Content uploaded by Alex Bowen
Author content
All content in this area was uploaded by Alex Bowen on Oct 16, 2014
Content may be subject to copyright.
P R W P
5990
‘Green’ Growth, ‘Green’ Jobs
and Labor Markets
Alex Bowen
e World Bank
Sustainable Development Network
Oce of the Chief Economist
March 2012
WPS5990
Public Disclosure AuthorizedPublic Disclosure AuthorizedPublic Disclosure AuthorizedPublic Disclosure Authorized
Produced by the Research Support Team
Abstract
e Policy Research Working Paper Series disseminates the ndings of work in progress to encourage the exchange of ideas about development
issues. An objective of the series is to get the ndings out quickly, even if the presentations are less than fully polished. e papers carry the
names of the authors and should be cited accordingly. e ndings, interpretations, and conclusions expressed in this paper are entirely those
of the authors. ey do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and
its aliated organizations, or those of the Executive Directors of the World Bank or the governments they represent.
P R W P 5990
e term ‘green jobs’ can refer to employment
in a narrowly dened set of industries providing
environmental services. But it is more useful for
the policy-maker to focus on the broader issue of
the employment consequences of policies to correct
environmental externalities such as anthropogenic
climate change. Most of the literature focuses on direct
employment created, with more cursory treatment of
indirect and induced job creation, especially that arising
from macroeconomic eects of policies. e potential
adverse impacts of green growth policies on labor
productivity and the costs of employment tend to be
overlooked. More attention also needs to be paid in this
is paper is a product of the Oce of the Chief Economist, Sustainable Development Network. It is part of a larger
eort by the World Bank to provide open access to its research and make a contribution to development policy discussions
around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. e author
may be contacted at a.bowen1@lse.ac.uk.
literature to how labor markets work in dierent types of
economy. ere may be wedges between the shadow wage
and the actual wage, particularly in developing countries
with segmented labor markets and after adverse aggregate
demand shocks, warranting a bigger and longer-lasting
boost to green projects with high labor content. In these
circumstances, the transition to green growth and job
creation can go hand in hand. But there are challenges,
especially for countries that have built their industrial
development strategies around cheap carbon-based
energy. Induced structural change, green or otherwise,
should be accompanied by active labor market policies.
‘Green’ growth, ‘green’ jobs and labor markets
Dr Alex Bowen
a
a
Principal Research Fellow, Grantham Research Institute on Climate Change and the
Environment and Centre for Climate Change Economics and Policy (CCCEP), London School
of Economics and Political Science.
JEL: J23, J24, Q01, Q54, Q56
Keyword: sustainable development, climate change, green job, green growth
Acknowledgements: This paper was produced as background material for the report on
Green Growth managed by the Office of the Chief Economist of the Sustainable
Development Network. I would like to thank Uwe Deichmann, Stéphane Hallegatte, Michael
Spackman and particularly Mike Toman for helpful comments. I remain responsible for all
errors and omissions.
2
1. Introduction
The challenge of human-induced climate change has led to renewed interest in sustainable
growth as a means to promote human development. Growth – of the right type – remains
an effective way of lifting people out of poverty (see, for example, Collier, 2007). Growth of
the wrong type will eventually be self-defeating, as environmental degradation reduces
productivity and welfare. The OECD has noted that “green growth is gaining support as a
way to pursue economic growth and development, while preventing environmental
degradation, biodiversity loss and unsustainable natural resource use” (OECD, 2010). There
is particular concern about the consequences of human-induced climate change. As the
World Bank’s World Development Report 2010 argues, “Economic growth alone is unlikely
to be fast or equitable enough to counter threats from climate change, particularly if it
remains carbon intensive and accelerates global warming. So climate policy cannot be
framed as a choice between growth and climate change. In fact, climate-smart policies are
those that enhance development, reduce vulnerability, and finance the transition to low-
carbon growth paths” (World Bank, 2009). Stern (2010) thinks that “the new industrial
revolution and the transition to low-carbon growth constitute a very attractive path. It is
likely to bring two or three decades of innovative and creative growth and large and growing
markets for the pioneers. Low-carbon growth, when achieved, will be more energy-secure,
cleaner, safer and more bio-diverse than its predecessors.”
But what are the consequences of ‘green’ growth for labor markets? UNEP (2011) has
argued that “the greening of economies is not generally a drag on growth but rather a new
engine of growth; that it is a net generator of decent jobs, and that it is also a vital strategy
for the elimination of persistent poverty.” The recent global economic downturn triggered
many proposals for ‘green’ fiscal stimuli to promote growth and, in particular, jobs (see, for
example, Pollin et al., 2008). The OECD has also suggested that investing in green activities
has significant job creation potential (OECD, 2011).
Yet some have claimed that the potential is overestimated and environmental policies may
have much less attractive labor market consequences (e.g. Morriss et al., 2009). Michaels
and Murphy (2009) argue that “it is highly questionable whether a government campaign to
spur ‘green jobs’ would have net economic benefits.” Hughes (2011) has written about the
‘myth’ of green jobs, arguing additionally that job creation has no merit as a basis for judging
3
policy. Policies to promote ‘green’ jobs have even been alleged to be “terribly economically
counterproductive” (Alvarez et al., 2010).
This paper explores the possible relationship between labor markets and policies to promote
sustainable growth – improving well-being with societies taking proper account of resource
depletion and environmental impacts. In Section 2, the concept of ‘green’ jobs is discussed.
In Section 3, some of the empirical estimates of job creation are reviewed, concentrating on
the issue of how labor markets are assumed to function. Most of the literature is focused on
developed industrial economies, but there have been attempts to estimate some of the
employment consequences for developing countries. However, it is not clear that these
make appropriate allowance for the range of ways in which labor markets function, an issue
discussed in Section 4. Section 5 draws attention to the heterogeneity of developing
countries with respect to the opportunities for job creation in green growth. In Section 6,
the question of the skills necessary to support green growth is considered. Finally, some of
the possible implications for policy design in developing countries, and for research
priorities, are considered.
2. The concept: What are ‘green’ jobs?
There is no single agreed definition of a ‘green’ job. That makes it hard to compare studies
of ‘green’ job creation (GHK, 2009) and has led some researchers to eschew use of the term
completely. In a loose sense, though, ‘green’ jobs can be regarded as those associated with
environmental objectives and policies.
Some definitions of ‘green’ jobs or related concepts focus on occupations and skills with an
identifiable environmental focus, but most focus on employment in industries (or specific
projects) the products of which are deemed to be of environmental benefit. Such benefits
can be defined more or less broadly – for example, some concentrate on renewable energy,
including or excluding biofuels, while others also include environmental services and/or
employment related to improving energy efficiency or developing less-carbon-intensive
products (e.g. building railways).
4
UNEP has adopted a definition that attempts to incorporate aspects of job content as well as
the characteristics of industry gods and services (UNEP/ILO/IOE/ITUC, 2008). It defines
‘green’ jobs as “work in agricultural, manufacturing, research and development (R&D),
administrative, and service activities that contribute substantially to preserving or restoring
environmental quality. Specifically, but not exclusively, this includes jobs that help to
protect ecosystems and biodiversity; reduce energy, materials, and water consumption
through high-efficiency strategies; de-carbonize the economy; and minimize or altogether
avoid generation of all forms of waste and pollution.” This definition takes a broad industry
perspective, extending beyond employment in narrowly defined environmental services. In
principle, it embraces employment in producing any goods and services that have smaller
adverse environmental impacts than existing close substitutes. UNEP also argues that there
is a spectrum of ‘greenness’: “There are different degrees to which technologies, products,
businesses, and business practices can be said to be green, ranging from reactive and
remedial measures on the one hand to proactive measures on the other.” In other words,
cleaning up pollution after the event is less green than stopping the pollution in the first
place. That appears to beg the question as to the most efficient way of dealing with an
environmental problem and would lead to the classification of many environmental services
as less green than, say, employment in the packaging industry.
However, UNEP adds the rider that “green jobs need to be decent work, i.e. good jobs which
offer adequate wages, safe working conditions, job security, reasonable career prospects,
and worker rights. People’s livelihoods and sense of dignity are bound up tightly with their
jobs. A job that is exploitative, harmful, fails to pay a living wage, and thus condemns
workers to a life of poverty can hardly be hailed as green.” UNEP suggest that, at a
conservative estimate, there were more than 2.3 million jobs in the renewable energy sector
around the world in 2006, and more in construction, providing improved energy efficiency in
buildings, in low-carbon transport and in other environmental activities. But that compares
with an employed labor force globally of around 1.8 billion.
Thus the UNEP definition also extends to characteristics of the jobs themselves. However,
their definition conflates different social objectives in one term. The rider is particularly
problematic in developing countries where more employment may be desirable for the relief
5
of poverty and an increase in overall productivity – even if the jobs created pay little more
than a subsistence wage or the employment is in less green industries and skill classes.
Some definitions focus on a subset of industries producing environmentally desirable
outputs. Thus some studies, notably by the European Commission’s Environment
Directorate, have used the OECD/Eurostat definition of the environmental goods and
services industry (OECD, 1999), comprising “activities which produce goods and services to
measure, prevent, limit, minimize or correct environmental damage to water, air and soil, as
well as problems related to waste, noise and eco-systems. This includes technologies,
products and services that reduce environmental risk and minimize pollution and
resources.” That covers pollution management (e.g. air pollution control) and resource
management (renewable energy plants and water supply). On this basis, green jobs
constitute a small but significant share of total employment – 1.7% of total paid
employment in Europe (EC, 2007). That is probably a higher fraction than a global estimate
along UNEP lines would suggest; as UNEP notes, most of the documented growth in green
jobs has so far occurred mostly in developed countries.
1
Jobs in the nuclear power sector
are not included, and these are not generally regarded as ‘green’, although they are in a low-
carbon industry. Employees in many jobs might find that their jobs are not counted as
‘green’ despite the nature of the goods and services that they help produce. For example,
jobs in the car industry are excluded, even though some may be devoted to developing low-
carbon vehicles.
Some studies have developed their own terminology, using data from detailed employment
statistics or detailed company databases. The Pew Center, for example, defines the ‘clean
energy economy’ as follows: “The ‘clean energy economy’ generates jobs, businesses and
investments while expanding clean energy production, increasing energy efficiency, reducing
greenhouse gas emissions, waste and pollution, and conserving water and other natural
resources” (Pew, 2009). It comprises five categories: Clean Energy; Energy Efficiency;
Environmentally Friendly Production; Conservation and Pollution Mitigation; and Training
and Support. Using data about individual companies, Pew estimates the USA’s ‘clean energy
economy’ to account for about half a percent of all US jobs.
1
At the same time, developed countries are responsible for by far the largest share of the stock of greenhouse gases in the
atmosphere. They have also probably made a disproportionate contribution to long-lived solid waste. Hence some of the
green jobs reflect the unsustainability of developed countries’ economies.
6
Some definitions start from a different analytical perspective and try to answer the question:
“What are the employment consequences of introducing ‘green’ policies (e.g. ‘cap and
trade’) relative to a baseline case?” This approach requires implicit or explicit economic
modeling of the policies. Some studies in this vein count only jobs directly created by the
policies (‘direct’ employment effects) while others include jobs created in the supply chain
for the products and services supported by ‘green’ policies (‘indirect’ employment effects).
Kammen at al. (2004) and Wei et al. (2010) review several studies that estimate direct
employment effects of promoting renewable and other low-carbon energy supply and
energy efficiency, focusing on the specific labor requirements of particular technologies
(‘bottom up’ estimates, using simple spreadsheet-based analytical models in conjunction
with engineering estimates). An important issue that arises is the timing and duration of job
creation; there is a key distinction between construction, manufacture and installation,
where jobs may be relatively short-lived, on the one hand and ongoing operation,
maintenance and fuel processing on the other, where the length of jobs depends on the
durability of the relevant plant.
They also consider studies that use input-output (I-O) tables to estimate both direct and
indirect employment effects, taking account, for example, of the jobs created in business
services provided to the renewable energy sector. These extend the scope of the estimates
while sacrificing the greater granularity derived from engineering studies of specific energy
projects. I-O based studies also fall prey to the usual criticisms of input-output models: that
they do not allow for changes in input-output coefficients induced for example by relative
price changes and technological progress; that they are often out-of-date; that they depend
on industrial classifications that do not distinguish some of the key sectors of interest; and
that they are highly aggregated. The meta-studies by Kammen and his associates attempt to
derive standardized measures to compare estimates of jobs created per average megawatt
over the life of an energy facility. The authors also explore the implications of various
scenarios of exogenous energy efficiency improvements and renewable energy portfolio
standards for US employment in total. As they take into account jobs destroyed when fossil-
fuel-based energy is displaced by low-carbon sources, their projections are for a net concept
of employment change, but they do not take into account general equilibrium effects
through relative wage changes.
7
Others go further still and include jobs created by the aggregate demand generated by the
extra direct and indirect employment (‘induced’ employment effects). This approach allows
jobs to be counted as ‘green’ if they are created by ‘green’ policies, even if they are in
sectors with no obvious direct relationship to environmental objectives (e.g. tobacco
processing) or only a secondary relationship (e.g. construction). A question arises as to
whether one should net off jobs destroyed in sectors disadvantaged by ‘green’ policies (e.g.
coal mining). This issue is less relevant if one is simply trying to enumerate jobs associated
directly with environmentally attractive goods and services. But it is crucial if one is trying to
evaluate the overall labor market impacts of environmental policies. Some studies finesse
this issue by focusing on the job creation implications of different fiscal stimulus packages
with greater or lesser reliance on ‘green’ spending, none of which are expected to destroy
jobs. Pollin et al. (2008) is a good example of this type of study, utilizing an estimate of the
macroeconomic multiplier effect of additional direct fiscal spending to calculate induced
employment creation. It has stimulated a lively debate about the merits of trying to create
‘green’ jobs by means of a fiscal stimulus (see Box 1).
Box 1: Myths or reality of green jobs?
This Box summarizes the debate between Morriss et al. (2009) and Pollin (2009) about green
jobs and comments on the implications for analysis in the context of developing countries.
The supposed myth
The riposte
Comment
Everyone understands what
a ‘green job’ is
One should focus on the
green economy and overall
employment effects, not
‘green’ jobs
It might be helpful if the
green transition in
developing countries creates
unskilled jobs, given the
extent of underemployment
Creating green jobs will
boost productive
employment
Properly designed spending
should stimulate the
production of useful goods
and services, not Keynesian
‘make-work’
Low carbon growth
strategies for developing
countries should emphasize
the potential for the supply
of clean energy – jobs in this
sector would be productive
but, given low wage rates,
jobs are quite properly likely
to be low productivity ones
Green jobs forecasts are
reliable
All forecasting is subject to
uncertainty; the linear input-
output approach is helpful as
long as the methodological
More emphasis needs to be
put on the macroeconomic
and general equilibrium
adjustments and job
8
caveats are recognized
displacement; modeling for
the US labor market at a
time of high unemployment
may not be very instructive
for developing countries
Green jobs promote
employment growth
Clean energy is more labor
intensive than fossil-fuel
industry; low productivity
jobs are better than zero
productivity unemployment;
the low-carbon transition
will create higher paid jobs
too
There is more scope for
employment creation in
developing counties given
their segmented labor
markets and low
productivity; but low-carbon
jobs may be low-wage jobs
The world economy can be
remade by reducing trade
and relying on local
production and reduced
consumption without
dramatically decreasing our
standard of
living
The low-carbon transition
would increase local content;
trade patterns would
change; the USA is running
too big a trade deficit
This is problematic for
emerging-market economies
that wish to export to the
USA; more analysis is
required of how world trade
patterns would change (e.g.
if biofuels displaced fossil
fuels)
Government mandates are a
substitute for free markets
Pricing an externality
requires public policy
intervention; but policies
should work with market
incentives
High-quality collective
decision-making and
intervention in markets are
challenges for many poorer
countries
Imposing technological
progress by regulation is
desirable
Many of the necessary
changes in production can be
carried out with existing
technologies; some support
for renewable energy R&D is
warranted
Innovation can be stimulated
by market-based incentives;
more support is needed for
research into low-carbon
energy and other products
particularly suitable for
developing countries given
their level of development
and endowments
Another approach considers different time horizons; the further away the time horizon, the
more economic variables can be adjusted. Fankhauser et al. (2008) consider
A short-term effect, when jobs are lost in sectors directly adversely affected by new
climate change policies and new ones are created in replacement industries. This they
label the direct employment effect.
A medium-term effect, when the impact of climate change policies diffuses throughout
the economy, creating and destroying jobs along the value chains of affected industries.
9
These they call the higher-order, economy-wide effects of climate policy. That
corresponds to indirect and at least some induced effects.
A long-term effect, when innovation and the development of new technologies create
opportunities for investment and growth. This they call the dynamic effect of climate
policy, a benign induced effect that has had less attention in the literature. But learning
by doing usually helps to increase labor productivity, so the impact on jobs of
introducing a new technology may be attenuated over time.
Ho et al. (2008) take a similar approach in their study of the impact of carbon price policies
on US industry, considering outcomes along four different time scales:
The very short run, where firms cannot adjust prices and profits fall accordingly.
The short run where firms can raise prices to reflect the higher energy costs, with a
corresponding decline in sales as a result of product or import substitution.
The medium run, when in addition to the changes in output prices, the mix of inputs
may also change, but capital remains in place, and economy-wide effects are
considered.
The long run, when capital may be reallocated and replaced with more energy-efficient
technologies.
They found that, over the short term, employment losses were likely to mirror output
declines, but, in the longer term, gains in other industries would fully offset those losses.
Finally, some studies attempt to take more thorough account of economy-wide ramifications
of ‘green’ policies such as carbon pricing by using some form of general equilibrium
modeling. This is implicit in estimates of induced employment, because some
macroeconomic theory is needed to determine what happens to aggregate demand. The
multiplier-based approach exemplified by Pollin et al. can be thought of as being based on a
simple fixed-price Keynesian view of the macro-economy with Keynesian unemployment
and some ‘leakage’ of injections of aggregate demand to exports from other countries.
Computable general equilibrium models in the neoclassical tradition paint a very different
picture. They usually assume complete markets and instantaneous Walrasian price
adjustment, so that there is no involuntary unemployment. In such models, implementing
carbon pricing will both tend to redistribute labor to low-carbon activities and reduce overall
10
labor supply, because the higher relative price of carbon-intensive goods and services
reduces the real wage of labor, thus encouraging households to work less. They recognize
that employment costs are generally a social cost, not a social benefit, an important point
often missing in green jobs studies. According to this framework, there can be net job
destruction, depending on how the revenues from carbon pricing are used. In the study by
Goettle and Fawcett (2009) of the potential implications of a cap-and-trade system for the
USA, for example, there are significant reductions in labor input in 29 out of 35 US industries
(if there is no revenue recycling). Box 2 gives an informal analysis of the importance of
economy-wide effects in a simple neoclassical setting.
Hence knowing how best to model how the aggregate labor market works – and, indeed,
how the macro-economy as a whole works – is crucial for a proper assessment of changes in
direct and indirect labor demand, and first-round and induced employment effects. There
are important differences across types of economy due to different industry structures,
labor market institutions and endowments. Probably the most useful concepts to which to
attach the label of green jobs are the gross and net numbers of jobs created as a result of
green policy implementation. In the case of climate change policies, it would be appropriate
to consider the net and gross labor market impacts not only of carbon pricing but also of
other efforts to correct market failures contributing to carbon emissions, such as market
failures afflicting innovation and the provision of energy infrastructure.
The importance of the implicit or explicit macroeconomic framework is illustrated by Babiker
and Eckaus (2007), who show how, in the presence of real wage rigidities or barriers to the
sectoral reallocation of labor, climate policy could increase unemployment. Overall labor
market impacts can also be influenced by how the revenues from carbon pricing or quota
auctions (or other environmental taxes) are used, as illustrated by the literature on the so-
called ‘double dividend’ from environmental taxation (see, for example, Fullerton and
Metcalf (1997) and Sartzetakis and Tsigaris (2007)). Studies tend to show that if tax
revenues are used to reduce payroll tax – a tax on labor supply – employment will fall by less
or even increase.
2
2
But they rarely ask why the tax authorities have been levying a payroll tax in the first place. It may be a second-best way of
raising tax revenue and still have a place in the second-best menu of taxes.
11
Box 2: Different measures of job creation – A diagrammatic illustration
Consider a simple textbook neoclassical economy. Aggregate labor demand is a declining
function of the real producer wage. Aggregate labor supply is an increasing function of the
real consumption wage. When carbon pricing is introduced, this introduces a wedge
between the producer wage and the consumption wage because the carbon content of
consumer goods is taxed. In the figure below, aggregate labor supply shifts from S
0
to S
1
.
Real
producer
wage
Labour supply;
labour demand
D
0
D
1
D
2
S
0
S
1
L
0
L
2
The introduction also reduces the profitability of production of carbon-intensive goods for a
given real producer wage, using currently deployed technologies. That will tend to reduce
aggregate supply at any given product wage. Aggregate labor demand shifts from D
0
to D
1
.
But it also makes it profitable for some companies to switch to low-carbon technologies,
which empirically appear to be more labor-intensive but less-capital intensive than their
predecessors. Over time, as new low-carbon plant and equipment are installed, the
aggregate labor demand schedule therefore shifts out again. In the figure, the schedule
moves out over time to D
2
, which is drawn to the right of D
0
. This shift can be thought of as
the increase in labor demand as a result of the change to low-carbon technologies at
constant (zero) pure profits, which is analogous to some of the estimates of green job
creation in the literature (but is likely to be less than the green jobs created holding output
constant, as one would expect some substitution away from the previously carbon-intensive
goods and services). Public spending on capital for low-carbon production could accelerate
12
the shift. But the shift from D
1
to D
2
may not be large enough to push aggregate labor
demand at any given product wage above its original level given by D
0
. That will depend on:
how labor intensive the new technologies are compared with the old;
how profitable the new technologies are at the chosen level of carbon pricing, which
itself depends on the importance of learning-by-doing and other forms of induced
technological progress;
what is done with the revenues from the carbon pricing; recycling of tax revenues to
households will increase aggregate product demand (and a reduction in payroll taxes
could shift aggregate labor supply back towards S
0
).
Even when D
2
is to the right of D
0
, the net effect on employment of the introduction of the
carbon tax is a reduction in employment from L
0
to L
2
. The horizontal shift from D
0
to D
2
is
an over-estimate of the job creation induced by the introduction of carbon pricing. Thus this
measure of green job creation is misleading about the macroeconomic effects of climate
policies, because it neglects jobs lost from the initial leftward shifts in both the labor
demand and labor supply schedules.
If the labor supply schedule S is vertical, the shifts in the aggregate labor demand schedule
induced by carbon pricing affect only the equilibrium real product and consumption wages,
not total employment. Net green job creation is zero regardless of the relative labor
intensity of high- and low-carbon technologies.
This informal analysis serves to illustrate the importance of knowing how the labor market
works at a macroeconomic level in order to work out the employment effects of climate
policies. Estimates of gross job formation from switching from high- to low-carbon
technologies can be very misleading, especially if it is assumed that output of the originally
carbon-intensive industry remains unchanged.
However, the simple neoclassical textbook case is usually not the most appropriate way of
modeling the aggregate labor market despite its utility in providing a benchmark. Other
assumptions are possible. For example, nominal wage or price rigidities may make the
effective aggregate labor supply curve horizontal while leaving some workers involuntarily
unemployed – the neo-Keynesian framework. Then questions arise as to whether carbon
pricing is or is not passed through to final goods prices and whether and how nominal wages
13
respond. Carbon pricing could in this world act initially like an adverse aggregate supply
shock creating involuntary unemployment (ruled out by assumption in the neoclassical case
above). But there would also be a case for Keynesian deficit-financed spending, including on
low-carbon plant and equipment, to increase employment.
3. Empirics: Estimates of potential ‘green’ job creation
A number of surveys have collated estimates of ‘job creation potential’ in green activities
variously defined and discussed their methodological strengths and weaknesses, including
GHK (2009), Fankhauser et al. (2008), GCN (2010) and the World Bank (undated). Kammen
at al. (2004), Wei et al. (2010), and the World Bank (undated) have reviewed the literature
on employment creation in the energy sector, including jobs in renewable energy. The
estimates covered use a range of methods, reflecting the different definitions of green job
creation discussed above. As GHK (2009) notes, “The estimates are not comparable due to:
Geographical differences – EU, US, Member State, global;
Sectoral differences – often focusing solely on individual sectors or sub-sectors;
Gross and net effects;
Inclusion / exclusion of whole value chain effects; and
Differing assumptions concerning economic growth and the effect of existing business as
usual policies.”
GHK tabulate results from 32 different studies; most used data from the EU or specific
member states, seven used US data and four global data. No studies explicitly covered
developing countries per se. Wei et al. (2010) cover 15 different studies, only one of which
explicitly mentioned a developing country (Brazil). The World Bank review of energy and
employment tabulates results from 33 studies, of which five focused on some aspect of
energy in developing countries. Five out of the 15 studies covered by Wei et al, 14 out of the
32 studies covered by GHK and 25 out of the 33 studies covered by the World Bank are
dated 2008 or later.
Given the heterogeneity of the studies, it is very difficult to draw broad conclusions. The
impression given by many of the findings is that climate-change policies in general and
renewable energy in particular can generate considerable extra employment. But many
14
studies ignored the potential for job destruction in non-green industries or implicitly or
explicitly assumed that there would be no crowding out of jobs via general equilibrium
effects. Most studies did not take into consideration general equilibrium effects at all while
some others offered a qualitative discussion. Few of the studies considered labor market
problems such as sector-specific human capital and job search costs that could slow the shift
of workers across sectors or out of long—term structural unemployment. Several of the
studies were in fact designed to assess the job creation possibilities from green policies at a
time of widespread involuntary unemployment – hardly surprising considering the severity
of the global downturn that has recently been experienced but not necessarily indicative of
green job prospects in moving from one longer-term growth path to another. However,
arguments might equally be advanced for many other employment generation policies.
Wei et al. (2010) are illustrative of the general thrust of the literature when they write:
“Our modelling approach yields the following key conclusions:
The renewable energy and low carbon sectors generate more jobs per unit of energy
delivered than the fossil fuel-based sector.
Among the common RPS [renewable portfolio standard] technologies, solar
photovoltaics (PV) creates the most jobs per unit of electricity output.
Energy efficiency and renewable energy can contribute to much lower CO
2
emissions
and significant job creation. Cutting the annual rate of increase in electricity generation
in half and targeting a 30% RPS in 2030 each generates about 2 million job-years through
2030 [in the USA].
A combination of renewable energy, EE [energy efficiency], and low carbon approaches
such as nuclear and CCS can yield over 4 million job-years through 2030 [in the USA]
with over 50% of the electricity supply from non-fossil supply sources.”
Pollin et al. (2008) similarly argue that a $100 billion fiscal stimulus in the USA spent on six
energy efficiency and renewable energy strategies would generate two million jobs. The
inclusion of indirect and induced employment means that many of the jobs created are not
in the conventionally defined ‘green’ industries (see above). Indeed, non-direct employment
creation is often larger than direct job creation given the assumptions about supply chains
and fiscal multipliers, reassuring authors that direct job losses due to the contraction of
fossil-fuel-intensive activities can be outweighed by the total gross job creation. Renewables
are more labor-intensive than conventional energy, especially at the construction,
15
manufacturing and installation stage (less so in operation and maintenance, partly because
fuel input management is not necessary).
Fankhauser et al. (2008) note two important caveats. First, higher labor intensity per unit of
energy capacity created implies lower labor productivity than in conventional energy
production. Renewable energy may be more expensive and less efficient than conventional
sources, with high capital as well as labor requirements, shorter-lived plant and more
intermittent energy production. Thus some studies of potential employment creation ignore
the question of whether it would be profitable for the private sector to adopt the programs
considered. Second, labor is relatively immobile in the short run, so policy changes may
create transitional frictional unemployment due to the structural change induced. Both
points reflect the tendency for employment creation studies to be vague about the
macroeconomic adjustment mechanisms implicit in their estimates.
In addition to those two caveats, studies rarely take into account the macroeconomic
consequences of higher real energy prices (or higher energy subsidies) on the overall price
level, labor supply, labor productivity and taxation. Converting countries’ energy sectors to
low-carbon technologies is likely to require increased investment. McKinsey & Company
(2009) estimates that the annual incremental investment costs required to get the global
economy on to an appropriate low-carbon trajectory would be EUR 320 billion by 2015. The
IEA (IEA, 2008) and UNFCCC (UNFCCC, 2007) suggest a similar figure for the incremental
costs of power generation in strong climate-change mitigation scenarios. However, the pace
of the transformation may be constrained by the cost and/or availability of financial capital
(given the risk characteristics of the investments involved) or incur a higher cost in terms of
investment and employment displaced elsewhere in the economy.
A further caveat is that studies tend to focus on alternative energy supply scenarios rather
than considering a broader range of alternatives, some of which could entail greater job
creation still. For temporary counter-cyclical employment creation, higher spending in
sectors with lower capital intensities than either conventional or renewable energy – such as
education and health services – may be more effective (although perhaps more difficult to
unwind when macroeconomic circumstances improve).
16
That is one reason why Strand and Toman (2010) express some skepticism about proposals
for a green fiscal stimulus, particularly in developing countries. Road-building, for example,
is relatively labor-intensive and can help to provide valuable infrastructure, but it is not
particularly green. They also note that many of the studies for developed countries do not
emphasize short-run employment creation, because of the lead times necessary for
replacing capital in the energy sector. In their view, “these studies strongly indicate that
government support to producing and developing renewable energies is not a very efficient
way of creating additional short-run employment in high-income countries; at least, not
when appropriately accounting for the opportunity cost of public funds going into such
support.”
Overall, Strand and Toman conclude that there are likely to be trade-offs for employment
generation, in the sense that programs that yield larger employment effects tend to lead to
more employment gains for largely lower-skilled workers, so that the long-term growth
effects are relatively small. Long-term development, including sustainable development,
requires more of a focus for public investment on growth-enhancing infrastructure, as well
as private sector investment, which is not necessarily labor-intensive. The argument for
investing in initially labor-intensive low-carbon technology in order to benefit from
experience, scale economies and learning by doing, thus driving costs down, is weaker for
developing countries because they are less likely to have a comparative advantage in low-
carbon innovation, at least in the energy sector.
Strand and Toman (2010) also review some of the small literature on green fiscal stimuli and
job creation specifically in developing countries, noting that the evidence is scant and spotty.
Studies largely focus on the direct effects of particular activities on employment. Barbier
(2009) investigates the South Korean green stimulus while the UNEP/ILO/IOE/ITUC (2008)
study looks at China’s experience. What is most striking is the large variation in employment
creation in jobs created per US$ billion spent. In South Korea, forest restoration is estimated
to be highly labor-intensive, generating nearly eight times as many jobs per dollar as the
least labor-intensive green objective, ‘vehicles and clean energy.’ In China, biomass
spending is thought to be nearly 30 times more effective in generating jobs per dollar spent
than wind power. That suggests that the focus on renewable energy and low-carbon
manufacturing prevalent in studies for Europe and the USA may miss the opportunities for
17
employment creation from changes in land management and agriculture in developing
countries, where these economic sectors are relatively more important.
Schwartz et al. (2009) considers direct employment effects of fiscal stimulus projects in
three Latin American countries, Honduras, Brazil and Peru. Again, the variation in jobs
created per dollar across the different projects considered is large. Water network
rehabilitation and expansion in Honduras is much more effective (by a factor of more than
ten) in creating jobs than hydroelectric schemes in Brazil, with rural electrification in Peru
falling in between (the Honduran projects seem to be a lot more labor intensive than water
resource management in South Korea).
Rutovitz (2010) carries out a detailed analysis of employment opportunities from a switch
towards renewable energy and greater energy efficiency in South Africa, in the spirit of
Rutovitz and Atherton (2009), a global study for Greenpeace. The focus is on direct
employment but allowance is made for learning effects with new technologies that increase
labor productivity (reduce the number of jobs supported) over time. Different scenarios are
considered for how much of the plant and equipment needed is manufactured in South
Africa and for how much South Africa supplies the rest of Africa with renewables
technologies. Job losses in the coal industry are factored into the projections. A business as
usual scenario from the International Energy Agency (IEA) is compared with (i) a South
African government ‘Growth Without Constraints’ (GWC) scenario designed to reflect South
Africa’s energy future in the absence of climate change, with no oil constraints, and if no
effort was made to internalize externalities, and (ii) an Energy [R]evolution scenario
designed to reduce South African emissions by 60% by 2050 (compared with 2005). The
authors estimate that the Energy [R]evolution scenario creates 27% more jobs than the IEA
reference scenario and 5% more than the GWC scenario. Macroeconomic feedbacks are not
analyzed and nor are costs. Yet the costs of job generation, including opportunity costs, can
be substantial once macroeconomic consequences are taken into account.
Upadhyay and Pahuja (2010) examine the employment potential of renewable energy,
especially wind and solar power, in India. Once again, its scope is limited to an estimate of
direct employment effects, because it is difficult to assemble input-output tables with an
appropriate breakdown of activities and to model induced macroeconomic effects. Unlike
the developing country studies discussed above, the authors concentrate on estimates of
18
jobs created per megawatt of energy generation capacity, similar to the work by Kammen
and others mentioned previously, drawing on India’s National Action Plan on Climate
Change. Such estimates are likely to show a bigger contrast between renewable and
traditional energy than do estimate of jobs per dollar spent, because the cost of an extra
megawatt of generating capacity is likely to be higher with the former. Unlike Strand and
Toman, where the focus is on evidence from advanced industrialized countries, they
conclude that low-carbon employment is one of the key co-benefits of promoting the
renewables sector in India. They find that solar power is more labor intensive than wind
power and better able to meet India’s requirements for small-scale, off-grid power; biomass,
‘green’ transport and public works in water and forest management are also seen as
attractive ways of achieving both employment and environmental objectives. One serious
problem with Upadhyay and Pahuja’s study is that the costs of different renewables
scenarios are barely considered; the perspective is more of a quantitative planner. It is also
interesting to note that the study explicitly rules out job losses in fossil-fuel power
generation, arguing that “investment in fossil fuels is going to be a mainstay” of the Indian
economy, at least in the nearer term, given the need to expand availability and reliability of
electricity supplies as well as to pursue environmental objectives. This, together with an
implicit assumption of surplus labor, allows the authors to dismiss job displacement as an
issue.
Upadhyay and Pahuja’s study is a contribution towards the Global Climate Network (GCN)
report Low Carbon Jobs in An Interconnected World (GCN, 2010), which also covers China,
South Africa, Brazil and Nigeria, as well as some developed countries. The study of China
emphasizes the potential employment losses from the planned sharp reduction in the
energy intensity of Chinese industry, but notes that this could be outweighed by increased
employment in renewables and – quantitatively, much more important – the shift of the
Chinese economy towards services and away from heavy industry. But the latter shift
cannot be seen as a specifically ‘green’ objective. The study of South Africa focuses on direct
job creation potential in renewables and draws the conclusion that “significant opportunities
for employment lie in clean-energy sectors and can be harnessed if the South African
government scales up its renewable energy ambitions.” But again, this result flows from the
fact that renewable energy production is more labor intensive than traditional energy
sources. The question of the labor productivity and costs of renewables compared with
traditional energy is not tackled.
19
The study of Brazil considers the employment consequences of fulfilling various targets for
renewable energy supply, drawing attention to employment opportunities in
hydroelectricity (notwithstanding Schwartz et al.’s finding that hydro is relatively capital
intensive and poor in creating jobs per dollar spent), biofuels, biomass and solar power. The
cost per job created is not calculated. The study concludes that renewable energy sources
have a stronger potential in Brazil than is currently envisioned in official studies and
government policies, both in terms of contributing to carbon dioxide mitigation and
generating jobs.
The study of Nigeria considers the employment impacts of small-scale hydroelectric power,
a key component of Nigeria’s 2005 Renewable Energy Master Plan, and greater utilization of
natural gas, a lower-carbon source of energy than oil and coal. Once again, estimates are
largely restricted to direct employment effects, although there is some effort to estimate the
number of jobs likely to be created in the supply chain of the natural gas sector.
Overall, GCN (2010) comes to the following conclusions:
(i) Clear, consistent and targeted government policy will help boost jobs numbers;
(ii) Finance is critical to the creation of low-carbon economic opportunities;
(iii) Training is critical to the development of low-carbon sectors;
(iv) Adjustment policies should also form part of the strategy; and
(v) More analysis is needed of how globalized markets will affect job creation.
The first conclusion is based largely on the fact that most of the policies analyzed in the
contributing studies generate additional direct employment, but little attention is paid to
employment displaced in high-carbon and other environmentally harmful activities or to
macroeconomic constraints and the possibility of crowding out jobs. However, it does
appear that several environmentally beneficial activities in developing countries could be
considerably more labor intensive than traditional fossil-fuel-based energy supply.
The second conclusion emphasizes the need for capital to finance green investment, noting
that the private sectors in many countries are currently not eager to lend. However, the
potential constraint imposed by potential lack of profitability of green investment, with or
without carbon pricing and other financial incentives, is not analyzed in depth.
20
The third conclusion, about training, is a useful reminder about possible bottlenecks in the
labor market.
The fourth conclusion addresses the structural change induced by environmental policies
and so acknowledges the need to take some account of the issue of job displacement.
The fifth conclusion very properly draws attention to possible leakages of green jobs and
spending to other countries, to an extent that depends on endowments of skills, existing
industry structure, the nature of the technologies newly deployed and the ways in which
comparative advantage is exploited – a reminder that general equilibrium effects matter.
Yet these are largely ignored in the green jobs literature. They may differ in developing
countries from the effects in developed countries, as the next section discusses.
4. Modeling labor markets in developing countries
It has long been recognized that the standard neoclassical or Keynesian models may not
adequately represent labor markets in developing countries. Thus the implicit assumption in
many green jobs studies that there is Keynesian unemployment and no crowding out of jobs
elsewhere in the economy by green fiscal stimuli may not be valid. But equally, the
objection that employment is determined in competitive markets, implying that green jobs
are likely to displace at least as many jobs elsewhere in the economy, may lack force
because the assumptions on which it is based do not hold.
This issue is important for evaluating green policy initiatives’ labor market implications and is
hardly a new one. For example, half a century ago Lewis drew attention to the possible
existence of surplus labor in developing countries, particularly in agriculture (Lewis, 1954).
In a surplus labor economy, there is much less risk of crowding out employment when green
projects are undertaken. Thus the estimates of direct employment creation in the green
jobs literature might be less misleading for developing countries than for industrial
economies close to full employment. But the situation is more complicated in ‘dual’
economies with modern and traditional sectors or three-sector economies with a traditional
rural sector and both formal and informal urban sectors, as has also been long understood
(Harris and Todaro, 1970; Fields, 1975; Mazumdar, 1976).
21
Other assumptions of the standard competitive labor markets model could also be relaxed.
In one of the few attempts to do this, Babiker and Eckaus (2007) explore the impact of
constraints on real wage adjustment and on inter-sectoral flows of labor in different regions
of the world and find that China and India could be the countries most adversely affected by
the structural change induced by carbon pricing, largely because these are the regions that
need most reallocation of labor as a result of environmental policy.
3
It would be very useful
to consider the implications of various models of labor market functioning more specific to
developing countries (e.g. as reviewed by Behrman, 1999) for the employment and wage
impacts associated with a shift towards green policies.
As an example, consider the model of Satchi and Temple (2009). As they point out, “At
present, we know relatively little about how changes in sectoral productivity translate into
labor market outcomes.” Yet green policies appear likely initially to lead to lower
productivity in the energy sector, with pervasive knock-on effects in the rest of the
economy, unless at the same time other market failures, such as those involved in the
provision of energy supply infrastructure, are corrected. Satchi and Temple set up a small-
scale general equilibrium model, in the tradition of labor market search models, designed to
characterize typical middle-income countries (their calibrations use Mexican data). There
are three sectors: a formal urban manufacturing sector, an informal urban sector, and a
rural sector. Underemployment – essentially, employment in the informal sector – arises
because of job/worker matching frictions in the formal sector. The urban wage (in the
formal sector) is endogenously determined, unlike in Harris and Todaro (1970). The rural
sector is assumed to be perfectly competitive with constant returns to scale and full
employment, thus ruling out the type of surplus labour considered in many low-income
countries.
The authors consider the impact of productivity shocks in different parts of the economy. If
green policies are thought of as having adverse impacts on energy sector productivity, it
matters where energy sector jobs are located. Suppose they are in the formal urban sector,
as would be the case if they are largely in regulated energy utilities. Then the size of the
informal sector may increase (and definitely does not decrease) as a result. The urban
3
The ways in which Babiker and Eckaus model real wage rigidity and limits on inter-sectoral labour flows can be argued to be
inappropriate over the very long run associated with integrated assessment models. But their study nevertheless
demonstrates the potential importance of these constraints.
22
unemployment rate goes up but rural-urban migration is reduced, pushing down agricultural
wages. Total unemployment does not necessarily increase, because the movement of
population to rural areas may reduce the denominator of the unemployment rate.
Green policies do not only change energy sector productivity. They may also lead to
increased direct employment in the formal sector if, for example, there is a green fiscal
stimulus aimed at the energy industry. In that case, one could simulate the impact in Satchi
and Temple’s model as a reduction in the size of the competitive formal sector (as workers
are siphoned off into the non-competitive state-financed projects). Re-establishment of
equilibrium would entail higher wages in the competitive formal sector, the growth of the
formal sector as a whole, and increased rural-urban migration, with ambiguous impacts on
the informal sector. That might in turn boost aggregate productivity, even if the state
projects generated jobs with lower productivity than the rest of the formal sector; in the
model, the marginal product of labor in the formal sector is 1.5 times that in the informal
sector and 2.7 times that in the rural sector.
Green policies might also induce inward flows of capital (e.g. through the Clean
Development Mechanism and other international climate finance arrangements). That
would tend to boost formal sector employment. One could carry out other thought
experiments in this vein. For example, adaptation to climate change in rural areas could be
modeled as an adverse productivity shock in the rural sector. The impact of green energy
policies on the price of energy (increasing them at least for a while) could be modeled as
reducing productivity throughout the economy. The key point is that the overall effects of
green policies on employment depend on the characteristics of the economy’s labor markets
and the nature of the policy interventions, including their funding, not just the input
requirements of rival energy technologies.
Indeed, it would be helpful to consider the implications of a wider range of theories of labor
market adjustment in different types of economy. After all, there still remains much debate
about how best to characterize business cycles and labor market adjustment in developed
countries. Are workers always on their labor supply curves, as in standard real business
cycle theory? What are the dynamics of job creation in a Mortensen-Pissarides world of job
search in a world of imperfect information and search externalities (Pissarides, 2009)? Is
Keynesian unemployment caused by wage and price rigidities and, if so, which matters
23
more, and is nominal or real rigidity more important? If the source of the market
malfunctioning at the macroeconomic level is not understood, it is difficult to work out what
effect a second-best employment creation policy such as new environmental investments
would have (as opposed to a first-best policy of removing the offending rigidity). All of these
questions are relevant to investigating the likely labor market consequences of green policy
initiatives ranging from carbon pricing to changes in the pattern of government subsidies to
energy to deficit-financed green infrastructure investments.
Understanding the interaction of business cycles and labor markets better would help to
establish what the scope is for green fiscal stimuli to counter increases in involuntary
unemployment and underemployment in developing countries. Although governments in
developing countries may on average have less of a capacity to time and manage fiscal
stimuli appropriately (Kraay and Servén, 2008), business cycles may be more of a problem in
developing countries, especially those engaged in international commodity trade, than in
developed economies (Kraay and Ventura, 1998; Otero, 2000). Arguably, development
economics has not taken sufficient account of the importance of fluctuations in aggregate
demand, instead paying more attention to capital accumulation and technological
innovation (Dutt, 2007).
Taking a much longer-term perspective, further analysis is needed of the scope for inducing
higher long-run growth in developing countries by raising the rate of innovation. That could
in principle speed up the re-allocation of labor inputs to more productive sectors of their
economies, reducing underemployment and increasing real wage growth (if the new skills
necessary are developed through appropriate training). Although the literature on green
jobs tends to focus on the current characteristics of renewable energy supply and energy
efficiency, integrated assessment modeling of climate-change mitigation is paying more
attention to induced technological progress in low-carbon technologies. Are developing
countries going to depend on spill-overs from R&D and learning in developed countries, as is
largely the case in the WITCH model (Bosetti et al., 2006), or can they benefit sooner, for
example, from advances in technologies more fitted to developing country circumstances
(e.g. biofuels in Brazil, concentrated solar thermal power in North Africa, forest
management in Indonesia)? Can developing countries with large manufacturing sectors
benefit from a new global Industrial Revolution driven by a Schumpeterian search for
competitive advantage (Perez, 2002)? Several of the jobs studies for countries such as South
24
Africa, Brazil, China and India allude to the possibility, but the long-run dynamics of green
growth are still not fully understood. There is of course a downside to gales of competitive
destruction and a risk that more rapidly shifting comparative advantage will give rise to
more abrupt structural change, putting strains on the operation of labor markets.
Besides innovation, there are other areas of policy, such as network infrastructure provision,
where there is also a presumption that private markets will provide less investment than is
socially desirable. Here too, the consequences of new policies for labor markets will be very
largely dictated by the way in which those markets function.
5. Heterogeneity of developing countries
The previous section of this paper noted the importance of recognizing the different
circumstances of developing countries compared with developed ones. But there is much
variation across developing countries, too, not least with respect to the likely ease of
transition to a world of low-carbon growth. For example, the scope for developing a
comparative advantage in the production of equipment for low-carbon electricity
production depends on the manufacturing base of the country concerned and whether
there are scale and learning economies in the particular technology. Thus the
Greenpeace/EREC (2011) study envisages that, in its Advanced Energy [R]evolution scenario,
20% of the growth in renewable capacity in the rest of Africa by 2020 will be supplied by
South African manufacturing, and by 2030 30%. Some countries have a comparative
advantage in particular renewable energy sources because of natural endowments. Brazil,
for example, has the right climatic conditions and soils to give it a substantial cost advantage
in biofuels, although other characteristics of the Brazilian economy also help (Kojima and
Johnson, 2005).
The developing countries that currently produce a high level of greenhouse gas emissions
per unit of GDP face a difficult challenge of structural adjustment. They are the ones in
which more labor is likely to have to be reallocated from currently GHG-intensive activities,
either by switching technologies within an industry or by moving labor between industry
sectors. Given the importance of carbon dioxide emissions from energy production, energy-
intensive economies will comprise a large part of this group. The following scatter diagrams
illustrate the wide variation across developing countries according to measures of carbon
25
dioxide emissions per unit of GDP and energy use per unit of GDP. Lower-income countries
in Europe, Central Asia and the Middle East take up places at the top of the rankings.
Figure 1: Carbon dioxide emissions intensity and per capita income
Figure 2: Energy intensity and per capita income
Endowments of fossil fuels combined with industrial development strategies that have
favored carbon-intensive industry make a transition to low carbon much more challenging
(EBRD, 2011). If such economies impose a carbon tax, the standard economic policy
0
2,000
4,000
6,000
8,000
10,000
0 200 400 600 800 1,000 1,200 1,400
GDP per capita
Energy intensity v income - 2005
Developing countries (GDP per cap. less than $10,000)
(GDP: PPP constant 2005 international $)
toe per million $ GDP
26
instrument to internalize the greenhouse gas externality, the relative returns to different
factors of production are likely to change. The few empirical studies focusing on how carbon
taxation might affect factor returns suggest that the incidence of a carbon tax is likely to be
regressive when emission abatement measures are capital-intensive (See Fullerton and
Heutel (2007) and Fullerton and Heutel (2010), who calibrate analytical general equilibrium
models for Japan and the United States respectively). Countries such as Kazakhstan and
Mongolia with a much larger-than-average proportion of the labor force in mining and
energy supply are more likely to suffer as a result of this adjustment and also from the
difficulties of reallocating displaced labor to other sectors emphasized by Babiker and Eckaus
(2007).
Similarly, some lower-income countries produce relatively high levels of methane from
agriculture per unit of GDP (Figure 3), although data are patchy and often out of date. The
substantial changes in agricultural practices and waste management that are needed pose
larger adjustment problems in these countries.
Figure 3: Methane emissions intensity and per capita income
6. Skills for green growth
27
The emphasis in this paper so far has been on changes in the quantity of labor demanded or
supplied as a consequence of adopting ‘green growth’ policies. But the skills needed in the
labor force are likely to be affected as well. If the availability of the skills required along
green growth trajectories is inadequate, that could place a major obstacle in the way of the
transition to green growth.
Fankhauser et al. (2008), in their review of the literature on climate policy and jobs, note
that there is relatively little information on the productivity of jobs created and destroyed by
climate-change policies. There is even less on their pecuniary and non-pecuniary attributes.
CEDEFOP (2009) complains that “We have not paid enough attention to the social dimension
of sustainable development: its implications for employment, training and skills.” Even in
countries with relatively good labor market data, it is difficult to identify the jobs the content
of which is most likely to be affected by green growth policies. For example, Hatfield-Dodds
et al (2008) note that in Australia “current information on green skills and workforce
capabilities is very poor;” there are some data for renewable energy and construction trades
but little for transport and agriculture.
Perhaps the most thorough study of green growth and skills so far is ILO/CEDEFOP (2011),
which reports the results of 21 country reviews and provides a synthesis of key findings. The
study notes that the demand for skills is being affected in three ways by the transition to
green growth.
First, there is induced structural change across industries, increasing the demand for the
skills specific to expanding industries such as renewable energy and reducing the demand
for skills such as those associated with coal-mining. This ‘green restructuring’ brings with it
the usual challenges to policy-makers trying to facilitate restructuring and to reduce the
labor market adjustment costs, including those resulting from a changing skill mix. Many of
the expanding industries are likely to be using new products and processes, reflecting the
transition to low-carbon technologies, so the generic skill requirements of many of the new
jobs created are likely to be higher than average as they have to allow for assimilation of
unfamiliar tasks and working methods and ‘learning by doing.’ It would appear, however,
that a larger proportion of jobs in the renewable energy sector and in energy efficiency are
low skilled than in the fossil-fuel energy sector (Pollin et al., 2009). The oil and gas industries
28
tend to have relatively well-paid workers and a high proportion of highly qualified engineers
and technicians. Energy efficiency improvements tend to require relatively unskilled
construction labor. But there is much heterogeneity, with, for example, smart grid
technology management requiring more input from high-level engineering services than do
building retrofits. In developing countries, re-afforestation is an important potential source
of jobs for low-skilled rural workers.
Second, some new occupations are emerging, such as photovoltaic fitters and carbon-
footprint assessors. But there appear to be relatively few unique green skills.
Third, the content of many existing jobs in existing industries is changing to reflect facets of
the transition to green growth, such as increasing emphasis on energy efficiency, switching
from fossil fuel sources to renewable energy and producing capital equipment for expanding
green industries. In agriculture, low- and no-till agriculture and reduced use of fertilizers and
pesticides will entail changes in farmers’ practices, as will increased production of biofuel
crops and efforts to increase forest cover. The ILO/CEDEFOP report argues that this third
channel is likely to have the most pervasive effects on labor markets and this seems
particularly likely to be the case in many developing countries. More generally, green jobs
appear to be very diverse in their skill requirements, both with respect to the sophistication
of the skills required and their degree of novelty.
Recent reports such as ILO/CEDEFOP (2011) and OECD (2011b) have found that skill
shortages are already impeding the transition to green growth. The OECD draws attention
to widespread skill shortages in energy-efficient construction and retrofitting; renewable
energy; energy and resource efficiency; and environmental services. Particular countries
have reported specific bottlenecks, such as the shortage of skilled PV workers in Germany
and the lack of design engineers for smart grids in the UK. ILO/CEDEFOP argues that there
are several reasons for these reported shortages, including the underestimation of the
growth of certain green sectors, the general shortage of scientists and engineers, the low
reputation and attractiveness of some sectors important for the green transition such as
waste management, and a shortage of teachers and trainers in environmental services. But,
as the OECD writes, “it is difficult to assess how general and severe these green skill
shortages are based on evidence from highly diverse case studies.”
29
The problems that can arise when training provision is not up to the challenge of the
induced structural change are illustrated by the Australian experience with a new Home
Insulation Program introduced in February 2009 as a key part of the Australian government’s
fiscal stimulus (Australian National Audit Office, 2010). The Program was designed partly to
generate jobs for lower-skilled workers in the housing and construction industries. At the
start of the program, only supervisors were required to satisfy one of three minimum
competences – prior experience in the insulation industry, qualifications in an approved
trade, or insulation-specific training. The program proved popular. At its peak, demand was
running at almost 2 ½ times the anticipated level and some 1.1 million roofs out of 2.7
million eligible were insulated. But fires, fitters’ deaths and reports of fraud undermined
public confidence and it was cancelled in February 2010. A subsequent sample of
inspections revealed that nearly 30% of installations had some level of deficiency.
Investigations showed that low skill levels in the industry, inadequate provision of training
and poor management of the program were among the factors responsible. The importance
of competent project management and national policy-making in this case is a reminder of
the key role of higher-level management and planning skills in a policy-induced transition to
green growth that is likely to take sustained effort and policy credibility over a long period.
In developing countries, it may be more difficult to fulfill the requirement suggested by
UNEP and others that green jobs should be ‘decent’ jobs. Upadhyay and Pahuja (2010)
make this point in the Indian context, noting that “jobs like that of unskilled labor in biofuel
and biomass production could be numerous but of low quality as they barely provide
subsistence wages and have difficult work conditions.” They suggest that environmental
activities under India’s National Rural Employment Guarantee Act could be particularly
beneficial in creating jobs for the unskilled. In the current circumstances of the Indian
economy, that is desirable for reasons of equity and poverty reduction even if it does not
increase the productivity of the employed labor force. But Schwartz et al (2009), in their
study of Peru, Brazil and Honduras, draw attention to the rather skills-intensive nature of the
projects undertaken as part of those countries’ fiscal stimuli, giving rise to a concern that a
more aggressive stimulus could run up against skills bottlenecks. GCN (2010) acknowledges
the need in South Africa to train workers in new skills if the renewables sector is promoted
as part of an aggressive green strategy and notes the risk that a lot of the new
manufacturing and construction jobs associated with expansion of renewable energy would
30
be located outside of South Africa, given its limited manufacturing base and availability of
appropriate skills.
The limited evidence in this area suggests that making green jobs decent is likely to require
greater outreach by training organizations and educators. The increasing appreciation of the
need to tackle environmental externalities of economic activity, especially climate change,
draws attention to pre-existing market failures in the provision of skills as well as in
stimulating innovation and satisfying infrastructure needs. Many of the skill shortages
already reported in connection with green growth strategies appear to result from generic
failings in education and training and reflect long-standing issues such as the lack of
incentives for employers to invest in developing the transferable skills of their workforces,
the lack of access to time and finance for training on the part of the disadvantaged and the
stickiness of relative pay rates. In many countries, public employment programs are likely
to be part of the solution (Lieuw-Kie-Song and Lal, 2010). Box 3 presents the special
measures that the ILO and CEDEFOP argue are necessary for labor markets to respond
effectively to green growth policies.
Box 3: Measures necessary to equip developing countries with the
skills for green growth
ILO and CEDEFOP (2011) argue that developing countries need special measures to equip
them with the skills for the transition to green growth:
(i) Capacity building for employers in the informal economy and micro and small
enterprises to enter green markets in localities where they are most needed
(ii) Entrepreneurship training and business coaching for young people and adults to start up
green businesses in conjunction with microfinance projects
(iii) Environmental awareness among decision-makers, business leaders and administrators
as well as institutions of formal and non-formal training systems
(iv) Capacity-building of tripartite constituents to strengthen social dialogue mechanisms
and to apply these to dialogue about accessibility to training for green jobs
(v) Increased capacity of formal education and training systems and institutions to provide
basic skills for all and to raise the skills base of the national workforce; this includes
improving apprenticeship systems and building synergies with NGOs that provide
education and training.
31
7. Conclusions
The term ‘green jobs’ means different things to different people. One serviceable definition
equates it with the employment in a narrowly defined set of industries providing
environmental services. Quantitatively, these industries amount to a small share of total
employment – of the order of 1%, but varying according to definition. This type of definition
provides a useful basis for benchmarking countries’ efforts in the environmental policy arena
and for understanding these industries better. But these objectives have not been at the
heart of the literature on green jobs and there is scant information on developing countries’
employment in these industries. Instead, studies have tended to focus on the employment
consequences of introducing public policies to correct environmental externalities,
especially anthropogenic climate change. This is more interesting from the policy-maker’s
perspective as it brings labor market behavior into the calculus when assessing the merits of
alternative courses of action against standard yardsticks of appraisal (e.g. measures of
productivity, well-being and equity). It is also in principle a broader approach, potentially
covering jobs created and destroyed across the whole economy – although many studies
take a much narrower approach in practice.
Renewable energy supply, an increase in which is a key component of global and local
climate-change policies, appears to be likely to be more labor intensive than traditional
fossil-fuel-based supply, per megawatt and per dollar, although there are questions about
how labor requirements are likely to evolve over a plant’s life and as technological progress
takes place. But it is not likely to be more labor intensive than several other activities on
which governments could spend in order to generate social benefits. Energy efficiency
improvements also appear to be labor-intensive, drawing heavily on relatively unskilled
labor in the construction sector. There are many attractive opportunities for developing
countries in both these areas. However, the implications of the current lower labor
productivity of these activities for public finances, aggregate productivity, energy prices and
the profitability of private-sector activity are not generally examined and need to be. The
32
fact that employment has a social cost (reflected in the economist’s concept of a shadow
price of labor) should be borne in mind.
There are important differences across types of economy due to different industry
structures, labor market institutions and endowments. The challenges for some developing
countries posed by the transition to green growth are particularly difficult and the prospects
for their labor markets more daunting, given the structural changes that will be needed.
Among these are the countries that at the moment have a comparative advantage in fossil-
fuel production, particularly those that have built their industrial development strategies
around cheap carbon-based energy.
Less attention has been paid to the labor market consequences of other environmental
policies, such as support for biofuels and low-carbon land use. These are likely to be
relatively much more important in low-income countries than in high-income ones. Some
evidence suggests that they are likely to be more labor intensive than renewable energy
supply too. More research is needed into their labor market implications, including the
possible impacts on rural-urban migration.
In practice, most of the literature has focused on direct employment created, with more
cursory or no treatment of indirect and induced jobs creation, especially that arising from
macroeconomic effects of policies. The perspective taken is often that of project-based
cost-benefit analysis where the project can be assumed to be too small to affect market
prices and labor supply can be assumed to be highly elastic. The costs and benefits are often
not fully enumerated either. This seems a major drawback in the ’green jobs’ debate, given
that nations are being encouraged to engineer non-marginal changes in the structures of
their economies, particularly those connected with energy supply and demand. A greater
emphasis on ‘top down’ modeling that incorporates important features of the macro-
economy is needed. This can be combined with the wealth of detailed microeconomic
information provided by ‘bottom up’ modeling from an engineering or project-based
perspective. Given the focus on environmental policies in the ‘green jobs’ literature, ‘top
down’ models that allow for the possibility of market failures are required. And given the
very distant time horizon for climate-change policies, examining the implications of
alternative theories of long-run growth is also desirable. Thus there should be synergies
33
with research on the sources of long-run growth in developing countries and the debate
about per capita income convergence.
Most important, more attention needs to be paid to how labor markets work at the
macroeconomic level. It is not surprising that this should be central to the determination of
the employment consequences of any given set of economy-wide policies. Several studies
have implicitly or explicitly used a framework of Keynesian involuntary unemployment. This
is not necessarily inappropriate for developing countries given their susceptibility to
business cycle shocks. But further thought about what modern macroeconomic theory
implies for the workings of labor markets is warranted. If, for example, involuntary
unemployment is the result of rigid real consumption wages, climate-change policies, by
increasing the price of carbon-intensive goods and services, could end up increasing
unemployment, notwithstanding the labor intensity of new investment projects.
Another aspect of the functioning of labor markets that needs to be taken into account is
human capital formation. There are generic problems afflicting the provision of skills, which
take on greater significance when policymakers are trying to induce a pervasive structural
change in the economy. Already skill shortages are being reported in industries and
occupations likely to benefit from green policies. Perhaps the most important of these are
the high-level skills necessary to manage large-scale green policy interventions and the
associated large-scale projects over a long period of time in a way that will build the
credibility of green growth aspirations while allowing for learning and policy improvement
over time. This is a big challenge for all countries but particularly for those in which
sustained constructive collective action has proved difficult in the past.
For developing countries, it is also desirable to consider the implications of labor market
models tailored to their particular circumstances – for example, surplus labor models and
multi-sector models where each sector’s labor market behaves differently. That may make
environmental policies look even more attractive than they do in developed countries,
where in some sense labor is relatively scarcer. It may also point to ways in which
environmental policies have different consequences in different sectors. Thus the general
equilibrium consequences of a given amount of direct job creation in a highly regulated
energy utility may be completely different from the consequences of the same amount of
direct job creation in rural land management.
34
What does this imply for governments in developing countries contemplating introducing
new environmental policies? Strand and Toman (2010) conclude that there are few obvious
candidates for triple-win policies that deliver simultaneously strong benefits for short-term
economic recovery, longer-term growth, and long-term environmental benefits.
4
Nevertheless, there is a case for opportunistically speeding up appropriate social projects
(environmental or otherwise) when resource constraints are relaxed, for example, because
of an adverse shock to private aggregate demand (Bowen and Stern, 2010). Such a shock
can increase the wedge between the shadow wage in an economy and the actual wage,
warranting more projects being undertaken by the public sector. However, it need not
necessarily do so. If the projects are difficult to finance because of constraints on
governments’ borrowing or tax-raising capacity, the shadow wage may not be lower than
the actual wage. The more labor-intensive green projects will naturally constitute a larger
share of the temporary incremental spending. But low-labor-productivity projects should
not be adopted unless they are profitable at the relevant shadow prices (including not only
the wage but the carbon price and the optimal subsidies for learning and R&D). The policy-
maker should not rank projects by employment per dollar spent, choosing simply those with
the highest rank. Such an approach risks policy-makers forgetting that employment costs
have a social cost associated with them.
More generally, there may be persistent wedges between the shadow wage and the actual
wage, particularly in developing countries with segmented labor markets, warranting a
bigger and longer-lasting boost to green projects with high labor content (even though the
shadow wage is positive). Whether that would lead to a larger rise in green spending than
other types of spending (e.g. health and education) is difficult to determine without more
detailed comparisons of technologies across a wider range of sectors. And there is a danger
that, without appropriate environmental pricing, counter-cyclical spending to increase
employment could lock in carbon-intensive technologies (e.g. spending on road-building and
fossil-fuel extraction). But the importance of climate change mitigation and the slow
development of low-carbon and other environmental policies up until now, together with
the relative labor intensity of at least some green investment, suggest that there are likely to
be many opportunities to make investments that have both environmental and labor market
4
Strand and Toman also invoke Tinbergen’s Principle, that if a government has n objectives it needs n policy instruments to
achieve them efficiently. At the same time, the optimal setting of an instrument is not necessarily independent of changes in
other instrument settings. Thus a change in the wedge between the shadow wage and the actual wage might, for example,
entail changes in carbon pricing or public infrastructure spending.
35
pay-offs. A policy-induced structural change of this sort, just like an exogenous structural
change (induced for example by globalization), requires attention to active labor market
policies with respect to training, education and transitional income support.
References
Alvarez, GC, Jara, RM, Juliá, JRR, and JIG Bielsa (2010): Study of the Effects on Employment
of Public Aid to Renewable Energy Sources, Procesos De Mercado. Vol. VII, No. 1,
Primavera.
Australian National Audit Office (2010): Report on the Home Insulation Program, Auditor
General Report No. 12, 2010-11, Canberra.
Babiker, MH, and RS Eckaus (2007): Unemployment effects of climate policy, Environmental
Science & Policy, Vol. 10, pages 600-609.
Barbier, EB (2009): A Global Green New Deal, United Nations Environment Programme,
Geneva.
Behrman, J (1999): Labor markets in developing countries, Chapter 43 in Ashenfelter, J, and
D Card (ed.s) (1999), Handbook of Labor Economics, Volume 3, pages 2859-2939.
Bosetti, V, Carraro, C, Galeotti, M, Massetti, E, and M Tavoni (2006): WITCH: A World
Induced Technical Change Hybrid Model, The Energy Journal, Special Issue on Hybrid
Modeling of Energy-Environment Policies; Reconciling Bottom-up and Top-down,
pages 13-37.
Bowen, A, and N Stern (2010): Environmental policy and the economic downturn, Oxford
Review of Economic Policy, Vol. 26, No. 2, pp.137-163.
CEDEFOP (2009): Future skill needs for the green economy, European Centre for the
Development of Vocational Training, Publications Office of the European Union,
Luxembourg.
Collier, P (2007): The bottom billion, Oxford University Press, Oxford.
EBRD (2011): The Low Carbon Transition Report, EBRD, London, April.
EC (2007): Facts and Figures: Links Between EU’s Economy and Environment, Office for
Official Publications of the European Communities, Luxembourg.
Dutt, AK (2007): Heterodox Development Economics and the Role of Aggregate Demand,
Prepared for presentation at Glasgow University’s International Conference on
Development Economics, Policies and institutions for growth and development: Is
there a ‘heterodox’ view? 26-27 June 2007, Glasgow, UK.
36
Fankhauser, S, Sehlleier, F, and N Stern (2008): Climate change, innovation and jobs, Climate
Policy, Vol. 8, pages 421-429.
Fields, GS (1975): Rural-urban migration, urban unemployment and underemployment, and
job search activity in LACS, Journal of Development Economics, Vol. 2, pages 165-187
Fullerton, D, and G Heutel (2007): Who bears the burden of a tax on carbon emissions in
Japan? Environmental Economics and Policy Studies, Vol. 8, pages 225-270.
Fullerton, D, and G Heutel (2010): Analytical General Equilibrium Effects of Energy Policy on
Output and Factor Prices, NBER Working Paper No. 15788, February.
Fullerton, D, and GE Metcalf (1997): Environmental taxes and the double-dividend
hypothesis: did you really expect something for nothing? NBER Working Paper No.
6199, September.
GCN (2010): Low-carbon Jobs in an Interconnected World, Global Carbon Network
discussion paper No. 3 and literature review, March.
GHK (2009): The Impacts of Climate Change on European Employment and Skills in the Short
to Medium-Term: A Review of the Literature, Final Report to the European
Commission Directorate for Employment, Social Affairs and Inclusion Restructuring
Forum, Vol. 2, May.
Goettle, RJ, and AA Fawcett (2009): The structural effects of cap and trade climate policy,
Energy Economics, Vol. 31, Special Supplement 2, pages 244-253.
Greenpeace/EREC (2011): The Advanced energy [R]evolution: a Sustainable Energy Outlook
for South Africa, Greenpeace and the European Renewable Energy Council, May.
Harris, JR, and MP Todaro (1970): Migration, unemployment, and development, American
Economic Review, Vol. 60, pages 126-142.
Hatfield-Dodds, S, Turner, G, Schandl H, and T Doss (2008): Growing the green collar
economy: Skills and labour challenges in reducing our greenhouse emissions and
national environmental footprint. Report to the Dusseldorp Skills Forum, CSIRO
Sustainable Ecosystems, Canberra, June.
Hughes, G. (2011): The myth of green jobs, The Global Warming Foundation Report No.3,
London.
IEA (2008): World Energy Outlook 2008, International Energy Agency, Paris.
ILO and CEDEFOP (2011): Skills for green jobs: a global view, ILO, Geneva.
Kammen, DM, Kapadia, K, and M Fripp (2004): Putting renewables to work: how many jobs
can the clean energy industry generate? Report of the Renewable and Appropriate
Energy Laboratory, University of California, Berkeley, April (corrected January 2006).
37
Kojima, M, and T Johnson (2005): Potential for biofuels for transport in developing countries,
Energy Sector Management Assistance Programme, The World Bank, Washington
DC, October.
Kraay, A, and L Servén (2008): Fiscal Policy Responses to the Current Financial Crisis: Issues
for Developing Countries, Research Brief, World Bank, Washington DC, December.
Kraay, A, and J Ventura (1998): Comparative advantage and the cross-section of business
cycles, Policy Research Working Paper 1948, World Bank, Washington DC, July.
Lewis, WA (1954): Economic development with unlimited supplies of labour, Manchester
School of Economic and Social Studies Vol. 22, pages 139-191.
Lieuw-Kie-Song, M, and R Lal (2010): Green jobs for the poor: why a public employment
approach is needed now, International Policy Centre for Inclusive Growth, Report
No. 105, Brasilia, February.
McKinsey & Company (2009): Pathways to a low-carbon economy, January.
Mazumdar, D (1976): The Rural-Urban Wage Gap, Migration, and the Shadow Wage, Oxford
Economic Papers, Vol. 28(3), pages 406-25.
Michaels, R, and RP Murphy (2009): Green jobs: fact or fiction? Institute for Energy
Research, Houston, Texas, January.
Morriss, AP, Bogart, WT, Dorchak, A, and RE Meiners (2009): Green jobs myths, University of
IllinoisLaw & Economics Research Paper No. LE09-001 and Case Western Reserve
University Research Paper Series No. 09-15, March.
OECD (1999): The Environmental Goods and Services Industry: Manual for Data Collection
and Analysis, OECD, Paris.
OECD (2011): Towards green growth, OECD, Paris, May.
Otero, JG (2000): Coffee, economic fluctuations and stabilisation: an intertemporal
disequilibrium model with capital market imperfections, Journal of Development
Economics, Vol. 62, pages 105–129.
Perez, C (2002): Technological Revolutions and Financial Capital: The Dynamics of Bubbles
and Golden Ages, Edward Elgar, UK.
Pew (2009): The Clean Energy Economy: Repowering Jobs, Businesses and Investments in
America, the Pew Charitable Trusts, Washington D.c. and Philadelphia, June.
Pissarides, CA (2009): Labour market adjustment, Cambridge University Press, Cambridge,
UK.
Pollin, R (2009): Response to “Seven Myths about Green Jobs” and “Green Jobs Myths”, PERI
Working Paper 198, University of Massachusetts, Amherst, USA.
38
Pollin, R, Garrett-Peltier, H, Heintz, J, and H Scharber (2008): Green recovery: a program to
create good jobs and start building a low-carbon economy, PERI and the Center for
American Progress, September.
Pollin, R, Heintz, J, and H Garrett-Peltier (2009): The economic benefits of investing in clean
energy PERI and the Center for American Progress, June.
Rutovitz, J (2010): South African energy sector jobs to 2030, prepared for Greenpeace Africa
by the Institute for Sustainable Futures, University of Technology, Sydney, Australia.
Rutovitz, J, and A Atherton (2009): Energy sector jobs to 2030: a global analysis, prepared
for Greenpeace International by the Institute for Sustainable Futures, University of
Technology, Sydney, Australia.
Sartzetakis, EF, and PD Tsigaris (2007): Uncertainty and the double dividend hypothesis,
Nota di Lavoro 99.2007, Fondazione Eni Enrico Mattei, November.
Satchi, M, and J Temple (2009): Labor markets and productivity in developing countries,
Review of Economic Dynamics, Vol. 12, pages 183–204.
Schwartz, JZ, Andres, LA, and G Draboiu (2009): Crisis in Latin America. Infrastructure
Investment, Employment and the Expectations of Stimulus, Policy Research Working
Paper WPS 5009, World Bank, Washington DC.
Stern, N (2010): China’s growth, China’s cities, and the new global low-carbon industrial
revolution, Policy Paper, Grantham Research Institute on Climate Change and the
Environment, LSE, November.
Strand, J, and M Toman (2010): “Green Stimulus,” Economic Recovery, and Long-Term
Sustainable Development, World Bank Policy Research Working Paper No. 5163,
Washington DC, January.
UNEP (2011): Towards a green economy: pathways to sustainable development and poverty
eradication - a synthesis for policy makers, UNEP, Geneva.
UNEP/ILO/IOE/ITUC (2008): Green Jobs: Towards a Decent Work in a Sustainable, Low-
Carbon World, UNEP/ILO/IOE/ITUC, Geneva.
UNFCCC (2007): Investment and financial flows to address climate change, United Nations
Framework Convention on Climate Change, Bonn.
Upadhyay, H, and N Pahuja (2010): Low-Carbon Employment Potential in India: A Climate of
Opportunities, Centre for Global Climate Research TERI and Global Climate
Framework Discussion Paper TERI/GCN – 2010:1, New Delhi.
World Bank (2009): World Development Report 2010, World Bank, Washington DC.
World Bank (undated): Issues in estimating the employment generated by energy sector
39
activities, Background paper for the World Bank Group energy sector strategy,
World Bank, Washington DC.
