The paper studies regional (spatial) inequality in five most populous countries in the world: China, India, the United States of America, Indonesia and Brazil in the period 1980-2000. They are all federations composed of entities (states or provinces) with substantial autonomy. Two types of regional inequalities are considered: Concept 1 inequality which is inequality between mean incomes (GDPs per capita) of states/provinces and Concert 2 inequality which is inequality between population-weighted regional mean incomes. The first inequality speaks to the issues of income convergence, the second, to the issue of overall inequality as perceived by citizens within a nation. China and India show rising inequality in terms of both concepts in the decade of the 1990’s; Indonesia, on the contrary, displays decreasing inequality in both from the early 1980’s up to the Asian crisis. Overall, we find that openness is negatively associated with Concept 1 regional inequality, and positively with Concept 2 inequality. Openness thus seems to help poorer regions (within nations) to catch up, but also leads to disparity in outcomes for populous states with some getting ahead and others falling behind. Maharashtra vs. Bihar, and Shandong vs. Sichuan provide nice examples of such outcomes in India and China. Higher inflation and higher real interest rate are also associated with greater Concept 2 regional inequality.
3. A brief review of regional (within-country) inequality studies6
The issues we address here—Concept 1 and Concept 2 inequality within nations—
have been, in slightly different contexts, addressed before. This was done in two contexts.
The first is the issue of regional inequality within countries. There are two views in the
literature that are often juxtaposed. The first is due to Williamson (1965) who argued
that in the early stages of economic development, regional inequality would tend to rise
as growth occurs in discrete locales, but that later inequalities will decline as equilibrating
forces such as better infrastructure, technological diffusion, decreasing returns to capital
in richer and high-wage areas, diseconomies of agglomeration etc. become stronger.
Thus, regional inequality is expected to follow an inverted U shape as income level
grows. Williamson’s reasoning is closely related to the idea of the Kuznets curve where
similar development although not in spatial terms produces first an increase and then a
decline in inequality. It is also based on the neoclassical (Solow-type) assumptions which
include decreasing marginal returns. A different view has been proposed more recently
within the context of the new economic geography school (Krugman and Venables,
1995) and endogenous growth (Romer 1986; see also a recent review by Easterly and
Levine 2002). There the argument is that that increasing returns to scale and advantages
of agglomeration of capital and knowledge will tend to perpetuate, or even increase,
spatial inequalities. Yet in Krugman and Venables (1995), decreasing transportation costs
may play an offsetting role: assume that transportation costs are zero, then the advantage
of cheap labor in less developed countries (or regions) will, to some extent, tend to offset
the advantages of increasing returns to scale.
The key difference between these two approaches seems not to lie in their view of the
short-run developments, where they all, including the earlier development theories such
as Myrdal’s (1957), Rosenstein-Rodan’s (1943), Hirschmann (1958) or Perroux (1970;
1988), seem to agree that growth is disequalizing, but in their view of the long-run
developments where either traditional neoclassical assumptions dominate—rendering
6 Since we deal with regional inequality, we do not review studies of the most common Concept 3 (interpersonal)
inequality.
10
growth ultimately equalizing in spatial terms as well—or where such assumptions are
rejected or made less potent thus weakening the forces which make for spatial equality. 7
Recently, the short- and long-run aspects have been combined in a paper by Petrakos,
Rodriguez-Pose and Roviolis (2003), which looks at the regional inequalities within the
European Union (with several regions defined within each country). The authors find
that that the short-term effects of growth are disequalizing in the sense that higher growth
rate tends to increase regional inequality (controlling for all other country-relevant
attributes), while higher income level is associated with lower regional inequality. The
authors interpret the second finding as implying the long-run equilibrating effects of
growth along the lines of the Solow and Williamson models. Their measure of regional
inequality, as in several other papers (e.g. Akita and Kataoka (2003) regarding Japan;
Akita and Kawamura (1992) regarding Indonesia and China; Bhalla, Yao and Zhang
(2003) and Kanbur and Zhang (2003) for China) is the population-weighted coefficient of
variation or population weighted Theil index. This is what we called Concept 2
inequality, and the justification for using Concept 2 (rather than Concept 1) inequality is
that it reflects regional inequality as experienced by an “average” person; in other words,
regional divergence which may be due to a few sparsely populated regions’ either very
fast or very slow growth is rather irrelevant for the actual feeling of spatial (horizontal)
inequality as experienced by the people in the country.
But the issue of regional inequality—using Concept 1 inequality—has also recently
received quite a lot of prominence due to the popularity enjoyed by the so-called growth
convergence literature. While the convergence issues have originally been defined and
studied at the level of countries (that is, convergence of national economies), they have
recently been studied at the level of subnational regions. For the countries included here,
such papers are Zhang, Liu and Yao (2001) for China, Azzoni (2001) for Brazil, Ram
(1992) for the United States, and Jha (2004) and Dreze and Srinivasan (2000) for India.8
The rationale for the interest in Concept 1 inequality is very different from the interest in
7 There is an obvious link between these views, as couched in terms of regional developments within
individual countries and regional developments in the world as in Krugman and Venables (1995) or
Krugman (1991).
11
Concept 2 inequality. The issue of convergence of (unweighted) regions within a country,
or (unweighted) countries in the world, addresses the problem of whether the same or
similar economic policies produce similar results or not. Consider the example of a
country which has a single national economic policy, that is where there is no significant
regional freedom of economic policy-making. Suppose that Concept 2 regional inequality
is decreasing. But if we still find that Concept 1 inequality is increasing, it raises the
interesting question of what characteristics enjoyed by some regions are responsible for
their not catching up (or for their growing too fast). Thus both Concept 1 and Concept 2
inequality are of interest.
We shall now briefly review some of key (representative) regional inequality studies
that deal with the five great federations included here. 9 The studies of China are the most
numerous. There are two reasons for this. First, the extremely fast growth of Chinese
economy over the last quarter century has been associated with increasing regional
inequalities. This has obvious implications both for political stability and for economic
theory, that is for figuring out why and how certain regions grow and others don’t, and
whether the dominant feature of China’s inequality is rural vs. urban inequality, or interprovincial
inequality. The consensus seems to be that it is the former. 10 For example,
Bhalla, Yao and Zhang (2003) calculate inequality in per capita consumption across
provincial and urban-rural partitions (that is, they use data for mean annual incomes for
rural and urban areas for each province, that is 28 provinces times 2 = 56 observations11)
and find that in 1995, more than ¾ of thus calculated Theil Concept 2 inequality is
accounted for by the rural-urban split.12 This is a result similar to the one obtained by
Kanbur and Zhang (2003) who find that the same urban-rural split (that is, the difference
8 For other countries, see Goerlich and Mas (2001) for Spanish provinces, Yemtsov (2002) for the
subjects of the Russian Federation.
9 We use the term ”federation” in a technical sense, to indicate that the constituent parts do have some
power of economic decision-making and represent meaningful entities in historic, ethnic or religious sense.
Not all of the five countries studied here are federations in a juridic sense of the word (see also Annex 1).
10 When it comes to inter-provincial inequality, it seems to be more the case of “club” inequality, that is of
three clubs (East, West, Center) diverging from each other (see Yao and Zhang 2001).
11 More exactly, they have the data on mean per capita consumption of peasants and non-peasants (by
province) as obtained from Chinese household surveys. They interpret peasant consumption to be rural, and
likewise non-peasant to be equivalent to urban (see Bhalla et al., p. 945).
12 It is notable that the share of the urban-rural difference in total appears constantly high, that is between
70 and 80 percent, from 1952 to the end of century (see Bhalla et al (2003, Table 2, p. 947)).
12
between mean urban and mean rural income by province) explains 56 percent of
Concept 2 inequality (calculated from the same 56 observations).13 The results of these
two studies and a few others are compared in Annex 2.
The second reason lies in the lack of individual-level data on income inequality in
China, that is lack of data on Concept 3 inequality. 14Concept 2 regional inequality can be
used, if our partitioning is sufficiently fine, to approximate the evolution in Concept 3
inequality. In other words, if we think that most of inequality is spatial, and use a very
fine partition (divide the country in meaningful but also numerous regional units) then
thus calculated Concept 2 inequality should approximate, if not necessarily the level then
the evolution, of Concept 3 inequality. To see this consider that, at the extreme, every
individual can be treated as a “region”: then Concept 2 inequality collapses into Concept
3 inequality. This was, for example, the approach underlying Kanbur and Zhang (2003)
paper on regional inequality in China. As in the Bhalla et al. (2003), Kanbur and Zhang
divide China into 28 provinces and each of the provinces into its rural and urban areas.
They have the data for mean incomes for each of thus defined 56 regions for the period
1952-2000. They calculate Concept 2 inequality from these means, find that the ruralurban
split accounts for the bulk of total inequality (much more than the inland-coastal
split)15 but use the Concept 2 inequality as a proxy for the Concept 3 inequality. Then
they try to relate changes in Concept 2 inequality levels over the last 50 years to various
policy episodes (Great Leap Forward, Cultural Revolution, agricultural liberalization,
urban and industrial liberalization etc.) They find that the Concept 2 inequality, with this
relatively fine partition, amounts in 2000 to a Gini of 37.2 which, of course, sets a rather
high lower bound to total personal income inequality in China.16
13 Why these two results are not the same is puzzling. A comparison of the results of these two studies and
a few others can be found in Annex 2 of the Internet version of this paper available at
http://econpapers.hhs.se/RAS/pmi44.htm
14 Of course, there are many studies of inter-personal inequality in China. They are however all
approximations based on the published group data from national surveys since Chinese authorities have
been unwilling to share micro (individual-level) data.
15 Although after 1993, there is a rapid increase in the within-urban and within-rural components indicating
that there are widening income differentiations within urban and within rural areas as well. (“Within” in this
context means “between mean incomes of rural (or urban) provincial incomes.”)
16 When we use regional GDP per capita as welfare indicators (as here) to derive Concept 2 inequality, its
value cannot be fully taken as the lower bound of Concept 3 inequality because of likely transfers between
13
As for other countries, Jha (2004), in a study of India’s inequality over the last fifty
years looks at the issue of Concept 1 convergence between the states and concludes that
divergence has been more common and has accelerated since the reforms in the early
1990’s. Ram (1992), and Barro and Sala-i-Martin (1992) have done similar analysis for
the United States. Ram (1992) finds a steadily decreasing inter-state inequality from 1950
to 1980 and an increase in the next decade. For Brazil, Azzoni (2001, p. 137) shows
decreasing Concept 1 inequality throughout the 1950’s and 1960’s and a stable one in the
last twenty years.
As this brief review shows, regional inequality studies fall into three categories that
closely match our three concepts of inequality. Many of them, in the wake of the
convergence literature, deal with Concept 1 inequality. Others, perhaps equally
numerous, deal with Concept 2 inequality—regional or horizontal inequality as actually
“experienced” by the population. Finally, some use regional partition (Concept 2
inequality) as a proxy for Concept 3 inequality.
It should be noted that the work on regional inequality is not facilitated by the
absence of an accepted or clear terminology. The results are often impenetrable because
the same term, say “regional inequality” may be used to represent Concept 1 or Concept
2 inequality. Even the term “region” is sometimes used for the smallest units (say, states
in a country) and, perhaps in the same paper, for the agglomeration of several such units
into a larger whole which is still not national level (thus, for example, authors often write
of China’s three regions: East, Center and West, and of China’s regions, meaning in the
latter case provinces). As a consequence, “regional inequality” might mean either one of
the four combinations: Concept 1 or Concept 2 inequality, or inequality between the
provinces or inequality between agglomerations of provinces, that is larger “regions.” In
addition, the share of the between (inter-regional) component is quite different—even if
the regions. This is different from the analysis on the global level when redistributive transfers between the
countries are minimal. I owe this point to Christian Morrisson.
5. Summary and Policy implications
In this paper we examine the link between national economic development and regional
inequalities for European regions and find strong evidence for a bell-shaped relationship betweenthese two elements. This evidence shows in particular that regional inequalities inevitably rise as economic development proceeds but then tend to decline once a certain level of national economicdevelopment is reached. While we try to provide an idea about the level of economic developmentneeded in order to observe a decline in regional inequalities, one must reckon that our results are tied to the particular case study considered here, namely the European Union. Despite this, we believe that our results are sufficiently general and robust in order to provide a general idea about the relationship between national development and regional inequalities and correspond pretty well
to the transition dynamics derived from a simple model of convergence with spillovers such as the one described by Lucas (2000). Several robustness checks are provided including other OECD countries and regions as well as alternative spatial units that tend to support our general results. Our findings have also important policy implications, especially for EU Cohesion policy,which is aimed at boosting convergence and catching-up of lagging EU regions and at reducing
regional inequalities across the EU. This is particularly true for the new EU member states given
that EU funds allocated to these countries could represent up to 4% of their national GDP and could
thus have substantial impact on the growth prospects of these countries. The evidence presented
here implies that some degree of regional inequality is hardly avoidable, at least at the initial stages
of development of countries starting from relatively low levels of GDP per capita such as the new
EU member states that entered the EU in 2004. The main reason for this is that growth is essentially
driven by innovation and technological progress which are unlikely to appear everywhere at the
same time. It follows that some degree of heterogeneity in regional economic development will
necessarily appear as countries are engaged into fast economic catching-up. Regional policy should
thus focus on boosting national growth in order to guarantee greater prosperity across all regions at
the expense of temporarily rising inequality, especially for the poorest new EU member states
starting both from very low levels of economic development and regional inequalities. In this sense,
our results tend to support the findings of a recent paper by de la Fuente (2004) who estimates that,
in the case of Spain, who has been largely benefiting from EU aid since the late 1980ies, the
allocation structural funds would have provided greater welfare through more concentration across
regions in order to favour nation-wide growth. As suggested by de la Fuente (2004), the cost of reshifting
funds toward the most dynamic regions is likely to be mitigated by national-level
interpersonal income redistribution mechanisms.
The dynamics of regional inequalities *
Salvador Barrios a and Eric Strobl a,b
This version: May 2005
Abstract
This paper analyses empirically the dynamics of regional inequalities in GDP per capita. Our
starting hypothesis is that the evolution of regional inequalities should follow an inverted u-shaped
curve depending on the level of national economic development. This hypothesis relates directly to
a number of theoretical findings put forward in the new growth literature, in particular, by Tamura
(1996) and Lucas (2000) who study the growth transition dynamics of regional/national economies
arguing that inequalities should first rise then decline depending on the total amount of knowledge
available in the economy which is directly linked to the level of economic development. These
arguments also correspond to earlier seminal papers by Kuznets (1955) and Williamson’s (1965)
who argued that national growth dynamics may drive, at least initially, to a rise in regional
inequalities. We test empirically these predictions by using regional data for a panel of European
countries and by making use of semi-parametric estimation techniques. Our results provide strong
support for an inverted U-shaped curve in the relationship between the national per capita income
level and the extent of regional inequalities independently of the time period and regional
administrative units considered. The nature of this non-monotonic relationship is not altered by the
inclusion of other possible determinants of regional inequalities.
JEL classification: R1, R5, D31
Keywords: economic development, growth, regional inequalities, transition dynamics, Europe
* Many thanks to Luisito Bertinelli, Fabio Canova, Bruno Cruz, Martin Hallet, Carole Garnier, Diego Martinez,
Antonio Teixeira and Jacques Thisse as well as participants to the CentrA workshop held in Seville and economic
seminar at the University of Nottingham for very helpful comments. Also many thanks to Paul Cheshire, Stefano
Magrini, Jim McKenna and Werner Roeger for help with the European data and Dana Weist and Ines Kudo for help
with the World Bank data. The views expressed by the authors are not those of the institutions they are affiliated with. a
European Commission, Directorate General for Economic and Financial Affairs b Université Paris-X Nanterre. Email:
salvador.barrios@cec.eu.int and eric.strobl@u-paris10.fr. A previous version of this paper was circulated under the title:
“Revisiting the link between national development and regional inequalities: Evidence for Europe”.
2
1. Introduction
This paper investigates empirically the way regional inequalities in GDP per capita are
influenced by nation-level economic development dynamics. Our analysis is grounded on a simple
theoretical framework derived from Lucas (2000) who describes the transition of regional
economies from stagnation to growth. Following this model, regional economic inequalities must
be seen as a by-product of the national development process. Accordingly, economic inequalities
among regions within the same country should first increase as the country they belong to starts
growing and then decrease after a certain level of national economic development is reached. Using
semi-parametric estimation techniques applied to a sample of EU countries we find strong evidence
for such an inverted U-shaped relationship between the national per capita income level and the
extent of regional inequalities.
Our results contribute to the existing empirical literature on growth and convergence in a
number of ways, see for instance Temple (1999) and Durlauf and Quah (1999) for comprehensive
reviews. In particular, while elements such as spillover effects and nonlinearities have become
prominent features of the new endogenous growth theory, empirical studies have, up to recently,
continued to stick to the linear specification derived from the neoclassical Solow types of models,
see Durlauf (2001) and Durlauf and Quah (1999) for critical reviews.1 Within this context, our
result may help understanding why, depending the sample of countries/regions and the time frame
considered, one may alternatively observe convergence or divergence in GDP per capita. Indeed,
when considering the theoretical literature on growth and convergence, a wide array of arguments
arise advocating either for the long-term reduction or, to the contrary, for the persistence and selfreinforcing
nature of economic inequalities across countries, see, for instance, Galor (1996) and
Prichett (1997). Independently of the type of result obtained, authors have increasingly focused on
the role played by knowledge and spillovers in order to explain countries’ growth differentials and
growth diffusion both across countries and regions, see Jones (2004) and Klenow and Rodriguez-
Clare (2004) for recent contribution on these issues. Knowledge spillovers would give rise to
substantial scale effects in productivity stemming from the non-rivalry nature of knowledge, and
this a central theme in the works of Romer (1990), Kremer (1993) and Jones (2001) among others.
Since knowledge and technological progress are often seen as the main engines of economic
development, the latter may inevitably increase rather than decrease inequalities since these two
elements are very unlikely to be evenly spread both across time and space. The latter means that
economic growth may, at least initially, foster divergence, rather than convergence across spatial.
One downside effect of rapid economic growth in China has been the ever rising inter-regional inequality. Foreign direct investment (FDI) has been blamed for driving the Chinese regions apart. It is difficult to reconcile the positive effect of FDI on economic growth with its potential 'negative' effect on regional inequality. Using the largest panel dataset for the Chinese regions over 1979-2003 and employing an augmented Cobb-Douglas production function, this paper proves that FDI has been an important factor of economic growth in China. It also suggests that it is the uneven distribution of FDI instead of FDI itself that has caused regional growth differences. Human Capital Formation and Economic Growth in India : A CGE Analysis4
V. P. Ojha5
B.K. Pradhan6
1. Introduction
It is well known that India’s transition to an outward-looking strategy is a delayed one. Compared to, say, China, India is almost a decade behind in launching its economic reforms program, which it did in 1991 as a response to the economic crises created by the chronic fiscal and trade imbalances of the eighties, rather than as a planned shift to outward orientation. Little wonder then, that India, again unlike China, was unprepared for the greater openness of the outward-oriented strategy. It had not gone through the internal adjustments and transformations which must ideally precede trade liberalization. In fact, India is still struggling to undergo the variety of internal economic reforms that are required to be able to face the challenges of globalization. Among these reforms are (i) de-bureaucratization and deregulation of the industrial environment, (ii) restructuring of the public sector, (iii) developing the agricultural and industrial infrastructure and (iv) promoting human development. It is not a matter of chance that the last one is not an integral component of the reform package of the government, but only a sort of add-on to the policy package. The underlying view is that policies for human development or social sector development, as it is referred to in the policy-making circles, are supplementary measures required to translate economic growth into an equivalent increase in human well being. While this view is not contestable, it is clearly insufficient. More specifically, it does not take into account the obvious lessons from the experience of high performance east Asian and the Chinese economies in the last two decades. The policy makers in these economies clearly regarded the causation between human development and economic growth as bi-directional. And in operational terms, they consciously developed the human resources to achieve higher economic growth.
The present study is motivated by a view similar to the one, which underlay the strategic planning of the East Asian Economies. It may also be mentioned here that, though human capital in
4 This study was undertaken at the National Council of Applied Economic Research (NCAER), New Delhi and sponsored by the South Asia Network of Economic Research Institutes (SANEI).
5 V P Ojha is a Senior Consultant at NCAER, New Delhi. E-mail : vpojha@ncaer.org
6 B K Pradhan is Chief Economist at NCAER, New Delhi. E-mail : bkpradhan@ncaer.org
3
an economy includes both the state of health and the educational levels of the the people, in this study the focus is exclusively on educational capital.
In sub-sections 1.1 – 1.3 that follow, we discuss the existing literature on human capital formation and economic growth. Finally, in sub-section 1.4 we outline the main objectives of the present study.
The rest of the paper is organised as follows. Section 2 presents the overall structure of the CGE model used in the present study, with special emphasis on the intertemporal dynamics which includes a mechanism through which public education expenditure augments the stock of human capital. Section 3 presents the main features, such as, GDP growth and growth of household incomes, of the base-line or the business-as-usual (BAU) scenario. In section 4, we report the simulation results of the three policy scenarios in comparison with the BAU scenario. Section 5 concludes and suggests policy implications of our results. In Appendix 1 we present the Social Accounting Matrix (SAM) which provides the benchmark equilibrium data set for the model. Appendix 2 gives the detailed set of equations of the model.
1.1 Studies on human capital formation
The literature on human capital formation is abound with partial equilibrium analyses of production and cost functions of education (see Shri Prakash and Chowdhury (1994), Tilak (1985) and Tilak (1988), as well as of determinants of household expenditure on education (see Tilak (2001a), Tilak (2001b)), Tilak (2001c) , Tilak (2002), and Shri Prakash and Chowdhury (1994) ). The studies dealing with the production function of education (say, for example, Shri Prakash and Chowdhury (1994)) measure output in terms of ‘enrolments’ and inputs in terms of ‘number of teachers employed’ and ‘value of non-teaching inputs’. Such production functions are obviously useful in determining whether the “production” of education is subject to increasing, constant or diminishing returns and the relationships between the marginal productivities of the teaching and non-teaching inputs. (The cost functions of education are essentially a ‘dual’ of the production function and serve the purpose of merely confirming the results obtained from the production functions). However, from these essentially technical descriptions of the ‘production’ of education no policy conclusion of consequence is derivable. In other words, in so far as these studies determine neither the private nor social returns to education, their policy significance is limited. The studies concerned with the determinants of household expenditure on education (for example, Tilak (2002) also treat education as an end in itself and fall short of explaining expenditure on education in terms of the expected private returns on education. Using state-wise cross-sectional state level data for his regressions,
4
Tilak (2002) explains household expenditure on education in terms of household incomes, and other household characteristics such as educational level of the head of the household, occupation, caste, religion.
The ‘general equilibrium’ studies on educational capital formation have a broader objective, namely, assessing the impact of investment on education on productivity (growth) and/or equity (wage-inequality). All these studies are based on the underlying assumption that public investment in education is a powerful policy instrument for inducing faster economic growth with an improved or a worsening income distribution. It needs to be stressed that a priori it cannot be known whether investment in education leads to growth with more or less wage inequality. Not surprisingly then, most of these studies are concerned with the impact of investment in education on changes in wage inequality over time. In a general equilibrium framework, there is multi-directional causation between investment in education and changes in the relative wages of skilled labor. On one hand, the increased investments in education lead to an increase in the relative supply of skilled labor, which in turn exerts a downward pressure on the relative wages of skilled labor. On the other hand, the technological changes and the changes in international terms of trade in favor of skill intensive goods, that necessarily accompany the growth process, push upwards the skilled wage rate relative to the unskilled wage rate by creating more demand for skilled labor. In short, relative factor supply and relative product price changes are both important in explaining the change in the relative return to skilled labor, and a general equilibrium model effectively captures the net impact of these factors on the relative wages.
Pradhan (2002) finds an interesting paradox in the growth process of the Indian economy, namely, that there is not much change in income inequality even though there are large changes in the educational levels of the population over time. He tries to resolve this paradox by using an applied general equilibrium model to simulate the impact of large changes in access to education on wage inequality. The model results clearly show that even for very large increases in access to education the wage inequality remains unchanged. Apparently, the dominant effect on the skilled labor wage rate is that of the changes in the relative product prices in the world market (i.e., the trade effect), rather than that of increased relative supply of educated labor ensuing from enhanced access to education. The trade effect on the relative demand for skilled labor has been shown to be very important for India by Wood and Calandrino (2000) also in a SAM (Social Accounting Matrix) based comparative analysis of the impact of trade liberalization on human resources in India and China. Gidling and Robbins (2001) analyze the patterns and sources of changing wage inequality in Chile and Costa Rica during structural adjustment, using an econometric decomposition technique which splits the effects of enhancement of human capital into the
5
‘education price’ and ‘education quantity’ effects. Their exercise shows that the education price effects varied across sectors on account of the variation in the sectoral rates of growth in the demand for educated workers, and this lead to an increase in inequality in Chile despite a large equalizing education quantity effect. Duflo (2002) in his paper on the effects of educational expansion in Indonesia shows a different impact on the relative wages of skilled labor. Using a two sector - formal and informal – econometric model, he shows that the skilled labor, employed exclusively in the formal sector, suffers a downward revision of relative wages, because the faster increase in human capital is not matched by a corresponding increase in physical capital in this sector. Interestingly, this paper indicates the possibility of there being competing demands of physical and human capital on the investible resources of the government for a mixed economy like India. That is to say, the public sector, which bases its investment decisions on long-term growth rather than on short-term profitability considerations, needs to define a trade-off between augmenting physical and human capital.
Most other general equilibrium studies on the shifts in the relative wages pertain to the U.S.A. Goldin and Katz (1999), Francois and Nelson (1998), Harrigan and Balaban (1999) and Baldwin and Cain (1997) are all concerned with explaining the “paradoxical” effect of educational expansion on the wage inequality – i.e., increased availability of education increases rather than decrease the relative wages for skilled labor. And, in fact, the paradox is resolved in almost all the cases by incorporating the effects of trade and technological changes on the relative demand for skilled labor.
1.2 Education and economic growth in India
The link between public spending on education and economic growth is by now well-established in the literature. Staring with the work of Schultz (1961) education has been viewed as investment in human capital rather than considered to be a consumption good under Keynes’ influence. Subsequently, Blaug et al (1969), Tilak (1987) and Psacharopoulos (1993) show that investment in education yields a higher rate of return than investment in physical capital. Romer (1986) and Lucas (1988) have propounded the new growth theories in which sustained long-run growth of per capita income is explained by the likelihood of investment in human capital generating constant or increasing returns. Empirical studies in the literature on education and economic growth also find compelling evidence for the hypothesis that a substantial proportion of the growth of the economies is attributable to the rise in the educational levels of the workforce. Lau et al (1993) attribute almost 25 percent of the economic growth in Brazil to the increase in the average education of the
6
workforce. The success stories of the East Asian miracle economies are also replete with references to mass primary education programmes pursued by their governments (World Bank, 1993). In India, Mathur (1993) has shown that a positive association exists between stocks of human capital and economic development and that the association becomes stronger at higher levels of education. Mathur and Mamgain (2002) find the influence of both technical and general education on per capita income to be positive with that of the former being more powerful. In agriculture, Chaudhri (1979) finds that primary schooling affects productivity positively, particularly in times of rapid technological change.
While the link between the spread of education and economic growth is regarded as undisputable, the preceding link between public education expenditure and the spreading of education has become a bit of a controversial area, especially in India. Empirical evidence in India in this regard is diverse – differing hugely across the states – and does not seem to corroborate the assumed positive linkage between public spending on education and the spread of education (Pradhan, Tripathy and Rajan (2000)). Various explanations are offered for the absence of a strong positive association between public education expenditure and educational outcome – leakages from the amount spent due to corruption, teacher absenteeism, non-motivated and discouraging teachers, ill-equipped schools and unwillingness of parents to send their children to schools due to economic or non-economic constraints. The conclusion sometimes drawn from all this is that public spending is not really instrumental in promoting education, and therefore should not be overdone. This is unfortunate especially because the diverse empirical evidence does not warrant this rather straightforward conclusion. A detailed examination of the question of the impact of public education expenditure on the quality of education and educational outcome, particularly enrolment, has been done by Pradhan and Singh (2004). Pradhan and Singh (2004) also do not find a strong influence of pubic expenditure per child and the rate of growth of expenditure on the enrolment rate for 16 major states of India. However, this is because the varying degrees of ‘efficiency’ of expenditure across states are not taken into account. The efficiency of expenditure is defined as the technical efficiency of the inputs – the number of schools and the number of teachers – in generating educational output, such as enrolment. Using Data Envelopment analysis (DEA), they rank the states by their levels of technical efficiency. Having thus ranked the states by their levels of technical efficiency, they a find stronger positive association between publc education expenditure and enrolment for the relatively efficient states as compared to the relatively inefficient states. In other words, once the efficiency of expenditure is taken into account, the effect of public education expenditure on enrolment is seen to be stronger. In general, it is arguable that states which employ better educational processes also demonstrate a stronger link between education
7
expenditure and educational outcome. By implication, the states in which the link between education expenditure and educational outcome is weak have to find ways and means to strengthen this link – i.e., control the leakage from the education expenditure, prevent teacher absenteeism improve infrastrucutre in schools, and, above all, take care of the economic and non-economic factors which are responsible for the lack of interest shown by households in providing education to their children. In short, the picture which emerges from the analysis of Pradhan and Singh (2004) is hardly the one which would undermine the importance of increasing public spending on education in India.
The share of expenditure on education in GDP in India has been continuously increasing from 1.19 percent in 1951 (not shown in table 1) to 3.98 percent in 1990-91, after which it suffered a decline till 1997-98. In 1998-99 it was restored to 3.90 percent, and in 1999-2000 it crossed the 4.0 percent mark. However, it may be noted that although education has always been given high priority by the government of India since independence, the public expenditure target of 6 percent of GDP is still nowhere in sight. Not surprisingly, even after 50 years of independence, the enrolment rates remain low in this country, particularly in case of poor and the inhabitants of rural areas. It follows that the role of public spending on education, though not complete per se, remains important in accelerating the growth in school enrolment. Besides, an expansion of public education expenditure is all the more desirable because of the externalities associated with education, such as, reduced population growth and better health care.
The sources of finance for education India are the central and the state governments, local bodies, consumers of education (fees etc.) and foreign aid. Primary among these are the state governments. However, as argued by Mehrotra (2004), given the serious fiscal deficits of the poorest states and the limited scope of inter-sectoral reallocation of expenditure towards education from other sectors and of intra-sectoral allocation within the education sector (from higher levels of education to lower levels), the only remaining option for financing further increases in public education expenditure is earmarked taxes for education, a source employed effectively by many countries, such as, Korea, China, Botswana and Brazil. Mehrotra (2004) also finds the successful example of Brazil, worth emulating for India. In Brazil, an education fund, FUNDEF, created by federal taxation, helps in the equalisation of expenditure capacity in education between poorer and richer states. He further recommends that in India, much like in Brazil, the central government, and not the state governments, should levy additional taxes and dedicate the revenue thus raised to the cause of education. The dedicated fund for education could then allocate resources to the states that are in greatest need and those that show the best performance. The initiative for additional taxation and the subsequent creation of the dedicated fund needs to be taken by the central government
8
because many of the state governments have been seen to be lacking in their commitment to elemenatry education.
Table 1 : Trends in Public Education Expenditure
Ed. Exp./
Total
Govt. Exp.
(in %)
Ed. Exp./
GDPMP
(in %)
1990-91 3.98 12.52
1991-92 3.84 12.15
1992-93 3.71 12.33
1993-94 3.64 12.16
1994-95 3.56 11.95
1995-96 3.57 12.58
1996-97 3.56 12.78
1997-98 3.57 12.48
1998-99 3.90 13.39
1999-00 4.44 14.21
2000-01 4.14 13.21
Source : Education expenditure : Analysis of Budget Expenditure in Education (various years).
New Delhi ,MHRD
Total government expenditure : Indian Public Finance Statistics, Government of India
GDP at market prices : National Accounts Statistics (various issues), CSO,
Government of India
Table 2 : Enrolment rates by
place of residence and
poverty category
Place of Residence Enrolment rate (in %)
All-India 76.2
BPL 65.6
APL 84.8
Rural-India 73.2
BPL 63.8
APL 81.1
Urban-India 86.7
BPL 72.4
APL 96.4
Note : APL : Above Poverty Line
BPL : Below Poverty Line.
Source : Pradhan and Roy (2003)
9
1.3 CGE analysis of the linkage between public expenditure on education and economic growth
The overall goal of this study is to investigate the linkage between human capital and economic growth. However, as mentioned before, in this study our focus is concentrated on only the educational aspect of human capital. Secondly, though the household expediture on education is by no means small (see Tilak (2002)), we are not treating it as a policy variable. It is infact largely dependent, among other things, on the government expenditure on education as shown by Tilak (2002). Having made these two assumptions, we narrow down the goal of the study to an investigation of the linkage between public investment in education and economic growth.
In a priori hypothesizing about the linkage between educational capital investment and economic growth, one tends to argue that investment in education increases the supply of educated (skilled) labour, which, on account of its higher productivity relative to non-educated (unskilled) labour leads to higher economic growth with lower relative wage for skilled labour. This is nothing but the standard one-sector endogenous growth theory line of reasoning, and need not hold in a multi-sector, multi-factor general equilibrium framework. What this line of reasoning overlooks is the fact that educational capital accumulation will in all likelihood be accompanied by a changes in demand pattern in favour of skill intensive goods. The (exogenous) international terms of trade will also most likely shift in favour of the skill intensive goods. All this will increase the relative demand for skilled labour exerting thereby an upward pressure on its relative wage. On the production side, there will not only be a restructuring of the composition of goods produced in favour of skill intensive goods, but also some resubstituting in favour of unskilled labour in the production processes. In short, changes in both the relative factor returns and the relative product prices play a role in determining the quantam of growth. It follows then that, how much the resultant growth will be is an empirical question best answered by a computable general equilibrium model.
A CGE analysis of the linkage between public expenditure on education and economic growth is conspicuous by absence in the scanty literature on human capital formation in India. For other countries also, CGE studies on the impact of public education expenditure on human capital formation are sparse. Suwa-Eisenmann, Zonzilos and Bourguignon (1995) assess the programs implemented in Greece under the Europeon Community Support Framework (1989-1993), for promoting growth through investments in infrastructure and human capital. The model used in this study is an extended version of the standard CGE model described in Dervis, de Melo and Robinson (1982) which incorporates a semi-Keynesian closure appropriate for the Greek economy.
10
However, the accumulation of human capital is treated in an exogenous manner. It is assumed that the expenditure on training programs has a direct positive effect on labour participation, thereby, increasing the absolute number of skilled workers, without changing that of the other (unskilled) types of labour. There is no transformation of unskilled labour into skilled labour envisaged in the model.
More closely related to the goal of this study is the paper by Jung and Thorbecke (2003). In this paper the impact of public education expenditure on human capital, the supply of different labour skills, and its macroeconomic consequences are analysed using a recursively dynamic multisectoral CGE model for two heavily indebted poor countries (HIPCs), Tanzania and Zambia. The CGE model used here is the standard neo-classical type described in Dervis, de Melo and Robinson (1982), Thorbecke (1992) and Robinson et al (1999), with the additional feature that three different types of labour - non-educated, primary-educated, and higher-educated labour - are combined in two stages in the production structure of the model, to reflect different levels of substitutability. The non-educated and the primary educated labour are combined within a Cobb-Douglas type Armington aggregation to produce an aggregate of unskilled labour. This unskilled-labour-aggregate is then combined with higher-educated labour within a CES type Armington aggregation to yield a composite labour measure. Profit maximizing firms employ the optimal amount of each type of labour given wage rates and the technical and budget constraints.
Another novel feature of the Jung and Thorbecke (2003) model is that its intertemporal dynamics includes a specific mechanism through which public education expenditure augments the stock of human capital. In other words, education expenditure provides additional educational capital to those who are in the educational pipeline. As these individuals come out of the educational pipeline, they acquire improved labour skills and, thereby, add to the stock of human capital.
The business-as-usual scenario or the base run of the Jung and Thorbecke (2003) model is generated under the assumption that each of the three types of labour grows at the given population growth rate. Subsequently, three alternative policy scenarios, each envisaging a 15 percent increase in real public expenditure on education over the base-run level under three different assumptions, are simulated. In the first scenario the supply of primary-educated and higher-educated labour are determined first in the model, and the supply of non-educated labour is determined residually, in such a manner that the total work force grows at the given population growth rate. The underlying assumption in this simulation being that the non-educated labour supply is not responsive to the wage rate. In the second simulation, the more realistic assumption of elastic labour supply is made. In this case, a rise in the wage rate results in a flow of the previously unemployed non-educated
11
workers into the labour market, with the flow ceasing when the wage rate of non-educated labour equals that in the base run. In the third simulation, the additional assumption made is that the increase in real education expenditure is directed exclusively to the poor household groups, so that the increase in the educated labour supply over that in the base-run comes entirely from the poor groups. In other words, this simulation provides for an increase in the endowment of human capital of the poor groups relative to that of the non-poor groups .
The higher education expenditure increases the labour supply growth rates for primary educated and higher educated labour to the same extent (0.4 percentage points) in all the three simulations7. However, in case of non-educated labour, the labour supply growth rate increases by 0.2 percentage point in simulations 2 and 3, but decreases by 2.1 percentage points in simulation 1, where labour supply of non-educated labour is determined residually.
The growth rates of the wages for both higher educated and primary educated labour decline by 0.3 percentage points in simulations 2 and 3, while in simulation 1, the wage growth rates for higher and primary educated labour decrease respectively by 0.6 and 0.3 percentage points. For the non-educated labour, the growth in wage level remains unchanged (at the base-run level) in simulation 2 and 3 by assumption, but increases by 2.3 percentage points in simulation 1. The average wage grows at the same rate in simulation 1 as in the base run . But the growth rate of the average wage declines by 0.3 percentage points for Tanzania and 1.1 percentage points for Zambia in case of simulations 2 and 3. The extent of physical capital accumulation in Zambia is much lower than Tanzania, on account of a lower saving ratio in the former country. Hence, the increase in the number of skilled workers is not complemented by an adequate increase in physical capital limiting the growth in labour productivity and consequently in wages.
With 15 percent increase in real public expenditure on education, GDP growth rate under simulations 2 and 3 increases by 0.2 and 0.1 percentage points respectively for Tanzania and Zambia. In simulation 1, however, GDP growth rate improves by only 0.1 percentage point for Tanzania and remains the same as in the base run in for Zambia. Moreover, in all the simulations for both the countries the capital income grows faster than the wage income. This is expected, as the supply of educated labour increases as a result of the expansion in public educational expenditure, and capital consequently becomes relatively more scarce.
Income distribution changes are not uniform either across the simulations or for the two countries. Under scenario 1, the growth rate of household incomes of the urban poor improves by
7 In the summary of results presented here, we mostly refer to the figures for Tanzania. Unless otherwise mentioned, the broad orders of magnitudes of the changes in the variables for the twocounties are the same.
12
0.1-0.2 percentage point, and declines by 0.1 percentage point for the ‘urban non-poor’ in case of Tanzania. However, it is the other way round for Zambia, where the growth rate of household incomes of the ‘urban poor’ declines by 0.1 percentage point, and improves by 0.1 percentage point for the ‘urban non-poor’. This happens because the urban poor household group is heavily dependent upon income from educated labour. Hence, an increase in public spending on education makes this group worse off by increasing the supply of educated labour. The rural household incomes grow relatively faster for the poor vis-a-vis the non poor in Tanzania, but not so in Zambia. Unlike in Tanzania, in Zambia, the educated workers are not concentrated within the non poor household groups, but dispersed among non poor and poor household groups.
Under scenario 2, in Tanzania, the growth rates of incomes of both the urban and rural poor improve by 0.3 percentage point, while those of urban and rural non-poor increase by only 0.1 percentage point. That is, the poor gain more than the non poor from the increase in public educational expenditure in Tanzania, but the opposite is true for Zambia. In Zambia, the improvement in the growth rates of incomes of the urban and rural non poor is greater than that for the urban and rural poor. Evidently, Zambia, as compared to Tanzania, has a larger proportion of educated workers within the poor households.
In simulation 3, the growth rates of incomes of the poor household groups improve significantly more than those of the non poor household groups resulting in a more equal distribution of income in Tanzania. However, in Zambia the income distribution does not improve. Here, both the poor and non poor groups improve their income growth rates relatively equally among the rural households, and, among the urban households the non poor groups, in fact, improve their income growth rates relatively more than the poor groups. Clearly, the difference in the endowment of human capital of poor households between Zambia and Tanzania matters in determining the impact of an increase in education expenditure on the income distribution in the two countries.
In short, the main conclusion that emerges from the counterfactual policy simulations of the Jung and Thorbecke (2003) model is that an increase in public education expenditure per se can contribute positively to GDP growth. Improved labour market flexibility will enhance the positive impact of an expansion in public educational expenditure on GDP growth. Furthermore, the rise in public expenditure on education should ideally be complemented with an increase in public investment on physical capital. And, finally, the increase in educational expenditure must be better targeted to poor households if any improvement in the income distribution is to be expected.
A crucial question on which the Jung and Thorbecke (2003) paper is silent is the following : how is the increase in public education expenditure to be financed or, to put it another way, what
13
will be the preferred mode of financing an expanded public education expenditure programme. In fact, the suggestion that the rise in public expenditure on education be matched with an increase in public investment on physical capital begs this question. If public investment cannot be curtailed (it might have to increase !), then some other adjustment has to be made in a resource constrained fiscal environment - either government expenditure in other sectors will have to be reduced or, if that is not feasible, taxation will have to be increased as suggested by Mehrotra (2004). We have considered the latter option in the present study.
1.4 The present study
In the present study, we have used a recursively dynamic multisectoral CGE model for the Indian economy. Our model has been formulated on the lines of the Jung and Thorbecke (2003) model to capture the impact of an increase in public education expenditure on GDP growth and income distribution across four rural and five urban household groups
As is usually done in a CGE modeling analysis, we first generate a base-line (business-as-usual) scenario, and then simulate alternative policy scenarios for assessing the consequences for growth and income distribution in India of an expansion in public education expenditure. The specific policy questions to which the policy scenarios are addressed are the following :
(i) What is the impact of an increase in public education expenditure financed by an increase in direct taxes on GDP growth and income distribution ?
(ii) What is the impact of an increase in public education expenditure concentrated in the secondary education sector financed by an increase in direct taxes on GDP growth and distribution ?
(iii) What is the impact of an increase in public education expenditure concentrated in the secondary education sector complemeted with an increase in public investment financed by an increase in direct taxes on GDP growth and distribution ?
14
2. Model Structure
Our model is a multisectoral, neo-classical type price driven CGE model, with the additional feature that it includes a mechanism through which public expenditure on education augments the supply of human capital (i.e., educated / skilled labour). The overall structure of our model is similar to the one presented in Jung and Thorbecke (2003). However, in formulating the details of the model, we follow an eclectic approach keeping in mind the institutional features peculiar to the Indian economy.
The model has 10 production sectors and three factors of production - land, capital and composite labour, which in turn, is a nested CES aggregation of non-educated, secondary-educated and higher-educated labour8. At the beginning of a period, the economy is endowed with a certain level of physical capital and human capital, in the form of stocks of different types of labour. In any given period the allocation of capital across production sectors is fixed, but labour is inter-sectorally mobile. Producers act as profit maximisers in perfectly competitive markets, i.e., they take factor and output prices (inclusive of any taxes) as given and generate demands for factors so as to minimise unit costs of output. The factors of production include intermediates and the primary inputs – capital, land and different types of labour. For households, the initial factor endowments are fixed. They, therefore, supply factors inelastically. Their commodity-wise demands are expressed, for given income and market prices, through the Stone-Geary linear expenditure system (LES). Also households save and pay taxes to the government. Furthermore, households are classified into four rural and five urban categories. The government is not asssumed to be an optimising agent. Instead, goverment consumption, transfers and tax rates are exogenous policy instruments. The rest of the world supplies goods to the economy which are imperfect substitutes for domestic output, makes transfer payments and demands exports. The standard small-country assumption is made, which implies that, India is a price-taker in import markets and can import as much as it wants. However, because the imported goods are differentiated from the domestically produced goods, the two varieties are aggregated using a constant elasticity of substitution (CES) function, based on the Armington assumption. As a result, the imports of a given good depends on the relation between the prices of the imported and the domestically produced varieties of that good. For exports, a downward sloping world demand curve is assumed. Furthermore, a constant elasticity of transformation (CET) function is used to define the output of a given sector as a
15
revenue-maximising aggregate of goods for the domestic market and goods for the foreign markets. This implies that the response of the domestic supply of goods in favour or against exports depends upon the price of those goods in the foreign markets vis-à-vis their prices in the domestic markets, given the elasticity of transformation between goods for the two types of markets. The model is Walrasian in character. Markets for all commodities and non-fixed factors - capital stocks are fixed and intersectorally immobile - clear through adjustment in prices. However, thanks to the Walras' law, the model determines only relative prices. The exchange rate is chosen as the numeraire and is, therefore, normalised to unity. The model determines endogenously the foreign savings in the external closure. Finally, because the aggregate investment is exogenously fixed, the model follows an investment-driven macro closure, in which the aggregate savings - i.e., the sum of household, government and foreign savings - adjusts, to satisfy the saving-investment balance.
Intertemporally, the model adjusts through changes in the stock of physical capital and the stock of human capital. Physical capital is increased by investment, which is exogenously given. Human capital is augmented by the new supply of educated labour, which in turn is a function of public education expenditure.
2.1 Sectoral disaggregation
Our model is based on the following ten sector disaggregation of the Indian economy :
1. Agriculture ( 1 to 7 ),
2. Mining ( 8 to 11 ),
3. Manufacturing-1 (12 to 24),
4. Manufacturing-2 (25 to 44),
5. Construction (45),
6. Electricity, gas and water supply, (46 to 47)
7. Transport, storage etc., (48 to 51)
8. Wholesale and retail trade etc, (52 to 53)
9. Finance, insurance, real estate etc., (54 to 56)
10. Community, social and personal services, (57 to 60)
Note that for each sector the constituents in terms of the 60-sector Central Statistical Organisation Input-Output Transaction Table (CSO-IOTT) is indicated in the parenthesis. Note also that each
8 In our classification of 3 types of labour in India, ‘secondary educated’ includes all those from 1st pass to 12th pass – i.e., ‘elementary’ + ‘secondary’ + ’higher secondary’ educated, and ‘higher educated’ includes
16
sector has 3 types of labour inputs – unskilled or non-educated labour, semi-skilled or secondary educated labour and skilled or higher educated labour – which sum up to what is called composite labour.
‘graduates ‘ + ‘higher-than-graduates’.
2.2 The production structure
Production technologies for all sectors are defined using nested CES functions as shown below :
Domestic Sectoral Gross Output
Intermediate Input Bundle Value Added (VA)
Composite Labour (CL) Capital (K)
Skilled Labour Composite (SLC) Non-educated Labour (LL1)
(i.e., Unskilled Labour)
Secondary-educated Labour (LL2) Higher-educated Labour (LL3)
(i.e, .Semi-skilled Labour ) (i.e, .Skilled Labour )
Note that vertical lines in the nesting diagram represent leontief combinations, while the slanting lines represent CES combinations of the inputs involved. For agriculture there is an additional branch in the nesting structure. In the agricultural sector, a cobb-douglas aggregation of land and capital produces composite capital which in turn is combined with composite labour to produce value added. At each level of the nested production function, the assumption of constant elasticity of substitution (CES) and constant returns to scale (CRS) is made. For every level , the producer’s problem is to minimise cost (or maximise profit) given the factor and output prices and express 17
demands for inputs. It follows that for every level, the following three relationships hold : the CES function relating output to inputs, the first order conditions, and the product exhausation theorem. For all the levels taken together, the production system thus determines the gross domestic output, the input demands, value-added as well as the demands for the various types of labour. (The capital stock in a particular period is given, so the first-order condition effectively determines the sectoral return on capital.)
2.3 Investment
Public and private investment are fed into the model as two distinct constituents of the total investment. There are fixed share parameters for distributing the aggregate investment across sectors of origin. However, the allocation mechanisms for sectors of destination are different in the two cases of public and private investment. For public investment there is discretionary allocation, and the allocation ratios are therefore set exogenously in the model in each period. On the other hand, for private investment the allocation ratios are given in a particular period, but are revised from period to period on the basis of the sectoral relative return on capital. The relative return on capital in any sector is given by the normalisation of the implicit price of capital in that sector to the economy-wide returns. Note that this rule does not imply full factor price equalisation, but only a sluggish reallocation of investment from sectors where rate of return is low to ones having higher rates of return.
Needless to say, all this bifurcation of total investment into its public and private components with their differing allocation mechanisms is an attempt to approximate the way investments are actually made in the Indian economy. Incidentally, it also allows for public investments to be directed towards “strategic” sectors disregarding short-run considerations of profit maximisation.
2.4 Factor markets
Labour is intersectorally mobile. Wages are flexible and adjust to equilibriate the demand and supply for each of the three types of labour – non-educated labour, secondary-educated labour and higher educated labour. There is no unemployment for any of the three types of labour. Cropping land in the agricultural sector is also fully utilised at the equilibrium rent. However, capital stocks are fixed sectorwise. The optimsing behaviour of producers therefore determines sector specific return on capital.
18
2.5 Household Income and consumption demand
There are nine household groups in the model - rural cultivator (RC), rural artisan (RATN), rural agricultural labour (RAL), rural others (RO), urban farmer (UF), urban non-agricultural self-employed, (UNASE), urban salaried (US), urban casual labourer (UCL), urban others (UO). The factor endowments for each household group are given. Households derive their income by selling the factors they own – land, labour (of 3 types) and capital. From these incomes, taxes are netted out and transfer payments by government and rest of the world are added to arrive at the household disposable incomes. The households are assumed to save a fixed fraction of their disposable incomes. The rest of it is spent on the consumption of goods. The consumption functions of the households are estimated by the most suitable Stone and Geary linear expenditure system (LES), which is widely used in India. Private corporate and public sectors do not have any consumption expenditure. They receive income from the rental values of non-land capital. Private corporate sector gets additional income from rental value of land and government transfer payments including interest payments.
2.6 Private corporate and public sector income
Private corporate sector income consists of its earning from factor incomes and transfers from government, which is equal to its savings. On the other hand, public sector income is defined as income from enterpreneurship (factor income from capital) that goes as transfers to government.
2.7 Household savings
The average propensity to save out of their disposable incomes is exogenously given for each of the four rural and five urban households. Households thus save a fixed part of their incomes. Total household savings in the economy is obtained by summing up the savings of all the nine household groups.
2.8 Government Savings
Government revenue originates from the following five sources : excise tax on production, sales tax on goods, import duties from imported goods and income tax from households. All the tax rates are exogenously given. Government income also includes the capital income and land rent from ownership of these factors, factor income from abroad and public sector income. Government expenditure takes place on account of government consumption and transfers to households and
19
firms, and public sector investment, all of which are exogenously fixed. Government savings is obtained as the difference between government income and expenditure.
2.9 Foreign Savings
Foreign savings in dollar terms is expressed in the model as the excess of payments for total imports over the sum of export earnings, net curent transfers and factor income from abroad. The latter two, it may be noted., are exogenously given values in the model.
2.10 Market equilibrium and macroeconomic closure
Market clearing equilibrium in the commodity markets is ensured by the condition that sectoral domestic supply must equal demand faced by that sector. The sectoral domestic supply, (i.e., domestic gross output) of a commodity is determined through the nested CES function in the production structure of the model. On the other hand, sectoral demand is a combination of domestic demand and export demand, based on a CET transformation function. In turn, the aggregate demand for a commodity – i.e., the sum of consumption, investment and government and intermediate demands - is equated to the demand for a composite commodity defined as an Armington type CES aggregation of domestic demand and imports.
The model is Walrasian in spirit with the sectoral prices being the equilibrating variables for the market-clearing equations. The Walras' law holds and the model is, therefore, homogeneous of degree zero in prices determining only relative prices. The exchange rate serves as the numeraire, and is, therefore, fixed at one.
Finally, note that although the model is neoclassical in nature, it follows investment-driven macro closure in which aggregate investment is fixed and the components of savings - household savings, government savings and foreign savings - are endogenous variables and adjust to equalize saving and investment.
2.11 Intertemporal adjustments
In the interim-period sub-model, the physical and human capital stocks are updated. Sectoral capital stocks are exogenously given at the beginning of a particular period. However, our model is recursively dynamic, which means that it is run for many periods as a sequence of equilibria. Between two periods there will be additions to capital stocks in each sector because of the investment undertaken in that sector in the previous period. More precisely, sectoral capital stocks
20
for any year t+1 are arrived at by adding the investments by sectors of destination, net of depreciation, in year t to the sectoral capital stocks at the beginning of the year t.
Between two periods there will be additions to human capital stocks also because of the public education expenditure undertaken in the previous period. More specifically, the output flow of labour of education level 'm', MSm , is an additive function of the education expenditure and the lifetime wage differential between wages at educational level 'm' and the next lower level 'l'9 . The function is specified as follows : ledtlρ−−)1t(l1( mW)++−−1t1g1
= β1 x + β2 x t GEDtWtr1mMS
where MSm : output flow of labour of education level m
GEDl : government education expenditure at level l,
Wm : Wage rate for labour of education level m
g : growth rate of the economy as a proxy for the growth rate of the wages.
r : discount rate
β1 , β2, ρedl : positive constants10
MLm : new labour supply of education level m
The flows of labour of different educational levels are interlinked with each other. From the pool of population growth (MS1), some proceed to secondary school (MS2), while others remain non-educated (ML1), and from secondary school, some advance to higher education (MS3), while others directly enter the labour market as secondary educated (ML2). Finally, higher-educated workers are produced and supplied (ML3). With the total increase of the labour force constrained to a fixed population growth rate, the new supply of non-educated labour (ML1) is determined residually. The labour flows are explained in the figure below :
9 For a detailed derivation of the function of the output flow of educated labour, see pages 704-708 of Jung and
Thorbecke (2003).
10 For β1 and β2 we have used the same values as Jung and Thorbecke (2003), which is 0.5 for each, and for ρedl we have used the values 0.44 and 0. 48 for secondary and higher education respectively.
21
ML3
MS3
ML2
MS2
ML1
MS1
Figure 1 : The labour flows
Note that the following relationships between the flows of labour types of different educational levels hold.
ML3 t = MS3 t ; ML2 t = MS2 t – MS3 t
ML1 t = n x Pt + ( dhΣ=31ll LSl t ) - ( ML2 t + ML3 t )
LSl (t+1) = LSl t ( 1- dhl ) + MLl t ; for l = 1,2,3 .
where P = population
n = labour participation rate
dhl = depreciation rate (retirees) of labour stock of educational level l
LSl = labour stock of educational level l
Higher-educated
Labour (level 3)
Population
Growth
Non-educated
Labour (level 1)
Primary
Education
Secondary-educated
Labour (level 2)
Higher
Education
22
3. The Base-Line Scenario
Our CGE model has been calibrated to the benchmark equilibrium data set represented in a SAM for the Indian economy for the year 1994-95. The SAM used for the present study is based on Pradhan, Sahoo and Saluja (1999). The SAM given here has been re-aggregated and modified to conform to the classification scheme of the production sectors, labour categories and the household groups, adopted in the model. The reaggregated and modified SAM is presented in Appendix 1.
Using the benchamark data set for the year 1994-95, we solve the CGE model first for the base-year, and, subsequently, using a time series of the exogenous variables of the model, we generate a sequence of equilibria for the period from 1994-95 to 2001-02. From the sequence of equilibria, the growth paths of selected (macro) variables of the economy are outlined to describe the base-line scenario.
3.1 Benchmark parameters
After having obtained the basic data set from the SAM, the CGE model is subjected to benchmark calibration. Calibration involves a deterministic approach to specifying parameter values in such a manner that the model solution replicates the base-year data (Shoven and Whalley (1992)). Calibration of the ‘shift’ and ‘share’ parameters of the production functions, CES aggregation function for imports and CET function for imports, however, require the elasticity parameters of these functions to be given. The elasticity parameters have been taken from different sources and are given below in table 3. Note that different types of labour are combined in two stages in the production structure to reflect different degrees of substituability. The skilled labour composite and non-educated labour are combined within a CES type Armington aggregation that has a small elasticity of substitution equal to 0.5 to yield composite labour. In turn, skilled labour composite is a CES Armington aggregation of secondary-educated and higher-educated labour based on a larger elasticity of substitution equal to 0.8. Through this labour aggregation scheme, the model is able to capture productivity growth caused by education. Note also that the higher wage income for the educated labourers results in higher share parameters for such workers in the calibration. Educated workers thereby contribute more to the composite labour. It follows that an increase in the supply of educated labour leads to a higher value for composite labour, resulting in higher production.
In table 4 we present the endowmnents of human capital across the nine household groups. It is interesting to note that most of the secondary and higher educated belong to the urban salaried
23
and urban non-agricultural self-employed groups. Almost 85 percent of higher-educated and 42 percent of secondary-educated workers come from these two groups. However, secondary-educated workers are more evenly spread over the urban and rural groups. Urban groups have 48.5 percent of the secondary-educated workers and rural groups have 52.5 percent of the educated workers. (It may be noted that, in our classfication of educated workers, secondary-educated includes elementary, secondary and higher- secondary educated. The disitribution of workers within these three levels of education is not shown in the table.)
Table 3 : Elasticity Parameters
ρ1 ρ2 ρ3 ρa ρc εex
s1 Agriculture 0.7800 0.5000 0.8000 1.1387 0.9200 0.8400
s2 Mining 1.3200 0.5000 0.8000 1.6195 0.4600 0.8600
s3 Manufacturing 1 0.7420 0.5000 0.8000 2.2470 1.7000 1.2300
s4 Manufacturing 2 0.9682 0.5000 0.8000 2.7368 1.3855 1.1739
s5 Construction 1.1000 0.5000 0.8000 0.0000 0.0000 0.0000
s6 Elec. Gas & W.S. 2.2600 0.5000 0.8000 0.0000 0.0000 0.0000
s7 Trans. & Stor. 1.4500 0.5000 0.8000 2.1450 0.9200 1.3200
s8 Whole & Ret.Trade 1.4500 0.5000 0.8000 2.1450 0.9200 1.2800
s9 Fin., Ins. & Real Es. 1.6500 0.5000 0.8000 2.1450 0.9200 1.3600
10 Comm.,Soc.& Per. Servs. 1.0800 0.5000 0.8000 0.7150 0.3067 0.6667
Note : ρ1 : elasticity of substitution between composite labour and capital.
ρ2 : elasticity of substitution between skilled labour composite labour and uneducted labour.
ρ3 : elasticity of substitution between secondary-educated labour and higher-educated labour.
ρa : elasticity of substitution between domestic demand and imports.
ρa : elasticity of substitution between domestic sales and exports.
εex : export demand elasticity
Source : Jung and Thorbecke (2003) and Chadha et al (1999) .
24
Table 4 : Resource endowment shares in percentages
Non-educated
labour Secondary-educated
labour Higher-educated labour Physical Capital
RC 20.34 13.98 2.65 27.34
RATN 19.54 4.33 0.63 10.06
RAL 31.02 11.59 0.32 0.33
RO 14.69 21.68 9.45 2.61
UF 1.37 0.50 0.00 1.00
UNASE 2.59 8.86 8.79 14.16
US 6.64 33.30 75.73 6.18
UCL 3.25 5.02 0.90 1.53
UO 0.55 0.75 1.54 3.64
100.00 100.00 100.00 66.86
Note : RC : Rural Cultivator ; RATN : Rural Artisan ; RAL : Rural Agricultural Labourer ;
RO : Rural Others ; UF : Urban farmer ; UNASE : Urban Non-agricultural Self-employed ;
US : Urban Salaried ; UCL : Urban Casual Labouer ; UO : Urban Others.
Physical capital endowment includes that of land. Capital column sums upto only 66.86% because
the remaining 33.14% accrues to private enterprise, public enterprise, government and the rest of world.
Source : Calculations from MIMAP India Survey, 1996, NCAER.
3.2 Labour supply and wage levels
In the base-line scenario, labour supply grows annually at the rate of 1.84 percent (table 5). Among the three types of labour, the supply of higher educated workers grows fastest at the rate of 4.94 percent, followed by secondary-educated workers’ supply which increase at the rate 3.66 percent. The supply of non-educated labour, which is determined residually, grows by only 1.04 percent annually. It would seem that the 8.31 percent and 9.34 percent annual growth in real public expenditure on secondary and higher education respectively is making a positive impact on the supply of educated workers.
Regarding wage levels, there is maximum improvement in the non-educated workers’ wage rate which increases by 3.86 percent annually. Education expenditure benfits the non-educated labour indirectly, by inducing a relative decrease in its supply. Secondary-educated workers’ wage rate also grows fast at 3.57 percent . The wage rate of higher-educated workers increases at only 3.07 percent per annum. The wage rates of secondary and higher educated workers rise despite the increase in their supplies because the techniques of production become more skill intensive as the economy grows over time (table 5).
25
Table 5 : Baseline : Labour supply, wage rates and public education expenditure
Average annual growth rates for 1994-95 to 2001–02 in percent
Labour Supply 1.84
Non-educated labour 1.04
Secondary-educated labour 3.66
Higher-educated labour 4.94
Wage rate (real) 4.55
Non-educated labour 3.86
Secondary-educated labour 3.57
Higher-educated labour 3.07
Public education expenditure (real) 8.47
Secondary education 8.31
Higher education 9.34
Table 6 : Baseline : Wage rate indexes
Wage rate as a multiple of non-educated worker’s wage rate
1994-95 2001-02
Wage rate (real)
Non-educated labour 1.00 1.00
Secondary-educated labour 1.98 1.95
Higher-educated labour 7.55 7.16
The higher rate of growth of the non-educated worker’s wage notwithstanding, the wage inequality across the three types of labour – particularly between non-educated and higher-educated labour - remains acute at the end of the seven-year period (see table 6). This is mainly due to the extreme inequality of wages of the three types of labour prevailing at the beginning of the period.
3.3 GDP and household income
Real GDP in the base-run grows at 5.99 percent per annum, with investment in physical capital being on an average 28.35 percent of GDP. The rate of growth of wage income is 2.45 percent higher than that of the capital income (table 7).
Household income as a whole grows at 5.64 percent per annum. But the rates of growth of incomes vary widely across the various household groups. The rate of growth of incomes of the urban salaried class is, expectedly, the highest – i.e., 7.35 percent. Urban salaried households are
26
the greatest beneficiaries from the spread of education. These households account for 75.75 percent of the higher-educated and 33.30 percent of the secondary-educated labour (see table 4). Urban non-agricultural self-employed improve their incomes at the rate of 5.17 percent per annum. This class also depends largely for its income on secondary and higher educated labour. Another group, not so expected, which benefits from the spread of education is rural others. This group is endowed with 21.68 percent of the secondary–educated workforce and 9.45 percent of higher-educated workforce. However, the non-beneficiaries of education – i.e., those having mainly non-educated labour as a source of their income – are also significantly better-off, thanks to the rise in the wage rate of non-educated labour. For example, household incomes of the rural agricultural labourers grow at 5.28 percent per annum. Urban casual labourers, who are to a large extent though not mainly dependent on non-educated labour, also increase their incomes by 5.49 percent per annum11.
Table 7 : Baseline : GDP and household income
Average annual growth rates for 1994-95 to 2001 –02
(in percent)
GDP (real) 5.99
Investment (% of GDP) 28.35
Wage Income (real) 6.57
Capital Income (real) 4.12
Household Income (real) 5.64
Rural Cultivator 4.70
Rural Artisan 4.71
Rural Agricultural Labour 5.28
Rural Others 6.07
Urban Farmers 4.66
Urban Non-ag. Self-Employed 5.17
Urban Salaried 7.35
Urban Casual Labourer 5.49
Urban Others 4.85
11 Note that wage income is allocated to each household group on the basis of the base-year endowment shares for all the years. That is, the flow of new labour types is distributed across household groups in the same way as the whole labour stock.
27
4. The Policy Simulations
We develop three alternative policy scenarios for an expansion in the public education expenditure. In all the the three simulations, the increase in public education expenditure is financed by an increase in the direct taxes – i.e., income and corporate tax. In fact, the increase in public education expenditure is implemented in a manner suggested by Mehrotra (2004). That is, we increase the income and corporate taxes by a specified percentage and dedicate the resulting additional revenue to public spending on education. The mode of financing remains the same in all the three simulations, but the mode of expenditure varies across them. In the first simulation, the additional expenditure on education is distributed between secondary and higher education in the same proportions as in the total expenditure of the base-line scenario. In the second scenario, the extra expenditure is directed exclusively towards secondary education. In the third policy scenario, the additional revenue from the specified increase in tax rates is shared equally between investment in physical capital and education expenditure concentrated in the secondary education sector.
4.1 Policy simulation 1
In this simulation, we increase the rates of income tax and corporate tax by 10 percent and use the additional revenue for increased public spending on secondary and higher education in the same proportions as in total public education expenditure of the base-run. By this mechanism, the 10 percent increase in the two direct tax rates, results in a 14.40 percent increase in real public education expenditure over the base-run. And public education expenditure as a percentage of GDP, increases by 0.43 percentage point compared to the base-run.
28
Table 8 : Simulation 1 : Labour supply and wage rates
Average annual growth rates for 1994-95 to 2001 –02
(in percent) Diff.from base-line
in %age points
Simulation 1 Baseline Simulation 1
Labour Supply 1.84 1.84 0.00
Non-educated labour 0.61 1.04 -0.43
Secondary-educated labour 4.01 3.66 0.35
Higher-educated labour 5.26 4.94 0.32
Wage rate (real) 4.57 4.55 0.02
Non-educated labour 5.13 3.86 1.27
Secondary-educated labour 3.02 3.57 -0.55
Higher-educated labour 2.57 3.07 -0.50
In policy scenario 1, the growth rate of secondary and higher educated labour supply goes up by 0.35 and 0.32 percentage points respectively, but that of the non-educated labour supply goes down by 0.43 percentage point, since it is determined residually. As a result non-educated workers become relatively more scarce and improve the growth rate of their wage rate by 1.27 percentage points. The secondary and higher educated workers are supplied more abundantly and, therefore, suffer a decline in the growth rates of their wage rates by 0.55 and 0.50 percentage points respectively (table 8). The inequality in the wages also narrows down a little, with the higher and secondary educated workers receiving wages which are respectively 6.35 times and 1.73 times the wage of the non-educated workers (table 9).
Table 9 : Simulation 1 : Wage rate indexes
Wage rate as a multiple of non-educated worker’s wage rate in 2001-02
Simulation 1 Baseline
Wage rate (real)
Non-educated labour 1.00 1.00
Secondary-educated labour 1.73 1.95
Higher-educated labour 6.35 7.16
With a 14.40 percent increase in public education expenditure, GDP growth rate improves by 0.17 percentage point. Investment as a percentage of GDP declines marginally, since its level is fixed exogeously and remains the same as in the base-run. As a result, capital, in comparison to educated labour whose supply increases, becomes more scarce. Hence, capital income growth rate increases by twice as many percentage points as the increase in the wage income growth rate (table
29
10). Household income also grows faster by 0.13 percentage points. An inter-group comparison of the household income growth rates reveals that all groups experience a faster growth in their incomes except, the urban salaried and the rural others, who suffer a decline in their income growth rates as a consequence of the fall in the growth rates of the wages of secondary-educated and higher-educated workers. It may be noted that these two groups are the ones experiencing the highest growth rates in their incomes in the business-as-usual scenario. Hence, a decline in their income growth rates in the face of a rise in the income growth rates of the remaining groups represents a distinct change towards greater equalisation of incomes.
Table 10 : Simulation 1 : GDP and household income
Average annual growth rates for 1994-95 to 2001 –02
(in percent) Diff.from base-line
in %age points
Simulation 1 Baseline Simulation 1
GDP (real) 6.16 5.99 0.17
Investment (% of GDP) 27.65 28.35 -0.70
Wage Income (real) 6.63 6.57 0.06
Capital Income (real) 4.24 4.12 0.12
Household Income (real) 5.77 5.64 0.13
Rural Cultivator 4.88 4.70 0.18
Rural Artisan 4.85 4.71 0.14
Rural Agricultural Labour 5.46 5.28 0.18
Rural Others 5.98 6.07 -0.09
Urban Farmers 4.59 4.66 -0.07
Urban Non-ag. Self-Employed 5.24 5.17 0.07
Urban Salaried 7.24 7.35 -0.11
Urban Casual Labourer 5.58 5.49 0.09
Urban Others 5.08 4.85 0.23
Note : The fast movers – i.e., those household groups having income growth rates higher than 6% in the
base-line - are shown in italics.
30
4.2 Policy simulation 2
In this simulation, we increase the rates of income tax and corporate tax by 10 percent and use the additional revenue for increased public spending exclusively on secondary education. By this mechanism, the 10 percent increase in the two direct tax rates, results in a 17.53 percent increase in real public expenditure on secondary education over the base-run. For public expenditure on education as whole the increase is of 14.47 percent. As a percentage of GDP, the increase in expenditure on elementary education is by 0.41 percentage point.
In policy scenario 2, supply of secondary-educated labour goes up while that of non-educated labour goes down like in simulation 1. But the order of magnitudes involved are higher in case of this simulation. In comparison to the base-run, the rate of growth of supply of secondary-educated labour increases by 0.52 percentage point, while that of non-educated labour declines by 0.48 percentage point. The growth rate of higher-educated workers also declines marginally. The improvement in the wages of the non-educated labour is, as compared to the base-run, much faster. That is, the rate of growth in their wages is 5.18 percent, whereas it was only 3.86 percent in the base-run. For secondary-educated labour, which is now more abundantly supplied, there is a fall in the growth rate of wages. It may be noted that in this scenario, there is a significant substitution in production in favour of secondary-educated labour vis-à-vis higher-educated labour. And this explains why there is a marginal decline in the growth rate of the higher-educated worker’s wage even as higher-educated labour becomes relatively more scarce. The wage rate inequality shows some improvement as the rate of growth of non-educated labour rises and that of the secondary-educated labour falls, but the higher-educated labour still earns a wage which is more than 6.5 times that of non-educated labour (table 12).
Table 11 : Simulation 2 : Labour supply and wage rates
Average annual growth rates for 1994-95 to 2001 –02
(in percent) Diff.from base-line
in %age points
Simulation 2 Baseline Simulation 2
Labour Supply 1.84 1.84 0.00
Non-educated labour 0.56 1.04 -0.48
Secondary-educated labour 4.19 3.66 0.52
Higher-educated labour 4.89 4.94 -0.05
Wage rate (real) 4.62 4.55 0.07
Non-educated labour 5.18 3.86 1.32
Secondary-educated labour 3.08 3.57 -0.49
Higher-educated labour 3.03 3.07 -0.04
31
Table
Subscribe to:
Post Comments (Atom)
Established in 1997, St. Gregorious Edu-Guidance is a leading education consultancy services providing exemplary service to students all over India. We deal in Admissions to all major professional courses in Premier Institutes across India. We are your one step solution for all career related needs, it may be MD, MBBS BE, BTech (ALL BRANCHES), , MDS, BDS, BPharm, BArch, MBA, MTech, MS, , PhD or any other courses. We provide personalized career solutions on an individual basis keeping in mind the aspirations of our client as well as the affordability factor.
ReplyDeleteFOR ALL CAREER RELATED NEEDS CONTACT US :
St. Gregorious Edu-Guidance,
#2, 2nd Floor,
J J Complex, Above Chemmannur Jewellers,
Marthahalli - P O,
Bangalore - 560037
Karnataka
e-mail :jojishpaily@gmail.com
Contact: +91 9448516637
+91 9886089896, +91 9449009983
080-32416570, 41719562
WEBSITE: www.stgregoriousedu.com