Growth without Development
Persistent Income Inequality in Latin America from a Dualistic Perspective
20-05-2005
Master Thesis General Economics by: Gaaitzen de Vries
supervisor: dr. M.P. Timmer at: University of Groningen Faculty of Economics
Preface
This thesis originated two years ago. At that time I visited Costa Rica and was perplexed by the very unequal standards of living. A year thereafter, at the time I wanted to start writing my master thesis, Marcel Timmer told me I could study upon this strange and unpleasant characteristic of not only Costa Rica, but Latin America in general. This marked the beginning of research I conducted, which eventually resulted in this thesis.
The data of my thesis, is collected during my internship at the United Nations Economic Commission of Latin America (UN ECLAC). This internship at the UN ECLAC has been very stimulating and highly worthwhile. Furthermore, the data collected has proven to be indispensable for the empirical part of this thesis.
I would like to thank Marcel Timmer for his continuous guidance, and the structure he gave to this thesis. We have had many long discussions and these I really enjoyed. I would also like to thank André Hofman, for providing me the opportunity of an internship at the UN ECLAC and being a splendid mentor during my stay in Santiago, Chile.
Finally, I would like to thank Corina Geerds for being a constant source of luck and joy, and I want to apologise to her that at many times I was “away in thoughts” about this thesis. Hopefully I will be able to make up for this as the summer is approaching!
Groningen, May 2005
Contents
Page
Preface 3
Chapter 1 – Introduction 7
1.1 Hypothesis and Methodology 9
1.2 Structure 11
Chapter 2 - Dualistic Economic Development 13
2.1 Structural Change 13
2.2 Dualistic Economic Development 14
2.3 Criticism 15
2.4 Dual-Dual Framework 17
2.5 Dualistic Development and the Income Distribution: Theory 19
2.6 Summary 23
Chapter 3 - The Informal Sector 25
3.1 Origins of the Urban Informal Sector in Latin America 25
3.2 Informal-Formal Sector relationships 26
Chapter 4 - Measurement and Data 31
4.1 Measuring Changes in the Industrial Structure 31
4.2 Measuring the Informal Sector 32
4.3 The Fields Decomposition 35
Chapter 5 – Results 36
5.1 Changes in the Industrial Structure 37
5.2 Enlargement effects 38
5.3 Enrichment effects 41
5.4 Fields’ Decomposition Analysis 42
Chapter 6 - An Exploration into the Economic Determinants 59
Chapter 7 – Conclusions 65
References 69
Chapter 1 – Introduction
Visiting Santiago, Chile, I became aware of some striking features of everyday life. The centre and east side of the city are like cities in the western world. Supermarkets, banks, parks, sport facilities, fast food restaurants and even huge golf courses are present. Yet, when I went from these areas to the airport in the west, I was confronted with vast slum areas. For the unfortunate who live in these quarters, each day is a survival. It is astonishing that some people living in these slums have no electricity and no sewer system while a couple of miles eastward, other people live in huge villas enjoying the pleasures life can give.
I found this marking disparity in living standards to be common throughout the whole Latin American region. Costa Rica, considered as one of the most egalitarian economies in the region, is not at all a positive exception. Costa Rica’s main capital San José has very extensive slums where many of the approximately one million refugees from Nicaragua (entailing almost 20% of the total population) live and treated as outcasts.
Differences in living standards in urban areas are also present in rural areas. Particular characteristic for rural areas are the disparities in plots of land. Todaro and Smith (2003) give some statistics that bear evidence on this inequality. Latifundios in Argentina (latifundios are farms with more than 12 employees) comprise 0.8% of the total number of farms. Yet they occupy 36.9% of the total cultivatable land in Argentina. For Peru, these figures are more drastic as 1.1% of the farms are latifundios, cultivating 82.4% of the available land. Hence, many farmers living in rural areas cultivate only a tiny piece of land.
Some countries have less severe disparities in plots of cultivatable land.1 However, as for the slums and villa neighbourhoods in urban areas, very small farms and massive haciendas are coexisting in rural areas and it is uniform throughout the Latin American region.
Latin American economies are a special case, as living standards are very unequal distributed in comparison with the rest of the world. Hofman (1998) presents Gini coefficients on pre-tax income of households. The arithmetic average Gini coefficient for the Latin American region is 0.528 in 1970 (ranging from 0.412 in Argentina to 0.630 in Brazil). For Korea, Taiwan, USA and the Netherlands the 1970 Gini coefficient is respectively 0.351, 0.396, 0.404 and 0.385. Evidently, Latin American income distribution is far more unequal.
1 This is for instance the case in Mexico, Bolivia and Cuba due to land reforms.
However, not only is the Latin American region marked by an extremely unequal income distribution, high inequality has also been very persistent (and even slightly rising) over time. This high and persistent inequality is displayed in figure 1.1. Data here is taken from Worldbank (2004) and Gini coefficients are calculated using a non-weighted distribution of household per capita income.
Figure 1.1 Gini coefficients, 1950-1992
35 40 45 50 55
1950 1960 1970 1980 1992
Year
Gini coefficient
Africa Asia Latin America Eastern Europe Developed Countries source: Worldbank (2004)
Studying upon income inequality is considered important for both normative and functional reasons.
At a normative level, there seems to be no ethical excuse why some are born in a very poor family and others in a wealthy family. Equal treatment for all humans is an ethical norm in various religions (Judaism and Christendom), but also appears more and more as a sociological norm and this is reflected in a leftist political tendency in Latin America where political leaders more and more express their willingness to make an end to high income inequality.
Inequality is also important with respect to its functioning in a society. There is for example growing evidence that inequality hinders economic growth (Barro, 1991, Alesina and Rodrik, 1994). Moreover, inequality determines the possibility for people to participate in the economy and inequality is therefore closely related to poverty.
Various studies tried to explain high and persistent income inequality in Latin America. From these studies, four main approaches can be distinguished: historical, regression analyses, technology and globalisation, dualistic theory. The historical approach traces current income inequality in Latin America back to the colonial period. An important element in the historical approach is the initial unequal distribution of assets that was inherited from the colonial period, thereby hinting at path dependency of inequality (Cardoso and Helwege, 1992, Worldbank, 2004). Another approach taken in
literature in order to explain inequality in Latin America, has been by means of econometric regression analysis. Basically, authors within this approach use Gini coefficients and test the sign of the parameter and the importance of several explanatory variables (Barro, 1999, De Javry and Sadoulet, 2000). A third approach, is with regard to technology and globalisation. It is argued, that technological developments and the increasing flow of goods and services led to a relative increase of demand for skilled labourers, and this increase in demand for skilled labourers caused wages to disperse because of the greater supply of low skilled labourers in Latin America (Cragg and Epelbaum, 1996, Esquivel and Lopez, 2003, Vries, 2004, Hanson, 2003). A fourth approach is by regarding the development of Latin America as dualistic. This thesis is embedded in that approach.
According to many authors (Lewis, 1972, Thorbecke, 1997, Tokman and Klein, 1996, Ranis, 2003), one of the most distinguishing characteristics of Latin America’s development is the dualistic nature of its growth path. Authors (Tokman 1978, PREALC, 1990) have noted, that alongside the development of an urban formal industrial sector, an urban informal sector appeared as well. This is in contrast to development in for example Asia (Chowdhury and Islam, 1995). With development, countries in Asia saw an initial rise in income inequality and a decline at a later stage. In Latin America, income inequality was high and persistent during 1950-2003. It therefore seems natural, to investigate the relationship between persistent high income inequality and dualistic economic development in Latin America. This is exactly what I will do in this thesis.
1.1 Hypothesis and Methodology
In this thesis, I will use the dual-dual framework developed by Santiago and Thorbecke (1984), and link this into a framework of income distribution proposed by Fields (1980). The crucial distinction to be made is between between-sector inequality (agriculture and industry) and within-sector inequality (formal and informal). Income earned in the formal sectors is generally much higher than in informal sectors, and income earned in industry is generally much higher than in agriculture. Hence, with a shift of labourers from the rural agricultural sector to the urban sector, there might be little change in income inequality in a country when most of the labourers end up in the urban informal sector.
Special emphasis in this thesis is put on dualistic economic development. First, because dualistic theory distinguishes a formal (modern) and an informal (traditional) sector. Traditional growth models only consider one sector (Heijdra and van der Ploeg, 2002) and are therefore of little use for this thesis. Second, an investigation into the determinants that dictate the pace and direction of the dualistic process, helps explain the persistence of the informal sector. Normally, dualistic theories describe the final disappearance of the traditional sector(s). Yet, they also encompass the possibility of structural
heterogeneity.2 If informal (traditional) sectors (characterized by low levels of labour productivity) are existing structurally, it provides a reason for the remaining high levels of income inequality in Latin America.
In particular, I want to study whether the formal and the informal sectors became larger or enriched in Latin America from 1950-2003. If the formal sectors became larger, economic growth occurred because of the higher income in the formal sectors, and with enlargement of the formal sector the economy also developed. However, when the formal sector only became richer while their employment share remained unchanged, the economy grew (due to increasing income in the formal sector), but without enlargement of the formal sector the economy did not develop.
• The main research question is: what development pattern characterises Latin America and how is this related to changes in the income distribution?
This research question entails finding an answer to what happened with between-sector inequality and with within-sector inequality. In order to study between-sector inequality, I have to analyse whether a shift from agricultural employment towards the industrial and service employment took place. For analysing within-sector inequality, I have to study the size and wage-income of the formal and informal sectors.
Therefore, in order to study upon this research question, I constructed:
• A sectoral database for Latin America, which enables me to analyse whether the industrial structure changed.
• Time series on the size and wage-income in the formal and informal sectors, which enables me to analyse whether the formal sector enlarged or became richer.
Beside these two main activities, I analyse the economic determinants that determine the direction of dualistic development and also structure the views upon the informal sector in a clarifying framework.
This thesis is a valuable addition to academic literature. The constructed sectoral database for Latin America, and the time series on the size and wage-income in the formal and informal sectors allow me to study economic development and the growth path, which characterized Latin America. It is this thesis, which is the first in examining Latin America’s growth path empirically.
2 The concept of structural heterogeneity was developed at the Economic Commission of Latin America and the Caribbean. Along with the notion of structural heterogeneity, ECLAC terms the absorption of informal sector workers into the formal sector as productive labour absorption. Of course, while being different in terminology, it is exactly this process that also became described in academic literature on dualistic theory.
1.2 Structure
The remaining part of this thesis is structured as follows. In chapter 2, dualistic economic development is studied upon. In that chapter, the typical pattern of changes in the industrial structure are studied upon from a dualistic perspective. Several critiques have been raised against dualistic theory. These are described and, when possible, answered. Also, the original dual model is translated into a dual- dual framework, which closer reflects Latin America’s reality. Finally, the relationship between dualistic economic development and the income distribution is studied upon in that chapter. Here I distinguish between four dualistic growth paths, with different implications upon the income distribution. Chapter 3 goes on to describe the informal sector and what views have been expressed upon the informal sector. In that chapter, I also study the implications upon the income distribution of the views upon the informal sector. Chapter 4 outlines the methodology to measure the size and wage- income in the formal and informal sectors. Furthermore, I examine similarity indexes. These similarity indexes give insight into the actual occurrence of a change in the industrial structure. In chapter 4, I also describe the methodology to study the dualistic growth path, which characterises Latin America’s economic development. The results of my empirical effort are presented in chapter 5. In that chapter I show, that in all Latin American countries the industrial structure changed. I also show that the informal sector continued to exist, and hence enlargement of the formal sector does not particularly characterise Latin America. Furthermore, it is shown that enrichment effects in the formal and informal sectors occurred in several countries. These findings of enlargement and enrichment, are related to income inequality in the countries studied in chapter 5. In chapter 6, I analyse the determinants which determine the direction of dualistic economic development. These determinants explain the persistence of informal sectors in Latin America. Finally, Chapter 7 presents conclusions and suggestions for further research.
Chapter 2 - Dualistic Economic Development
The process of structural change is the main focus of attention in this chapter. Paragraph 2.1 starts with a description of the general process of structural change, occurring as an economy grows. A dualistic view on this changing structural composition of an economy, was first described by Lewis (1954). In paragraph 2.2 the establishment of this view is described. It has become known as a theory of dualistic development. Criticism upon dualistic theory and further extensions are presented in paragraph 2.3. In paragraph 2.4, the traditional dualistic model is applied to the situation in Latin America. Finally, the relationship between dualistic growth paths and income inequality is established in paragraph 2.5.
2.1 Structural Change
Many studies document the process of a changing structural composition of the economy with economic development (Fisher, 1948, Maddison, 1980, Mulder, 2002, Timmer, 1999). Mulder (2002) argues, that Sir William Petty already in 1691 acknowledged the changing labour force composition in England. Fisher (1948) was a pioneer in explaining this changing composition. In essence, the process of structural change is one in which labour moves from agriculture into manufacturing and (increasingly) services. In a study on the development of 16 OECD countries during 1870-1970, Maddison (1980) describes some general trends in structural change. For the 16 OECD countries studied, employment shows a continuous declining share in agriculture. On the other hand, the employment share in services is continuously rising. The employment share of industry differs over the course of development. First the share rises, but at a certain point its share starts to decrease. This might be considered as a bell-shaped pattern of the employment share of industry. This pattern mainly happens due to Engel-effects. When people become richer, additional demand first shifts from agricultural products towards industrial products, but more and more towards services.
Structural change will contribute to economic growth if the expanding sectors are associated with a higher labour productivity. Usually the industrial- and service sectors (at least) initially have higher labour productivities then agriculture,3 and therefore the process of structural change is normally considered as economic growth in itself. This growth enhancing process, due to the expansion of industrial and service sectors, was fundamental to the theory of Lewis on dualistic economic development. The theory of dualistic economic development is studied in paragraph 2.2.
3 Mulder (2002) presents evidence for a convergence in sectoral labour productivity for Brazil and Mexico during the course of development. For further empirical evidence, see chapter 5.
2.2 Dualistic Economic Development
Early models from the 1920s and 1930s on economic development did not take the process of structural change into account. Models developed considered only one sector.4 Lewis (1954) acknowledged the growth potential associated with structural change and analysed the economy from a two-sector perspective. Yet, Lewis’ departure from one-sector models originated from his dissatisfaction with the neoclassical view on the functioning of the labour market. He took his standing in the works from classical authors (Smith, Ricardo and Marx) stating: “This essay is written in the classical tradition, making the classical assumption, and asking the classical question.”5 With this classical standing he departed from neoclassical economists.
The core of his departure from neoclassical economy was formed by his view on the functioning of the labour market. This resulted in a different view on the marginal productivity of labour. Neoclassical economists assume the labour market operating under competitive conditions and hence wages are equal to the marginal product of labour. According to Lewis this view on the functioning of the labour market might be applicable to “the European parts of the world”, but is not for “the greater part of Asia.” On the contrary, Asian countries are characterized by a huge supply of labour that could function as a reservoir for workers: an unlimited supply of labour.
This unlimited supply of labour is created by population growth and female household members, which enter the labour market. Population growth is high because many countries are in the second demographical transition phase where birth rates are high, and death rates drastically drop due to increased health.6 Female household members is another source that could be transferred from households into employment.7 This transfer is to some extent an economic gain, because many activities by females in the household can be done better and cheaper outside. A strict neoclassical consequence of an unlimited supply of labour would be that labour is used until the marginal productivity is zero and therefore workers would work without any wage. To meet this troublesome outcome, Lewis came to state the dualistic nature of developing countries.
His theory is denoted dualism, as two distinguishable sectors initially characterize an economy. First there are people who live in the subsistence (agricultural) sector. This subsistence sector is where the reservoir of labour is located and people in the subsistence sector earn a wage that is just enough to
4 These one-sector models are for example the Solow-Swan model, the Harrod-Domar model and the Ramsey- model of economic growth (Heijdra and van der Ploeg 2002).
5 Lewis (1954), p. 139.
6 In the first phase both birth and death rates are high, whereas in the third phase birth and death rates are low
7 In 1954 this was still an important source for economic development in England even though the industrial revolution started here.
stay alive: a subsistence wage. The subsistence wage is the average of income generated by families.
Various family workers deliver products or services and income accruing due to these activities are shared among the family members. According to Lewis, productivity in this subsistence sector is generally low. Second, there is a capitalist (industrial) sector.8 The capitalist sector is distinguishable from the subsistence sector because it maximizes profit and uses reproducible capital. Furthermore, productivity is higher in the capitalist sector. The capitalist sector employs labour that is extracted from the subsistence sector with an “unlimited supply of labour,” and pays these workers the subsistence wage plus a margin.9
The core aim of a dualistic model is to describe the process by which the capitalist (industrial) sector expands and eventually exhausts surplus labour in the subsistence (agricultural) sector. At early stages the economy is considered as being traditionally agricultural and the modern industrial sector is virtually non-existent. If the industrial sector is able to expand and eventually catches up with the supply of labour, the economy no longer corresponds to the classical propositions on the functioning of the labour market. Here “we are back in the world of neo-classical economics.”10 Both capital and labour are scarce and characterized by an inelastic supply. Dualism has come to an end.
Dualistic theory is heavily criticized. The next paragraph studies the major lines of critique and important extensions of dualistic theory.
2.3 Criticism
Researchers put serious doubt upon some fundamental concepts of dualistic theory described in paragraph 2.2 (Ghatak, 1995). Five main areas of critique stand out in literature: the saving pattern, the role of education, the role of agriculture, technology, and the universal applicability. They are treated in turn.
The saving pattern
The marginal propensity of saving in the capitalist sector is assumed to be close to one in dualistic theory. However, a rise in (capitalistic) profits is always accompanied by a more than proportional
8 A remark is in place for Lewis noted a subsistence sector and a capitalist sector. However, the subsistence sector as being a sector that does not use reproducible capital has been rather unsatisfyingly described by Lewis.
Therefore it was tempting to adopt an agriculture-industry distinction (Leeson 1979). This view on the dualistic nature was adopted by Ranis and Fei (1961), who extended and enhanced the theory of dualistic economic development. Later, Lewis (1972) also went along with the division of sectors into agriculture and industry, but he also introduced another terminology. Lewis (1979) then referred to the two sectors as a modern and a traditional sector.
9 This margin is paid to compensate workers. For example for the negative effect of pollution and congestion in cities.
10 Lewis (1958), p. 26.
increase in consumption and therefore the total increase in saving is less than the increase in profits.
Furthermore, recent writers on savings behaviour have proposed that the marginal rate of saving is highest for the middle class (as compared to savings by the poorest and the richest). High savings result from high aspirations by the middle class (Ray, 1998).11 This saving pattern would have its repercussions on the development of the capitalist sector. If the middle class saves, it is this class that is needed for the expansion of the capitalist sector. This naturally raises the question what might be considered the middle class. The middle class of Latin America primarily seems to consist of administrative employees, technicians and less-educated professionals, whose income is between those with very low incomes (e.g. workers in commerce and machine operators) and very high incomes (e.g.
senior professionals, managers and senior officials, Social Panorama (1999)).
Education and skills
Workers move from agriculture to industry, which is according to dualistic theory a smooth process.
However, industry requires different types of labourers. Most workers have to be trained before being able to work in industry. Therefore investments in education should occur, and because of this, it is not a smooth process and neither is it costless (Ghatak, 1995).
The role of agriculture
As dualistic theory applied the agriculture-industry distinction, it did not take full account of the seasonality of labour demand in agriculture. Agricultural demand is higher during sowing and harvesting periods (if crops are cultivated). It is in these peak periods where the marginal productivity of labour in agriculture might be positive and therefore a transfer of labour to industry reduces agricultural output. Jorgenson (1961) undertook a neo-classical attack on the classical dualistic theory along this line. His approach differed from classical dualistic theory as Jorgenson assumed the marginal product of labour never to be equal to zero. The fundamental difference, however, is that Jorgenson turns classical dualistic theory around in terms of the importance of the sectors (Dixit, 1973). In the classical dualistic models, capital accumulation in the capitalist (industrial) sector is the main determinant. Mainly capital accumulation determines the rate at which labour is absorbed in industry. The neo-classical model by Jorgenson turned this around as he states that agricultural surplus is fundamental for economic development. It is this surplus in agriculture that allows labour to be released. Hornby (1968) argued that food requirements of the capitalist (industrial) sector might be met by imports. This eases the marketable surplus requirement in the agricultural sector set forth by Jorgenson (1961), and also the argument of the seasonality of labour demand in agriculture. Import of food can then be balanced against the export of industrial products.
11 Because of the limited scope for the poorest to improve their living conditions, they are not inspired to save.
Technology
Dualistic theory assumes accumulation of profits in the capitalist (industrial) sector to raise employment. However, there might be no enlargement of the capitalist sector if the capitalist sector chooses to invest in technology which is intensive in capital and not labour. Investing would then cause labour productivity to rise, but not employment.
Both Lewis (1954) and Ranis and Fei (1961) expressed a somewhat simplistic view on technological progress. In their dualistic models, technological progress runs in the same direction as capital accumulation. According to them, at least in their early works, technological progress leads to an expanding formal sector in terms of employment. Dixit (1973) refined their view in two ways:
First, technological progress in agriculture might not only lead to an increase in surplus-labourers, but may also lead to an increase in the marginal product. This could cause the surplus-labour phase to end without an absolute decline in agricultural employment. Second, necessary for the capitalist sector to catch up with population growth, is labour-augmenting technological progress. This second refinement from Dixit (1973) is also studied by Kelley et.al. (1972). According to Kelly, technology in industry is imported, and because technology is imported from labour-scarce countries, it is labour-saving technology that is adopted in industry. In agriculture, technology is endogenously determined. Here, use is made of the abundant factor labour and labour-augmenting technological progress takes place.
Thereby, the process of dualistic development not simply depends upon technological improvements, but upon the particular form. In order for the capitalist sector to enlarge, technology has to be relatively intensive in the use of labour.
Is dualistic theory universally applicable?
It is often argued that the dual economy model performs quite well in describing the economic development of some countries in Asia, but it’s applicability to Latin America is doubted (see for references on this critique Ranis 2003). Critics sometimes do acknowledge the development process in Asia as being a process in which the modern industrial sector emerged and in which the traditional agricultural sector was the main supplier of labour and food for the emergence of the industrial sector.
It is however not a general model, these critics argue. Rather it is a view on economic development in Asia. Even more, founders of dualistic models also seem to focus their attention to Asia.12 The next section establishes a dualistic framework which applies to Latin America.
2.4 Dual-Dual Framework
The dualistic approach on Asia can be translated into a view on dualistic development for Latin America. This translation is for example found in the studies of Thorbecke (1984, 1997, 2003).
12 See for example Lewis (1954, p.139), the study of Ranis and Fei on Taiwan (1972), and also Jorgenson (1961, p. 311 footnote 1).
Thorbecke splits an economy on regional and organisational grounds. By doing this, Thorbecke distinguishes between a formal and an informal sector in both rural and urban areas. This approach is known as the dual-dual approach, as there is dualism in urban areas and dualism in rural areas. It is the heterogeneity noted in the introduction, being both in rural and urban areas, that seems to be much more in line with this dual(urban)-dual(rural) approach to Latin America in stead of a dual approach.
The remaining part of this paragraph further describes the dual-dual framework.
Dualism in urban areas
The industrial sector in dualistic theory distinguishes itself from agriculture, because it is concentrated in urban areas and is formally organised. There has been critique upon this stringent view of dualism.
This critique became significant with the famous Harris-Todaro model of rural-urban migration developed in the late 1960s.13 The basic equation from their model is the equalisation of the agricultural wage to the average wage in the urban formal and urban informal sector. The Harris- Todaro model gives rise to a new perspective on the interplay of agriculture and industry. Differences in urban and rural wages dictate migration flows and because the industrial sector is unable to fully absorb this labour flow, migrants end up in the urban informal sector (see 3.1 for a further elaboration upon the informal sectors’ origins). Although this process changed economists’ view on economic development, it does not necessarily deny dualistic economic development (Khandker and Rashid, 1995).
The inability of the urban formal sector to fully absorb migratory labour flows created an urban informal sector. Dualism emerged in urban areas as an urban formal sector and an urban informal sector originated.
Dualism in rural areas
For rural areas, huge differences between the agrarian systems of Asia and Latin America can be found. The Latin American countries, for example, have a much higher land-ownership inequality, due to the latifundio-minifundio system. This latifundio-minifundio system causes agricultural dualism. In Argentina 0,8% of the farms are latifundios, but they occupy 36.9% of the land (Todaro, 2003).14 In Peru 1.1% of the farms are latifundios, occupying 82.4% of the land.15 Rural dualism is less severe in some countries (for example Mexico, Bolivia and Cuba, due to land reforms), but huge land- ownership inequalities are uniform in Latin America.
13 E.g. Harris and Todaro (1970).
14 Minifundios comprise 43.2% of the farms and occupy 3.4% of the land.
15 Minifundios comprise 88% of the farms and occupy 7.4% of the land.
Summarising
It is argued, that fitting dualism to Latin American reality requires two transitions. First, dualism is extended to the urban sector where a formal and an informal sector are found. Second, dualism in the rural agricultural sector has to be accounted for.
There are various growth scenarios possible with dualistic development. These scenarios are outlined in the next paragraph and a relationship with income inequality is established.
2.5 Dualistic Development and the Income Distribution: Theory
Lewis (1954) described dualistic economic development by employing a two-sector approach.
Basically he argued, that agricultural forms of production initially characterize an economy. As an economy grows, people move from the rural agricultural sector towards the industrial sector. Kuznets (1955) firmly related this process of dualistic economic development to income inequality. Kuznets argued that because of wage differences and the relatively small urban industrial sector at initial stages of economic development, income inequality would rise at early stages of economic development and decline further up in the development process as the industrial sector enlarges. The hypothesis by Kuznets, of an early rise in income inequality and a decrease in income inequality at later stages of industrial sector enlargement, is known as the inverted U-hypothesis. The inverted U-hypothesis is pictured below in figure 2.1. Because an industrial sector with high incomes is distinguished from a agricultural sector with low incomes, the inverted U-hypothesis is on between-sector inequality.
Figure 2.1 Inverted U hypothesis, between-sector inequality
Income inequality
between-sector
inequality
size of the industrial sector
However, there might also be differences in income inequality within the industrial and the agricultural sector. In other words, there might be a formal and an informal sector within urban industrial areas and a formal and an informal sector within rural agricultural areas as in a dual-dual framework. Within-sector inequality, for example, increases if income inequality within the urban industrial sectors is higher than in agriculture and the urban industrial sectors enlarge. Therefore,
changes in income inequality due to changes in between-sector inequality, can be reinforced or off set by changes in within-sector inequality (Chowdhury and Islam, 1995). A possible pattern of between and within-sector inequality and the total effect upon income inequality is pictured below in figure 2.2.
Here the inverted U-hypothesis is restated, but due to the off setting effect of within-sector inequality, there is no or less change in income inequality in part of the scenario depictured in figure 2.2.
Figure 2.2 Between and within-sector inequality
income
inequality Income inequality Within-sector inequality
Between-sector inequality
size of the industrial sector
I therefore need a framework to examine possible patterns of dualistic development and study its implications upon the income distribution. Particular helpful in analysing the relationship between dualistic economic development and the income distribution is a study by Fields (1980). Fields analyses the dualistic development model with Lorenz curves and gives four possible paths of dualistic development with different consequences upon the income distribution. These four possible growth scenarios are: formal-sector enlargement, formal-sector enrichment, informal-sector enlargement, and informal-sector enrichment. They are described in turn. A fundamental assumption by Fields (1980) is that productivity is higher (hence also income is higher) in the formal sector. When a sector enlarges, the employment share of the sector in total employment rises. When a sector enriches, the employment share in total employment does not change but wages relative to the other sector rise. Lorenz curves are indicative of income inequality in an economy. Income inequality in an economy can be estimated by measuring the distance between the 45° line and the Lorenz curve (Ray 1998). A rise in income inequality coincides with an outward move of the Lorenz curve.
The first possible growth path is one, which is characterized by formal-sector enlargement. With formal-sector enlargement, the employment share of the formal sector rises, which causes national income to rise. Lorenz curves with formal-sector enlargement are displayed in figure 2.3. Initially, at
t=0, only an informal sector exists and the Lorenz curve is equal to the 45° line. With enlargement of the formal sector, the Lorenz curve moves outward due to higher wages in the formal sector. Hence income inequality rises (t=1). However, at a certain stage during formal sector enlargement, income inequality starts to decrease. In figure 2.3, at t=2 the informal sector no longer exists and only a formal sector exists. As at the initial stage, the Lorenz curve now coincides with the 45° line. Between t=1 and t=2, income inequality decreased. This is in line with Kuznets’ inverted U-hypothesis.
Figure 2.3 Modern-Sector Enlargement
Cumulative income 100%
t=0, t=2
50%
t=1
45°
50% 100%
cumulative
population
Second, an economy can face formal-sector enrichment. Then the employment share of the formal and informal sector remain constant. Here national income also rises, but this growth is intensive in the use of capital and therefore the share of workers in the formal sector does not change. As a consequence, the size and wages of the informal sector are not affected. If formal-sector enrichment takes place, individuals with higher incomes in the formal sector receive the benefits of economic growth and the income distribution unambiguously worsens. It is in this scenario where the economy grows without developing. Figure 2.4 depicts the downward shifting Lorenz curve.
Figure 2.4 Modern-Sector Enrichment
Cumulative income 100%
50%
t=0
t=1 45°
50% 100%
cumulative
population
A third path of economic development is called informal-sector enrichment. National income rises, as was the case in the first two patterns of dualistic economic development, but contrary to formal-sector enrichment growth the benefits of economic growth now accrue to the people who live in the informal sector. This results unambiguously in a more equal income distribution, which is depicted in figure 2.5 by the upward move of the Lorenz curve.
Figure 2.5 Traditional-Sector Enrichment
Cumulative income 100%
50%
t=1
t=0
45°
50% 100%
Cumulative
Population
A fourth possible growth path of an economy is informal-sector enlargement. Fields (1980) does not pay attention to this possibility, but I would argue that the effects upon the income distribution are exactly the opposite of modern-sector enlargement and the change in income inequality depends upon the employment share of the formal sector. Although Fields does not consider this scenario, it is certainly not unrealistic. An economic crisis could lead to job firings in the formal sector where after
workers move into the urban informal sector. Here I touch upon a whole area of research, which is concerned with the cyclicity of the informal sector. The main research question in this area is, whether the informal sector shrinks with economic growth and expands during a crisis. I do not elaborate upon this, but there are many parallels with this thesis and therefore it could be an interesting direction of future research, using the findings from this thesis.16
The four growth scenarios have different implication upon the income distribution. The inverted U- hypothesis corresponds to modern-sector enlargement. With modern-sector enlargement, income inequality first rises, but falls at a later stage. Various authors (Lewis, 1972, Prebisch, 1988, Ranis, 2003) argued, that the growth scenario of Latin America is one which is characterised by modern- sector enrichment. However, with modern-sector enrichment, income inequality unambiguously rises.
Therefore, both modern-sector enlargement and enrichment do not explain stagnant inequality in Latin America described in (see chapter 1). It therefore seems important to look at an extended dual-dual framework rather than dual only.
Stagnant inequality could be explained by a combination of growth scenarios. I distinguish between a formal urban, an informal urban, a formal rural and an informal rural sector. Therefore I have 8 possible growth scenarios. This implies, that counteracting tendencies might occur. Consider for example the counteracting effect of formal urban sector enrichment and informal rural sector enrichment. In fact, within a dual-dual framework very few cases exist where changes in income inequality are unambiguous. An example of an unambiguous deterioration of income inequality is the a rise in income in the urban formal sector while employment share of all sector remain unchanged. In paragraph 5.2 and 5.3, I empirically examine the significance of enlargement and enrichment effects.
The relationship between dualistic development and income inequality outlined in this paragraph is with regard to the personal income distribution. That is due to the analysis with Lorenz curves. A Lorenz curve considers individuals. However, I distinguish between four sectors (and therefore 8 growth scenarios) in this thesis. Hence, my analysis of dualistic development and inequality is with regard to the sectoral income distribution. The only difference between the sectoral and personal income distribution in this thesis is, that normally the Lorenz curve is shaped by the size and wage- income of individuals whereas in this thesis the Lorenz curve is shaped by the size and wage-income of sectors.
2.6 Summary
16 De Janvry and Sadoulet (2000) could touch upon traditional-sector enlargement when they find an increase in inequality during recessions.
This chapter starts with a description of the general process of structural change. In essence, the process of structural change is one in which labour moves from agriculture into manufacturing and (increasingly) services. Lewis (1954) proposed a perspective on the development process in which a capitalist urban (industrial) sector expands at the cost of a subsistence rural (agricultural) sector.
Along with the critique upon dualistic theory, the dualistic model is enhanced in order to closer approach Latin American reality. A formal and an informal sector is present in both urban and rural areas. Therefore I adopted a dual-dual framework.
The relationship between dualistic economic development and inequality, is studied by Fields (1980).
There are four possible growth scenarios with dualistic development: formal sector enlargement, formal sector enrichment, informal sector enlargement, and informal sector enrichment. These growth scenarios have different implications upon the income distribution. If the formal sector enlarges, income inequality first rises but starts to fall after the formal sector reaches a certain size. If there is formal sector enrichment, the income distribution worsens. If the informal sector enlarges, a backward working inverted U-hypothesis might occur and if the informal sector enriches, the income distribution improves.
In a dual-dual framework, this analysis can be expanded. Now I have 8 possible growth scenarios. I will provide a decomposition of Latin America’s growth in chapter 4 and empirically examine in chapter 5 which development path characterises growth in Latin American countries during 1950- 2003. First I will pay closer attention to the informal sector in urban areas and the views upon the informal sector in chapter 3.
Chapter 3 - The Informal Sector
In the previous chapter it is argued, that classical dualistic models have to be enhanced in order to correspond to current Latin American reality. Enhanced, because a substantial informal sector emerged in urban areas. Rather than viewing the agricultural sector as traditional and the industrial sector as modern, it is argued, that both the rural agricultural sector and the urban industrial sector consist of formal and informal firms. However, so far only a marginal description is given on what might be considered the urban informal sector in Latin America. Getting grip on the urban informal sector is this chapter’s aim. This chapter then first describes the origins and characteristics of the urban informal sector in Latin America in paragraph 3.1. Paragraph 3.2 broadens the perspective on the informal sector by studying different views upon the relationship between the formal and informal sector.
3.1 Origins of the Urban Informal Sector in Latin America
By the early seventies, many researchers noticed that open unemployment in urban areas explained just a minor part of the labour market problem in Latin America. Among others, Souza and Tokman (1976) notice that “ very largely owing to rural-urban migration, the supply of labour in Latin America increased much faster during the post-war period than it did in the more developed economies, and a considerable proportion of this growing labour force had no alternative but to create low-productivity jobs for itself.” (p. 355). During these years the Regional Employment Programme for Latin America and the Caribbean (PREALC) recognised that these low-productive labourers “formed a sector that was not governed by the laws of capitalistic accumulation in dependent economies, but was characterised by very limited access to capital and productive resources other than raw labour, and survived in competitive markets or in the competitive base of an oligopolistic pyramid” (PREALC, 1990 p. 1). Similar findings of an emerging low-productive urban sector were presented in a study on Kenya (ILO 1972). In this study the sector was termed the informal sector and subsequent studies of the low productive urban sector adopted this term.
Distinguishing characteristics of the urban informal sector are mostly related to market principles and the labour force. With respect to the first characteristic, the urban informal sector consists of a great number of very small enterprises. These enterprises usually have a very small amount of physical capital, use very simple technologies and are informally organised. Due to these operating conditions, access to the urban informal sector is generally easy. Hence, enterprises in the informal sector are operating under very competitive conditions and are usually unable to fix prices or make exceptional
profits. Tokman (1978, 1980, 1989) argues that in analysing the informal sector, one has to study informal firms operating in a competitive market besides studying informal firms that operate in an oligopolistic system. I will elaborate upon his arguments below, but notice here that when informal firms operate in the “competitive base of an oligopolistic pyramid” (PREALC, p.1, see citation above) they again are unable to fix prices or make profits.
With regard to the labour force, PREALC conducted various case studies in Latin American cities (Souza and Tokman (1976) provide an overview). PREALC found female employment in the informal sector relative to the formal sector to be much higher.17 The relative share of female employment has also increased during the period 1960-1980 and Tokman (1989) notices, that for Chile in 1980 between 72 and 76 percent of informal employees are women. Recently this has been confirmed in a study conducted at ECALC (Panorama Social, 1997). The relative importance is reflected in the finding that the informal sectors’ female employment share in total female employment is about 50%.
Another characteristic is the high concentration of younger and older workers in the informal sector.
Nevertheless, there isn’t any concentration of labourers having its second job in the informal sector.
This might suggest, that many young and/or female heads of the family are working in the informal sector. Also, it was found that the informal sector contains a high proportion of very low educated and illiterate people. Finally, the informal sector contained people who had migrated to the cities. Many people moved from rural areas to the city and ended up in the informal sector. Very often, they remained in this sector. In San Salvador18, it was found that 50% of those that had been working in the informal sector for more than ten years where migrants. 70% of those who had arrived in the past year were employed in the urban informal sector (Souza and Tokman, 1976).
When researchers became aware of this urban informal sector coexisting beside the urban formal sector, they started to discuss the position of the informal sector and its relationship with the formal sector. That discussion is studied in the next paragraph.
3.2 Informal-Formal Sector relationships
In the 1970s, two different opinions about the informal sector emerged. Tokman (1978) reviews the literature upon the informal sector and puts it into perspective. He terms the first approach a “ benign relationship” approach as authors within this approach assume that the informal sector encompasses the potential for growth. The second approach allocates a subordinate or dependent character to informal activities and is called the “subordinated relationship” approach.
Authors within both approaches disagree about the degree of integration of the informal sector. Some consider the informal sector as autonomous, while others consider it as being integrated with the rest
17 This fact remained true when domestic services, arguably occupying more female workers, were excluded.
18 El Salvador’s main capital.
of the economy. This degree of integration implies different anticipated trends about the informal sector.
Tokman (1978) provides another view on the functioning of the informal sector. According to him the informal sector is subordinated, but partially linked with the formal sector. This linkage depends upon the operating conditions of informal firms. Tokman (1978) terms his view heterogeneous subordination. Due to his acceptance of subordination, this line of thinking is categorized within the subordinated relationship approach.
Recently, researchers at the Worldbank acknowledged the existence of an informal sector and started to argue that the informal sector consists of many dynamic and efficient small firms. Their reasoning is comparable with authors within the benign relationship.
Two approaches and various opinions about conditions of integration within both approaches can now be distinguished. An overview of the approaches and its perspectives is provided below in table 3.1. In the remainder of this paragraph, these approaches are studied.
Table 3.1 Informal-Formal Relationships
Benign relationship approach
Subordinated relationship Approach
Conditions of autonomy Dualism Dualistic Marginality
Conditions of integration Complementary Exploitation
Partly integrated Heterogeneous subordination
The Benign Relationship Approach
Hart (1970) and the ILO study on Kenia (1972) introduced the informal sector. Both studies argued about a high potential for growth within the informal sector. Their expectation was based on the finding that people in the informal sector manage to survive under very unfavourable circumstances.
Dualism
Some authors (e.g. Sethuraman, 1976) see the informal sector as being autonomous. This amounts to strict dualism, as two sectors coexist without being linked to each other in any way. These authors consider the sector as being characterized by a lack of links with the rest of the economy. However, the efficient use of production factors (principally labour) generates surpluses and if reinvested, the informal sector grows.
Complementary
If the informal sector is integrated with the rest of the economy, the growth potential is reinforced. The linkages between the formal and the informal sector are then assumed to be complementary. Weeks (1973, 1975) considers the informal sector as an important provider of products, but especially services. The informal sector delivers services, such as transport in Peru where there is insufficient supply (de Soto, 1989). Growth of the informal sector is then stimulated through its access to expanding markets.
Recently, various authors at the Worldbank (Cunningham and Maloney, 1998, Maloney, 1999 and Maloney, 2003) set forth a line of thinking in which the informal sector principally encompasses efficient and competitive small firms. Their findings are mainly based upon studies of micro enterprises in Mexico. In their view, small firms in developing countries do not differ greatly in operating conditions from small firms in developed countries. It is the great degree of independence (and possible higher incomes) that attracts entrepreneurs. Young workers are mostly voluntarily19 employed in the informal sector and will enter the formal sector at a later stage. Older workers having difficulties within the formal sector move to the informal sector. This view regards the labour market as highly integrated, but divides formal and informal firms on capital intensity, size and formality.
According to me, recent thinking at the Worldbank corresponds closely to the ideas of complementary and the view of potential growth in the informal sector.
Authors arguing about dualism and complementary consider the informal sector as encompassing the potential for growth; evolutionary growth. If the informal sector is considered autonomous, development can not be predicted a priori. If the sector is considered as partly integrated in the economy, the informal sector is expected to grow pari passu with formal sector growth.
The Subordinated Relationship Approach
Authors within this approach (e.g. Prebisch, 1963 and Pinto, 1965) take “ the existence of the informal sector as the national assimilation of the prevailing unequal international system” (Tokman, 1978 p.
1067). Prebisch (1963, 1988) argues about unequal trade20 and price relationships. Pinto (1965) argues about the unequal transfer and incorporation of technology. The main argument is that productivity gains in the modern sector are retained and productivity gains in the informal sector are appropriated.
Hence the informal or peripherical sector is subordinated and has little capacity for growth.
19 This finding is based on an micro-enterprise survey. In this survey people where asked why the left their previous job in the formal sector. 35% responded, they wanted to have more independence. 32% gave higher pays as a reason, whereas only 29% stated that it was involuntary.
20 Prebisch (1963) expects and finds a downward trend in the terms of trade to the disadvantage of Latin American countries.
Dualistic Marginality
If the informal sector is considered subordinated, but also autonomous (Quijano, 1974), there is dualism and the informal sector is marginal. According to Quijano (1974), the informal sectors’ labour force is permanently excluded from the formal sector. This implies a prevailing labour surplus. Several arguments come in play for this seemingly strange situation. A first argument are institutional factors.
These are for example trade unions. Second, firms operating in the formal sector wish a stable labour force for themselves for which they are willing to pay higher wages. Due to subordination, there is also a lack of access to inputs and products. Formal firms monopolize upon basic resources, and the informal sector is only able to operate on residual sources and hence also lacks possibilities for technological improvement.
Exploitation
Again, as within the “benign relationship approach” to the informal sector, the processes are reinforced if the informal sector is integrated with the rest of the economy. Market subordination and the lack of access to inputs and products are reinforced and surplus is extracted by the formal sector.
This situation justifies to talk about exploitation of the informal sector. The process is reinforced, because dependency in the case of an integrated informal sector operates on the sources of supply and the markets for its products as well as on prices. Prices of products purchased are generally higher because informal firms can only buy small quantities and they do not have access to credit facilities, while the prices for its products, mostly services, are lower because of the market they depend on.
Markets are controlled outside the informal sector (Bienefeld and Godfrey, 1975).
Here, both lines of thinking consider the informal sector as encompassing no potential for growth.
Authors within the subordination approach rather consider involutionary growth to be the case.
According to them, the market is unable to expand and because of an increasing labour force, incomes are constantly depressed.
Heterogeneous subordination
According to Tokman (1978), the informal sector should be seen as having strong links with the formal sector, while also being autonomous to a considerable degree. As in the subordinated relationship approach, Tokman (1978) also assumes subordination but leaves room for evolutionary growth as the kind of growth process depends upon the fierceness of subordination. This degree of subordination depends upon market conditions and the products or services being produced and delivered. Tokman distinguishes between three different market circumstances in which informal firms operate. Informal firms operate under conditions of oligopoly, perfect competition, or monopolistic competition. These conditions are examined respectively.
