Low Paid Employment in Britain: Estimating State‐Dependence and Stepping Stone Effects
| Date | 01 April 2018 |
| Author | Lixin Cai,Kostas Mavromaras,Peter Sloane |
| Published date | 01 April 2018 |
| DOI | http://doi.org/10.1111/obes.12197 |
283
©2017 The Department of Economics, University of Oxford and JohnWiley & Sons Ltd.
OXFORD BULLETIN OF ECONOMICSAND STATISTICS, 80, 2 (2018) 0305–9049
doi: 10.1111/obes.12197
Low Paid Employment in Britain: Estimating
State-Dependence and Stepping Stone Effects*
Lixin Cai,†Kostas Mavromaras†, ‡ and Peter Sloane†, ‡, §
†National Institute of Labour Studies, Flinders University, South Australia
(e-mail: Lixin.c.Cai@gmail.com)
‡IZA, Bonn,
§WELMERC, School of Business & Economics, Swansea University,
Abstract
Using 18 wavesof the British Household Panel Study, this paper examines state-dependence
and stepping stone effects of low pay. The results show that both state-dependence and
stepping stone effects of low pay are present. However, there is no evidence to support a
low-pay no-pay cycle. The introduction of the national minimum wage does not appear to
have affected state-dependence and stepping stone effects of low pay.
I. Introduction
Early work on the dynamics of paysuggested that being in low paid employment increases
the chances of future low paid employment (i.e. state-dependence of low pay) and perhaps
even unemployment (Stewart and Swaffield, 1999, Stewart, 2007). However, this literature
does not consider the possibility that low pay may also lead to higher pay. As we show
below, following individuals over time between 1991 and 2008 using data from the British
Household Panel Study (BHPS), about 26% (40%) of low-paid women(men) are in higher
paying jobs in the subsequent year.Despite these high percentages, presently, the literature
has paid little attention to the empirical possibility that low payitself may have a low-pay to
higher-pay stepping stone effect.The core objective of this paper was to develop and esti-
mate a dynamic multinomial logit model of employmentwhich will allow us to examine the
dynamic relationship between no-pay, low-pay and higher-pay outcomes simultaneously.1
For the sake of completeness we also include in the analysis self-employment, an em-
ployment status typically ignored in the relevant literature. The policy contribution of this
JEL Classification numbers: J24, J31, I21.
*Data from the British Household Panel Study and the Understanding Society have been used in this study; and
they are provided by the UK DataArchive.
1While the modelling framework in this study has the advantage of being ableto incor porate transitions between
low pay and no pay,it also has limitations. In par ticular, the modelling framework cannot examine the effects of job
characteristics on the dynamics of low pay because information on these variables is missing for those who are not
employed.This issue may be picked up in future research using a different modelling framework.
284 Bulletin
paper is that it will inform the debate surrounding welfare policies designed to facilitate
the transition from welfare to work.
While whether low pay has a stepping stone effect, exhibits state-dependence, or leads
to no pay, is an empirical issue, there are some theoretical underpinnings based on human
capital and signalling of individual productivity. If work experience, evenif low paid, helps
accumulate human capital which is valued by employers, then low paid employmentwould
lead to higher pay in the future. However, if low paid employment does not help human
capital accumulation or may even cause deterioration of human capital perhaps because
low paid jobs are of lowquality, past low pay could lead to future low pay or evento no pay.
Similarly, if low paid employment is viewed by employers as a signal of low productivity
of the worker, past low pay could also increase the chance of the worker remaining in low
pay or moving into non-employment. In such context McCormick (1990) discusses why
redundant skilled workers may be reluctant to accept interim low skilled jobs.
We show that both state-dependence and stepping stone effects of low pay are present
among British workers after observed and unobserved individual heterogeneity has been
accounted for. Our results also show that, other things being equal, people who are on low
pay are more likely to be in employment in the future than those who are either unemployed
or not in the labour force (NILF). However, we also show that people on low pay are not
more likely to become jobless in the future than their higher paid counterparts. Simply
put, the paper looks for but does not find any evidence for a low pay–no pay cycle among
British workers.
The paper is arranged as follows: Section II briefly reviews the literature; section III
describes the econometric model and estimation strategy; section IV discusses the data
and model specification; section V presents the estimation results; section VI tests for and
discusses panel attrition bias and section VII sets out the conclusion.
II. Literature review
There is a sizable body of literature examining low paid employment with a focus on state-
dependence of low pay – that is, whether and to what extent current low paid employment
increases the probability of remaining in low pay in the future (see for instance, Sloane and
Theodossiou, 1998; Stewart and Swaffield, 1999; Cappellari, 2002, 2007; Cappellari and
Jenkins, 2008a, b; Clark and Kanellopoulos, 2013; Fok, Scutella and Wilkins, 2015). The
interest in state-dependence of low pay arises from a concern that with increasing earnings
inequality, if there is state-dependence of low pay (i.e. low pay is persistent), life-time
earnings inequality will increase as well. Indeed, state-dependence of low pay has been
found in a number of studies (among them, Stewart and Swaffield, 1999; Cappellari, 2002;
Clark and Kanellopoulos, 2013; Fok et al., 2015) even after individual heterogeneity is
controlled for.
However, there is another possible effect of low pay, to which earlier studies have paid
little attention – that is, the effect of current low pay on influencing the probability of
moving to higher pay in the future. We will refer to this possibility as a stepping stone
effect of low pay. To be consistent with the earlier literature, we will continue to use the
term state-dependence to refer to the first type of state-dependence of low pay (i.e. its
persistence).
©2017 The Department of Economics, University of Oxford and JohnWiley & Sons Ltd
Low paid employment in Britain 285
Answers to the question whetherand to what extent low paid employment has a stepping
stone effect are particularly relevantto policy makers. From a welfare policy perspective, if
low paid employment acts as a stepping stone to higher pay, welfare reforms that promote
employment,even if it is low paid, such as the work-first approach to welfarerecipients, have
a chance of improving the financial well-being of welfare recipients over time and should
therefore be considered as potentially welfare-improving policies. This study extends the
literature by estimating a dynamic multinomial logit model to examine both the state-
dependence and the stepping stone effects of low pay.
It appears that there are only two studies that take a similar modelling approach to the
analysis contained in this paper, namely Uhlendorff (2006) and Fok et al. (2015). Using
the German Socio-Economic Panel Study (SOEP) waves 1998–2003, Uhlendorff (2006)
examines low pay dynamics of German men and finds that there exists genuine state-
dependence of low pay as wellas of non-employment. However, unlike Uhlendorff (2006)
who treats unemployment and NILF as one labour force state (i.e. non-employment), our
present study models the two non-employment states separately. The distinction between
NILF and unemployment is particularly important in estimating the stepping stone effect
of low pay since the stepping stone effect may differ, depending on whether low paid
employment is compared with NILF or with unemployment.
Fok et al. (2015) examine the dynamics of low paid employment in Australia, using
the Household, Income and Labour Dynamics in Australia (HILDA) Survey. Both state-
dependent and stepping stone effects of low pay are found in that study. Although that
study uses an extended definition of unemployment to include those who are marginally
attached to the labour market in the analysis, it excludesthose who NILF and not marginally
attached to the labour market, as well as those who are self-employed, which may lead to
sample selection bias in their estimation.
In a dynamic probit model framework and using the German SOEP, Knabe and Plum
(2013) examine the stepping stone effect of lowpay relative to unemployment by including
both lagged unemployment and lagged low pay as the explanatory variables. Theyfind that
low pay can act as a stepping stone to better paid employment, particularly for those who
do not have a college degree, who have been unemployed more often in the past and whose
low paid job carries relatively high social status. While their model takes into account
potential endogeneity of initial low pay,initial unemployment is assumed to be exogenous.
Given that their estimation results show that initial low pay is not in fact exogenous, it is
likely that initial unemploymentis endogenous. Consequently, the estimates of their model
are likely to be biased.
A related theme of research on low pay dynamics examines whether low paid employ-
ment and unemployment are inter-related.This question arises due to the concer n that low
paid workers may cycle between low pay and unemployment (or non-employment) with
little hope of moving up the labour market ladder. For example, descriptive analyses tend
to show that low paid workers are more likely than higher paid workers to move into job-
lessness in the future (e.g. Stewart and Swaffield, 1999; Cappellari and Jenkins, 2008a, b).
But evidence from modelling results is still mixed.
Cappellari and Jenkins (2008b) find that for the UK men, low pay experience has
only a modest (and statistically insignificant) effect on the probability of experiencing
unemployment in the future when individual heterogeneity is accounted for. This result
©2017 The Department of Economics, University of Oxford and JohnWiley & Sons Ltd
To continue reading
Request your trial