DUAL LABOUR MARKETS AND THE POTENTIAL EARNINGS OF THE UNEMPLOYED
| Author | Kenneth Clark,David Blackaby,Derek Leslie |
| DOI | http://doi.org/10.1111/j.1467-9485.1995.tb01144.x |
| Date | 01 February 1995 |
| Published date | 01 February 1995 |
Sconish
JOU~M~
of
Polifical
Economy.
Vol.
42.
No.
I.
February
1995
6
Scottish
Economic
Society
1995.
Published
by
Blackarc11
Publishers.
108
Cowley
Road,
Oxford
OX4
IF,
UK
and
238
Main
Smcf.
Cambridge.
MA
02142.
USA
DUAL LABOUR MARKETS AND THE
POTENTIAL EARNINGS
OF
THE UNEMPLOYED
David Blackaby, Kenneth Clark and Derek Leslie'
I
INTRODUCTION
Dual labour market theory has recently been integrated into the mainstream of
labour market theory and is no longer regarded as a fringe idea. The theory
suggests that workers with more
or
less the same characteristics earn a lower
wage if attached to the secondary sector and that such secondary sector workers
are more unemployment prone.
This
idea of duality has been supported by a
variety of theoretical justifications and empirical evidence. Efficiency wage
theory and insider-outsider theory
are
two recent theoretical innovations
consistent with the existence of a high wage low unemployment primary sector
and
a
low wage high unemployment secondary sector. They stand comfortably
within the older more anecdotal tradition of a dual
or
segmented labour market.'
In
the United States there is a large and growing empirical literature
supporting the idea of labour market segmentation and that observed patterns of
wages cannot comfortably be explained
in
any other way.2 By contrast, evidence
for Britain is more limited and most studies find little support for segmentation
the~ry.~ The present paper takes as its starting point the idea that workers can
be
split into these two broad categories. We follow the terminology of Atkinson
and Micklewright (1991) and describe as regular workers those attached to the
primary sector and as marginal workers those attached to the secondary sector.
Our key finding is that the potential earnings of the employed are substantially
higher (on average 29.8% over the period 1979-86) compared to the unem-
ployed. Roughly two-thirds of
this
difference is due to the unemployed having
less favourable characteristics and one-third due to the fact that given character-
istics are less well rewarded. We interpret
this
finding as being supportive of a
dual labour market. Given the lower attachment
of
marginal workers to the
labour market, we would expect a greater concentration of regular workers
'
The definitive statement of the original dual labour market framework
is
Doeringer and
Piore (1971). which can
be
traced back to the
19th
century with Mill (1900). Bulow and
Summers (1986) is one example exploiting efficiency wage theory and Lindbeck and Snower
(1988) summarises the insider-outsider approach.
'Dickens and Lang (1985, 1987, 1992), Dickens and Katz (1987). Krueger and Summers
(1987, 1988) are some examples of mainstream econometric investigation of the
dual
labour
market approach.
Sloane
et
al.
(1993) is
a
recent example and
also
provides a useful summary of the state of
play of
this
approach. Mayhew and Rosewell (1979), McNabb and Psacharopoulous (1981).
McNabb (1987) find little support. McNabb and Ryan (1990) and Bosanquet and Doeringer
(1973) provide some modest support.
Department of Economics, Manchester Metropolitan University
37
38
DAVID BLACKABY. KENNETH CLARK AND DEREK LESLIE
among the employed compared to the unemployed.
This
explains both the less
favourable characteristics and lower rewards to the unemployed.
The paper is structured
in
the following way. Section
II
presents a simple
theoretical framework within the dual labour market tradition where a sharp
distinction between regular and marginal workers is drawn. Although the
primary purpose of the paper is to present evidence favouring duality in labour
markets, one implication is discussed to illustrate the fact that duality, if true,
does make an important difference. The example concerns the measurement of
replacement ratios and shows how these can
be
considerably underestimated.
Section
lTI
discusses the data and presents some key results in defence of the
theoretical model.
An
alternative explanation, namely that lower earnings
among the unemployed reflects quitting by
underrewarded
(for given human
capital characteristics) employed workers is considered and rejected.
II
THEFRAMEWORK
There
are
two groups. First, there are regular workers with actual earnings
described by:
(1)
where
Y;"
is the log of earnings.
XI
is a vector of characteristics,
B
is a vector of
associate3 coefficients and
u;"
is a disturbance term. Regular workers will from
time to time find themselves unemployed and equation
(1)
assumes that
(log)
potential earnings of an unemployed regular worker is given by the mean of the
wage distribution conditional on characteristics, that
is
PX,.
Earnings of
marginal workers are similarly described by:
(2)
The important point is that the rewards to characteristics of regular workers
described by
/3
need not equal the rewards to marginal workers described by
y.
Earnings functions are replete with examples where differences in rewards to
characteristics are the
rule
when samples are split according to some ~riterion.~
It would
be
surprising if there were not also significant coefficient differences
between marginal and regular workers. It would also be reasonable to assume
that average characteristics would differ between the two groups. The various
theories of labour market segmentation strongly suggest these
two
features will
be present.
Equations
(1)
and
(2)
will not
be
directly observed, rather what will be
observed
are
employed workers and unemployed workers. Employed workers
consist of a mixture
of
regular and marginal workers, described by:
Y;"
=
B'X,
+
u;"
i
=
1,
...,
m
Yy
=
y'X,
+
u;"
i
=
m
+
1,
...
,
T.
Y~=(D1~+[l-Dl]y')X,+Dlu;"+(l-D,)u;"
i=l,
...,
K
(3)
where
D,
is a zero-one dummy
=
1
for the case of an employed regular worker,
'Three important examples are
the
unionlnon-union,
male/female and black/white
earnings
differentials.
(0
Sconish
Economic
Society
1995
To continue reading
Request your trial