to be expected
if
the aggregates, which form the basis
of
the comparison, are comprised
of
individuals who differ with respect to their acquisition
of
human capital. We
find, however, that while controlling
for
these characteristics reduces the mag-
nitude of derived industry-regional wage premia, they are insufficient to elim-
inate them altogether. Accordingly,
a
fuller understanding
of
why such premia
arise demands a wider perspective which should examine the role played by
a
variety
of
factors that economic theory suggests influence the wage structure.
*
We are grateful
to
E.
T.
Nevin,
K.
D. George, L. C. Hunt, D. Leslie,
L.
Mainwaring,
D.
N.
Manning and R.
F
Elliott, and an anonymous referee of this journal for helpful com-
ments
on
previous drafts
of
this paper. We would also like to extend
our
thanks to Fiona
Johnson for invaluable computational assistance. Naturally all remaining errors are
our
own. Material from the General Household Survey, made available through the Office
of
Population Censuses and Surveys and the ESRC Data Archive, has been used by permission
of
the Controller
of
HM Stationary Office, with the usual disclaimer.
'
Admittedly much
of
this difference was due to the influence
of
two
regions, the
GLC
and
East Anglia, which when excluded, cut the range
of
variation down to
10%
of
the national
average. Nevertheless, as Harris points out, this still represents a significant average
hourly
earnings premium at approximately
f0.38
per hour and
f0.27
per hour for males and females
respectively in
1982.
Date
of
receipt
of
final manuscript: 14 June
1990
142
INDUSTRY CHARACTERISTICS AND WAGE DIFFERENCES
143
A
competitive labour market model offers several possible explanations for
the existence
of
human capital adjusted industry-regional wage differentials.
On the one hand they can arise from differences in the quality
of
employment
that affect the utility
of
workers (Rosen, 1974) and which are themselves corre-
lated with industry-regional status. On the other hand industry-regional mark
ups may reflect
a
‘transitory’ phenomenon that is related to those differential
shifts in either labour demand or supply which sometimes affect particular seg-
ments
of
the labour market. Such mark-ups may show
a
degree
of
persistence,
however, where the short-run mobility
of
labour is impeded (Hughes and
McCormick, 1981) as this can prevent the system from adjusting quickly to its
new equilibrium position.
Mark-ups on wages can also arise under non-competitive conditions. Trade
unions, and in particular the type
of
bargaining arrangements to which workers
adhere can affect the wage structure where these are not uniformly distributed
across industry and regional boundaries. Alternatively, efficiency wage
theories-surveyed by Stiglitz (1986) and Yellen (1984)-suggest other reasons
why firms find it profitable to pay wages above market clearing levels.
Specifically, in the ‘shirking’ and ‘turnover’ versions
of
this theory it is
expected that those factors which affect either the likelihood
of
an unemployed
worker obtaining employment, or the level and cost of labour turnover to the
firm will have some influence over the process
of
wage determination. Again,
where such factors have
a
particular industry-regional dimension, this will be
reflected in the observed industry-regional wage structure.
The present paper seeks to analyse regional wage differentials-these being
defined as the variations in average wage rates
for
any given industry across
different regions
of
the
UK.
Using data drawn from two primary sources, the
New Earnings Survey (NES) and the General Household Survey (GHS), we find
that a significant proportion
of
the variation in industry-regional wage premia
can be explained by our analysis. Unfortunately, given the nature
of
the data,
and the non-nested nature
of
many
of
the theories
of
wage determination we
outline, it is impossible to come to any firm conclusions as to which single
theory best explains the existence
of
such wage premia.
The plan
of
the paper is as follows. In Section
I1
we report
on
the spread
of
industry-regional wage premia, and the extent to which these are either
diminished or eliminated once allowance is made for the differing human
capital characteristics
of
the sample. Section
I11
details various competitive and
non-competitive theories
of
wage determination that offer an explanation for
the existence of industry-regional wage premia. Section
IV
details the results
of
the estimated relationship between the industry-regional fixed effects that are
identified and those variables whose potential roles are highlighted in Section
111.
Finally, Section
V
gives some concluding remarks and suggestions for
further research.
To continue reading
Request your trial