The English Language Fluency And Earnings Of Ethnic Minorities In Britain
| DOI | http://doi.org/10.1111/1467-9485.00242 |
| Author | Joanne Lindley |
| Date | 01 September 2002 |
| Published date | 01 September 2002 |
THE ENGLISH LANGUAGE FLUENCY AND
EARNINGS OF ETHNIC MINORITIES IN
BRITAIN
Joanne Lindley*
ABSTRACT
This study addresses two issues. First it estimates how much of the male and female
ethnic earnings gap is the result of an advantage in the English language and
whether there is an earnings penalty to non-whites, over and above this. Lack of
fluency is shown to have a highly significant impact on the earnings of ethnic
minorities in Britain, although the language penalty is much greater for women
than it is for men. Moreover, only foreign born non-white males that have arrived in
Britain between 1970 and 1994, exhibit lower earnings once language fluency is
taken into consideration, whilst British born females exhibit higher earnings. So the
evidence here suggests that non-white earnings are assimilating towards those of
whites and that lower female non-white earnings are a direct result of a lack
of fluency rather than ethnicity. Secondly, the study will try to measure any
endogenous bias associated with the non-fluency earnings penalty. Controlling for
the endogeneity between language fluency and earnings is shown to be problematic.
Estimates suggest that single equation earnings functions slightly underestimate
the true language fluency penalty for males, and slightly overestimate the fluency
penalty for females. Finally, education and fluency are not surprisingly shown to be
complementary.
II
NTRODUCTION
Following, the pioneering work of Chiswick (1978), the impact of language
fluency on economic outcomes has been explored fairly extensively. Blackaby et
al. (2001) explored the impact of fluency on the male ethnic wage gap in Britain.
They find only a small language penalty for males, which is significantly less
than the ethnic penalty and any other characteristic effect. It has been argued
however, that individuals have an economic incentive to acquire language skills.
Following Chiswick and Miller (1995), Shields and Price (2002) tried to correct
for this endogeneity by using instrumental variables. They find that fluency is
the second most important determinant of occupational success amongst British
Scottish Journal of Political Economy,Vol.49,No.4,September2002
#Scottish Economic Society 2002,Publ ishedby Blackwell Publishers Ltd, 108 Cowley Road, Oxford OX4 1JF, UK and
350 Main Street, Malden, MA 02148, USA
467
*University of Nottingham
immigrant men. This paper updates earlier work by estimating separate earnings
functions for males and females, including both British born and immigrant
non-whites, and addressing the issue of language endogeneity. First, single
equation earnings functions are estimated for both whites and non-whites.
Secondly, a simultaneous equation model is estimated for non-whites only. This
involves a two stage least squares approach using an earnings equation and a
probit equation for English language proficiency. Finally a language fluency
selection model is estimated.
The paper proceeds as follows. Section II provides an insight into the theory
of language and develops the estimation equations. Section III describes the
data source and presents some descriptive statistics. Section IV analyses how
language proficiency affects the earnings of non-whites relative to whites, using
single equation estimates. Sections V and 6 address the issue of endogeneity
between language and earnings. Section V adopts a simultaneous estimation
approach, whilst Section VI reports the selection equation estimates. The final
section summarises the results obtained.
II THE THEORY OF EARNINGS AND LANGUAGE FLUENCY
The factors that determine language acquisition have been documented by
Chiswick and Miller (1995). These are collated into three categories. The first
set of factors increase immigrant efficiency in learning to speak the host
language. These in clude ability, ag e, colonial ties of t he native country wi th
the host country and the linguistic distance between native tongue and host
language. The second set of factors correspond to the immigrants exposure to
the host language before and after migration. These include pre-migration and
post-migration qualifications, whether the host language was frequently used
in the country of origin, time spent in the destination country and whether the
immigrant lives in an area that is highly concentrated in terms of members of
the same ethnic group. Finally, there are economic incentives. If language
ability affects economic outcomes such as lower unemployment and higher
earnings, then there are economic incentives for individuals to acquire
language skills. As a result language proficiency and earnings might be
endogenous.1
In the single-equation model there is always a potential bias which could be
the result of the endogeneity between earnings levels and English language
fluency. Using single equation estimates may produce biased estimates of the
true effect of language on earnings. This was first pointed out by McManus et al.
(1983), but has been the subject of subsequent discussion, notably by Chiswick
and Miller (1995), Dustmann and van Soest (1998) and Shields and Price (2002).
To investigate this issue two different approaches are adopted. The first involves
a simultaneous equation model, whilst the second involves the estimation of a
1See Leslie and Lindley (2001) for a theoretical framework on the acquisition of language
skills.
468 JOANNE LINDLEY
#Scottish Economic Society 2002
To continue reading
Request your trial