minimum wages in sub-saharan africa: a primer
TRANSCRIPT
Minimum Wages in Sub-SaharanAfrica: A Primer
Haroon Bhorat, Ravi Kanbur, and Benjamin Stanwix
The fraction of workers currently covered by minimum wages in Sub-Saharan Africa
(SSA) is small, but as formality and urbanization increase, wage regulation will become
increasingly relevant. In this analysis, we find that higher minimum wage values are as-
sociated with higher levels of GDP per capita, in both SSA and non-SSA countries. Using
two measures to assess the level at which minimum wages are set, we find that minimum
wages in SSA countries are on average lower—relative to average wages—than most
other comparable regions of the world. Thus, SSA as a whole reflects no particular bias
toward a comparatively more pro–minimum wage policy. Within SSA, however, we ob-
serve that low-income countries set relatively higher minimum wages than middle- or
upper-income countries. We find significant variation in the detail of minimum wage re-
gimes and schedules in the region, as well as large variations in compliance. Notably, sev-
eral countries in SSA have relatively complex minimum wage schedules, and on average
we find high levels of noncompliance among covered workers. We also summarize the
limited research on the employment effects of minimum wages in SSA, which are consis-
tent with global results. By and large, introducing and raising the minimum wage ap-
pears to have small negative employment impacts or no statistically significant negative
impacts. There are country studies, however, where substantial negative effects on em-
ployment are reported—often for specific cohorts. The release of country-level earnings
and employment data at regular intervals lies at the heart of a more substantive,
country-focused minimum wage research agenda for Africa.
Legislated minimum wages apply to many millions of workers around the world.
Over the latter half of the 20th century, almost all countries in Sub-Saharan
Africa (SSA) introduced some form of minimum wage legislation. In many cases,
The World Bank Research ObserverVC The Author 2017. Published by Oxford University Press on behalf of the International Bank for Reconstruction andDevelopment / THE WORLD BANK. All rights reserved. For Permissions, please e-mail: [email protected]:10.1093/wbro/lkw007 Advance Access publication January 3, 2017 32:21–74
these laws apply to workers in specific industries or occupations, and in instances
where wage levels are not set by the state, collectively bargained agreements may
fix wages for specific sectors or occupations. Broadly, however, the introduction of
national laws governing wages is part of an observed regulatory revival in low-
and middle-income countries, where a range of labor regulations aimed at protect-
ing low-paid workers have gradually been introduced (Piore and Schrank 2008).
As will be documented in the paper, there has been widespread adoption of mini-
mum wage legislation as a policy tool in SSA. And yet there has been little work
on the nature, scope, and impact of such laws in the region.
A key distinction must be made between the usually small, formal, wage-
earning sector and usually large, informal, non-wage-earning sector in most
African countries.1 This has implications for minimum wage policy and research.
It is well known that in the overwhelming majority of SSA economies’ subsistence
agriculture and, more recently, urban informal employment dominates the labor
market. In many such countries, wage-earning employees only make up a small
portion of the labor force. As figure 1, below, indicates for a selection of SSA coun-
tries, it is only South Africa where formal salaried employees constitute the
Figure 1. Wage-Earning Employees as a Percentage of Total Employment
020
%40
%60
%80
%
Perc
enta
ge o
f w
ag
e-e
arn
ing
em
plo
ye
es
Mali
Ta
nza
nia
Ug
an
da
SS
A A
ve
rag
e
Za
mbia
Ken
ya
Gro
up
Ave
rag
e
Na
mib
ia
Sou
th A
fric
a
Sources: South Africa: Labour Market Dynamics Study (2013); Kenya: Kenya Integrated Household Budget
Survey (2005–6); Uganda: Uganda National Panel Survey (2012); Mali: Rani et al. (2013); Zambia: Living
Conditions Monitoring Survey (2010); Tanzania: Integrated Labour Force Survey (2005–6); Namibia: Labour
Force Survey (2012); Bhorat, Naidoo, and Pillay (2015).
22 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
majority of the labor force. In the six other countries represented, salaried em-
ployees account for less than half of the labor force, while the sample average is
close to 20 percent.
Minimum wages only apply to wage-earning employees and, in some cases,
only to formal sector wage earners. Wage regulations thus cover only a minority
of the total workforce in most SSA economies. This may go some way to explaining
the dearth of existing research. However, (a) as the urban and formal sector grows,
the experience of wage regulation in the currently covered areas will become sig-
nificant, (b) there can be spill over or “lighthouse” effects on uncovered sectors,2
and (c) as we shall see, minimum wage policy is particularly progressive in the
covered sectors in Sub-Saharan Africa. For these reasons, we believe that an em-
pirical overview of minimum wages in the region is important for the current pol-
icy and analytical discourse.
It is a stylized fact that a large gap exists between de jure and de facto labor reg-
ulation in most low- and middle-income (LMI) countries globally. Simply put, the
levels of noncompliance with minimum wage laws are high. This has been shown
in several studies (Bhorat et al. 2012; Rani et al. 2013; and Almeida and Ronconi
2012) and is supported by the evidence presented in this paper. While it could be
argued that high levels of noncompliance make wage regulation irrelevant, the
reasons for high levels of noncompliance are not yet adequately understood. A study
of the enforcement of minimum wages is of interest in its own right but can also
contribute to a broader research agenda on questions related to the rule of law.
There is a paucity of available household survey data for many SSA countries,
and particularly an absence of reliable information on earnings. This makes rigor-
ous study on the impact of minimum wages difficult, particularly because any at-
tempts to measure minimum wage impacts require pre- and postintervention data,
and this is almost always absent for countries in SSA. The problem is brought into
sharp relief when analyzing one of the most extensive and centralized bodies of in-
formation on global minimum wages—the International Labour Organisation’s
(ILO) TRAVAIL database and the ILO global wage database. While these databases
contain detailed information on minimum wage frameworks, the method of setting
wages in each of the ILO’s member states, and data on enforcement practices for
almost every country, there is very limited information on wages, coverage, or lev-
els of compliance for countries in Africa. This is because for many countries in the
region reliable data on wages either do not exist or are not accessible.
It is thus the purpose of this piece to provide a basic overview of minimum wage
regimes in SSA, where data are available, as well as some more detailed work for
several countries using household survey data. We begin with a brief literature re-
view of more recent minimum wage work and build on Neumark and Wascher’s
(2007) meta-review to provide an updated overview of the empirical wage-
employment trade-off estimates, focused on LMI countries. We then document the
Bhorat et al. 23
level of minimum wages in the region, present simple Kaitz ratios (the Kaitz ratio is
the ratio of the minimum wage to the average wage) and use this to make compari-
sons with wage levels and ratios elsewhere in the world. We proceed to examine com-
pliance levels in seven SSA countries using household survey data and again
compare the outcomes to estimates for a selection of non-SSA LMI countries. We end
by reflecting on possible early lessons for wage-setting regimes in the region and on a
research agenda to inform policy makers on key trade-offs in minimum wage setting.
The Developing Country Literature: A Brief Overview
In most SSA countries, minimum wage laws are used as a policy tool to achieve a
number of objectives, and while there are large cross-country differences, wage
legislation reveals considerable overlap. The stated objectives usually focus on pro-
tecting vulnerable workers from extreme levels of low pay, addressing poverty by
redistributing income from employers to low-wage employees, and encouraging la-
bor productivity. It is well known that there are also risks associated with institut-
ing wage floors, and in most cases country-level legislation recognizes the possible
trade-offs. The costs can include increased unemployment in certain settings, ad-
justed hours of work that disadvantage workers, and the movement of workers
from formal to informal employment—where livelihoods are often more precari-
ous. Rising wages as a result of minimum wage policy, while having positive ef-
fects through driving demand, may also have some impact on the cost of living in
the medium term if firms’ cross-price elasticity response to higher labor costs is to
raise output prices.3 Beyond the various costs and benefits, which are the focus of
a well-established body of literature for developed countries, the attendant issues of
enforcement and compliance are central to any discussion of minimum wages in
SSA. Indeed, it is arguable that most of the impact of a minimum wage policy is
contingent on enforcement and compliance in LMI country settings. We proceed
to briefly analyze the minimum wage literature here.
The empirical work on minimum wages constitutes a large field that is now well
established, with several recent books and papers dedicated to reviewing the main
findings of this literature.4 Studies in the United States were the first to suggest
that the textbook wage-employment trade-offs do not always hold in practice—
that an increase in the minimum wage will not always perfectly predict a decrease
in employment (the seminal paper in the new minimum wage literature being
Card and Krueger 1994). In addition to the potential employment effects, mini-
mum wage laws have been shown to induce adjustments in hours of work and
nonwage benefits, as well as having possible “nonstandard effects” such as influenc-
ing educational decisions and reservation wages.5 The limited but fast-growing body
of work focused on LMI countries has emphasized that the impact of introducing, or
24 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
increasing, a minimum wage can have mixed impacts that are often crucially con-
tingent on a variety of factors, including the level at which the minimum wage is
set, broader economic conditions, the nature of the minimum wage intervention, po-
litical economy factors, the enforcement regime, and so on. Existing studies provide
evidence of negative employment effects in some cases but also evidence of no em-
ployment declines in others, with a range of adjustments observed for hours of
work, wages, and nonwage benefits.
In Neumark and Wascher’s (2007) meta-review, they include 15 studies focused
on eight developing countries over the period 1992 to 2006.6 The authors caution
that studying minimum wage effects in developing countries is complicated, partly
due to issues we have already mentioned above that relate to data availability and
quality, but also because the results are often not easily generalizable across countries
or sectors. The majority of findings reviewed by the authors reveal either no effects or
small negative employment effects of minimum wages in LMI country settings. In
Brazil, for example, disemployment effects and reduced work hours are seen to be mi-
nor or nonexistent overall, but more pronounced for individuals with low skills and
lower wages (Fajnzylber 2001; Lemos 2004, 2006, 2007; and Neumark et al.
2006). In Chile, increases in the minimum wage had negative employment effects for
youth and unskilled workers but led to an increase in the employment of women
(Montenegro and Pages 2004). In Colombia, research suggests that disemployment
effects were present, and these were higher for low-skilled workers (Bell 1997; and
Maloney and Nu~nez Mendez 2004). In Costa Rica, increases in minimum wage also
decreased employment and reduced hours worked by employees in covered sectors,
especially those in the lower half of the skill distribution (Gindling and Terrell 2005,
2007). Similar results are found in Indonesia and Trinidad and Tobago (Alatas and
Cameron 2003; Rama 2001; Suryahadi et al. 2003; Hyslop and Stillman 2004;
Comola and De Mello 2011; and Strobl and Walsh 2003).
In the years following the publication of Neumark and Wascher’s (2007) review
paper, a growing body of research has focused on minimum wage effects in LMI
countries (see table A4 in the Appendix for a detailed literature summary). In East
Asia, a number of papers provide new evidence on minimum wage impacts: In
Thailand, Del Carpio, Messina, and Sanz-de-Galdeano (2014) found small disem-
ployment effects on female, elderly, and less-educated workers and large positive
effects on the wages of prime-age male workers; in Vietnam, Del Carpio and Liang
(2013) and Nguyen (2010) estimates showed that low-wage formal sector employ-
ment is negatively affected. In the Phillipines, Lanzona (2012) and Del Carpio,
Margolis, and Okamura (2013) observed disemployment effects only in sectors
with high levels of compliance; and in Indonesia, two papers (Comola and De
Mello 2011; and Harrison and Scorse 2010) found disemployment effects. In
China, Wang and Gunderson (2011) and Fang and Lin (2013) provided some of
the first estimates of minimum wage effects—which are generally negative. In
Bhorat et al. 25
Latin America, several new papers built on existing work to provide new sector-
specific results on employment, but also a nuanced focus on informality, social se-
curity, and poverty (Ham 2013; Gindling 2014; and Kharmis 2013).
The literature on SSA, however, remains rather limited, with published work ex-
isting only for four countries, namely Ghana (Jones 1997), Kenya (Andal�on and
Pages 2008), Malawi (Livingstone 1995), and South Africa (Hertz 2005;
Dinkelman and Ranchhod 2013; Bhorat et al. 2013, 2014a; Bhorat et al. 2014b;
and Garbers 2015). The most comprehensive literature in the SSA region exists for
South Africa and shows that while the introduction of minimum wages had a neg-
ative impact on employment in agriculture, in all other covered sectors no employ-
ment decreases were evident (Dinkelman and Ranchhod 2013; Bhorat et al. 2013;
and Nattrass and Seekings 2014).
In figure 2(a) and (b) below, we construct a graph of employment elasticities es-
timated in the minimum wage literature described above. This includes the 98
Figure 2(a). Minimum Wage-Employment Elasticities for Low- and Middle-Income Countries
-1.5
-1-.
50
.51
Ela
sticity
Indo
ne
sia
Ho
nd
ura
sB
razil
Pue
rto
Ric
oIn
do
ne
sia
Bra
zil
Indo
ne
sia
Bra
zil
Ch
ina
Indo
ne
sia
Indo
ne
sia
Th
aila
nd
Sou
th A
fric
aIn
do
ne
sia
Bra
zil
Bra
zil
Indo
ne
sia
Indo
ne
sia
Bra
zil
Co
sta
Ric
aG
han
aP
ue
rto
Ric
oIn
do
ne
sia
Indo
ne
sia
Cze
ch R
epu
blic
Gh
an
aC
olo
mbia
Ho
nd
ura
sS
ou
th A
fric
aM
exic
oP
ue
rto
Ric
oIn
do
ne
sia
Note: The upper dotted line is the median elasticity (-0.08); the lower dotted line is the mean elasticity (-
0.11).
Source: Neumark and Wascher (2007) and authors’ calculations.
26 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
papers reviewed in Neumark and Wascher’s (2007) work and 17 more recent
studies focused on LMI countries not included in Neumark and Wascher (2007).
The results include aggregate impacts for all workers but also the employment im-
pacts for specific demographic groups, geographic locations, and sectors. Put differ-
ently, where a study produced elasticity estimates for more than one cohort of
workers, we include each estimate separately. Unlike in the Neumark and
Wascher (2007) review, we have only included those estimates that were statisti-
cally significant, and it is worth noting that 55 percent of reported elasticity esti-
mates reviewed were not statistically significant. A detailed description of each of
these studies is provided in the Appendix.
In figure 2(a), the median elasticity is -0.08, represented by the upper dotted
line on the figure, while the mean is -0.11, shown as the lower dotted line. In fig-
ure 2(b), the averages are slightly higher and the upper dotted line represents a
Figure 2(b). Minimum Wage-Employment Elasticities for High-Income Countries
-4-2
02
Ela
sticity
Un
ited S
tate
s
Ca
na
da
Un
ited S
tate
s
Sw
ed
en
Fra
nce
Gre
ece
Un
ited S
tate
s
Un
ited S
tate
s
Au
stra
lia
Un
ited S
tate
s
Un
ited S
tate
s
Ca
na
da
Fra
nce
Un
ited S
tate
s
Un
ited S
tate
s
Un
ited S
tate
s
Ca
na
da
Un
ited S
tate
s
Un
ited S
tate
s
Un
ited S
tate
s
Un
ited S
tate
s
Un
ited S
tate
s
Un
ited S
tate
s
Un
ited S
tate
s
Au
stra
lia
Un
ited S
tate
s
Un
ited S
tate
s
Un
ited S
tate
s
Un
ited S
tate
s
Un
ited S
tate
s
Un
ited S
tate
s
Un
ited S
tate
s
Un
ited S
tate
s
Ca
na
da
Un
ited S
tate
s
Po
rtug
al
Ca
na
da
Ca
na
da
Un
ited S
tate
s
Un
ited S
tate
s
Un
ited S
tate
s
Ca
na
da
Un
ites
Sta
tes
Un
ited S
tate
s
Ca
na
da
Un
ited S
tate
s
Au
stra
lia
Fra
nce
Ca
na
da
Un
ited S
tate
s
Au
stra
lia
Ca
na
da
Fra
nce
Un
ited S
tate
s
Fra
nce
Au
stra
lia
Ca
na
da
Ca
na
da
Fra
nce
Note: The upper dotted line is the median elasticity (-0.23); the lower dotted line is the mean elasticity (-
0.37).
Source: Neumark and Wascher (2007) and authors’ calculations.
Bhorat et al. 27
median elasticity of -0.23, while the mean for upper-income countries is -0.37,
shown as the lower dotted line. Overall, estimated employment elasticities range
from 2.17 (Katz and Kreuger 1992) to -4.6 (Abowd, Kramarz, and Margolis 1999)
in both samples. The mean and median elasticities suggest that on average the im-
pacts of minimum wage hikes in the countries under review have been marginally
negative. From the sample of 59 developed and 32 developing country estimates,
81 percent of the elasticities were negative while 19 percent were positive. It is im-
portant to note that the absolute value of these coefficients on average is small.
This would suggest that in general the minimum wage has either benign or
slightly negative employment effects. However, the question of impact requires
careful focused empirical investigation on the basis of each minimum wage inter-
vention. In addition, it is crucial to note that in the developing country context dis-
employment effects may be further mitigated due to signficant levels of
noncompliance. We discuss this latter issue in greater detail below.
There are of course a number of studies in both developed and developing coun-
tries that do find substantial negative employment effects relative to the average
elasticities presented above. These outlier results should not be dismissed. As
Neumark and Wascher (2006) note, a study on Canadian teenagers by Baker
et al. (1999) reports an average employment elasticity of -0.25, while work by
Abowd et al. (1999) on France finds that the minimum wage had high negative
impacts on minimum wage earning men aged 25–29 years (who are not protected
by employment promotion contracts), where the employment elasticity was -4.6
relative to a comparable group of men who earn just above the minimum wage. In
Indonesia, Suryahadi et al. (2003) find an overall negative employment elasticity
of -.06, and this is more pronounced for women, where the elasticity is estimated
at -.16. An interesting paper by Feliciano (1998) shows that a decline in real mini-
mum wages in Mexico led to an increase in employment, particularly among
women, where the elasticity ranges from -0.58 to -1.25. In Honduras, Gindling
and Terrell (2007) find a wage employment elasticity of -.46 in the private sector,
while in South Africa Bhorat et al. (2014) find a decrease in agricultural employ-
ment of over 5 percent.
In addition, the figures above underscore the fact that there is a range of poten-
tial impacts of minimum wages on employment. The heterogeneity of outcomes, in
LMI countries in particular, suggests that a variety of context-specific factors inter-
act with the minimum wage. These could include: the level of the minimum wage
relative to average wages, the size of the minimum wage increase, the sector under
consideration (for example, whether it is a tradeable or nontradeable sector), the
timing of wage changes, the change postlaw in the level of worker productivity,
the enforcement regime, and the extent of compliance. While past increases in
minimum wages have generally not had a large negative affect on employment, it
is not the case that such positive outcomes will persist regardless of the level to
28 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
which a minimum wage is raised. There is a level beyond which a minimum wage
will begin to negatively affect employment, and this level may differ across geo-
graphic regions, sectors, and firms.
Figure 3, below, visually represents this idea in a basic theoretical construct by
presenting two different impact scenarios. Each of the two lines can be thought of
as labor demand functions representing the employment response to increases in
the minimum wage. The level of employment is on the y-axis, and the level of a
minimum wage increase on the x-axis. It is clear that in the case of the upper line
the minimum wage can be raised to WH (the high wage threshold) without any
impact on employment. Beyond WH, however, any increase in the hypothetical
minimum wage will begin to result in job losses. In the second case, the lower line,
the broad relationship between the minimum wage and employment stays the
same, but, crucially, employment is more sensitive to changes in the minimum
wage where wages can only be raised to WL(the low wage threshold) before they
begin to impact employment negatively.
We could interpret the upper line, for example, as representing a nontradeable
sector that can absorb wage increases more robustly, while the lower line would
represent a sector more sensitive to changes in input costs and perhaps with more
capital/labor replacement possibilities. These elements, together with the list of
abovementioned factors that influence wage-employment trade-offs, all play a role
in determining where the wage threshold lies for a given set of firms, workers, or
sectors.
Figure 3. The Relationship between Minimum Wage Adjustments and Employment
Bhorat et al. 29
One distinct possibility is that minimum wages have larger negative effects
when they are higher relative to average wages. In order to briefly examine this re-
lationship further, we plot the statistically significant point estimates of employ-
ment effects (shown in figures 2[a] and [b]) against the Kaitz ratio for each
country in the study. The figure is shown in the Appendix (figure A1). While this
is a fairly crude approach, the relationship in this case is not statistically signifi-
cant. This could be for several data-specific reasons, such as, for example, that the
employment effects are often calculated for specific subgroups or sectors within a
country for which it is difficult to calculate Kaitz ratios. However, more impor-
tantly, what this statistically insignificant relationship suggests is that the relative
level at which the minimum wage is set is not necessarily the key, or indeed the
only, factor that may impact the employment effects emanating from the mini-
mum wage—a point we emphasize in figure 3 above.
The functions above also suggest a key insight into estimated wage-employment
elasticities: namely that they can be nonlinear. Put differently, if a 50 percent in-
crease in the minimum wage results in a 10 percent drop in employment (where
the wage-employment elasticity is thus -0.2), it is not the case that a 100 percent
increase will result in a 20 percent drop in employment—the elasticity will not re-
main at -0.2. The literature suggests that where increases in a minimum wage are
large and immediate this can result in employment losses, especially for unskilled
workers. More modest increases usually have very few observably adverse effects
and may have large positive impacts on wages and a range of other labor market
outcomes. The positive wage impacts in a legislated increase appear to hold even
in situations of weak enforcement (Dinkelman and Ranchhod 2013; Bhorat et al.
2015).
Minimum Wages in Sub-Saharan Africa
Minimum wage systems in SSA, as elsewhere in the world, can be helpfully classi-
fied into three broad categories: national, sectoral (occupational), and some combi-
nation of the two (hybrid). Needless to say, a national minimum wage system is a
single wage rate that applies to all workers, a sectoral system is one in which there
are separated determinations for workers in particular sectors and/or occupations,
and a combination or hybrid system can exist when the body that decides on mini-
mum wage levels also decides to whom the wage applies and when this should
change. Table A1 (see the Appendix) groups selected African countries into these
three categories. The table serves to show the variety of systems that exist in the
region. In a report by the ILO (2013), the minimum wage frameworks on the con-
tinent are categorized in a similar manner, revealing that a majority of countries
in Africa have adopted some form of sectoral or occupational minimum wage
30 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
structure with multiple wage rates rather than a single national minimum wage—
approximately 30 percent versus 61 percent, respectively. Indeed, the percentage
of African countries with multiple minimum wage systems is higher than any of
the other world regions.7 A comprehensive account of wage-setting systems
around the world can be found in Eyraud and Saget (2005).
Beyond the diverse regulatory frameworks, the level at which the wage is set
also varies substantially by the income group that a country falls into. We show
this by grouping SSA countries into low-income (LI), lower-middle-income (LMI),
and upper-middle-income (UMI) categories. We compare 37 SSA countries in these
three groups to a comparable set of 67 non-SSA countries similarly grouped.
Figure 4, below, presents average minimum wage levels by income group for SSA
and non-SSA countries.
The figure reveals substantial differences both across the three country income
groups within SSA and compared to non-SSA countries. As expected, minimum
wage levels are positively correlated with GNI per capita levels. In particular, for
lower-income countries (LICs) in Africa with a GNI per capita of US$1,045 and be-
low, mean minimum wages stood at an average of US$119 (PPP 2013). This in-
creases by 19 percent to $142 for LMI economies and then further by 218 percent
Figure 4. Monthly Minimum Wage Levels by Country Income Group, SSA and Non-SSA
Countries (US$ PPP)
010
020
030
040
050
0M
inim
um
wag
e le
vel (U
S$
PP
P)
LI SSA LI non-SSA LMI SSA LMI non-SSA UMI SSA UMI non-SSA
Notes: LI stands for low income (in black), LMI for lower-middle income (in dark gray), and UMI for upper-
middle income (in light gray).
Sources: ILO global wage database, World Bank WDI.
Bhorat et al. 31
to $366 for UMI African economies. This relationship is explored in further detail
below. In addition, our results suggest that minimum wage levels in SSA are on
average lower than elsewhere in the world for countries in the same income group
(it must be noted that in the LI group our sample of non-SSA countries is limited to
four economies, while in the UMI group our sample of SSA countries is limited to
five countries). In the LMI category, the difference in minimum wage levels is large
and significant at the 5 percent level.
In order to advance this notion of the relationship between country income lev-
els and minimum wage levels, we produce two more figures (5[a] and [b], below).
The figures plot levels of GDP per capita against the level of minimum wages for
the same group of SSA and non-SSA countries used above. We find a relationship
between GDP and the level of the minimum wage that is not dissimilar to what
has been found for the relation between the poverty line and levels of consumption
by Ravallion et al. (2009). The latter noted that based on cross-country evidence,
the value of the national poverty line rises as average consumption levels rise
Figure 5(a). Monthly Minimum Wages and GDP Per Capita (US$ PPP), Africa
AngolaBeninBotswana
Burkina Faso
Burundi
Cameroon
Chad
Comoros
CongoCôte d'Ivoire
Democratic Republic of the Congo
Equatorial Guinea
Ethiopia
Gabon
GambiaGhana
Guinea-Bissau
Kenya
Lesotho
Liberia
Madagascar
Malawi
Mali
Mauritania
Mauritius
Mozambique
Niger
Nigeria
Senegal
South Africa
SudanTanzania, United Republic of
Togo
Uganda
Zambia
010
020
030
040
0M
inim
um
wag
e (
US
$ P
PP
)
5 6 7 8 9 10Log of GDP per capita (US$ PPP)
Fitted values Minimum wage (US$ PPP)
Note: Sample based on 37 African economies, where the latest available data for each country was utilized.
Sources: ILO global wage database, World Bank WDI.
32 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
across economies. Levels of economic development are thus positively related to
the country-based poverty line. This relationship appears to hold for minimum
wage levels. As we illustrate then, for a sample of African economies and non-
African developing countries, minimum wage levels are adjusted upwards along
with the increases in GDP per capita.
Specifically, in figure 5(a), the coefficient for the underlying relationship be-
tween the log of GDP per capita and the minimum wage level in SSA is 59.42
(where the level of the minimum wage is on the y-axis). This relationship is similar
but larger for the 67 non-SSA countries presented in figure 5b, where the coeffi-
cient is 125.14. This suggests that across countries minimum wage levels in non-
SSA countries are not only higher relative to levels of GDP compared to minimum
wage levels in SSA countries, but also more responsive to increases in GDP relative
to SSA countries.
Figure 5(b). Minimum Wages and GDP Per Capita (US$ PPP), Non-SSA Developing Countries
Afghanistan
Bangladesh
Haiti
Nepal
Tajikistan
Armenia
Bhutan
Bolivia
El Salvador
Georgia
Guatemala
Guyana
Honduras
IndonesiaIndia
Kyrgyzstan
Lao PDR Moldova
Mongolia
Morocco
Nicaragua
Pakistan
Papua New Guinea
Paraguay
Philippines
Samoa
Solomon Islands
Sri Lanka
Ukraine
Uzbekistan
Vietnam
Albania
Azerbaijan
Belarus
Belize
Bosnia and Herzegovina
Brazil
Bulgaria
China
Colombia
Costa Rica
Cuba
Dominica
Dominican Republic
Ecuador
Fiji
Hungary
Iran, Islamic Rep.
Jamaica
Jordan
Kazakhstan
Lebanon
Libya
Macedonia, FYR
Malaysia
Mexico
MontenegroPanama
Peru
Romania
Serbia
Thailand
Tunisia
Turkey
Turkmenistan
Venezuela, RB
020
040
060
080
010
00
Min
imu
m w
ag
e (
US
$ P
PP
)
6 7 8 9 10Log of GDP per capita (US$ PPP)
Fitted values Minimum wage (US$ PPP)
Note: Sample based on 67 non-SSA developing economies, where the latest available data for each country
was utilized.
Sources: ILO global wage database, World Bank WDI.
Bhorat et al. 33
In attempting to disaggregate minimum wage trends within SSA in greater de-
tail, table 1 below presents country-level data for 21 countries. We focus on the
average level of the minimum wage, the mean wage, and the Kaitz ratio (mini-
mum-to-mean wage). Our estimates of country-based minimum wages represent
the average of all schedules for those economies with more than one minimum.
The Kaitz ratio, in turn, provides an indication of how high the minimum wage is
set relative to average wages. Ideally we would present median wages and use the
median wage to calculate the Kaitz ratio, but, unfortunately, data on median
wages for countries in SSA is rare and would limit our sample even further.
Table 1 reports the most recently available minimum wage rates, grouped ac-
cording to country incomes. We convert all the data into current US$ (PPP) for
the sake of comparison. The first column of minimum wage rates again makes it
clear that despite substantial variation across individual economies, there is a clear
trend showing that minimum wage levels covary positively with country income
group. The group mean and median for UMI countries is thus more than double
those of LI countries. For example, then, UMI economies such as Algeria, Gabon,
and South Africa have legislated average minimum wage levels in excess of $400
per month. LI countries, on the other hand, have promulgated minimum wage lev-
els as low as $26 a month (Burundi).8 The country GDP and minimum wage cor-
relation, though, must be emphasized as an average effect, and from the table it is
clear that there are outliers in each income group.
The mean wage figures in column 2 reveal important cross-country differences,
even within the country groups. Hence, as expected we observe a variance in the
mean wages by country—which is broadly consistent with the GNI per capita of
the country under scrutiny. There remains within this data, however, an interest-
ing approach to measure the tendency of an economy, in a comparative sense, to-
ward setting a higher minimum wage relative to other economies. One can think
of this notion in the following manner: if the ratio of the minimum wage in coun-
try i to the highest minimum wage country, country max, is higher than the ratio
of the mean wage in country i to the highest mean wage country, country max—
then this would reflect a relatively pro–minimum wage policy environment rela-
tive to other countries. Simply put, we are measuring:
Wp ¼Wm
i
Wmmax
Wli
Wlmax
24
35
where if Wp > 1, it reflects a relatively pro–minimum wage policy environment
compared to other economies, whereas if Wp < 1, we suggest a minimum wage
policy environment which is relatively benign. In the figure below, we provide
estimates of Wp for a sample of 66 developing countries. An example from the
34 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
data: Chad’s minimum wage mean is 22 percent of the region’s maximum mini-
mum wage. Yet, its mean wage nationally is only 15 percent of the region’s maxi-
mum mean wage. This suggests a relatively pro–minimum wage policy in Chad
when compared with its mean wage differential relative to other developing coun-
tries. Put in numerical terms using the formula presented above, Chad’s Wp value
is 1.45, where a value over 1 suggests a pro–minimum wage policy.
The data suggest that the majority of countries in our sample of 66 have a Wp
ratio below 1, and countries in SSA do not appear to be large outliers in this re-
gard. Indeed, the average figures for each region (presented on the right) suggest
that the EAP, LAC, and ECA regions all have higher Wp ratios than SSA. The over-
all sample average is 0.80.
Figure 7, below, focuses only on the SSA region and presents the mean and me-
dian Wp figures by country income group. There are two notable features here.
Figure 6. Degree of Relative Minimum Wage Policy Bias by Country
0.5
11.5
22.5
mea
n o
f W
p
Cu
ba
Bu
rund
iU
zbe
kis
tan
Kyrg
yzst
an
Sw
azila
nd
Bo
tsw
ana
Ve
ne
zuela
, R
BG
eorg
iaU
gan
da
Se
ne
ga
lM
au
ritius
So
uth
Afr
ica
Ga
bo
nD
om
inic
an
Rep
ub
lic
Ka
zakh
sta
nB
ang
lad
esh
Bo
snia
an
d H
erz
eg
ovin
aM
aced
on
ia, F
YR
M
on
tene
gro
Mexic
oT
anzan
ia
Vie
tnam
Jam
aic
aM
ala
ysia
Azerb
aija
nM
old
ova
Ch
ina
Se
rbia
Pa
na
ma
Mon
go
liaK
enya
Fiji
Nic
ara
gua
India
Mala
wi
Co
sta
Ric
aE
l S
alv
ado
rA
rmen
iaZ
am
bia
Bu
lga
ria
Gh
an
aIn
do
ne
sia
Ukra
ine
Eth
iopia
Be
laru
sB
razil
Alb
an
iaP
eru
R
om
an
iaA
lgeri
aT
ajik
ista
nP
hili
pp
ine
sT
haila
nd
Le
soth
oH
ond
ura
sC
ong
o, R
ep
.E
cua
do
rC
had
Pa
kista
n
Gu
ate
ma
laB
urk
ina
Faso
Mad
ag
asca
rA
fgh
an
ista
nB
oliv
iaC
ong
o, D
em
. R
ep
Arg
entina
SA
Me
an
SS
A M
ea
nE
AP
Mea
nLA
C M
ea
nE
CA
Mea
n
Notes: See Mean on far right for legend: South Asia (SA), Sub-Saharan Africa (SSA), East Asia and Pacific
(EAP), Latin America and the Caribbean (LAC), Europe and Central Asia (ECA). Sample based on 66 countries,
where the latest available data for each country was utilized.
Sources: ILO Global Wage database, World Bank WDI.
Bhorat et al. 35
Firstly, the figure shows that the ratio of minimum wages in Africa to the maxi-
mum minimum wage on the continent is greater than the similar ratio in relation
to mean wages. Secondly, the LIC mean and median values are higher than both
the LMI and UMI sample of countries. This would suggest that despite absolute
minimum wage levels increasing with GNI per capita, low-income African econo-
mies are relatively more pro–minimum wage in their policy setting when com-
pared with the mean relative wages on the continent.
Reverting back to table 1, we provide in column three, the common in-country
measure of the extent to which the minimum wage has “bite”—namely the Kaitz
ratio. The ratio is simply that of the minimum wage to the mean wage.
The Kaitz ratio values range from 0.03 in Burundi to a value of 1.27 in the
Democratic Republic of the Congo. The overall SSA mean is instructive and of use in
and of itself: for the region, minimum wages are 37 percent of mean wages and 28
percent of the median wage. This ratio varies by income classification; so while the
Kaitz is 0.28 at the mean for UMI African economies, it is 0.31 for LMI economies
and 0.46 for LICs. We observe substantially higher ratios for LI countries, where the
group median is 0.16 points above the median for LMI countries in the region and
0.26 points above the UMI median. Simply put, low-income African countries are
setting higher minimum wages relative to their domestic mean wages, compared
with both UMI and LMI economies in the region. When compared with other
Figure 7. Average Wp Values for SSA by Country Income Group0
.51
1.5
Wp
UMI median UMI mean LMI median LMI mean LIC median LIC mean
Sources: ILO Global Wage database, World Bank WDI.
36 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
Table 1. Monthly Average Minimum Wage Estimates: Sub-Saharan Africa
Country Minimum wage (US$ PPP) Mean wage (US$ PPP) Kaitz ratio
Low-income economies
Burkina Faso 138 210 0.66
Burundi 26 256 0.10
Chad 239 371 0.64
Congo, Dem. Rep 68 53 1.27
Ethiopia 77 175 0.44
Madagascar 128 183 0.7
Malawi 49 368 0.13
Tanzania 149 624 0.24
Uganda 65 464 0.10
Group mean 104 300 0.46
Group median 77 256 0.44
Lower-middle-income economies
Congo, Rep. 145 526 0.28
Ghana 128 469 0.27
Kenya 331 979 0.34
Lesotho 242 377 0.64
Senegal 148 983 0.15
Swaziland 94 815 0.12
Zambia 98 252 0.39
Group mean 169 629 0.31
Group median 145 526 0.28
Upper-middle-income economies
Algeria 531 1,003 0.53
Botswana 148 1,287 0.12
Gabon 418 2,356 0.18
Mauritius 218 1,424 0.15
South Africa 517 1,251 0.41
Group mean 366 1,464 0.28
Group median 418 1,287 0.18
Total SSA mean 188 687 0.37
Total SSA median 145 469 0.28
Other regional averages
LAC mean 369 937 0.46
(1.24)*
LAC median 289 859 0.37
(1.32)*
EAP mean 317 884 0.38
(1.03)*
EAP median 284 739 0.32
(1.14)*
SA mean 233 386 0.63
(1.70)*
continued
Bhorat et al. 37
regions, however, the data show that mean and median Kaitz ratios are lower in
SSA. Indeed, the mean and median Kaitz ratios of other regions relative to that of
SSA show that it is only the ECA mean that has a lower Kaitz than that of the SSA
region. This provides some support for the evidence presented in figure 6.
Ultimately, then, the cross-country evidence on minimum wages in Sub-Saharan
Africa indicates firstly that if we simply group countries according to broad income
levels, minimum wages in Sub-Saharan Africa appear in general to be set lower
than those in the rest of the developing world. Secondly, the evidence illustrates a
positive and linear relationship between GDP per capita and the level at which the
minimum wage is set. Consistent with the positive country poverty line—GDP rela-
tionship, and in keeping with other developing countries—African economies reveal
an upward adjustment in the value of the minimum wage as the economies in the
region grow and develop. Thirdly, when assessing the extent to which African gov-
ernments may be more pro–minimum wage, or not, in setting wage levels compared
to other regions, our cross-country evidence indicates that there is no tendency to-
wards progressive minimum wage policy in our sample. Indeed, the regional Kaitz
ratios presented in table 1 are lower for SSA than any other region except ECA.
Finally, it is clear again that when examining the Kaitz index, while SSA economies
overall are setting minimum wages at just over a third of the mean wage, substan-
tially higher ratios are observed for low-income countries in the region.
Variations in African Minimum Wage Schedules:A Country-Level Overview
For the purposes of a cross-country comparison, the above estimates have aggregated
across the different minimum wage schedules that exist within many economies.
Table 1. Continued
Country Minimum wage (US$ PPP) Mean wage (US$ PPP) Kaitz ratio
SA median 255 368 0.59
(2.11)*
ECA mean 325 1,136 0.3
(0.81)*
ECA median 344 1,183 0.28
(1.00)*
Notes: SSA aggregate estimates based on 21 countries, EAP sample based on eight countries, LAC sample based
on 16 countries, SA sample based on four countries, ECA sample based on 17 countries. The latest available data
for each country was utilized. Wages are monthly, and where there is more than one minimum wage schedule,
we have estimated the average minimum wage across the within-country schedules. All estimates are in current
US$ PPP. Asterisk (*) indicates ratio of measure to SSA mean or median.
Sources: ILO Global Wage database, World Bank WDI.
38 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
There is, however, much granularity and nuance that is overlooked when presenting
average minimum wage estimates by country in this fashion. We return here, then,
to a more detailed consideration of minimum wage schedules, focusing on seven SSA
economies for which we have the appropriate microdata.
We have alluded to fact that many minimum wage systems in the region in-
clude a range of sectoral and occupational schedules, and it is these complex
schedules that make a single figure appear blunt as it masks substantial heteroge-
neity at a country level. As an example, the case of Kenya’s minimum wage regime
is instructive in terms of the complexity of minimum wage systems: Kenya has had
an active minimum wage policy since it achieved independence in 1964.
Minimum wages are set by Ministerial order following recommendations by a tri-
partite council and public consultation (ILO 2014). The wage schedule that is pro-
duced is intricate. Wages are set at different rates for agricultural and
nonagricultural occupations, and within these categories there are geographical
and occupational distinctions, each with a unique set of wages. Table A2 presents
daily minimum wage rates for the agricultural sector, which is broken down into
10 categories according to the specific type of employee.9 Within the agricultural
sector, daily wages can range from 203 Kenyan Shillings for a general worker to
370 Kenyan Shillings for a lorry or car driver. Table A3 details monthly minimum
wages for the nonagricultural sector, which is firstly disaggregated by geographic
region into three categories: cities, municipalities and town councils, and other
areas. Within each of these three geographical areas, wages are further delineated
across 15 different employee types. Monthly minimum wages for nonagricultural
workers range from 5,217 to 22,070 Kenyan shillings. In total, Kenya has 55 sep-
arate minimum wage rates.
It is clear how the complexity of Kenya’s wage determination system makes ob-
taining accurate aggregate estimates difficult. But it must be emphasized that
Kenya is not an exception in this regard. South Africa’s legislated minimum wage
schedule, for example, is even more complex for certain sectors, and in total the
country currently has 124 different minimum wage rates. This excludes minimum
wages agreed upon within Bargaining Councils, which would push the number of
specific wage rates in the thousands.
In table 2, below, we provide a brief overview of the number of minimum wage
schedules for 12 countries in SSA. The data show that seven of the 12 countries in
the sample have 10 or more wage schedules, with South Africa being an outlier in
this sample. A differentiated, or complex, wage schedule can be useful in that it
takes account of variations across worker skill levels, geographic regions, and sec-
tors. Yet increasing levels of complexity also makes wage setting, as well as en-
forcement and compliance, more difficult. The ILO (2014) suggests that in general
the complexity of a wage schedule should be commensurate with the county’s re-
source availability, where simpler wage schedules are more suitable if the
Bhorat et al. 39
resources dedicated to minimum wage systems are few. A complex schedule such
as Kenya’s certainly creates a more challenging set of rates to establish each year
and enforce. Labor market regulations in several SSA countries would probably
benefit from a re-assessment of current minimum wage systems in this regard,
with a view toward greater simplification.
Given the complexity of schedules, it is useful to look beyond cross-country ag-
gregates, which, while useful, cannot provide a detailed picture or give any indica-
tion toward levels of compliance with minimum wage laws. In order to do this, we
make use of household survey data for seven SSA countries (South Africa,
Uganda, Kenya, Zambia, Tanzania, Mali, and Namibia).
As tables 2, A2, and A3 indicate, however, there remain significant complexities
within several of these country minimum wage systems—many of which go be-
yond the detail provided by a basic labor force survey. To deal with this, for the
countries with more than one minimum wage we select and focus on two key sec-
tors. The first, which we call the “lower floor” is a low-paid, unskilled sector cov-
ered by a general minimum wage (usually agriculture or “general workers”),
while the second sector we select is higher-paid and medium-skilled (such as retail
trade or working as a clerk), and we call this the “upper floor.”10 This allows us to
present the granularity of minimum wage schedules while at the same time
enabling us to estimate more accurate Kaitz ratios (for specific sectors), as well as
exploring levels of compliance (see the section Minimum Wage Compliance in Sub-
Saharan Africa). To begin, we present kernel density estimates that provide a basic
picture of where specific minimum wages are set relative to the wage distributions
Table 2. Number of Minimum Wage Schedules by African Country
Country Number of wage schedules
Uganda 1
Mali 1
Ghana 1
Malawi 1
Nigeria 2
Botswana 10
Zambia 10
Tanzania 29
Namibia 32
Kenya 55
Ethiopia (public sector) 57
South Africa 124
Average 27
Source: ILO TRAVAIL Database.
40 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
Figure 8. (a) Distribution of Wages, Zambia (2010); (b) Distribution of Wages, Tanzania
(2011); (c) Distribution of Wages, South Africa (2013); (d) Distribution of Wages, Namibia
(2012); (e) Distribution of Wages, Uganda (2012); (f) Distribution of Wages, Kenya (2005/6).
(a) (b)
(c) (d)
(e) (f)
Source: Zambia: Living Conditions Monitoring Survey (2010); Tanzania: Integrated Labour Force Survey
(2005/06); South Africa: Labour Market Dynamics Study (2013); Namibia: Labour Force Survey (2012);
Uganda: Uganda National Panel Survey (2012); Kenya: Kenya Integrated Household Budget Survey (2005-
06).
Bhorat et al. 41
in each country.11 We include all employed wage earners in the sample, regardless
of occupation, sector, or whether they are covered by a minimum wage or not.
In all cases, the distributional graphs indicate the lack of a traditional “spike” at
either of minimum wage schedules. This is interesting in two respects: Firstly, un-
like most of the developed country literature, this six-country sample of African
economies shows no evidence of a spike in the wage distribution relevant to the
level of the minimum. Secondly, though, this lack of a spike is also consistent with
fairly fat tails to the left of both minimum wage schedules. Put differently, there is
a significant share of earners not adhering to the minimum wage. This incidence
and deficit of noncompliance is taken up in the next section. Ultimately, the notion
of a subset of African economies displaying relatively high minimum wage levels
that in turn coexist with a fair degree of independence from the country’s self-
same wage distribution is a key result.
Table 3, below, provides more detailed information on minimum wages for the
six SSA countries in the figures above, as well as for Mali. In two of the seven
countries in the table, there are single national minimum wages (Mali and
Uganda), while in the other five there are detailed sectoral and occupational wage
schedules. As noted above, for each of these five countries we select a lower and
upper wage floor and identify the employees covered by these sectoral wage levels.
For the two countries with a single national minimum wage, we simply include all
workers classified as wage-earning employees in our analysis. The table provides
an overview of wage levels, minimum wage levels, coverage, and we also draw at-
tention to the difference between the level of the lower and upper floors, where ap-
plicable. What emerges firstly and most obviously is the within-country variation
in minimum wage levels.
Among the seven countries profiled in table 3 and across the lower and upper
floors in the same country, large variations are apparent both in the level of legis-
lated minimum wages and in mean and median earnings. Column five shows that
minimum wage levels differ substantially across the lower and upper floor categories
within the same country, with wages for workers in lower floor, or unskilled catego-
ries, set at 54 percent of wage levels for upper floor categories on average. While the
heterogeneity in minimum wage levels makes for a complex wage schedule, it also
appears to be reflective of the substantial intersectoral wage inequality, which is evi-
dent in the differences in average wages for the same country. It is also worth noting
the large difference between the mean and median earnings across any one category
of workers, who are assigned here to a lower or upper floor. These estimates are sug-
gestive of high levels of intrasector wage inequality—where mean wages are signifi-
cantly larger than median wages. On average across the group of countries, mean
wages (US$457) are 55 percent higher than median wages (US$251).
Importantly, though, the data reveals an important characteristic of minimum
wage setting in Africa—namely that in general these sectoral wages, given their
42 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
specific targeting—will not cover a large share of workers. Apart from Uganda and
Mali, where a national minimum wage exists, coverage of most sectoral minimum
wages is fairly low. In the pursuit of higher wages for vulnerable workers, it is pos-
sible that many African economies do not end up covering a large share of the
wage employed.
To go beyond focusing on the variation in wages and wage levels, we examine
the Kaitz ratios using both mean and median wages for workers in the lower and
upper floor categories. These ratios give an indication of how high minimum wages
are set relative to average wages within in each category and also provide suggestive
evidence regarding compliance, which we take up in more detail shortly.
Figure 9, below, presents the Kaitz ratios for mean wages across categories and
for each country. The ratios vary widely across countries, as well as across the
lower and upper floor categories. In Namibia, for example, for workers in the lower
floor category (“general workers”), the minimum wage is set at 1.5 times the
mean wage, while in Zambia this figure is 0.5.
Overall, however, the average ratio is 0.93 among lower floor workers and 0.63
among upper floor workers. As a point of comparison, the average level of
Table 3. Monthly Minimum Wages (US$ PPP) In Seven African Economies: Upper and LowerFloors
Countries Sector % of totalemployees
Minimum wage Lower/upperfloor
Meanwage
Medianwage(US$ PPP)
South Africa (2013) Lower floor 4 441 0.65 558 329
Upper floor 13 680 1,475 526
Kenya (2005) Lower floor 7 116 0.44 146 96
Upper floor 2 264 470 353
Zambia (2010) Lower floor – 69 0.38 124 78
Upper floor 7 185 417 284
Tanzania (2007) Lower floor 61 162 0.59 100 52
Upper floor 2 274 190 87
Namibia (2012) Lower floor – 388 0.63 257 174
Upper floor 4 615 1,057 677
Uganda (2012) National 100 65 N/A 227 107
Mali (2013)*** National 100 132 N/A – –
Mean 0.3 283 0.54 457 251
Notes: Figures only include those classified in household surveys as being wage-earning employees. The
Ugandan minimum wage has not been updated since 1984, thus we take the 1984 rate and adjust for inflation to
obtain a comparable 2012 figure. The data from Mali is taken from Rani et al. (2013), and we are unable to calcu-
late mean or median wages. Wages are in current monthly US$ PPP.
Sources: South Africa: Labour Market Dynamics Study (2013); Kenya: Kenya Integrated Household Budget
Survey (2005–6); Uganda: Uganda National Panel Survey (2012); Mali: Rani et al. (2013); Zambia: Living
Conditions Monitoring Survey (2010); Tanzania: Integrated Labour Force Survey (2005–6); Namibia: Labour
Force Survey (2012).
Bhorat et al. 43
minimum-to-mean wages for nine LMI countries outside of SSA12, presented in
Rani et al. (2013), was 54 percent.
Figure 10 presents the Kaitz ratio using median instead of mean wages, and the
different results are stark, as table 3 suggested. The ratios are significantly higher,
Figure 9. Ratio of Minimum to Mean Wages, Seven African Economies0
.51
1.5
Uganda Zambia Mali South Africa Kenya Namibia Tanzania Average
Lower floor Upper floor
Source: Authors’ calculations.
Figure 10. Ratio of Minimum to Median Wages, Seven African Economies
01
23
Uganda Mali Zambia Kenya South Africa Namibia Tanzania Average
Lower floor Upper floor
Source: Authors’ calculations.
44 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
with the average figures above 1, and the black dotted line makes this clear.
Again, comparing these estimates against the sample of non-SSA countries from
Rani et al. (2013), where the average ratio is 0.76, shows that indeed the figures
are high.
Overall, the Kaitz estimates for the SSA region highlight two main aspects of
minimum wages: firstly, the country-specific nature of minimum wage frame-
works, the level at which wages are set and the relation of this level to average
wages for that covered group; and secondly, given that we are only focusing on
covered workers in the two figures above, the data show that minimum wages are
set high relative to average wages and are suggestive of significant noncompliance.
This leads us to explore levels of noncompliance in more detail.
Minimum Wage Compliance in Sub-Saharan Africa
While the Kaitz ratios presented above are indicative of widespread noncompliance
with minimum wage laws, a more robust method is required to investigate further.
We apply an Index of Violation13 from Bhorat, Kanbur, and Mayet (2012) to cal-
culate the level and depth of noncompliance, which provides a more
Figure 11. Average Compliance Rates (V0), African and Developing Country Comparison
0.2
.4.6
.8
Mali
Za
mbia
Ug
an
da
Ken
ya
Sou
th A
fric
a
Na
mib
ia
Ta
nza
nia
SS
A A
ve
rag
e
Vie
t N
am
Mexic
o
Bra
zil
Peru
Co
sta
Ric
a
India
Phili
pp
ines
Tu
rkey
Indo
ne
sia
No
n-S
SA
Avera
ge
Source: Authors’ calculations and Rani et al. (2013).
Bhorat et al. 45
comprehensive picture. To do this, we combine the lower and upper floor catego-
ries presented above where applicable, and we compare the resulting estimates
with those of Rani et al. (2013). The Index of Violation allows us to calculate the
level of noncompliance, or V0, which is simply the percentage of workers who earn
below the minimum wage that applies to them. It also allows us to go beyond this
to calculate the depth of this noncompliance, or V1, which measures how far below
the minimum wage these workers earn on average.
The data in figure 11 clearly show that for most countries in the SSA region, for
the sectors we include, noncompliance is widespread. On average, 58 percent of
workers earn below the minimum wage legislated for them. This is compared to an
average of 30 percent for the non-SSA countries in the figure. This relationship
holds for the mean and median, despite outlier economies such as Tanzania. The
figure also highlights the cross-country differences in levels of noncompliance. In
Zambia, approximately 36 percent of workers earn subminimum wages, while this
rises to 80 percent in Tanzania. These numbers reinforce the picture provided by
the Kaitz ratios above where minimum wage rates were shown to be set high rela-
tive to average wages.
Figure 12 moves beyond the level of compliance to explore how far below the
minimum wage workers are on average. This gives a more nuanced picture of
Figure 12. Average Depth of Noncompliance (V1), African and Developing Country
Comparison
0.2
.4.6
V1
Za
mbia
Ug
an
da
Ken
ya
Sou
th A
fric
a
Na
mib
ia
Mali
Ta
nza
nia
SS
A A
ve
rag
e
Tu
rkey
Vie
t N
am
India
Co
sta
Ric
a
Mexic
o
Phili
pp
ines
Indo
ne
sia
Bra
zil
Peru
No
n-S
SA
Avera
ge
Source: Authors’ calculations and Rani et al. (2013).
46 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
noncompliance and how severe noncompliance levels are when examining the av-
erage distance below the minimum wage that workers are earning.
The results suggest an interesting switch: while noncompliance levels were
higher in the African sample of economies, relative levels of noncompliance are in
fact higher in the non-African sample. Hence, our estimates show that the SSA re-
gional average V1 estimate stands at 0.30 while the corresponding figure for the
non-SSA countries is 0.35. This suggests, at least on the basis of this subsample of
developing countries, that while Africa yields to higher levels of noncompliance
than other developing countries, those who are below the minimum wage face a
greater disadvantage in non-African economies. Put simply, absolute levels of non-
compliance are higher in Africa, while relative levels of noncompliance are higher
in non-African developing countries.
Conclusion
Most countries in Sub-Saharan Africa (SSA) have adopted minimum wage regula-
tion. Although the sectors and fraction of workers covered are small given the low
rates of formality and urbanization in SSA, as the number of covered workers
grows, wage regulation will become increasingly significant for the economy as a
whole. In addition, there can be spillover effects in uncovered sectors, which can
be exacerbated, as we show in the paper, when wage regulation is particularly pro-
gressive. Current experience with minimum wages in SSA is thus relevant for ana-
lysts and policy makers as economies may consider redesigning their minimum
wage architecture in particular, and active labor market policy in general.
In examining the variety of minimum wage frameworks in the region, it is clear
that the typology of minimum wage schedules, as well as the levels at which these
minima are set, varies considerably across countries in Africa. Our evidence shows
that higher minimum wage values are associated with higher GDP per capita. In
addition, utilizing two different pieces of evidence, we illustrate that SSA does not
seem to reflect a particular bias towards a more pro–minimum wage policy relative
to other regions of the developing world. Importantly, however, we find that mini-
mum wages in low-income countries in SSA are set at values relative to the coun-
try’s mean wage that are higher than those in lower- and upper-middle income
countries in Africa.
There is limited research on the employment effect of minimum wages in SSA,
but the findings for the four countries (Ghana, Kenya, Malawi, and South Africa)
are consistent with the broad summary of global research. By and large, introducing
and raising the minimum wage has a small negative impact or no measurable nega-
tive impact. However, there is significant variation around this average finding, and
some country studies do present evidence of employment losses in excess of 10
Bhorat et al. 47
percent in certain cases. In short, employment elasticities are neither constant nor
linear. Where increases in a minimum wage are large and immediate, this can result
in employment losses, but more modest increases usually have few observably ad-
verse effects on employment and may have positive impacts on wages.
The great variability in findings on employment can be explained partly by the
great variation in the detail of the minimum wage regimes and schedules country
by country, but also by the variations in compliance. Our data on minimum wage
compliance in Africa illustrates that the continent has higher levels of absolute
noncompliance when compared with other developing countries, but lower levels
of relative noncompliance. However, there is considerable variation across coun-
tries. We find that higher Kaitz indices are associated with higher levels of non-
compliance, but more detailed explanation of noncompliance is an important item
on the research agenda.
This paper has provided an empirical overview of minimum wages in Sub-
Saharan Africa. It is evident that research in this area is still at a very early stage,
with this paper in essence attempting to craft the broad contours of the nature and
extent of minimum wage setting on the continent. While work on minimum wages
is fairly mature in many OECD countries, our understanding of minimum wage
policy and its impact in SSA is not. This is in large part due to a severe lack of nec-
essary data. The release of country-level earnings and employment data at regular
intervals lies at the heart of a future country-focused minimum wage research
agenda for Africa.
Notes
Haroon Bhorat is the Director of the Development Policy Research Unit, School of Economics,University of Cape Town, South Africa; Ravi Kanbur is the T. H. Lee Professor of World Affairs,International Professor of Applied Economics and Management, and Professor of Economics, CornellUniversity; and Benjamin Stanwix is a Senior Researcher in the Development Policy Research Unit,University of Cape Town, South Africa. The authors of this paper acknowledge Benjamin Jourdan forproviding a number of graphs and tables used for the study.
1. In Sub-Saharan Africa, data indicates that approximately 19 percent of the labor force is inwage employment, while 74 percent is in agricultural or nonfarm self-employment (Bhorat, Naidoo,and Pillay 2015). In West Africa, for example, the informal sector accounts for approximately 50 per-cent of national output, over 80 percent of employment and 90 percent of new jobs (Benjamin,Golub, and Mbaye 2015).
2. While there have been several studies analyzing the lighthouse effect of minimum wage lawsin many countries around the world, we know of no such work done for African countries. Relatedresearch on the impact of minimum wage laws in sectors with weak enforcement has been done inSouth Africa (see Dinkelman and Ranchhod 2013 and Bhorat, Kanbur, and Stanwix 2015).
3. Indeed, many countries have wage-setting legislation that takes cost of living measures intoaccount when updating minimum wages.
4. Belman and Wolfson (2014); Neumark, Salas, and Wascher (2014); Schmitt (2013);Neumark and Wascher (2007, 2008).
48 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
5. See, for example, Agell and Lommerud (1997) or for a discussion of the possible impacts ofminimum wages from a behavioral perspective, see Falk, Fehr, and Zehnder (2006).
6. The country studies focus on Brazil, Chile, Colombia, Costa Rica, Indonesia, Mexico, PuertoRico, and Trinidad and Tobago. Neumark and Wascher (2007) use the term “developing country” todescribe what we have referred to as low- and middle-income countries.
7. The table below shows the percentage split across world regions between setting a nationalminimum wage versus some form or regional, sectoral, or occupational minimum wage. The percent-ages do not add up to 100 percent given that not all countries in a region have minimum wages.Source: ILO (2013).
8. It should be noted that in the case of Burundi the minimum wage has not been updated since1988; the figure presented here is adjusted for inflation to 2012, priced and converted in US$ PPP.
9. Unskilled employee; stockman, herdsman, and watchman; skilled and semiskilled employees;house servant or cook; farm foreman; farm clerk; section foreman; farm artisan; tractor driver; com-bined harvester-driver; lorry driver or car driver.
10. The specific sectors for each country are as follows:South AfricaLower floor – AgricultureUpper floor – Wholesale and retailKenyaLower floor – AgricultureUpper floor – Employee type [G], see table A2ZambiaLower floor – General workerUpper floor – ClerksTanzaniaLower floor – AgricultureUpper floor – HospitalityNamibia (wages set by collective bargaining)Lower floor – General workersUpper floor – Clerks
11. We only do this for six countries as the data for Mali are taken from Rani et al. (2013).12. Brazil, Costa Rica, India, Indonesia, Mexico, Philippines, Peru, Turkey, Viet Nam.13. If we consider a distribution of wages F(w), where the density function is f(w) and the mini-
mum wage is Wm; then an index of violation can be calculated as follows:
V að Þ ¼ðWm
0
Wm �Wi
Wm
� �a
f wð Þ;
where Wi are individual wages and a is the “violation aversion” parameter such that when: a¼0,V(a) measures the percentage of workers below the minimum; when a¼1, V(a) measures the aver-age gap between Wm; and when a¼2, V(a) is the squared violation gap. This index allows us to cal-culate a family of measures that capture both absolute (V0) and relative (V1, V2) levels ofnoncompliance.
Bhorat et al. 49
References
Abowd, J., F. Kramarz, and D. Margolis. 1999. “Minimum Wages and Employment in France andthe United States.” Working Paper 6996. National Bureau of Economic Research, Cambridge,MA.
Agell, J., and K. E. Lommerud. 1997. “Minimum Wages and the Incentives for Skill Formation.”Journal of Public Economics 64 (1) : 25–40.
Alatas, V., and L. Cameron. 2003. “The Impact of Minimum Wages on Employment in a Low IncomeCountry: An Evaluation Using the Difference-in-Differences Approach.” Working Paper 2985.World Bank, Policy Research Department, Washington, DC.
Almeida, R., and L. Ronconi. 2012. “The Enforcement of Labor Law in the Developing World: SomeStylized Facts from Labor Inspections.” Paper Presented at the 7th Institute for the Study of Labor/World Bank Conference on Employment and Developing, 2012–11, New Delhi, India.
Andal�on, M., and C. Pages. 2008. “Minimum Wages in Kenya.” Discussion Paper 3390. Institute forthe Study of Labor, Bonn, Germany.
Bell, L. A. 1997. “The Impact of Minimum Wages in Mexico and Colombia.” Journal of LaborEconomics 15 (S3) : S102–35.
Belman, D., and P. J. Wolfson. 2014. “What Does the Minimum Wage Do?” WE Upjohn Institute,Kalamazoo, MI.
Benjamin, N., S. Golub, and A. A. Mbaye. 2015. “The Informal Sector, Employment, and EconomicGrowth: Recommendations for Effective Policies.” In E. Zedillo, O. Cattaneo, and H. Wheeler eds.,AFRICA at a Fork in the Road. 163–77. New Haven, CT: Yale Center for the Study ofGlobalization.
Bhorat, H., R. Kanbur, and N. Mayet. 2012. “Minimum Wage Violation in South Africa.”International Labour Review 151 (3) : 277–87.
———. 2013. “The Impact of Sectoral Minimum Wage Laws on Employment, Wages, and Hours ofWork in South Africa.” IZA Journal of Labor and Development 2 (1) : 1–27.
Bhorat, H., R. Kanbur, and B. Stanwix. 2014. “Estimating the Impact of Minimum Wages onEmployment, Wages, and Non-Wage Benefits: The Case of Agriculture in South Africa.” AmericanJournal of Agricultural Economics 96 (5) : 1,402–19.
———. 2015. “Minimum Wages in Sub-Saharan Africa: A Primer.” Discussion Paper 9204.Institute for the Study of Labor, Bonn, Germany.
Bhorat, H., K. Naidoo, and K. Pillay. 2015. “Growth, Poverty and Inequality Interactions in Africa:An Overview of Key Issues.” Unpublished mimeograph. University of Cape Town, DevelopmentPolicy Research Unit, Cape Town, South Africa.
Card, D., and A. Krueger. 1994. “Minimum Wages and Employment: A Case Study of the Fast-FoodIndustry in New Jersey and Pennsylvania.” The American Economic Review 772–93.
Comola, M., and L. De Mello. 2011. “How Does Decentralized Minimum Wage Setting AffectEmployment and Informality? The Case of Indonesia.” Review of Income and Wealth 57 (S1) :S79–99.
Del Carpio, X. N., and C. W. Liang. 2013. “The Impacts of Minimum Wages on Employment, Wagesand Welfare: The Case of Vietnam.” Working Paper. World Bank, East Asia and Pacific SocialProtection and Labor Unit, Washington, DC.
Del Capio, X., D. Margolis, and Y. Okamura. 2013. “Effects of Minimum Wages on Labor Marketsand Welfare in the Philippines: A Micro-Simulation-Based Estimation.” Working Paper. WorldBank, East Asia and Pacific Social Protection and Labor Unit, Washington, DC.
50 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
Del Carpio, X., J. Messina, and A. Sanz-de-Galdeano. 2014. “Minimum Wage: Does It ImproveWelfare in Thailand?” Discussion Paper 7911. Institute for the Study of Labor, Bonn, Germany.
———. 2014. “Minimum Wage: Does It Improve Welfare in Thailand?” Discussion Paper 7911.Institute for the Study of Labor, Bonn, Germany.
Dinkelman, T., and V. Ranchhod. 2012. “Evidence on the Impact of Minimum Wage Laws in anInformal Sector: Domestic Workers in South Africa. Journal of Development Economics 99 (1) :27–45.
Dykstra, S., B. Dykstra, and J. Sandefur. 2014. “We Just Ran Twenty-Three Million Queries of theWorld Bank’s Website.” Working Paper 362. Center for Global Development, Washington, DC.
Eyraud, F., and C. Saget. 2005. The Fundamentals of Minimum Wage Fixing. Geneva, Switzerland:International Labour Organization.
Fajnzylber, P. 2001. “Minimum Wage Effects throughout the Wage Distribution: Evidence fromBrazil’s Formal and Informal Sectors.” Unpublished manuscript. Universidade Federal Do BeloHorizonte, Department of Economics and CEDEPLAR, Minas Gerais, Brazil.
Falk, A., E. Fehr, and C. Zehnder. 2006. “Fairness Perceptions and Reservation Wages: The BehavioralEffects of Minimum Wage Laws.” The Quarterly Journal of Economics 121 (4) : 1,347–81.
Fang, T., and C. Lin. 2013. “Minimum Wages and Employment in China.” Discussion Paper 7813.Institute for the Study of Labor, Bonn, Germany.
Garbers, C. 2015. “The Impact of the Agricultural Minimum Wage on Farmworker Employment inSouth Africa: A Fixed Effects Approach.” Unpublished masters thesis. Stellenbosch University,Cape Town, South Africa.
Gindling, T. H. 2014. “Does Increasing the Minimum Wage Reduce Poverty in DevelopingCountries?” Working Paper. World of Labor, Institute for the Study of Labor, Bonn, Germany.
Gindling, T. and K. Terrell. 2005. “The Effect of Minimum Wages on Actual Wages in Formal andInformal Sectors in Costa Rica.” World Development 33 (11) : 1,905–21.
———. 2007. “The Effects of Multiple Minimum Wages throughout the Labor Market: The Case ofCosta Rica.” Labour Economics 485–511.
Ham, A. 2013. “Revisiting the Effects of Minimum Wages in Developing Countries: Evidence from aParticular Policy Change in Honduras.” University of Illinois at Urbana–Champaign.
Hyslop, D., and S. Stillman. 2004. “Youth Minimum Wage Reform and the Labour Market.Discussion Paper 1091. Institute for the Study of Labor, Bonn, Germany.
Harrison, A., and J. Scorse. 2010. “Multinationals and Anti-Sweatshop Activism.” American EconomicReview 100 (1) : 247–73.
Hertz, T. 2005. “The Effect of Minimum Wages on the Employment and Earning of South Africa’sDomestic Service Workers.” Working Paper 05-120. Upjohn Institute, Kalamazoo, MI.
ILO. 2013. Global Employment Trends for Youth, 2013: A Generation at Risk. Geneva, Switzerland:International Labour Organisation.
Jones, P. 1997. “The Impact of Minimum Wage Legislation in Developing Countries Where CoverageIs Incomplete.” Working Paper WPS/98-2. University of Oxford, Center for the Study of AfricanEconomies, Oxford, UK.
Katz, L., and A. Krueger. 1992. “The Effects of the Minimum Wage on the Fast Food Industry.”Industrial and Labor Relations Review 6–21.
Kharmis, M. 2013. “Does the Minimum Wage Have a Higher Impact on the Informal than on theFormal Labour Market? Evidence from Quasi-Experiments.” Applied Economics 45 (4) : 477–95.
Bhorat et al. 51
Lanzona, L. 2012. “The Impact of Minimum Wages on Output and Employment in Labor-IntensiveManufacturing Industries in the Philippines.” Background Paper for the Philippine DevelopmentReport, 2013. World Bank, Manila, Philippines.
Lemos, S. 2004. “Minimum Wage Policy and Employment Effects: Evidence from Brazil.” Economia219–66.
——— 2006. “The Minimum Wage Effects in a Developing Country.” Discussion Paper 01/06.University of Leicester, Leicester, UK.
——— 2007. “Minimum Wage Effects across the Private and Public Sectors in Brazil.” Journal ofDevelopment Studies 700–20.
——— 2009. “Minimum Wage Effects in a Developing Country.” Labour Economics 16 (2) : 224–37.
Livingstone, I. 1995. “Minimum Wage Policy, Rural Development and Poverty Alleviation inMalawi.” Development and Change 26 (2) : 729–52.
Maloney, W., and J. Nu~nez Mendez. 2004. Measuring the Impact of Minimum Wages: Evidence fromLatin America. Chicago: Chicago University Press.
Montenegro, C., and C. Pages. 2004. “Who Benefits from Labor Market Regulations? Chile 1960–1998.” In J. Heckman, and C. Pages, eds., Law and Employment: Lessons from Latin America and theCaribbean, 401–34. Chicago: University of Chicago Press.
Nattrass, N., and J. Seekings. 2014. “Job Destruction in Newcastle: Minimum Wage-Setting and Low-Wage Employment in the South African Clothing Industry.” Transformation: Critical Perspectives onSouthern Africa 84 (1) : 1–30.
Neumark, D., W. Cunningham, and L. Siga. 2006. “The Effects of the Minimum Wage in Brazil onthe Distribution of Family Incomes: 1996–2001.” Journal of Development Economics 80 (1) :136–59.
Neumark, D., and W. Wascher. 2006. “Minimum Wages and Employment: A Review of Evidencefrom the New Minimum Wage Research” Working Paper W12663. National Bureau of EconomicResearch, Cambridge, MA.
———. 2007. “Minimum Wages and Employment.” Discussion Paper 2570. Institute for the Studyof Labor, Bonn, Germany.
———. 2008. Minimum Wages. Cambridge, MA: MIT Press.
Neumark, D., J. I. Salas, and W. Wascher. 2014. “Revisiting the Minimum Wage—EmploymentDebate: Throwing out the Baby with the Bathwater?” Industrial and Labor Relations Review 67 (3Suppl) : 608–48.
Nguyen, C. V. 2010. “The Impact of a Minimum Wage Increase on Employment, Wages andExpenditures of Low-Wage Workers in Vietnam.” Archive Paper 36751. Munich Personal RePEc,Munich, Germany.
Piore, M. J., and A. Schrank. 2008. “Toward Managed Flexibility: The Revival of Labour Inspectionin the Latin World.” International Labour Review 147 (1) : 1–23.
Rama, M. 2001. “The Consequences of Doubling the Minimum Wage: the Case ofIndonesia.” Industrial and Labor Relations Review 54 (4) : 864–81.
Rani, U., P. Belser, M. Oelz, and S. Ranjbar. 2013. “Minimum Wage Coverage and Compliance inDeveloping Countries.” International Labour Review 152 (3–4) : 381–410.
Ravallion, M., S. Chen, and P. Sangraula. 2009. “Dollar a Day Revisited.” World Bank EconomicReview 23 (2) : 163–84.
Schmitt, J. 2013. “Why Does the Minimum Wage Have No Discernible Effect onEmployment?” Center for Economic and Policy Research 22 : 1–28.
52 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
Strobl, E., and F. Walsh. 2003. “Minimum Wages and Compliance: The Case of Trinidad andTobago.” Economic Development and Cultural Change 51 (1) : 427–50.
Suryahadi, A., W. Widyanti, D. Perwira, and S. Sumarto. 2003. “Minimum Wage Policy and ItsImpact on Employment in the Urban Formal Sector.” Bulletin of Indonesian Economic Studies 39 (1): 29–50.
Wang, J., and M. Gunderson. 2011. “Minimum Wage Impacts in China: Estimates from a PrescribedResearch Design, 2000–2007.” Contemporary Economic Policy 29 (3) : 392–406.
Appendix
Figure A1. Wage-Employment Elasticities and Kaitz Ratios
-4-2
02
Ela
sticity
.2 .3 .4 .5 .6 .7Kaitz Ratio
Elasticity Fitted values
Source: Neumark and Wascher (2007) and authors’ calculations.
Table A1. Minimum Wage System Classification for Selected African Countries
National Sectoral/occupational Hybrid
Algeria Botswana Burkina Faso
Cameroon Kenya Angola
Cote d’Ivoire Malawi
Egypt Mozambique
Ghana Rwanda
Mali South Africa
Uganda Tanzania
Source: ILO TRAVAIL database.
Bhorat et al. 53
Table A2. Agricultural Minimum Wages, Kenya, 2013
Type of employee Monthly wage (shillings)
Unskilled employee 4,854.35Stockman, herdsman, and watchman 5,606.05House servant or cook 5,541.55Farm foreman 8,757.20Farm clerk 8,757.20Section foreman 5,669.20Farm artisan 5,802.05Tractor driver 6,152.70Combined harvester-driver 6,778.10Lorry driver or car driver 7,113.25
Source: Ministry of Labour Social, Security and Services (2013).
Table A3. Nonagricultural Monthly Minimum Wages, Kenyan Shillings, Kenya, 2013
Type ofemployee
Nonagricultural industry in cities(Nairobi, Mombasa, and Kisumu)
Nonagricultural industry atall municipalities and town councils
Nonagricultural industryat all other areas
a 9,780.95 9,024.10 5,217.95b 10,563.60 9,372.20 6,028.95c 10,911.70 10,116.15 6,223.65d 11,085.70 10,316.00 8,361.35e 12,654.90 11,838.70 9,679.05f 13,201.55 12,184.25 10,071.05g 15,064.65 13,772.75 11,743.30h 16,602.85 15,259.35 13,606.40i 18,329.25 17,101.80 15,434.70j 20,283.90 18,940.40 17,644.60k 22,070.95 20,769.95 19,474.20l 13,201.55 12,184.25 10,071.05m 16,602.85 15,259.35 13,580.60n 17,932.15 17,101.55 15,434.70o 22,070.95 20,769.95 19,474.20
Notes:aGeneral labor, including cleaner, sweeper, gardener, children’s ayah, house servant, day watchman, messenger.bMiner, stone cutter, turn boy, waiter, cook, logger line cutter.cNight watchman.dMachine attendant, sawmill sawyer, machine assistant, mass production machinist, shoe cutter, bakery worker, bakery assis-
tant, tailor’s assistant.eMachinist (made-to-measure), shoe upper preparer, chaplis maker, vehicle service worker (petrol and service stations), bakery
plant hand, laundry operator, junior clerk, wheeled tractor driver (light).fPrinting machine operator, bakery machine operator, plywood machine operator, sawmill dresser, shop assistant, machine
tool operator, dough maker, table hand baker or confectioner, copy-typist, driver (cars and light vans).gPattern designer (draughts-man), garment and dress cutter, single hand oven man, charge-hand baker, general clerk, tele-
phone operator, receptionist, storekeeper.hTailor, driver (medium-sized vehicle).iDyer, crawler tractor driver, salesman.jSaw doctor, caretaker (buildings).kCashier, driver (heavy commercial vehicle), salesman-driver.lUngraded artisan.mArtisan grade III.nArtisan grade II.oArtisan grade I.
Source: Ministry of Labour Social, Security and Services.
54 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
Ta
ble
A4
.In
tern
ati
on
al
LM
IC
ou
ntr
yS
tud
ies
on
Min
imu
mW
ag
es
Stu
dy
Co
un
try
Pa
ram
eter
so
fst
ud
yE
mp
loy
men
t/h
ou
rsw
ork
edF
orm
al/
info
rma
lse
cto
rsh
ifts
Wa
ges
Kh
arm
is(2
01
3)
Arg
enti
na
Da
te:
19
92
–2
00
5
Da
ta:
Per
ma
nen
tH
ou
seh
old
Su
rvey
(EP
H)
for
the
yea
r1
99
3
an
dth
eC
on
tin
uo
us
Per
ma
nen
tH
ou
seh
old
surv
ey
(EP
H-C
)fo
rth
ey
ear
20
04
Info
rma
lw
ork
ers,
wo
rker
sw
ith
ou
t
soci
al
secu
rity
con
trib
uti
on
,
exp
erie
nce
dsi
gn
ifica
nt
wa
ge
incr
ease
s
wh
enth
em
inim
um
wa
ge
wa
sra
ised
,
wh
ile
form
al
wo
rker
sd
idn
ot.
Fa
jnzy
lber
(20
01
)
Bra
zil
Da
te:
19
82
–1
99
7
Da
ta:
Bra
zil’
sM
on
thly
Em
plo
ym
ent
Su
rvey
,
ind
ivid
ua
lsa
ged
15
to6
5y
ears
Est
ima
tes
sug
ges
tth
at
emp
loy
men
tel
ast
icit
ies
are
neg
ati
ve
for
mo
st
low
-wa
ge
wo
rker
s,b
ein
g
low
erin
ab
solu
tev
alu
e
for
form
al
sala
ried
wo
rker
s(a
rou
nd
-0.1
at
the
bo
tto
mo
fth
ew
ag
e
dis
trib
uti
on
)th
an
for
low
-wa
ge
info
rma
l
sala
ried
an
dse
lf-
emp
loy
ed(b
etw
een
-0.2
5
an
d-0
.35
).
Sig
nifi
can
tm
inim
um
wa
ge
effe
cts
acr
oss
the
wh
ole
wa
ge
dis
trib
uti
on
an
db
oth
inth
efo
rma
la
nd
the
info
rma
lse
cto
rs.
We
als
ofi
nd
tha
tth
e
tota
lim
pa
cto
fm
inim
um
wa
ges
on
wo
rker
s’ea
rnin
gs
(der
ived
fro
m
curr
ent
an
dla
gg
edef
fect
s)is
po
siti
ve
bu
tsm
all
erth
an
the
con
tem
po
ran
eou
s
on
e.O
ther
resu
lts
incl
ud
eh
igh
er
earn
ing
sel
ast
icit
ies
for
men
,a
du
lts,
an
dh
ead
so
fh
ou
seh
old
sth
an
for
wo
men
,te
ena
ger
s,a
nd
no
nh
ead
s,
resp
ecti
vel
y.
Fo
gu
el,
Ra
mo
s,
an
dC
arn
eiro
(20
01
)
Bra
zil
Da
te:
19
82
–1
99
9
Da
ta:
Mo
nth
lyE
mp
loy
men
t
Su
rvey
(Pes
qu
isa
Men
sal
de
Em
pre
go
/IB
GE
),o
ffici
al
min
imu
mw
ag
era
tes
fro
mth
e
Min
istr
yo
fL
ab
or,
Bra
zili
an
Inst
itu
teo
fG
eog
rap
hy
an
d
Sta
tist
ics
Incr
ease
sin
the
va
lue
of
the
offi
cia
lm
inim
um
wa
ge
ten
dto
dec
rea
se
form
al
emp
loy
men
t(-
0.0
01
to-0
.02
4)
an
d
incr
ease
info
rma
l
emp
loy
men
t(0
.00
04
to
0.0
03
).
A1
0p
erce
nt
incr
ease
in
un
emp
loy
men
tca
use
sa
1.2
per
cen
t
dro
pin
the
earn
ing
so
fin
form
al
wo
rker
s,a
so
pp
ose
dto
afa
llo
fo
nly
0.9
per
cen
tin
the
wa
ges
of
form
al
wo
rker
s.
con
tin
ued
Bhorat et al. 55
Ta
ble
A4
.C
onti
nu
ed
Stu
dy
Co
un
try
Pa
ram
eter
so
fst
ud
yE
mp
loy
men
t/h
ou
rsw
ork
edF
orm
al/
info
rma
lse
cto
rsh
ifts
Wa
ges
Lem
os
(20
04
)B
razi
lD
ate
:1
98
2–
20
00
Da
ta:
Bra
zil’
sM
on
thly
Em
plo
ym
ent
Su
rvey
an
d
Bra
zili
an
La
bo
ur
Min
istr
yd
ata
Sm
all
neg
ati
ve
effe
cts
on
emp
loy
men
t(-
0.0
5),
an
d
itis
do
min
ate
db
yth
e
ho
urs
effe
ct,
or
red
uct
ion
inh
ou
rsw
ork
ed.
An
incr
ease
inth
em
inim
um
wa
ge
stro
ng
lyco
mp
ress
esth
ew
ag
e
dis
trib
uti
on
.
Lem
os
(20
06
)B
razi
lD
ate
:1
98
2–
20
00
Da
ta:
Bra
zil’
sM
on
thly
Em
plo
ym
ent
Su
rvey
an
d
Bra
zili
an
La
bo
ur
Min
istr
yd
ata
Min
imu
mw
ag
eh
as
no
ad
ver
seef
fect
on
emp
loy
men
tin
Bra
zil
bet
wee
n1
98
2a
nd
20
00
,
des
pit
eth
esi
zea
ble
wa
ge
effe
cts
fou
nd
inb
oth
the
form
al
an
din
form
al
sect
ors
.
Inth
efo
rma
lse
cto
r,a
1p
erce
nt
incr
ease
inth
em
inim
um
wa
ge
incr
ease
sth
ew
ag
eso
fth
ose
inth
e
25
thp
erce
nti
leb
y0
.33
per
cen
ta
nd
of
tho
sein
the
50
thp
erce
nti
leb
y0
.10
per
cen
t(e
va
lua
ted
at
the
av
era
ge
“fra
ctio
na
t”1
1.6
per
cen
t).
Inth
e
info
rma
lse
cto
r,it
incr
ease
sth
ew
ag
es
of
tho
sein
the
25
thp
erce
nti
leb
y0
.31
per
cen
ta
nd
of
tho
sein
the
50
th
per
cen
tile
by
0.4
8p
erce
nt.
Lem
os
(20
07
)B
razi
lD
ate
:1
98
2–
20
00
Da
ta:
Bra
zil’
sM
on
thly
Em
plo
ym
ent
Su
rvey
an
d
Bra
zili
an
La
bo
ur
Min
istr
yd
ata
No
evid
ence
of
ad
ver
se
emp
loy
men
tef
fect
sin
eith
erth
ep
ub
lic
or
pri
va
tese
cto
rsa
tth
e
ag
gre
ga
tele
vel
or
for
vu
lner
ab
leg
rou
ps
such
as
teen
ag
ers,
wo
men
,
an
dth
elo
wed
uca
ted
.
Min
imu
mw
ag
ep
oli
cies
inB
razi
la
pp
ear
tob
ea
po
ten
tia
lly
via
ble
an
tip
ov
erty
inst
rum
ent.
Inth
ep
riv
ate
sect
or,
a1
0p
erce
nt
incr
ease
inth
ere
al
min
imu
mw
ag
e
incr
ease
sth
ew
ag
eso
fth
ose
inth
e
20
thp
erce
nti
leb
y3
.41
per
cen
ta
nd
of
tho
sein
the
30
thp
erce
nti
leb
y2
.55
per
cen
t.In
the
pu
bli
cse
cto
r,it
incr
ease
sth
ew
ag
eso
fth
ose
inth
e
20
thp
erce
nti
leb
y1
.79
per
cen
ta
nd
of
tho
sein
the
30
thp
erce
nti
leb
y1
.14
per
cen
t.
con
tin
ued
56 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
Ta
ble
A4
.C
onti
nu
ed
Stu
dy
Co
un
try
Pa
ram
eter
so
fst
ud
yE
mp
loy
men
t/h
ou
rsw
ork
edF
orm
al/
info
rma
lse
cto
rsh
ifts
Wa
ges
Lem
os
(20
09
)B
razi
lD
ate
:1
98
2–
20
04
Da
ta:
Bra
zil’
sM
on
thly
Em
plo
ym
ent
Su
rvey
an
d
Bra
zili
an
La
bo
ur
Min
istr
yd
ata
No
effe
ctfo
un
din
eith
er
the
form
al
or
info
rma
l
sect
or
(0.0
80
to0
.35
8).
Min
imu
mw
ag
eco
mp
ress
esth
ew
ag
e
dis
trib
uti
on
of
bo
thth
efo
rma
la
nd
info
rma
lse
cto
rsin
Bra
zil
inM
ay
19
95
.
McI
nty
re
(20
06
)
Bra
zil
Da
te:
19
81
–1
99
9(e
xce
pt
19
91
an
d1
99
4)
Da
ta:
Pes
qu
isa
Na
cio
na
ld
e
Am
ost
rad
eD
om
icil
ios
(PN
AD
)
Est
ima
tes
rev
eal
tha
tth
e
min
imu
mw
ag
ein
Bra
zil
do
esn
ot
incr
ease
un
emp
loy
men
t;ra
ther
,it
rais
esfo
rma
lity
.
Ma
nd
ate
dn
on
wa
ge
ben
efits
an
dth
e
min
imu
mw
ag
ela
w
ha
ve
no
effe
cto
n
emp
loy
men
tb
ut
do
enco
ura
ge
info
rma
lity
an
d
low
erto
tal
com
pen
sati
on
.
Lo
wer
min
imu
m
wa
ges
enco
ura
ge
wo
rker
sto
form
ali
ze
thei
rb
enefi
ts:
a1
0
per
cen
td
ecre
ase
in
the
min
imu
mw
ag
e
incr
ease
sb
y1
.9
per
cen
tth
en
um
ber
of
wo
rker
sp
ay
ing
all
pa
yro
llta
xes
.T
he
av
era
ge
form
ali
ty
pre
miu
mis
hig
hes
t
Lo
wer
min
imu
mw
ag
esa
nd
lax
er
enfo
rcem
ent
of
the
law
bo
thin
crea
se
wa
ges
am
on
gth
elo
w-s
kil
led
an
d
dec
rea
sew
ag
ein
equ
ali
ty.
con
tin
ued
Bhorat et al. 57
Ta
ble
A4
.C
onti
nu
ed
Stu
dy
Co
un
try
Pa
ram
eter
so
fst
ud
yE
mp
loy
men
t/h
ou
rsw
ork
edF
orm
al/
info
rma
lse
cto
rsh
ifts
Wa
ges
am
on
gth
ele
ast
edu
cate
d.
Neu
ma
rk,
Cu
nn
ing
ha
m,
an
dS
iga
(20
06
)
Bra
zil
Da
te:
19
96
–2
00
1
Da
ta:
Bra
zili
an
Mo
nth
ly
Em
plo
ym
ent
Su
rvey
(Pes
qu
isa
Men
sal
do
Em
pre
go
,o
rP
ME
)
Neg
ati
ve
emp
loy
men
t
effe
cts
(-0
.07
).
Est
ima
tes
pro
vid
en
oev
iden
ceth
at
min
imu
mw
ag
esin
Bra
zil
com
pre
ss
the
inco
me
dis
trib
uti
on
—li
ftin
gfa
mil
y
inco
mes
at
the
low
erp
oin
tso
fth
ein
-
com
ed
istr
ibu
tio
n—
an
d,
ifa
ny
thin
g,
som
etim
esin
dic
ate
tha
tm
inim
um
wa
ges
ha
ve
the
op
po
site
effe
cto
fre
-
du
cin
gfa
mil
yin
com
esin
the
low
erta
il
of
the
dis
trib
uti
on
.
Mo
nte
neg
roa
nd
Pa
ges
(20
04
)
Ch
ile
Da
te:
19
60
–1
99
8
Da
ta:
Ho
use
ho
ldsu
rvey
sfr
om
the
Un
iver
sity
of
Ch
ile’
s
Eco
no
mic
sD
epa
rtm
ent
Res
ult
ssu
gg
est
tha
tb
oth
min
imu
mw
ag
esa
nd
job
secu
rity
reg
ula
tio
ns
red
uce
the
emp
loy
men
t
op
po
rtu
nit
ies
of
the
yo
un
g,
un
skil
led
,a
nd
pa
rtic
ula
rly
un
skil
led
yo
uth
wh
ile
pro
mo
tin
g
the
emp
loy
men
tra
tes
of
skil
led
an
do
lder
wo
rker
s.
We
ha
ve
als
ofo
un
d
ind
ica
tio
ns
tha
tjo
b
secu
rity
reg
ula
tio
ns
ma
y
forc
eso
me
wo
rker
s,
con
tin
ued
58 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
Ta
ble
A4
.C
onti
nu
ed
Stu
dy
Co
un
try
Pa
ram
eter
so
fst
ud
yE
mp
loy
men
t/h
ou
rsw
ork
edF
orm
al/
info
rma
lse
cto
rsh
ifts
Wa
ges
pa
rtic
ula
rly
wo
men
an
d
the
un
skil
led
,o
ut
of
wa
ge
emp
loy
men
ta
nd
into
self
-em
plo
ym
ent.
Fa
ng
an
dL
in
(20
13
)
Ch
ina
Da
te:
20
04
–0
9
Da
ta:
Urb
an
Ho
use
ho
ldS
urv
ey
(UH
S)
an
dm
inim
um
wa
ge
da
ta
coll
ecte
da
tth
eco
un
tyle
vel
Min
imu
mw
ag
ech
an
ges
ha
ve
sig
nifi
can
ta
dv
erse
effe
cts
on
emp
loy
men
tin
the
Ea
ster
na
nd
Cen
tra
l
reg
ion
so
fC
hin
aa
nd
resu
ltin
dis
emp
loy
men
t
for
fem
ale
s,y
ou
ng
ad
ult
s,a
nd
low
-sk
ille
d
wo
rker
s.
Yo
uth
:-0
.13
6to
-0.1
56
At-
risk
gro
up
s:-0
.26
5to
-
0.3
40
.
Wa
ng
an
d
Gu
nd
erso
n
(20
11
)
Ch
ina
Da
te:
20
00
–0
7
Da
ta:
Ch
ina
Po
pu
lati
on
Sta
tist
ic
Yea
rbo
ok
Neg
ati
ve
emp
loy
men
t
effe
cts
insl
ow
er-g
row
ing
reg
ion
s(-
0.1
56
to-
0.1
78
);la
rger
neg
ati
ve
effe
cts
inn
on
-sta
te-
ow
ned
org
an
iza
tio
ns
tha
t
ten
dto
be
mo
re
resp
on
siv
eto
ma
rket
pre
ssu
res;
mu
chla
rger
lag
ged
effe
cts
refl
ecti
ng
the
tim
en
eed
edfo
r
ad
just
men
tsto
occ
ur;
no
ad
ver
seem
plo
ym
ent
effe
cts
inth
ep
rosp
ero
us
an
dg
row
ing
.
con
tin
ued
Bhorat et al. 59
Ta
ble
A4
.C
onti
nu
ed
Stu
dy
Co
un
try
Pa
ram
eter
so
fst
ud
yE
mp
loy
men
t/h
ou
rsw
ork
edF
orm
al/
info
rma
lse
cto
rsh
ifts
Wa
ges
Bel
l(1
99
7)
Co
lom
bia
Da
te:
19
77
–1
98
7
Da
ta:
Co
lom
bia
’sA
nn
ua
l
Ind
ust
ria
lS
urv
ey
Su
bst
an
tia
l
dis
emp
loy
men
tef
fect
so
f
min
imu
mw
ag
esa
re
fou
nd
,w
her
eth
eim
pa
ct
ises
tim
ate
da
tro
ug
hly
2
per
cen
tto
12
per
cen
t
du
rin
g1
98
1–
87
.Im
pli
ed
Ela
stic
itie
ssu
gg
est
tha
t
the
incr
ease
inth
ere
la-
tiv
ev
alu
eo
fth
em
ini-
mu
mw
ag
ein
Co
lom
bia
fro
m1
97
7to
19
87
(ro
ug
hly
15
per
cen
t)h
ad
the
effe
cto
fre
du
cin
g
ma
nu
fact
uri
ng
emp
loy
-
men
tb
y5
per
cen
to
ver
this
per
iod
.
Min
imu
mw
ag
esh
av
ea
stro
ng
imp
act
on
wa
ges
jud
gin
gb
yth
eir
pro
xim
ity
to
the
av
era
ge
wa
ge
an
dti
me
seri
es
esti
ma
tes.
Ma
nlo
ney
an
d
Nu
~ nez
Men
dez
(20
04
)
Co
lom
bia
Da
te:
19
97
an
d1
99
9
Da
ta:
Na
tio
na
lS
tati
stic
al
Ag
ency
(DA
NE
)a
nd
Na
tio
na
l
Ho
use
ho
ldS
urv
ey(E
NH
)
Ari
sein
the
min
imu
m
wa
ge
ha
sa
sta
tist
ica
lly
ver
ysi
gn
ifica
nt
imp
act
on
the
pro
ba
bil
ity
of
bec
om
ing
un
emp
loy
ed
tha
ta
ga
ind
ecre
ase
sw
ith
ari
sin
gp
osi
tio
nin
the
Th
eef
fect
on
rea
lw
ag
eso
fa
cha
ng
ein
the
rea
lm
inim
um
wa
ge
ish
igh
for
tho
seea
rnin
g7
0p
erce
nt
to9
0
per
cen
to
fth
em
inim
um
wa
ge;
87
per
cen
to
fth
eri
sein
min
imu
mw
ag
es
isco
mm
un
ica
ted
tow
ag
es.
con
tin
ued
60 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
Ta
ble
A4
.C
onti
nu
ed
Stu
dy
Co
un
try
Pa
ram
eter
so
fst
ud
yE
mp
loy
men
t/h
ou
rsw
ork
edF
orm
al/
info
rma
lse
cto
rsh
ifts
Wa
ges
wa
ge
dis
trib
uti
on
(-
0.1
61
to-0
.35
6).
El
Ha
mid
ia
nd
Ter
rell
(20
01
)
Co
sta
Ric
aD
ate
:1
97
6–
19
92
Da
ta:
Ho
use
ho
ldS
urv
eyo
f
Em
plo
ym
ent
an
d
Un
emp
loy
men
t,M
inis
try
of
La
bo
rW
ag
ed
ata
Incr
ease
inth
em
inim
um
wa
ge
rela
tiv
eto
the
av
era
ge
wa
ge
is
ass
oci
ate
dw
ith
an
incr
ease
inth
ele
vel
of
cov
ered
sect
or
emp
loy
men
tb
y0
.56
per
cen
t,b
ut
no
effe
cto
n
the
nu
mb
ero
fse
lf-
emp
loy
edo
ver
tim
e,a
nd
an
incr
ease
inth
e
av
era
ge
nu
mb
ero
fh
ou
rs
wo
rked
per
wee
kb
y0
.14
per
cen
tin
the
cov
ered
sect
or
an
d0
.34
per
cen
t
inth
eu
nco
ver
edse
cto
r.
Th
ese
fin
din
gs
ma
yb
e
inte
rpre
ted
as
sup
po
rtin
g
the
mo
no
pso
nis
tic
mo
del
,
wh
ich
pre
dic
tsth
at
incr
ease
sin
wa
ges
can
incr
ease
emp
loy
men
tu
p
toth
ep
oin
tw
her
eth
e
ma
rgin
al
cost
iseq
ua
lto
the
ma
rgin
al
rev
enu
e.
Gin
dli
ng
an
d
Ter
rell
(20
05
)
Co
sta
Ric
aD
ate
:1
98
8–
20
00
Da
ta:
Leg
al
min
imu
mw
ag
ed
ata
fro
mth
eM
inis
try
of
La
bo
r,
Ho
use
ho
ldS
urv
eys
for
Mu
ltip
le
A1
0p
erce
nt
incr
ease
in
min
imu
mw
ag
eslo
wer
s
emp
loy
men
tin
the
cov
ered
sect
or
by
1.0
9
Leg
al
min
imu
mw
ag
esh
av
ea
sig
nifi
can
tp
osi
tiv
eef
fect
on
the
wa
ges
of
wo
rker
sin
the
cov
ered
sect
or
(wit
h
an
ela
stic
ity
of
0.1
0),
bu
tn
oef
fect
on
con
tin
ued
Bhorat et al. 61
Ta
ble
A4
.C
onti
nu
ed
Stu
dy
Co
un
try
Pa
ram
eter
so
fst
ud
yE
mp
loy
men
t/h
ou
rsw
ork
edF
orm
al/
info
rma
lse
cto
rsh
ifts
Wa
ges
Pu
rpo
ses
by
the
Co
sta
Ric
an
Inst
itu
teo
fS
tati
stic
sa
nd
Cen
sus,
an
din
du
stry
da
tafr
om
the
Co
sta
Ric
an
Cen
tra
lB
an
k
per
cen
ta
nd
dec
rea
ses
the
av
era
ge
nu
mb
ero
fh
ou
rs
wo
rked
of
tho
sew
ho
rem
ain
inth
eco
ver
ed
sect
or
by
ab
ou
t0
.6
per
cen
t.D
esp
ite
the
wid
e
ran
ge
of
min
imu
m
wa
ges
,th
ela
rges
tim
pa
ct
on
the
wa
ges
an
d
emp
loy
men
to
fco
ver
ed
sect
or
wo
rker
sis
inth
e
low
erh
alf
of
the
dis
trib
uti
on
.
wa
ges
of
wo
rker
sin
the
un
cov
ered
sect
or.
Eri
kss
on
an
d
Py
tlo
ko
va
(20
04
)
Cze
ch
Rep
ub
lic
Da
te:
19
98
–2
00
0
Da
ta:
Av
era
ge
Ea
rnin
gs
Info
rma
tio
nS
yst
em
Neg
ati
ve
emp
loy
men
ta
nd
wo
rkin
gh
ou
rsef
fect
so
f-
14
.4p
erce
nt
an
d5
.1
per
cen
t,re
spec
tiv
ely
,in
the
yea
rfo
llo
win
gth
e
19
98
min
imu
mw
ag
e
incr
ease
.N
ega
tiv
e
emp
loy
men
ta
nd
wo
rkin
gh
ou
rsef
fect
so
f-
5.1
per
cen
ta
nd
5.4
per
cen
t,re
spec
tiv
ely
,in
the
yea
rfo
llo
win
gth
e
19
99
min
imu
mw
ag
e
incr
ease
.
Th
ela
rger
the
pro
po
rtio
no
flo
w-p
aid
wo
rker
sin
afi
rm,
the
hig
her
the
incr
ease
inth
efi
rm’s
av
era
ge
wa
ge.
Jon
es(1
99
7)
Gh
an
aD
ate
:1
97
0–
19
91
Da
ta:
Yea
rbo
ok
of
La
bo
ur
Pro
vid
esfa
irly
stro
ng
evid
ence
tha
tth
e
min
imu
mw
ag
ein
Imp
lied
ela
stic
itie
s
sug
ges
tth
at
ala
rge
pro
po
rtio
no
fth
e
con
tin
ued
62 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
Ta
ble
A4
.C
onti
nu
ed
Stu
dy
Co
un
try
Pa
ram
eter
so
fst
ud
yE
mp
loy
men
t/h
ou
rsw
ork
edF
orm
al/
info
rma
lse
cto
rsh
ifts
Wa
ges
Sta
tist
ics
(IL
O),
Inte
rna
tio
na
l
Fin
an
cia
lS
tati
stic
s(I
MF
),
Afr
ica
nE
mp
loy
men
tR
epo
rt
(19
90
),P
enn
Wo
rld
Ta
ble
s,
Wo
rld
Ba
nk
So
cia
lIn
dic
ato
rso
f
Dev
elo
pm
ent
(19
95
),W
orl
d
Ba
nk
’sR
egio
na
l
Pro
gra
mo
nE
nte
rpri
se
Dev
elo
pm
ent
(19
92
)
Gh
an
ah
ad
an
ega
tiv
e
imp
act
on
emp
loy
men
t
inth
efo
rma
lp
riv
ate
sect
or
(-0
.12
ela
stic
ity
)
an
dfo
rma
lp
ub
lic
sect
or
(-0
.17
ela
stic
ity
).
Info
rma
lem
plo
ym
ent
incr
ease
da
sa
resu
lto
f
incr
ease
dm
inim
um
wa
ge.
Em
plo
ym
ent
of
wo
men
.
pu
bli
cse
cto
r
wo
rker
sd
isp
lace
db
y
the
min
imu
mw
ag
e
shif
ted
into
info
rma
l
sect
or
emp
loy
men
t.
Gin
dli
ng
an
d
Ter
rell
(20
07
)
Ho
nd
ura
sD
ate
:1
99
0–
20
04
Da
ta:
Ho
nd
ura
sM
inim
um
Wa
ge
Dec
rees
an
dth
eP
erm
an
ent
Ho
use
ho
ldS
urv
eys
for
Mu
ltip
le
Pu
rpo
ses
Dis
emp
loy
men
tef
fect
sin
the
pri
va
tese
cto
r(-
0.4
6).
Wa
ge
ela
stic
ity
of
0.2
9o
ver
all
.
Ho
wev
er,
the
wel
fare
—th
eto
tal
earn
ing
s—o
flo
w-p
aid
wo
rker
sin
the
larg
efi
rm–
cov
ered
sect
or
fall
sw
ith
hig
her
min
imu
mw
ag
es.
Ala
tas
an
d
Ca
mer
on
(20
03
)
Ind
on
esia
Da
te:
19
90
–1
99
6
Da
ta:
An
nu
al
Su
rvey
of
Ma
nu
fact
uri
ng
Fir
ms
(Su
rvei
Ta
hu
na
nP
eru
sah
aa
nIn
du
stri
,
SI)
So
me
evid
ence
of
a
neg
ati
ve
emp
loy
men
t
imp
act
for
sma
lld
om
esti
c
firm
sb
ut
no
emp
loy
men
t
imp
act
for
larg
efi
rms—
fore
ign
or
do
mes
tic.
Co
mo
laa
nd
De
Mel
lo(2
01
1)
Ind
on
esia
Da
te:
19
96
–2
00
4
Da
ta:
Ind
on
esia
nS
tati
stic
s
Bu
rea
u’s
Na
tio
na
lL
ab
or
Fo
rce
Su
rvey
(Sa
ker
na
s)w
ork
ing
ag
e
(15
to6
5y
ears
),N
ati
on
al
Eco
no
mic
Su
rvey
(Su
sen
as)
,
Ind
ust
ria
lS
urv
ey(S
urv
ei
Ind
ust
ri)
Min
imu
mw
ag
eh
ikes
des
tro
yfo
rma
lse
cto
rjo
bs
(-0
.05
6),
bu
tth
ese
job
s
loss
esa
rem
ore
tha
n
com
pen
sate
dfo
rb
yth
e
exp
an
sio
no
fth
ein
form
al
sect
or
(0.0
74
),
sug
ges
tin
gth
at
min
imu
mw
ag
e
Th
en
ega
tiv
ea
nd
sig
nifi
can
t
coef
fici
ent
on
un
emp
loy
men
t
seem
sto
sug
ges
t
tha
tth
ed
ecre
ase
in
form
al-
sect
or
emp
loy
men
td
ue
to
ari
sein
the
rela
tiv
e
con
tin
ued
Bhorat et al. 63
Ta
ble
A4
.C
onti
nu
ed
Stu
dy
Co
un
try
Pa
ram
eter
so
fst
ud
yE
mp
loy
men
t/h
ou
rsw
ork
edF
orm
al/
info
rma
lse
cto
rsh
ifts
Wa
ges
leg
isla
tio
nis
hu
rtin
g,
inst
ead
of
pro
tect
ing
vu
lner
ab
lew
ork
ers.
va
lue
of
the
min
imu
mw
ag
e
shif
tsw
ork
ers
fro
m
“qu
euin
g”
un
emp
loy
men
tto
the
ina
ctiv
e
po
pu
lati
on
of
the
info
rma
lse
cto
r.
Del
Ca
rpio
,
Ng
uy
en,
an
d
Wa
ng
(20
12
)
Ind
on
esia
Da
te:
19
93
–2
00
6
Da
ta:
An
nu
al
Ma
nu
fact
uri
ng
Su
rvey
(Su
rvei
Ind
ust
rio
rS
I)
an
dth
eN
ati
on
al
So
cio
-
Eco
no
mic
Su
rvey
(Su
sen
as)
Th
eem
plo
ym
ent
effe
cts
of
min
imu
mw
ag
esa
re
sig
nifi
can
ta
nd
neg
ati
ve
am
on
ga
llfi
rms
(-0
.02
33
to-0
.05
42
),b
ut
mo
re
pre
do
min
an
tin
sma
ll
firm
sa
nd
less
edu
cate
d
wo
rker
sa
nd
less
am
on
g
larg
efi
rms
an
dw
ork
ers
wit
hh
igh
sch
oo
l
edu
cati
on
.
Dis
emp
loy
men
tef
fect
s
are
stro
ng
erfo
r
no
np
rod
uct
ion
wo
rker
s
(�0
.05
4)
an
dw
om
en
tha
nfo
rp
rod
uct
ion
wo
rker
s(�
0.0
23
).
Min
imu
mw
ag
esa
rem
ore
bin
din
gin
sma
llfi
rms
tha
nin
larg
efi
rms.
Ha
rris
on
an
d
Sco
rse
(20
10
)
Ind
on
esia
Da
te:
19
90
–1
99
6
Da
ta:
An
nu
al
Ma
nu
fact
uri
ng
Su
rvey
,B
ad
an
Pu
sat
Sta
tist
ik
Res
ult
ssu
gg
est
tha
tth
e
min
imu
mw
ag
ein
crea
ses
led
toem
plo
ym
ent
loss
es
for
pro
du
ctio
nw
ork
ers
acr
oss
all
sect
ors
in
ma
nu
fact
uri
ng
.
A1
per
cen
tin
crea
sein
the
rea
lv
alu
eo
f
the
min
imu
mw
ag
ew
as
ass
oci
ate
d
wit
ha
0.6
75
per
cen
tin
crea
sein
the
rea
lu
nsk
ille
dw
ag
e.
Ind
on
esia
Da
te:
19
93
–2
00
0
con
tin
ued
64 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
Ta
ble
A4
.C
onti
nu
ed
Stu
dy
Co
un
try
Pa
ram
eter
so
fst
ud
yE
mp
loy
men
t/h
ou
rsw
ork
edF
orm
al/
info
rma
lse
cto
rsh
ifts
Wa
ges
Ma
gru
der
(20
13
)
Da
ta:
Ind
on
esia
nF
am
ily
Lif
e
Su
rvey
(wa
ves
1,
2,
an
d3
),
Sta
tist
ics
Ind
ust
ry(S
I)
Du
rin
gth
e1
99
0s’
ma
ssiv
efo
reig
n
inv
estm
ent
an
dra
pid
eco
no
mic
gro
wth
,w
ag
es
wer
ein
crea
sed
an
da
“big
pu
sh”
inth
e
eco
no
my
incr
ease
d
do
mes
tic
dem
an
d.
Ob
serv
ing
on
eo
ut
of
the
33
dis
tric
tsin
Ind
on
esia
,
form
al
emp
loy
men
t
incr
ease
da
nd
info
rma
l
emp
loy
men
td
ecre
ase
d
on
lyin
loca
lin
du
stri
es;
tra
dea
ble
ma
nu
fact
uri
ng
firm
ssa
wn
og
row
thin
emp
loy
men
t,a
nd
un
tra
dea
ble
bu
t
no
nin
du
stri
ali
zab
le
serv
ices
saw
an
incr
ease
inin
form
al
emp
loy
men
t.
Sh
ift
fro
min
form
al
to
form
al
emp
loy
men
t
inlo
cal
ind
ust
ries
;
tra
dea
ble
ind
ust
ry
saw
no
mo
vem
ents
;
un
tra
dea
ble
,
no
nin
du
stri
ali
zab
le
serv
ices
saw
ari
sein
info
rma
l
emp
loy
men
t.
Th
eb
ott
om
qu
art
ile
of
the
wa
ge
dis
trib
uti
on
exp
erie
nce
sm
ass
ive
wa
ge
ga
ins
wh
enm
inim
um
wa
ges
gro
w,
so
tha
ta
do
ub
lin
go
fm
inim
um
wa
ges
is
ass
oci
ate
dw
ith
a1
50
per
cen
tto
16
0
per
cen
tin
crea
sein
the
25
thp
erce
nti
le
of
the
wa
ge
dis
trib
uti
on
.
Ra
ma
(20
01
)In
do
nes
iaD
ate
:1
99
3
Da
ta:
La
bo
rF
orc
eS
urv
ey
(Sa
ker
na
s)
Urb
an
un
emp
loy
men
t
dec
rea
sed
by
0p
erce
nt
to
5p
erce
nt.
Th
e
emp
loy
men
tef
fect
s,
ho
wev
er,
va
ried
sub
sta
nti
all
yb
yfi
rmsi
ze:
sma
llfi
rms
ap
pa
ren
tly
Av
era
ge
wa
ges
incr
ease
db
y5
per
cen
t
to1
5p
erce
nt.
con
tin
ued
Bhorat et al. 65
Ta
ble
A4
.C
onti
nu
ed
Stu
dy
Co
un
try
Pa
ram
eter
so
fst
ud
yE
mp
loy
men
t/h
ou
rsw
ork
edF
orm
al/
info
rma
lse
cto
rsh
ifts
Wa
ges
exp
erie
nce
dsu
bst
an
tia
l
dec
rea
ses
in
emp
loy
men
t,w
her
eas
som
ela
rge
firm
sa
ctu
all
y
saw
thei
rem
plo
ym
ent
incr
ease
.W
ork
ers
in
thes
ela
rge
firm
s,th
e
au
tho
rco
ncl
ud
es,
are
the
evid
ent
win
ner
sfr
om
the
min
imu
mw
ag
eh
ike.
Su
rya
ha
di
eta
l.
(20
03
)
Ind
on
esia
Da
te:
19
88
–2
00
0
Da
ta:
Na
tio
na
lL
ab
ou
rS
urv
ey
(Sa
ker
na
s)
Th
eim
po
siti
on
of
min
imu
mw
ag
esh
as
a
neg
ati
ve
an
dst
ati
stic
all
y
sig
nifi
can
tim
pa
cto
n
emp
loy
men
tin
the
urb
an
form
al
sect
or.
Th
e
dis
emp
loy
men
tim
pa
ctis
gre
ate
stfo
rfe
ma
le,
yo
un
ga
nd
less
edu
cate
d
wo
rker
s,w
hil
eth
e
emp
loy
men
tp
rosp
ects
of
wh
ite-
coll
ar
wo
rker
sa
re
enh
an
ced
by
incr
ease
sin
min
imu
mw
ages
(-0
.11
2).
So
me
wo
rker
sw
ho
lose
job
sin
the
form
al
sect
or
an
d
ha
ve
tore
loca
teto
the
info
rma
lse
cto
r
face
low
erea
rnin
gs
an
dp
oo
rer
wo
rkin
g
con
dit
ion
s.
Th
em
inim
um
wa
ge
bec
am
em
ore
bin
din
ga
nd
imp
act
ful
on
the
wa
ge
dis
trib
uti
on
inIn
do
nes
iafr
om
19
88
to
20
00
.
An
da
l� on
an
d
Pa
ges
(20
08
)
Ken
ya
Da
te:
19
98
–1
99
9
Da
ta:
Cen
tra
lB
ure
au
of
Sta
tist
ics’
Inte
gra
ted
La
bo
ur
Fo
rce
Su
rvey
Est
ima
tes
ind
ica
teth
at
a
10
per
cen
tag
ep
oin
t
incr
ease
inth
em
inim
um
tom
edia
nw
ag
era
tio
cou
ldb
ea
sso
cia
ted
wit
h
ad
ecli
ne
inth
esh
are
of
form
al
emp
loy
men
to
f
Min
imu
mw
ag
esw
ere
po
siti
vel
y
ass
oci
ate
dw
ith
wa
ges
of
low
-ed
uca
ted
wo
rker
sa
nd
wo
men
in
no
na
gri
cult
ura
la
ctiv
itie
s,w
hil
en
o
such
rela
tio
nsh
ipis
fou
nd
for
wo
rker
s
ina
gri
cult
ure
.
con
tin
ued
66 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
Ta
ble
A4
.C
onti
nu
ed
Stu
dy
Co
un
try
Pa
ram
eter
so
fst
ud
yE
mp
loy
men
t/h
ou
rsw
ork
edF
orm
al/
info
rma
lse
cto
rsh
ifts
Wa
ges
bet
wee
n1
.2p
erce
nt
an
d
5.6
per
cen
ta
nd
an
incr
ease
of
bet
wee
n2
.7
per
cen
ta
nd
5.9
per
cen
tag
ep
oin
tsin
the
sha
reo
fse
lf-e
mp
loy
men
t.
Bel
l(1
99
7)
Mex
ico
Da
te:
19
84
–1
99
0
Da
ta:
Mex
ico
’sA
nn
ua
lIn
du
stri
al
Su
rvey
,M
exic
an
Ex
ues
ta
Na
tio
na
led
eE
mp
leo
Min
imu
mw
ag
esh
ad
vir
tua
lly
no
effe
cto
n
emp
loy
men
tin
the
form
al
sect
or.
Min
imu
mw
ag
esh
ad
vir
tua
lly
no
effe
ct
on
wa
ges
inth
efo
rma
lse
cto
r.
Sig
nifi
can
tn
um
ber
so
fw
ork
ers
are
pa
ida
to
rb
elo
wm
inim
um
wa
ges
.
Bo
sch
an
d
Ma
na
cord
a
(20
10
)
Mex
ico
Da
te:
19
89
–2
00
0
Da
ta:
Mic
rod
ata
,E
ncu
esta
Na
cio
na
ld
eE
mp
leo
Urb
an
o
Ad
ecli
ne
inth
ere
al
va
lue
of
the
min
imu
mw
ag
eex
pla
ins
av
ery
sig
nifi
can
tin
crea
sein
ineq
ua
lity
ob
serv
edin
Mex
ico
bet
wee
nth
ela
te
19
80
sa
nd
late
19
90
s.
Fel
icia
no
(19
98
)M
exic
oD
ate
:1
97
0,
19
80
,1
99
0
Da
ta:
Ind
ust
ria
lC
ensu
s,M
on
thly
Ind
ust
ria
lS
urv
ey
La
rge
red
uct
ion
sin
the
min
imu
mw
ag
ew
ere
fou
nd
toin
crea
se
emp
loy
men
to
ffe
ma
les
ag
ed1
5to
64
yea
rs
(�0
.58
to�
1.2
5).
Dem
an
ded
shif
tsa
wa
y
fro
mo
lder
skil
led
ma
les
tow
ard
less
-sk
ille
dm
ale
wo
rker
s.
Ca
stil
lo-
Fre
ema
na
nd
Fre
ema
n
(19
92
)
Pu
erto
Ric
oD
ate
:1
95
6–
19
87
Da
ta:
U.S
.D
epa
rtm
ent
of
La
bo
r
Min
imu
mW
ag
eIn
du
stry
Stu
die
s,C
ensu
sP
op
ula
tio
nfo
r
Pu
erto
Ric
o,
Cu
rren
t
Po
pu
lati
on
Su
rvey
(CP
S)
for
Pu
erto
Ric
o
Th
eim
po
siti
on
of
the
US
-
lev
elm
inim
um
wa
ge
to
Pu
erto
Ric
od
isto
rted
the
Pu
erto
Ric
an
earn
ing
s
dis
trib
uti
on
,su
bst
an
tia
lly
red
uce
dem
plo
ym
ent
on
the
isla
nd
(0.2
0to
-
Th
ree
da
tase
tso
fea
rnin
gs
sho
w
rem
ark
ab
lesp
ikes
at
the
rele
va
nt
min
ima
inea
chd
istr
ibu
tio
n,
imp
lyin
g
tha
tth
em
inim
um
wa
ge
law
isa
ma
jor
det
erm
ina
nt
of
act
ua
lw
ag
esp
aid
.
con
tin
ued
Bhorat et al. 67
Ta
ble
A4
.C
onti
nu
ed
Stu
dy
Co
un
try
Pa
ram
eter
so
fst
ud
yE
mp
loy
men
t/h
ou
rsw
ork
edF
orm
al/
info
rma
lse
cto
rsh
ifts
Wa
ges
0.9
1),
rea
llo
cate
dla
bo
r
acr
oss
ind
ust
ries
,a
nd
aff
ecte
dth
e
cha
ract
eris
tics
of
mig
ran
tsto
the
Un
ited
Sta
tes.
Eri
kss
on
an
d
Py
tlo
ko
va
(20
04
)
Slo
va
k
Rep
ub
lic
Da
te:
19
98
–2
00
0
Da
ta:
Av
era
ge
Ea
rnin
gs
Info
rma
tio
nS
yst
em
Neg
ati
ve
effe
cto
n
emp
loy
men
tin
19
98
aft
erfi
rst
wa
ge
hik
e,b
ut
no
effe
ctin
19
99
aft
er
seco
nd
wa
ge
hik
e.(V
ery
sma
lln
um
ber
of
ob
serv
ati
on
sth
at
ma
y
no
tp
rese
nt
ad
equ
ate
resu
lts.
)
Av
era
ge
wa
ge
of
firm
sem
plo
yin
g,
rela
tiv
ely
,m
an
yw
ork
ers
fro
mth
e
low
eren
do
fth
ew
ag
ed
istr
ibu
tio
nis
rais
edm
ore
as
aco
nse
qu
ence
of
hik
es
inth
em
inim
um
wa
ge.
Din
kel
ma
na
nd
Ra
nch
ho
d
(20
12
)
So
uth
Afr
ica
Da
te:
Sep
tem
ber
20
01
,M
arc
h
20
02
,S
epte
mb
er2
00
2,
Ma
rch
20
03
,S
epte
mb
er2
00
3,
Ma
rch
20
04
Da
ta:
So
uth
Afr
ica
nL
ab
ou
r
Fo
rce
Su
rvey
s
No
sig
nifi
can
tef
fect
so
n
emp
loy
men
to
rh
ou
rso
f
wo
rkfo
rd
om
esti
c
wo
rker
sin
the
info
rma
l
sect
or.
Th
ere
wa
sa
larg
e,si
gn
ifica
nt
incr
ease
inw
ag
esa
fter
the
min
imu
mw
ag
ew
as
incr
ease
db
etw
een
18
.9a
nd
21
.7
per
cen
t.
Bh
ora
t,K
an
bu
r,
an
dS
tan
wix
(20
14
)
So
uth
Afr
ica
Da
te:
Sep
tem
ber
20
00
to
Sep
tem
ber
20
07
Da
ta:
So
uth
Afr
ica
nL
ab
ou
r
Fo
rce
Su
rvey
s
Res
ult
ssu
gg
est
a
sig
nifi
can
tem
plo
ym
ent
red
uct
ion
ina
gri
cult
ure
fro
mth
em
inim
um
wa
ge—
an
din
pa
rtic
ula
r
an
oti
cea
ble
mo
ve
aw
ay
fro
mem
plo
ym
ent
of
pa
rt-
tim
ew
ork
ers—
an
in-
crea
sein
wa
ges
on
av
er-
ag
ea
nd
ari
sein
no
nw
ag
eco
mp
lia
nce
.
Su
bst
an
tia
lin
crea
se
inco
ntr
act
cov
era
ge
for
farm
wo
rker
s
inS
ou
thA
fric
a.
Th
e
nu
mb
ero
fw
ork
ers
wit
ha
wri
tten
emp
loy
men
t
con
tra
ctin
crea
sed
to
rea
ch5
2p
erce
nt
in
20
07
.
Su
bst
an
tia
lin
crea
sein
farm
wo
rker
s’
wa
ges
by
ap
pro
xim
ate
ly3
0p
erce
nt.
con
tin
ued
68 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
Ta
ble
A4
.C
onti
nu
ed
Stu
dy
Co
un
try
Pa
ram
eter
so
fst
ud
yE
mp
loy
men
t/h
ou
rsw
ork
edF
orm
al/
info
rma
lse
cto
rsh
ifts
Wa
ges
Ov
era
lla
ver
ag
eo
fh
ou
rs
wo
rked
fell
inth
ep
ost
law
per
iod
,su
gg
esti
ng
tha
t
emp
loy
ers
ad
just
edto
som
eex
ten
to
nth
ein
ten
-
siv
em
arg
in,
an
dit
ap
-
pea
rsth
at
ho
urs
of
wo
rk
incr
ease
db
ym
ore
in
are
as
wh
ere
wa
ges
wer
e
low
erin
the
pre
law
per
iod
—d
riv
enla
rgel
yb
y
the
fall
inp
art
-tim
e
emp
loy
men
t.
Her
tz(2
00
5)
So
uth
Afr
ica
Da
te:
Sep
tem
ber
20
01
to
Sep
tem
ber
20
04
Da
ta:
So
uth
Afr
ica
nL
ab
ou
r
Fo
rce
Su
rvey
s
Ho
urs
of
wo
rkp
erw
eek
dec
rea
sed
(-0
.47
for
wo
men
an
d-0
.28
for
men
),a
nd
emp
loy
men
t
fell
(bet
wee
n-0
.19
an
d-
0.3
3)
for
do
mes
tic
serv
ice
wo
rker
s.
Av
era
ge
wa
ges
of
tho
seem
plo
yed
incr
ease
db
ya
pp
rox
ima
tely
20
per
cen
t.
Bh
ora
t,K
an
bu
r,
an
dM
ay
et
(20
12
)
So
uth
Afr
ica
Da
te:
20
00
–0
7
Da
ta:
So
uth
Afr
ica
nL
ab
ou
r
Fo
rce
Su
rvey
s
No
clea
rev
iden
ceth
at
the
intr
od
uct
ion
of
min
imu
m
wa
ge
law
sh
ad
a
sig
nifi
can
tim
pa
cto
n
emp
loy
men
tin
ag
iven
per
iod
for
the
reta
il,
do
mes
tic
wo
rk,
tax
i,
secu
rity
,a
nd
fore
stry
sect
ors
.W
ork
ers
inth
e
reta
il(-
4.5
per
cen
t),
secu
rity
(-4
.5p
erce
nt)
,
Ev
iden
ceo
fa
sig
nifi
can
tin
crea
sein
rea
l
ho
url
yw
ag
esin
the
po
stla
wp
erio
din
the
reta
il,
do
mes
tic
wo
rk,
an
dse
curi
ty
sect
ors
exa
min
ed.
Th
eta
xi
an
d
fore
stry
sect
ors
did
no
tex
per
ien
cea
n
incr
ease
inw
ag
es.
con
tin
ued
Bhorat et al. 69
Ta
ble
A4
.C
onti
nu
ed
Stu
dy
Co
un
try
Pa
ram
eter
so
fst
ud
yE
mp
loy
men
t/h
ou
rsw
ork
edF
orm
al/
info
rma
lse
cto
rsh
ifts
Wa
ges
an
dd
om
esti
cw
ork
(-7
.7
per
cen
t)se
cto
rs
exp
erie
nce
da
red
uct
ion
of
ho
urs
,b
ut
incr
ease
sin
wa
ges
ou
twei
gh
edth
ese
effe
cts.
Mu
rra
ya
nd
va
n
Wa
lbee
k
(20
07
)
So
uth
Afr
ica
Da
te:
Jan
ua
ryto
Ma
rch
20
05
Da
ta:
10
3se
mis
tru
ctu
red
inte
rvie
ws
of
larg
e-a
nd
med
ium
-sca
lefa
rmin
g
emp
loy
ers
No
larg
ed
isem
plo
ym
ent
effe
cts
occ
urr
edfo
rfa
rm
wo
rker
s,b
ut
ther
ew
as
som
ein
dic
ati
on
tha
t
emp
loy
ers
sub
stit
ute
da
t
the
low
ersk
ills
ma
rgin
as
wel
la
sa
dju
sted
at
the
inte
nsi
ve
ma
rgin
of
lab
or
tore
du
cew
eek
lyw
ag
e
cost
s.O
na
ver
ag
e,
wo
rker
sh
ou
rsw
ere
red
uce
db
etw
een
27
an
d
35
ho
urs
per
wee
k,
as
op
po
sed
toth
est
an
da
rd
45
-ho
ur
wo
rkw
eek
.
Co
nra
die
(20
04
)S
ou
thA
fric
aD
ate
:2
00
4
Da
ta:
Su
rvey
of
19
0w
ine
an
d
tab
leg
rap
efa
rmer
s
Dis
emp
loy
men
tef
fect
sa
re
fou
nd
wit
hta
ble
gra
pe
farm
wo
rker
s(-
0.5
9)
an
d
win
efa
rmw
ork
ers
(�0
.33
).
Na
ttra
ssa
nd
See
kin
gs
(20
14
)
So
uth
Afr
ica
Da
te:
20
10
–1
1
Da
ta:
Na
tio
na
lB
arg
ain
ing
Co
un
cil
for
the
Clo
thin
ga
nd
Ma
nu
fact
uri
ng
Ind
ust
ry
Th
eN
ati
on
al
Ba
rga
inin
g
Co
un
cil
for
the
Clo
thin
g
an
dM
an
ufa
ctu
rin
g
Ind
ust
ryla
un
ched
a
con
tin
ued
70 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
Ta
ble
A4
.C
onti
nu
ed
Stu
dy
Co
un
try
Pa
ram
eter
so
fst
ud
yE
mp
loy
men
t/h
ou
rsw
ork
edF
orm
al/
info
rma
lse
cto
rsh
ifts
Wa
ges
wa
ge
com
pli
an
ced
riv
ein
20
10
,w
hic
hre
sult
edin
the
clo
sin
go
ffo
ur
fact
ori
esth
at
thre
ate
ned
at
lea
st2
0,0
00
job
sin
New
cast
le,
So
uth
Afr
ica
.
Ga
rber
s(2
01
5,
un
pu
bli
shed
)
So
uth
Afr
ica
Da
te:
19
97
–2
00
7
Da
ta:
Oct
ob
erH
ou
seh
old
Su
rvey
an
dL
ab
ou
rF
orc
eS
urv
ey
Fin
din
gs
ind
ica
teth
at
form
al
un
skil
led
farm
emp
loy
men
td
ecre
ase
d
by
ap
pro
xim
ate
ly1
6
per
cen
ta
sa
resu
lto
fth
e
20
03
ag
ricu
ltu
ral
min
imu
mw
ag
e
reg
ula
tio
n,
of
wh
ich
7.5
per
cen
tis
dir
ectl
y
att
rib
uta
ble
toh
igh
er
un
skil
led
lab
or
cost
s
resu
ltin
gfr
om
the
wa
ge
flo
or.
Th
ere
isa
lso
evid
ence
of
skil
la
nd
cap
ita
lin
ten
sifi
cati
on
resu
ltin
gfr
om
the
min
imu
mw
ag
e.
Del
Ca
rpio
,
Mes
sin
a,
an
d
Sa
nz-
de-
Th
ail
an
dD
ate
:1
99
8–
20
10
Da
ta:
La
bo
rF
orc
eS
urv
eya
nd
Ho
use
ho
ldS
oci
o-E
con
om
ic
Su
rvey
Min
imu
mw
ag
ein
crea
ses
ha
ve
sma
ll
dis
emp
loy
men
tef
fect
so
n
fem
ale
,el
der
ly,
an
dle
ss-
Insp
ite
of
sub
sta
nti
al
no
nco
mp
lia
nce
,
the
min
imu
mw
ag
ein
Th
ail
an
dis
bin
din
g,
an
dit
ha
sa
bea
rin
go
na
ctu
al
wa
ge
(0.3
6).
con
tin
ued
Bhorat et al. 71
Ta
ble
A4
.C
onti
nu
ed
Stu
dy
Co
un
try
Pa
ram
eter
so
fst
ud
yE
mp
loy
men
t/h
ou
rsw
ork
edF
orm
al/
info
rma
lse
cto
rsh
ifts
Wa
ges
Ga
ldea
no
(20
14
)
edu
cate
dw
ork
ers
an
d
larg
ep
osi
tiv
eef
fect
so
n
the
wa
ges
of
pri
me-
ag
e
ma
lew
ork
ers
(�0
.05
5).
As
such
,in
crea
ses
inth
e
min
imu
mw
ag
ea
re
ass
oci
ate
dw
ith
incr
ease
s
inh
ou
seh
old
con
sum
pti
on
per
cap
ita
ing
ener
al,
bu
tth
e
con
sum
pti
on
incr
ease
is
gre
ate
sta
mo
ng
tho
se
ho
use
ho
lds
aro
un
dth
e
med
ian
of
the
dis
trib
uti
on
.In
fact
,ri
ses
inth
em
inim
um
wa
ge
incr
ease
din
equ
ali
tyin
con
sum
pti
on
per
cap
ita
wit
hin
the
bo
tto
mh
alf
of
the
dis
trib
uti
on
.
Del
Ca
rpio
,
Ma
rgo
lis,
an
d
Ok
am
ura
(20
13
)
Th
e
Ph
ilip
pin
es
N/A
Rea
lm
inim
um
wa
ge
incr
ease
sh
ad
neg
lig
ible
effe
cts
on
ov
era
ll
emp
loy
men
t,o
win
gto
the
lim
ited
cov
era
ge
of
min
imu
mw
ag
eru
les
an
dh
igh
no
nco
mp
lia
nce
.
con
tin
ued
72 The World Bank Research Observer, vol. 32, no. 1 (February 2017)
Ta
ble
A4
.C
onti
nu
ed
Stu
dy
Co
un
try
Pa
ram
eter
so
fst
ud
yE
mp
loy
men
t/h
ou
rsw
ork
edF
orm
al/
info
rma
lse
cto
rsh
ifts
Wa
ges
La
nzo
na
,2
01
2T
he
Ph
ilip
pin
es
N/A
Sec
tors
wit
hh
igh
cov
era
ge
an
dco
mp
lia
nce
exp
erie
nce
dn
ega
tiv
e
emp
loy
men
tef
fect
s.
Str
ob
la
nd
Wa
lsh
(20
03
)
Tri
nid
ad
an
d
Ta
ba
go
Da
te:
19
96
–1
99
8
Da
ta:
Co
nti
nu
ou
sS
am
ple
Su
rvey
of
Po
pu
lati
on
Bo
thla
rge
an
dsm
all
emp
loy
ers
inso
me
case
s
resp
on
ded
toth
e
min
imu
mw
ag
eb
yla
yin
g
off
wo
rker
s.
No
nco
mp
lia
nce
by
emp
loy
ers
wa
sp
rov
ento
be
asu
bst
an
tia
lis
sue
in
imp
lem
enti
ng
incr
ease
d
min
imu
mw
ag
es.
Ma
les
wo
rkin
gin
larg
efi
rms
ten
ded
to
ha
ve
thei
rw
ag
ein
crea
sed
toa
tle
ast
the
min
imu
mle
vel
;so
me
fem
ale
sin
bo
thla
rge
an
dsm
all
firm
sex
per
ien
ced
aw
ag
ein
crea
sed
ue
toth
em
inim
um
wa
ge.
Del
Ca
rpio
an
d
Lia
ng
(20
13
)
Vie
tna
mN
/AL
ow
-in
com
eo
ro
ther
wis
e
vu
lner
ab
lew
ork
ers
(in
clu
din
gw
om
en,
yo
uth
,re
cen
tla
bo
r-
ma
rket
entr
an
ts,
the
low
-
skil
led
,n
on
ma
na
ger
ial
no
np
rod
uct
ion
wo
rker
s
such
as
clea
ner
so
r
gu
ard
s,el
der
lyw
ork
ers,
an
dth
ose
emp
loy
edb
y
sma
llfi
rms)
are
pa
rtic
ula
rly
lik
ely
to
bes
hu
to
ut
of
the
form
al
con
tin
ued
Bhorat et al. 73
Ta
ble
A4
.C
onti
nu
ed
Stu
dy
Co
un
try
Pa
ram
eter
so
fst
ud
yE
mp
loy
men
t/h
ou
rsw
ork
edF
orm
al/
info
rma
lse
cto
rsh
ifts
Wa
ges
lab
or
ma
rket
as
are
sult
of
ov
erly
hig
hm
inim
um
wa
ges
.
Ng
uy
en(2
01
0)
Vie
tna
mD
ate
:2
00
4a
nd
20
06
Da
ta:
Vie
tna
mH
ou
seh
old
Liv
ing
Sta
nd
ard
Su
rvey
s
Min
imu
mw
ag
ein
crea
se
red
uce
dem
plo
ym
ent
of
low
-wa
ge
wo
rker
sin
the
form
al
sect
or.
Ho
wev
er,
wo
rker
sw
ho
lost
form
al
sect
or
job
sw
ere
ab
leto
fin
djo
bs
inth
ein
form
al
sect
or.
Th
eef
fect
of
the
min
imu
mw
ag
e
incr
ease
on
wa
ges
an
dex
pen
dit
ure
so
f
wo
rker
sis
no
tst
ati
stic
all
ysi
gn
ifica
nt.
74 The World Bank Research Observer, vol. 32, no. 1 (February 2017)