Download Big push or big failure? On the effectiveness of industrialization

J. Japanese Int. Economies 26 (2012) 129–141
Contents lists available at SciVerse ScienceDirect
Journal of The Japanese and
International Economies
journal homepage: www.elsevier.com/locate/jjie
Big push or big failure? On the effectiveness
of industrialization policies for economic development q
Kjetil Bjorvatn a,⇑, Nicola Daniele Coniglio b
a
b
NHH Norwegian School of Economics, Helleveien 30, 5045 Bergen, Norway
University of Bari, Piazza Umberto I, 1-70121 Bari, Italy
a r t i c l e
i n f o
Article history:
Received 13 June 2010
Revised 16 July 2011
Available online 12 August 2011
JEL classification:
O14
O20
O33
O38
Keywords:
Industrialization
Big push
State owned enterprises
Government failure
Patronage employment
a b s t r a c t
Bjorvatn, Kjetil, and Coniglio, Nicola Daniele—Big push or big
failure? On the effectiveness of industrialization policies for economic development
The role of the government in industrialization is heavily debated.
Some claim that extensive government involvement is key to initiate a sustainable development process, others see the government
as an obstacle to it, pointing to the importance of government failure. We formulate a model, which explains why even a highly inefficient industrial policy can successfully promote big-push
development. Moreover, we show that extensive government
intervention is more likely to be successful when the initial level
of development is low. J. Japanese Int. Economies 26 (1) (2012)
129–141. NHH Norwegian School of Economics, Helleveien 30,
5045 Bergen, Norway; University of Bari, Piazza Umberto I, 170121 Bari, Italy.
Ó 2011 Published by Elsevier Inc.
1. Introduction
What role should the government play in the process of industrialization? Active state intervention
appears to have been an important ingredient in the development of many of today’s rich countries,
including the Western European economies, Japan, as well as the first generation of the East Asian
Tiger economies. Rodrik (1995) argues that a large component of the ‘‘economic miracles’’ of Taiwan
and Singapore should be ascribed to the active role of the governments in implementing big-push
q
We would like to thank an anonymous referee for extremely useful suggestions. The usual disclaimer applies.
⇑ Corresponding author.
E-mail addresses: [email protected] (K. Bjorvatn), [email protected] (N.D. Coniglio).
0889-1583/$ - see front matter Ó 2011 Published by Elsevier Inc.
doi:10.1016/j.jjie.2011.07.005
130
K. Bjorvatn, N.D. Coniglio / J. Japanese Int. Economies 26 (2012) 129–141
policies aimed at removing coordination failures in investment.1 Cassen and Lall, 1996, in a critical
comment of the East Asian Miracles study by the World Bank, argue that selective industrialization policies, and not just export promotion, were key to the rapid economic growth experienced in the region in
the 1970s and 1980s, and moreover that there are important lessons to be learned for other countries
from this experience. Similarly, Kohli (2004) analyzes the growth experience of Korea, Brazil, India
and Nigeria. He emphasises the key role that State intervention has played in promoting industrialization
in the least developed countries (LDCs) by supporting the profitability of private investments in early
phases of economic development. However, it is fair to say that the focus amongst economists and donors in recent decades has shifted from the importance of market failure to that of ‘‘government failure’’ .
Indeed, the World Bank and the IMF have emphasized the need for privatization and deregulation as conditions for their continued support of developing countries.2 Consequently, we have witnessed a trend of
overall reduction in the relative (and often absolute) weight of the State in the economies of many developing countries. For instance, the share of public investment over total GDP fell in developing countries
from approximately 10% in the early 1980s to 5% in 2000 (Chang, 2007).
Several observers have raised concerns about the potentially negative effects of reduced State
intervention on economic growth in developing countries. For instance, Stiglitz (2002, p. 55) argues
that the IMF simply assumed that markets arise quickly to meet every need, when in fact, many government activities arise because markets have failed to provide essential services. Similarly, Stein
(1992, p. 83) claims that the World Bank/IMF model is likely to de-industrialize the existing manufacturing base without encouraging any significant replacement. Clearly, the presence of government
failure is not by itself a justification for reduced government intervention.
A quick look at the data lends support to these concerns. Fig. 1 illustrates the change in government
intervention from the 1980s to the 1990s (where a positive value means less intervention) and the
average GDP per capita growth rate in these two decades for 81 developing countries.3
Evidently, a reduction in the role of government (i.e., a positive change in the Intervention Index in
the period) is negatively associated with economic performance (the slope of the curve is 0.21 and
statistically significant at 10% level, p-value = 0.06.). A closer inspection of the data shows that the
negative correlation is driven mainly by developments in the 1990s.
Although this quick look at the data provides no conclusive evidence about the (causal) relationship between government intervention and economic growth, the simple correlation at least suggests
that developing countries that liberalized their industrial policies did not necessarily experience more
rapid economic growth. Fig. 1 is thus broadly consistent with concerns raised by Stiglitz (2002) on the
risks of removing government intervention in order to let the markets do the job more efficiently. The
data presented above are also consistent with more in depth empirical research on today’s developing
countries and on the long-term development of today’s developed countries. For instance, Atukeren
(2005) finds that the likelihood that public investments (including those carried out through stateowned enterprises) crowd out private investment is lower in relatively less stable and less developed
economies. Evidence of a non-linear effects of government intervention on growth also emerges from
the study by Grossman (1988), who studies the development of the US economy in the period 1929–
1982. The author finds that the effect of government spending on growth is initially positive but that
‘‘the positive contributions of an increasing size of government are virtually offset by the inefficiencies
created by the provision process’’ (Grossman, 1988, p. 199).4
1
See also Wade and Robert (1990) and Greenwald and Stiglitz (2006). For a critical survey, see Pack and Saggi (2006).
For a critical assessment of the so-called ‘‘Washington consensus’’ policies in developing countries, see Gore (2000) and Rodrik
(2006).
3
As measure of government intervention we consider the Index of Government Enterprises and Investment (which we shall
simply refer to as the Intervention Index) from the Economic Freedom Index developed by Gwartney and Lawson (2009). The
Intervention index ranks countries from zero (‘‘The economy is dominated by SOEs and government investment exceeds 50% of
total investment’’) to 10 (‘‘Only few SOEs operate and government investment is generally less than 15% of total investment’’). See
the Appendix A for a further description of the Intervention Index and Appendix B for the intervention and growth data for the
countries in our sample.
4
For a survey of the relationship between the size of the government, and specific areas of government intervention, such as
education and infrastructure, and long-run economic performance, see Poot (2000).
2
K. Bjorvatn, N.D. Coniglio / J. Japanese Int. Economies 26 (2012) 129–141
131
Fig. 1. Government intervention and growth.
We take these empirical observations as a starting point for our theoretical analysis of industrial
policy in developing countries. The ambition of the present paper is to demonstrate how even highly
imperfect government policies can play a positive role in the industrialization process. Our analysis is
based on the seminal contribution by Murphy et al. (1989), which, in turn, was inspired by the original
thoughts of Rosenstein-Rodan (1943, 1961). Murphy et al. (1989) focus on demand side externalities,
and demonstrate that, in a closed economy, the government may lead the economy out of a low-tech,
low-income trap, by stimulating investment in high-tech firms. Rodrik (1996) presents a big-push
model focusing on supply side externalities. In his setting, intermediate inputs are non-traded, and
the production of a high-tech final good requires the availability of a critical number of varieties of
intermediate inputs. Rodrik demonstrates that a country may be caught in a low-income trap, with
no intermediate input production and hence no high-tech final goods production. A big-push policy
that stimulates the establishment of a critical mass of intermediate inputs, may lead the economy
to a high-wage, high-tech equilibrium. Trindade (2005) presents a big-push model in an open economy setting, focusing on the role of trade policy in the industrialization process. He shows that by supporting exports of intermediate goods, a developing country can attract the investment of a final goods
producer, which in turn increases the profitability of intermediates production in this country. Gans
(1998) describes a situation with coordination failure in investment. In a model with vertical linkages,
he derives the critical number of upstream investment needed to guide the economy towards the
high-income equilibrium. Da Rin and Hellmann (2002) discuss the role that the banking system has
played in promoting industrialization in many countries by coordinating investments. The authors
present a big-push model where the role of banks in inducing the economy to move out of a poverty
trap is analyzed.
The main difference between these papers and the one we present here, is that we focus on government failure in the policy intervention. Government failure arises due to pressures for patronage
employment in firms supported by or owned by the government, leading to excessive costs and thus
lower profits. However, our main goal is to demonstrate that despite their inherent inefficiency, such
policy interventions may still play a key role in an industrialization process. Hence, while in isolation it
may be tempting to conclude that, in order to reduce economic inefficiency due to government failure,
privatization and liberalization should be carried out, our analysis shows that this policy advice may in
fact be misguided.
Moreover, we analyze the conditions under which state-led development is likely to contribute to
economic efficiency, and when it is likely to fail to do so. We show that an extensive government
132
K. Bjorvatn, N.D. Coniglio / J. Japanese Int. Economies 26 (2012) 129–141
intervention can be justified when the level of development is very low. Intuitively, the economy in
this case needs a big push to lift it from the low-income trap. In contrast, a more developed economy
needs a smaller push: Indeed, extensive intervention in this environment would probably ‘‘overshoot’’
relative to the efficient policy, crowding out more productive private firms and thus depressing aggregate income. The remainder of the paper is organized as follows. Section 2 outlines the theoretical
model. The analysis is reported in Section 3, while Section 4 concludes.
2. Model
Along the lines of Murphy et al. (1989), we consider a closed economy with a single factor of production, labor.5 There is a large number of goods in the economy. To simplify notation, the mass of both
labor and goods are given by unity. Goods can be produced by means of two technologies, traditional and
modern. Traditional production is carried out by small-scale entrepreneurs in the informal sector. The
government may support the establishment of modern production through industrial policies. Here,
these policies take the form of subsidies to cover the fixed costs of production. For concreteness, we shall
refer to modern firms that receive government support as ‘‘state-owned enterprises’’ (SOEs), and modern
firms that do not receive this support as ‘‘private firms’’ . While private owners maximize profits, SOEs
are assumed to be unprofitable. In particular, SOEs that would have been profitable in private hands are
forced to hire labor in excess of the profit maximizing level, thus crowding out any profits. Similarly,
SOEs that are technically unprofitable (which, as will be clarified below, amounts to operating profits
not being able to cover the minimal level of fixed costs), receive government support, allowing them
to break even. We can think of the difference between private and state owned enterprises as caused
by political pressure on governments from key interest groups, seeking employment in SOEs for their
members. In this sense, the government is weak and subject to regulatory capture: It does not have full
autonomy in the definition and implementation of industrialization policies, for instance because these
decisions typically involve multiple levels of government. We can refer to this institutional framework as
one characterized by an ‘‘incompetent benevolent government’’.6
The traditional technology is constant returns to scale, with the following labor input requirement:
l
crs
¼ xcrs :
ð1Þ
Alternatively, the same goods can be produced by means of modern technology, which exhibits
increasing returns to scale. The labor input requirement of a firm i = p, s (private and state-owned,
respectively) in this case can be described as follows:
irs
li ¼ f i þ
xirs
;
b
ð2Þ
where fi is the fixed cost of production and where b > 1 represents the productivity markup in modern
production over traditional production. We can think of fi as ‘‘administration’’ . Private firms only employ administrative staff up to the minimum level required for production, fp = f. State-owned firms, in
contrast, are subject to political pressure to hire people in the administration beyond the efficient level, hence fs > f. The level of fs is endogenously determined in the model, given by the zero-profit constraint, and will be derived later. Note that the fact that excess labor in state-owned enterprises is
employed in the administration implies that these workers are entirely unproductive. This highlights
the underlying assumption that these workers are employed for patronage purposes rather than
production.7
5
The closed economy assumption can be justified by reference to high fixed costs of entering international markets, as
documented by for instance Das et al. (2007) using data from Colombia. For less developed countries in particular, the relevant
market for the majority of firms is the local market. However, big-push models can be framed also in an open economy setting (see
eg Trindade, 2005). Indeed, the key results of our model do not rely on the closed economy assumption.
6
We would like to thank an anonymous referee for suggesting this characterization.
7
We do not consider the case of loss-making SOEs. In other words we limit the degree of inefficiency, in terms of ‘‘excessive
size’’ , of a single SOE. This assumption does not alter the qualitative results of the analysis.
K. Bjorvatn, N.D. Coniglio / J. Japanese Int. Economies 26 (2012) 129–141
133
We shall interpret f as a proxy for the level of development of the economy, determined by the
quality of infrastructure, the level of education, research and development, etc. The more advanced
is the economy, the lower is f, and hence the lower are the barriers to profitable production, ceteris
paribus. In our analysis of industrialization policies, we shall focus on how the effectiveness of these
policies depends on the level of f and hence the level of development.
We follow the standard assumption in big-push models by assuming that workers in the modern
sector are paid a markup w > 1 over the wage in the traditional sector.8 Traditional sector workers, in
turn, receive a wage given by the value of their marginal product. Normalizing the supply price from the
traditional sector to unity, the traditional wage is wcrs = 1. Using (2), production costs using modern technology are therefore given by:
xirs
irs
cirs
¼
wl
¼
w
f
þ
;
i
i
i
b
ð3Þ
The fixed costs in modern production are assumed to be sufficiently high to allow at most one domestic producer of a manufactured good. When in operation, the profit maximizing strategy of a modern
firm is limit pricing, i.e., choosing a price (marginally below) traditional producers’ supply price of
unity. In this way, a modern firm captures the entire market for its product. Its monopoly power is,
however, limited by the potential entry of small scale producers. Profits by a modern sector firm
can be expressed as:
pi ¼
yðb wÞ
wfi ;
b
ð4Þ
where y is the aggregate income in the economy. It is assumed that consumers share their budget
equally across all the unit-mass of goods produced in the economy. We start out considering an economy using only traditional technology. Due to a coordination failure in investment, the economy is
caught in a low-income trap. We then address the issue of industrialization through government
intervention.
2.1. Coordination failure
The initial state of the economy is one involving only traditional production. Let’s ignore for the
moment any government intervention. When making an investment decision, a private investor takes
the income level in the economy as given. Thus, he does not internalize the effect of his own investment on aggregate income. Neither does he internalize the effect of his own investment on the investment decisions of other entrepreneurs. Hence, there will be a difference between perceived profits and
realized profits, and this is a source of inefficiency in the investment choices. Since in a traditional
economy y0 = 1, perceived profits can be expressed as:
pp0 ¼
ðb wÞ
wf :
b
ð5Þ
The condition that investment is unprofitable in this initial state can be expressed as:
pp0 < 0 ) f >
bw
:
bw
ð6Þ
In the following we assume that this inequality holds. This is a necessary condition for efficiency
improving government intervention. The second necessary condition for government intervention
to play a positive role is that income under full industrialization is sufficient to make each private firm
profitable. With full industrialization, y = w + pp, and equilibrium profits can be found from (4) as:
pp1 ¼ bð1 f Þ w;
ð7Þ
8
As shown in Murphy et al. (1989), a wage markup in the modern sector is necessary to create the demand externality
underlying the coordination failure.
134
K. Bjorvatn, N.D. Coniglio / J. Japanese Int. Economies 26 (2012) 129–141
which inserted in income identity yields total income
y1 ¼ bð1 f Þ:
ð8Þ
We observe that
pp1 > 0 ) f <
bw
;
b
ð9Þ
which we assume holds. Combining (6) and (9) we can conclude that our basic assumption of the existence of a coordination failure is based on the following condition:
bw
bw
<f <
:
bw
b
ð10Þ
We now turn to the question of government intervention.
2.2. Government intervention
In the framework outlined above, modernization in some sectors increases demand for all the varieties produced in the economy. Demand externalities in manufacturing constitute a rationale for policies of industrialization. Such policies can of course be of various kinds, but for concreteness we focus
on the establishment of SOEs. However, the qualitative results of our analysis are naturally not limited
to this specific type of intervention.
As discussed above, SOEs are inherently inefficient, in the sense that they never generate any profit.
We argued that this could be due to political pressure from special interest groups to find employment
for their members in the SOEs. Overstaffing among SOEs in developing countries is a common phenomenon. For instance, World Bank (1995) reports that: ‘‘In Bangladesh in 1992, the state sugar milling monopoly had twice as many office workers as it needed, or about 8000 extra employees.’’ (page
14). In general, this World Bank report claims that up to 90% of the employees may not be needed in
some SOEs (page 35).
In the same way that a government may face pressure for job creation in any given SOE, it may also
face pressue to create new SOEs. Hence, we consider a government with limited autonomy in the planning and implementation of its industrialization policy. Accordingly, lack of autonomy may lead to
excessive intervention, while lack of financial capacity may lead to a sub-optimal level of intervention.
In this section, we first define the zero profit condition for the SOE, and then analyze the potential for
efficiency improving industrial policy based on this sub-optimal policy tool.
From (4) we find that the zero-profit condition p p = 0 as:
f0 ¼
yðb wÞ
:
bw
ð11Þ
For f0 < f, the gap between the fixed costs necessary to start up modern production and the level of
fixed costs that would lead to break-even, i.e., w(f f0), is covered by subsidies (conditioned, of course,
upon availability of government funding). In this case, there is no room for patronage driven excess
employment, since the non-negative profit constraint is binding. When the level of income (y) is high
relative to the minimal level of fixed costs (f), such that f0 > f, the situation changes. Now, the SOEs are
in principle making positive profits, but due to patronage pressure, the profits are crowed out by excess employment. Hence, fixed costs for the SOE are given by:
fs ¼
f0
for f 0 > f ;
f
for f 0 6 f :
ð12Þ
Total income in an economy where n SOEs have been established, and where no private firms operate,
can be written as:
irs
1 þ nðw 1Þ f þ xb nwðf f0 Þ for f 0 6 f ;
yn ¼
irs
1 þ nðw 1Þ f0 þ xb
for f 0 > f :
ð13Þ
K. Bjorvatn, N.D. Coniglio / J. Japanese Int. Economies 26 (2012) 129–141
135
In both income expressions, the first term is labor income of unity, and the second is the wage markup
irs
(w 1)for all workers employed by the n SOEs, namely n times ls . For f0 6 f, the minimum administrative requirement f is binding, and the third term is the transfer necessary to allow the SOE to break
even. For f0 > f, subsidies are not necessary, but the administrative input now exceeds the minimum,
crowding out all profits. Using the fact that xirs = y and f0 from (11), we find the equilibrium level of
income in an economy with n SOEs as:
yn ¼
bð1fnÞ
bnðb1Þ
for f 0 6 f ;
w
wnðw1Þ
for f 0 > f ;
ð14Þ
where it is clear that yn > 1 for n > 0. We observe that:
Proposition 1. The establishment of large-scale, but unprofitable, state-owned enterprises increases
income in a traditional economy.
Proof. First, for f0 6 f ;
@yn
@n
ðbð1f Þ1Þ
¼ ðbbnðb1Þ
> 0 if b(1 f) P 1, which must hold, given the condition (9) and
Þ2
@y
since w > 1. Second, for f0 > f ; @nn ¼ wðw 1Þ=½w nðw 1Þ2 > 0 since w > 1. h
Aggregate income increases in n despite the inherent inefficiency of this policy intervention. This is
due to the positive demand externalities; while the SOEs themselves do not generate any profits, by
paying the wage markup to their employees, they do contribute positively to overall income. Clearly,
as long as f0 6 f, firms depend on subsidies to cover their costs, and hence private investment is not
profitable. For f0 > f, on the other hand, private investment will take place. Using y = yn in (4), given
f0 > f, we find the profitability of private investment, as a function of the number of SOEs (n), as:
pp ðnÞ ¼
wðb wÞ
wf :
bðw nðw 1ÞÞ
ð15Þ
The critical number of SOEs required to trigger private investment is given by pp(n) = 0, resulting in:
n¼
wðf b þ 1Þ b
n :
f bðw 1Þ
ð16Þ
Given (10), we observe that n⁄ 2 (0, 1). This is the optimal push, in the sense that n⁄ defines the minimum effort required to induce private investment in modern-technology manufacturing. Note in particular that:
Proposition 2. The critical ‘‘push’’ required to take the economy to the high-income equilibrium is larger
the less developed is the economy.
bw
Proof. This is clear from the fact that @n
¼ f 2 bðw1Þ
> 0, since pi > 0 M b > w. h
@f
Hence, the more advanced the economy, the smaller is the push, in the form of establishing SOEs,
required to achieve the goal. Ideally, once n⁄ has been established, direct government involvement can
be phased out, for instance in the form of privatization of SOEs. But clearly, overall economic concerns
are not the only guide to the level of SOE involvement in the economy. As argued above, capital constraints may limit the scale of industrial policy to a level below n⁄, or pressure from special interest
groups may lead to n > n⁄. Hence, in our analysis, we regard n as an exogenous variable, which for various reasons may deviate from n⁄.
3. Analysis
To evaluate the efficiency or inefficiency of industrial policy, consider first the case of overshooting,
where n > n⁄. In this case, the state maintains its ownership of firms even though it could successfully
have privatized them, and is thus an expression of what we have referred to as the incompetentbenevolent government. This ‘‘mixed ownership’’ is inefficient, the sources of inefficiency stemming
from the loss of potential output due to patronage jobs and also from the excessive number of SOEs.
136
K. Bjorvatn, N.D. Coniglio / J. Japanese Int. Economies 26 (2012) 129–141
Now, in addition to the income with only SOEs described in (13) for f0 > f, income consists of added
labor income and profits from private firms; i.e., ym = yn + (1 n)(w 1)(f + xirs/b) +
(1 n)(xirs(b w)/b wf), where the second term on the right hand side is additional wage income
in private firms, and the third term is profits in the private sector. Using the fact that xirs = y , income
where the SOEs and private firms coexist is given by:
ym ¼
wb½1 f ð1 nÞ
;
w þ nðb wÞ
for n > n :
ð17Þ
There are several interesting observations we can make from (17). First, we can state that:
Proposition 3. Increasing state ownership beyond the minimum push leads to a reduction in aggregate
income in an economy with mixed ownership.
ðwbð1f ÞÞ
Proof. @y@nm ¼ wb
< 0 since from (9), b(1 f) > w. h
ðwþnðbwÞÞ2
Second, observe that:
Proposition 4. Even excessive state ownership generates more aggregate income than the purely
traditional economy, i.e., ym > y0 for all n > n⁄.
Proof. We know that @ym/@n < 0. Hence, aggregate income with excessive state ownership reaches its
lowest level for n = 1. Note, however, that for n = 1, ym = w > y0 = 1. The proposition then follows. h
Hence, even if increasing state ownership beyond n⁄ reduces the income level in the economy, an
economy dominated by state ownership is still better than one with only traditional manufacturing.
This is the case also for a patronage maximizing government which would result in the complete dominance of SOEs.
The income gap between the fully industrialized economy with private firms (y1), as shown in (8),
and the economy with mixed ownership (ym) is a measure of the inefficiency associated with excessive
state ownership. This gap can be found as:
y1 ym ¼
nbðbð1 f Þ wÞ
:
w þ nðb wÞ
ð18Þ
From (9) we know that b(1 f) > w, and hence y1 > ym for any n 2 [n⁄, 1]. It is interesting to point out
that the magnitude of the inefficiency of excessive state intervention depends upon the level of development of the economy as proxied in the model by f. We observe that:
Proposition 5. The inefficiency of excessive state ownership is larger the more advanced is the economy.
2
nb
Proof. We find that @ ðy1@fym Þ ¼ wþnðbwÞ
< 0, since pi > 0 M b > w. h
To illustrate some key relationships of the model, consider first the potentially hump-shaped effect
of government intervention on aggregate income. Fig. 2 illustrates9.
The optimal policy for this country is to implement a policy such that n⁄ firms are established by
direct control of the state. This is sufficient to trigger private investment, represented by the discrete
jump in income at this point. Once private investment is in place, the state can pull back from its active
role in the economy. Going beyond n⁄, however, implies policy overshooting, with state ownership
crowding out rather than crowding in private investment. Since state-owned enterprises by assumption are less efficient than private enterprises, this leads to a reduction in aggregate income, as can be
observed from the falling ym-curve.
The optimal policy depends inter alia on the level of development in the economy, in our model
captured by f. As shown in Proposition 2, n⁄ is an increasing function of f. In other words, the less
9
Fig. 2 is based on w = 1.5, f = 0.5, b = 4, implying that n⁄ = 0.5.
K. Bjorvatn, N.D. Coniglio / J. Japanese Int. Economies 26 (2012) 129–141
137
y
Fig. 2. Optimal policy, overshooting, and income.
y
n1*
n*2
Fig. 3. Industrial policy and level of development.
developed the country is, the larger is the push needed to phase in private investment. Fig. 3 illustrates
the situation for two countries at different levels of development, where the more developed country
has a lower f and hence a lower n⁄, in the figure given by n1 , while the less developed country has a
higher f and hence a higher n ; n2 .
For n < n1 , the establishment of SOEs raises income in both countries. Similarly, for n > n2 , it reduces income in both countries (as more profitable private firms are crowded out). However, for
n1 < n < n2 industrial policy has a different impact on the two countries. For the more developed
country, it reduces income (along the ym-curve), while for the less developed country, it increases income (along the yn-curve). The main insight is that in a less developed country, the coordination failure in investment is severe, and an extensive government intervention, even if imperfect (eg due to
patronage employment) can be justified. In contrast, in a more developed country, the coordination
failure is less severe. In this case, the optimal intervention is more likely to be modest, and extensive
government ownership would result in lower aggregate income.
138
K. Bjorvatn, N.D. Coniglio / J. Japanese Int. Economies 26 (2012) 129–141
4. Conclusion
The issue of the appropriate role of the government in the industrialization process is a long-standing and highly controversial one in the political and economic debate. Many economists would argue
that picking-winner strategies are likely to fail, in particular in developing countries, because of ‘‘government failures’’. Governments may fail in the difficult process of choosing the right policy objectives,
selecting the appropriate policy tools and efficiently implementing these policies.
While many observers agree that, in principle, the existence of a wide variety of coordination failures justify the use of industrial policy, there is also a widespread scepticism regarding the ability of
governments to replicate the detailed intervention carried out by the newly industrialized countries.
An important reason for this scepticism is the risk that, in a less developed institutional environment,
such policies may be captured by interest groups and give rise to costly rent seeking (see Bjorvatn and
Coniglio, 2006, and for empirical evidence on industrial policy and corruption, Ades and di Tella,
1997).
However, the potential hazard of government failure does not imply that less intervention should
always be the preferred choice. A trade-off exists between government failure and market failure. This
trade-off is particularly relevant for the poorest among developing countries which are often overwhelmed by non-functioning markets. It is far from obvious that reducing the level of state intervention in these countries will make private sector manufacturing blossom.
We develop a theoretical model that analyzes the effectiveness of industrialization policy in setting
where the policy as such is hampered by inefficiencies. The main conclusion from the theoretical model is that ambitious policies of state sponsored investment could be necessary to stimulate industrialization, and that such policies are likely to be more successful in the least developed countries. In more
developed countries, however, industrial policies may do more harm than good. In this way, industrialization policies may lead to big push or big failure depending on the circumstances. An interesting
policy implication from our research is thus that poor countries should think twice before implementing policies of deregulation and privatization.
In our model, the policy inefficiency takes the form of patronage employment, which crowds out
any profits in state-owned enterprises. Clearly, adding further inefficiencies, such as allowing for
loss-making SOEs or long-term negative productivity effects of SOEs, would make industrial policy
less attractive, but would not change the main message of the present paper, namely that even relatively inefficienct industrial policies can be better than no industrial policy to stimulate the transition
from a traditional to a modern economy.
Appendix A
The Appendix describes the Intervention index and gives summary statistics of the countries used
for Fig. 1. The source of the Intervention index is Gwartney and Lawson (2009).
The Intervention index:
The Index of Government enterprises and investment is one component of the Economic Freedom
Index developed by Gwartney and Lawson (2009) and is based on data on the number, composition,
and share of output supplied by State-Owned Enterprises (SOEs) and government investment as a
share of total investment. The index ranges from zero-to-10 ratings. Lower ratings are associated to
countries with more government enterprises and government investment. In more detail, ratings
are assigned on the basis of the following criteria:
Rating = 10. Only few SOEs operate and government investment is generally less than 15% of total
investment.
Rating = 8. SOEs are mainly involved in industries where economies of scale reduce the effectiveness of competition (e.g., power generation); outside these sectors only a limited number of SOEs
operates. Government investment is between 15% and 20% of the total investment.
Rating = 7. SOEs are mainly involved in industries where economies of scale reduce the effectiveness of competition (e.g., power generation). Government investment is between 20% and 25% of
the total.
K. Bjorvatn, N.D. Coniglio / J. Japanese Int. Economies 26 (2012) 129–141
139
Rating = 6. SOEs are present in the energy, transportation, and communication sectors and government investment is between 25% and 30%.
Rating = 4. A large number of SOEs operate in many sectors of the economy, including manufacturing. In addition government investment is generally between 30% and 40% of total investment.
Rating = 2. SOEs are widespread in many sectors, including retail sales and government investment
is between 40% and 50% of the total.
Rating = 0. The economy is dominated by SOEs and government investment exceeds 50% of total
investment.
Appendix B
Countries
Algeria
Argentina
Bahamas
Bahrain
Bangladesh
Barbados
Belize
Benin
Bolivia
Botswana
Brazil
Burundi
Cameroon
Central Afr.Rep.
Chile
China
Colombia
Congo, Dem. R.
Congo, Rep. Of
Costa Rica
Cote d’Ivoire
Cyprus
Dominican Rep.
Ecuador
Egypt
El Salvador
Fiji
Gabon
Ghana
Guatemala
Haiti
Honduras
Hong Kong
India
Indonesia
Iran
Government Intervention
Index
GDP per capita average yearly growth rate (%)
2000
1980
2000–1980
2000–1980
4
10
7
4
4
10
4
2
6
2
8
0
10
2
6
0
0
10
4
6
6
8
8
10
0
8
4
10
2
8
10
7
10
4
7
4
4
4
7
6
2
8
0
0
0
2
2
0
7
0
4
0
2
0
2
4
2
8
4
2
4
7
2
8
4
4
8
4
10
2
2
2
0
6
0
2
2
2
4
2
6
0
6
0
3
2
2
0
2
10
2
2
4
0
4
8
4
1
2
2
2
4
2
3
0
2
5
2
0.22
0.10
0.73
0.33
1.98
1.33
2.63
0.53
0.29
5.55
0.23
1.06
0.47
1.19
3.40
8.48
1.11
5.25
0.49
1.22
2.08
3.88
2.20
0.27
2.51
0.85
0.69
1.24
0.31
0.10
2.87
0.15
3.87
3.46
3.57
0.39
(continued on next page)
140
K. Bjorvatn, N.D. Coniglio / J. Japanese Int. Economies 26 (2012) 129–141
Appendix B (continued)
Countries
Jamaica
Jordan
Kenya
Korea, South
Kuwait
Madagascar
Malawi
Malaysia
Mali
Mauritius
Mexico
Morocco
Namibia
Nepal
Nicaragua
Niger
Nigeria
Oman
Pakistan
Panama
Pap. New Guinea
Paraguay
Peru
Philippines
Rwanda
Senegal
Sierra Leone
South Africa
Sri Lanka
Syria
Thailand
Togo
Trinidad & Tob.
Tunisia
Turkey
Uruguay
Venezuela
Zambia
Government Intervention
Index
GDP per capita average yearly growth rate (%)
2000
2000–1980
2000–1980
8
0
6
1
5
2
0
2
3
0
6
6
4
2
6
0
2
0
4
6
6
4
1
3
4
3
4
1
4
0
0
8
4
2
5
0
0
0
1.24
0.45
0.34
6.28
0.82
1.73
0.35
3.78
0.22
4.47
0.75
1.14
0.52
2.39
1.26
2.41
0.91
3.08
2.45
1.08
0.03
0.45
0.48
0.06
0.94
0.02
3.14
0.68
3.53
0.51
4.68
1.55
0.50
2.07
2.54
0.85
0.94
2.10
10
2
6
6
7
2
0
2
4
4
8
8
4
2
6
0
2
0
4
10
8
4
8
4
4
7
0
8
4
0
4
8
6
2
7
6
2
0
1980
2
2
0
7
2
0
0
4
7
4
2
2
0
4
0
0
0
0
0
4
2
8
7
7
0
4
4
7
0
0
4
0
2
0
2
6
2
0
References
Ades, A., di Tella, R., 1997. National champions and corruption: some unpleasant interventionist arithmetic. Econ. J. 107 (443),
1023–1042.
Atukeren, E., 2005. Interactions between public and private investments: evidence from developing countries. Kyklos 58 (3),
307–330.
Bjorvatn, K., Coniglio, N.D., 2006. Policy design and rent seeking: targeted versus broad based intervention. Rev. Devel. Econ. 10
(4), 577–585.
Cassen, R., Lall, S., 1996. Lessons of East Asian development. J. Japanese Int. Economies 10, 326–334.
K. Bjorvatn, N.D. Coniglio / J. Japanese Int. Economies 26 (2012) 129–141
141
Chang, H.-J., 2007. State-owned Enterprises Reform, United Nations Department of Economic and Social Affairs (UNDESA).
National Development Strategies: Policy Notes, United Nations.
Da Rin, M., Hellmann, T., 2002. Banks as a catalysts for industrialization. J. Finan. Intermediation 11, 366–397.
Das, S., Roberts, M.J., Tybout, J.R., 2007. Market entry costs, producer heterogeneity, and export dynamics. Econometrica 75 (3),
837–873.
Gans, J.S., 1998. Industrialization policy and the big push. In: Arrow, K.J. et al. (Eds.), Increasing Returns and Economic Analysis.
Macmillan, London, Chapter 13.
Gore, C., 2000. The rise and fall of Washington Consensus as a paradigm for developing countries. World Devel. 28 (5), 789–804.
Greenwald, B., Stiglitz, J., 2006. Helping infant economies grow: foundation of trade policies for developing countries. Amer.
Econ. Rev. 96 (2), 141–146.
Grossman, P., 1988. Government and economic growth: a non-linear relationship. Public Choice 56, 193–200.
Gwartney, J.D., Lawson, R., 2009. Economic Freedom of the World: 2009 Annual Report, Economic Freedom Network.
<www.fraserinstitute.org and www.freetheworld.com>.
Kohli, A., 2004. State-Directed Development. Political Power and Industrialization in the Global Periphery. Cambridge University
Press.
Murphy, K.M., Shleifer, A., Vishny, R., 1989. Industrialization and the big push. J. Polit. Economy 97, 1003–1026.
Pack, H., Saggi, K., 2006. The case for industrial policy: a critical survey. World Bank Res. Observer 21 (2), 267–297.
Poot, J., 2000. A synthesis of empirical research on the impact of government on long-run growth. Growth Change 31 (Fall 2000),
516–546.
Rodrik, D., 1995. Getting intervention right: How South Korea and Taiwan grew rich. Econ. Pol. 20, 55–107.
Rodrik, D., 1996. Coordination failures and government policy: a model with applications to East Asia and Eastern Europe. J. Int.
Econ. 40, 1–22.
Rodrik, K., 2006. Goodbye Washington consensus, hello Washington confusion. J. Econ. Lit. 44, 973–987.
Rosenstein-Rodan, P.N., 1943. Problems of industrialization of Eastern and South-Eastern Europe. Econ. J. 53, 202–211.
Rosenstein-Rodan, P.N., 1961. Notes on the theory of the big push. In: Ellis, H.S., Wallich, H.C. (Eds.), Economic Development for
Latin America. St. Martins, New York.
Stein, H., 1992. Deindustrialization, adjustment, the World Bank and the IMF in Africa. World Devel. 20 (1), 83–95.
Stiglitz, J., 2002. Globalization and its Discontents. The Penguin Press.
Trindade, V., 2005. The big-push industrialization and international trade: the role of exports. J. Devel. Econ. 78, 22–48.
Wade, Robert, 1990. Governing the Market: Economic Theory and the Role of Government in East Asia Industrialization.
Princeton University press, Princeton: New Jersey.
World Bank, 1995. Bureaucrats in Business: The Economics and Politics of Government Ownership. World Bank Policy Research
Report, Washington, DC.