|How Urbanization Affects the Inequality in Developing Countries: A Critique of Kuznets Curve
According to the Kuznets’s U-curve theory, urbanization increases inequality in developing countries during the first phase of industrialization. This argument is based on two assumptions. Firstly, the increase in productivity is greater for the industrial population; secondly, inequality in urban areas is greater than or equal to inequality for rural areas. However, the Kuznetsian argument could be false for many cases including the Turkish case, in which rural inequality was relatively higher due to the social structure of the country. Therefore, in many developing countries the relation between inequality and income may be negative or be only slightly positive for the first phase of industrialization.
This paper aims to prove that: (1) The argument suggested by Kuznets may not explain real changes in income distribution for many cases including Turkey (2) The long-run fall in Gini coefficient in Turkey may be highly related to the migration process within the country.
In developing countries, the growth process not only alters the level of output or technology, but also changes the economic relations and the social structure. The development process involves significant changes in the levels of income of individuals; however, a part of the society usually benefits relatively more from the rising incomes. Therefore, in many cases the development process changes the income distribution; the income of a specific part of the society increases relatively more compared to the rest of the society.
The first stage of the development process is usually associated with industrialization. As a country industrializes; individuals in urban areas benefit from rising productivities and the gap between per capita incomes of urban and rural areas increases. The growing benefits in the urban areas also alter the structure within the society. Many individuals in rural areas decide to migrate to urban areas to take advantage of the industrialization and rising urban incomes.
A part of the migration stream may be formed by lower-income individuals that are mostly unskilled and illiterate. These individuals mostly have individual concerns and they mostly move to the informal urban sector for earning wages. The higher income individuals also migrate for exploiting the advantages of urban society; however, these individuals concern more about their family requirements and aim to generate income, education or remittances that could be useful for the whole family (Lipton, 1982). In any case, urbanization preceded by industrialization changes economic structure and gains in the society. As the urbanization process alters the incomes of individuals, it will create significant changes in the income distribution both in the short and the long run.
The impact of migration on inequality is explored by Kuznets (1955) in his AER article “Economic Growth and Income Inequality”, which forms a basis for Kuznets’s inverted U-curve theory. In his article, Kuznets shows how demographic changes followed by industrialization alter income distribution within a country. We should notice that in his theory, Kuznets is only concerned about the one way relation from development to inequality, he never mentions anything on the impact of inequality on growth. In fact, Kuznets (1963) says that: “All we can say is that the unequal distribution of income in the earlier decades in the presently developed countries did not prevent economic growth. But…we cannot say that a somewhat or less (or more) unequal size distribution might not have contributed to even faster growth.”. Therefore, our focus on this study will be on how development and urbanization affect the income distribution.
The development process can change the overall inequality by affecting three factors. Firstly, as the urban share increases, the urban inequality starts to make a greater contribution to overall inequality. Secondly, the industrialization could widen the urban-rural gap; however, the impact of industrialization would be offset by factors like technological spillovers, changes in the terms of trade and rising land per capita. Thirdly, the development could change both within rural and urban inequalities.
In Kuznetsian framework, development may bring unequal gains and the first phases of industrialization are associated with higher inequality. However, in some cases the urbanization followed after the early stage of industrialization may have instant positive effects on inequality. This instant positive effect may even offset the unequalizing effect of the early industrialization. Thus, the start of industrialization does not guarantee a rise in inequality.
Development, urbanization and the change of inequality in Kuznetsian framework
The Kuznetsian approach is considered as one of the most important approaches for explaining, how development affects inequality. According to Kuznets (1955), urbanization followed by industrialization is an important factor in the shift of inequality. Kuznets says “an invariable accomplishment of growth in developing countries is the shift away from agriculture, a process usually referred to as industrialization and urbanization”. Thus, Kuznets sees industrialization and urbanization as processes that are mutually developing at the same time. Kuznets claims that urbanization increases inequality in the first stages of industrialization and he makes two crucial assumptions for strengthening his claim:
“a) the average per capita income of the rural population is usually lower than that of the urban; b) inequality in the percentage shares within the distribution for the rural population is somewhat narrower than in that for the urban population –even based on annual income; and this difference would probably be wider for distributions by secular income levels.”
With the given assumptions Kuznets claims that increasing the urban population means increasing the share of the more unequal component of the income distribution. Therefore, according to Kuznets urbanization pursued by the migration process inherently raises inequality during the first stages of industrialization. Secondly, the emergence of industrialization raises the per capita income gap between the urban and rural population, until the benefits of industrialization are also shared by the rural population. Thus, “the relative difference in per capita income…is stable at best, and tends to widen because per capita productivity in urban pursuits increases more rapidly than in agriculture”.
The Kuznetsian framework relies on the Lewis (1954) model that was outlined in Lewis’s classic article “Economic Development with Unlimited Supplies of Labor”. Lewis claims that development is driven by rising profits and accumulation in the capitalist sector. The saving rates for the capitalist class are relatively higher compared to the working class, since the working class could only save for its essential expenditures like housing, education etc. Therefore, according to Lewis, the amount of savings grows as the relative income of the capitalists rises. In his model, savings are realized as the major source of capitalist accumulation, so Lewis concludes that “the central fact of economic development is that the distribution of incomes is altered in favor of the saving class”.
The capitalist group accumulating in Lewis’s context is mainly the capitalist industrial class. Lewis argues that merchants use their profits mainly for speculation and peasants would prefer to spend his profits for enlarging his land, not his capital. Only the class of industrialists reinvests its profit productively; only the industrialists have incentives “towards using profits to create a bigger empire of bricks and steel”.
Lewis argues that the subsistence sector, which is mostly associated with the agricultural sector, cannot stimulate development, since the subsistence sector cannot produce reproducible capital. During the development process, the incomes of capitalist industrialists will increase, whereas incomes in the subsistence agricultural sector stay relatively stagnant, unless the subsistence sectors start benefiting from the development of the capitalist sector. Eventually, the emergence of development in the Lewis model is associated with higher urban inequality and a rising income gap between urban and rural areas. The development of the capitalist sector will be supported with the labor supplied by the subsistence agriculture sector, which will lead to urbanization as it is claimed by Kuznets. As a result, the urban population grows and higher inequality will be expected at least in the first stage of Lewisian development.
Lewis says the subsistence sector could be more productive, if the peasants start to imitate capitalist techniques by using new seeds and fertilizers or if they start to benefit from capitalist investments in irrigation, in transport facilities or in electricity. However, in his statement Lewis does not say whether increase in productivity in the subsistence sector will be bigger or less than the increase in productivity in the capitalist sector. Thus, he does not imply anything about the changing income gap between the urban and rural sectors. He only argues that higher productivity of the subsistence sector will lead to higher real wages in the capitalist sector and reduce the capitalist surplus and accumulation. Hence, Lewis focuses on the relationship between the development of subsistence sector and the income distribution within the capitalist sector.
The Kuznets process is depicted in some studies by using modeling. The model proposed by Robinson (1976) claims that the Kuznets process holds even if the rural inequality is greater than the urban inequality. However, he assumes that both within rural and urban inequalities and rural-urban income gap are constant, while developing his model. Robinson’s assumption is not applicable for many cases; the changes that may occur in inter- and intrasectoral inequalities could falsify Robinson’s model.
Another model that explains the Kuznets process is developed by Anand and Kanbur (1993). Consistent with the Kuznetsian approach, Anand and Kanbur assume that population moves to the urban sector, where inequality is larger. The model decomposes the inequality into within-sector and between-sector component. Anand and Kanbur first assume that means and the within-group inequalities of the urban and rural sectors are constant and the change in inequality is occurring only due to population shifts. They show that the within-sector component should be increased due to the greater population in the more unequal urban sector and therefore the between-sector inequality should be reduced for the U curve to be formed under their assumptions. However, empirically they show that their assumptions do not hold and the model should allow the sectoral means and sectoral inequalities to shift over time. The model concludes that inequality changes not only due to the population shifts, but also due to the changes in sectoral means and within sectoral inequalities. This result is consistent with the Kuznetsian approach that explains the mechanisms that reduces urban inequality. Hence, explaining the changes in inequality merely by urbanization is not enough; the results of urbanization on both sectoral mean incomes and inequalities should also be explored.
Urban-Rural Income Gaps
The urbanization changes the overall inequality in a country by raising the impact of urban inequality on the overall distribution. However, the industrialization and urbanization processes can affect the overall inequality not only by shifting the sectoral shares, but also by changing the intra and intersectoral inequalities. As a country develops, the intersectoral inequality between the urban and rural sectors changes due to first and secondary effects of industrialization.
According to Kuznets (1955), the early effects of industrialization could raise the income gap between the urban and rural sectors. In his AER article, Kuznets (1955) claims that:
“The relative difference in per capita income between the rural and urban populations does not necessarily drift downward in the process of economic growth: indeed, there is some evidence to suggest that it is stable at best, and tends to widen because per capita productivity in urban pursuits increases more rapidly than in agriculture.”
The intuition behind this argument could come from the Lewis model (1954). In Lewis model, the development is mainly driven by the industrial sector. However, Lewis also examines the secondary effects of development on the subsistence sector (mainly agriculture). Firstly, Lewis mentions that the subsistence sector could also increase its productivity by imitating the techniques in the capitalist sector. For example, peasants could get new seeds or they could use fertilizers for raising their productivities. He also says that the subsistence sector could benefit from the capitalist investments like irrigation works, transport, and electricity.
Lewis secondly claims that if the capitalist and subsistence sectors specialize in different products (like agriculture and industry), the urbanization could lead to greater demand and relative shortage of agricultural commodities. As a result, the agricultural output will not be enough to feed the industrial workers. The terms of trade will worsen for the industrial sector, which would partially close the urban-rural income gap. Lewis also mentions that the terms of trade effect might not be seen, if the capitalist sector invests on agriculture or the productivity in the subsistence sector increases significantly.
Ranis and Fei’s (1961) arguments in their article “The Theory of Economic Development” could be considered as a complement to Lewis model. In their study, Ranis and Fei claim that “Lewis…has failed to present a satisfactory analysis of the subsistence or agricultural sector.” Thus, they propose a more detailed analysis for examining the secondary effects of industrialization and urbanization on agriculture.
In their model, Ranis and Fei (1961) claim that the economy could follow a balanced growth, in which time to time deviations are seen. As in the Lewis model, in the case of overinvestment in the industrial sector, the shortage of food could lead to deterioration of the terms of trade in the industrial sector and industrial wages will rise due to growing food prices. As a result of this, the industrial investments will be discouraged; hence, there will be more incentives to invest on agriculture. Thus, the actual growth path could return to the balanced growth path and the urban-rural income gaps could narrow down.
The urbanization could also increase the income in rural areas through the remittances. The urban settlers can transfer money to their relatives, which could generate an extra income for the rural dwellers (Lall, Selod and Shalizi, 2006). Remittances also create an extra economic activity in the rural sector, when the new urban settlers spent their income for investments like housing in their origins (HDR, 2009). Many studies show that remittances increase the level of education and improve health of origin families (Becker, 2007). Improved human capital could also contribute to rural productivity positively.
There are also studies that show how urbanization can change the inequality by altering the land per capita. In fact, many studies (Sen, 1996; Griffin, Khan and Ickowitz, 2002) claim that the land productivity might be lower in small scale farms since family farms have advantages in monitoring and effort. However, the migration of rural individuals could still increase the rural income per capita (Bourguignon and Morrisson, 1998), although it might not lead to higher land productivity, as in the case of shift from small to large scale farming. In addition, urbanization might not lead to a transformation from small to large scale farming. The family farming structure could continue, only the family members that cultivate the land could be reduced due to the migration. In this case, it is more probable that the rural income per capita will increase with urbanization.
Stiglitz (1982) also focuses on increases in agricultural productivity; however, he makes an alternative interpretation by using his efficiency wage model. According to Stiglitz, the migration of a family member changes the marginal product of the other members that stay in the rural sector. If output is proportional to effort, equal share of output between many family members could reduce output by lowering incentives for supplying effort. This might cause a loss of efficiency in agricultural production. In such cases, the migration of an individual within the rural family could increase the efficiency by bringing better incentives to the peasants for supplying more effort. Therefore, migration process could close the urban-rural income gaps by not only increasing average rural incomes, but also by raising total productivity in the planted areas.
In his article, Kuznets (1955) also claims that the inequality will decrease in the later stages of development. However, unlike the studies mentioned above (Ranis and Fei , 1961; Bourguignon and Morrisson, 1998; Stiglitz, 1982), Kuznets focuses on the changes in urban inequality rather than the narrowing intersectoral income gap. He uses the long term effects of urbanization for making his analysis and perceives urbanization as a process that would reduce inequality in the long run.
According to Kuznets, raising inequality will be narrowed, mostly due to the declining inequality within the urban groups. Kuznets claims that, within years the economic positions of new urban dwellers and their descendants improve. The social mobility equalizes economic differences, and hence inequality follows a declining path. The mechanism behind the social mobility is explained by Kuznets. Firstly, as the development process continues a larger share of the urban population becomes “native” urban dweller. Thus, a larger proportion of population benefits from the advantages of the city life. Secondly, in democratic societies the growing political power of the lower-income groups resulted in changing legislation and new policies that counteracts against the negative consequences of rapid industrialization and urbanization.
Thirdly, the forces of “freedom of individual opportunity” enable development of new fields that bring opportunities for new entrepreneurs. New and profitable industries will be run by the new entrepreneurs, unless the descendants of upper-income groups do not shift to the new industries. However, Kuznets says that “the successful great entrepreneurs of today are rarely sons of the great and successful entrepreneurs of yesterday”. Therefore, Kuznets claims that the development of new industries stimulates social mobility by creating opportunities for different income groups.
Lastly, the improvement of the service sector reduces income inequality. The service sector is expected to grow as a result of the development process and rising GDP. In the earlier stages of development, the proportion of workers in industrial sector increases as the proportion of workers in agriculture declines. However, as economic growth continues the services sector starts to expand, while the agricultural sector continues to decline (Syrquin, 1988).
According to Kuznets, the growth of the service sector decreases inequality for several reasons. First, Kuznets argues that service incomes are mostly earned due to the individual excellence; thus, the higher levels of service incomes are not necessarily pursued by the descendants of the wealthier individuals. Kuznets also claims that the possibilities of rising income are limited for people who are already in high-income occupations. Hence, incomes of the lower-income workers in services are more likely to increase. Therefore, incomes are expected to be more equal within the services sector compared to industry.
In summary, Kuznets explains the change in income distribution as increasing in the first stage of development and declining in the later stages. With this explanation, Kuznets describes an inverse U curve relation between income and inequality. Urbanization is pursued together with industrialization; therefore, migration from rural to urban areas inherently increases inequality by raising the proportion of population of the more unequal part of the country. As the development process continues the income inequality declines, due to the factors like social mobility and improving service sector.
There are many extensions and criticisms that could made for Kuznets’s arguments on urban inequality. Firstly, the pace of social mobility is highly related with structural factors in the urban sector. Amongst the structural factors, the distinction between the urban formal and urban informal sectors could be considered as important.
For examining the migration process, unlike Harris and Todaro (1970), Cole and Sanders (1985) use a model that distinguishes the urban modern and urban subsistence sectors. According to Cole and Sanders (1985), it is more likely that the new arrivals would enter to the urban subsistence sector (mostly informal), in which barriers to employment are few and average income is lower compared to the urban modern sector. Many other studies (Benarjee, 1983; Rauch, 1993) claim that wages in the urban informal sector are equal to or slightly higher than the rural wages. However, the many migrants shift to the urban informal sector, preserving their hopes for finding a formal job in future.
The shares of informal-formal sectors change during the different phases of urbanization. According to Rauch (1993), the share of being “underemployed” in the informal sector follows an inverted U path as the economy urbanizes. Likewise the “Todaro paradox”, the growth of the urban formal sector attracts rural workers and leads to a greater growth in the urban informal sector. However, as urbanization proceeds, pressure on the land decreases and the agricultural income rises. The agents become less willing to leave the rural sector and be “underemployed” in the informal urban sector. Thus, the share of informal urban sector employment will start to decline. Since, the wages are generally lower in the urban informal sector; the inverted U curve of the informal urban sector share could also form an inverted U curve between the level of urbanization and within urban inequality.
The migrants could enter to the informal and formal sectors due to several factors. A study made for Bolivia (Pradhan and Van Soest, 1993) show that the education level is an important determinant of entering the formal sector. Pradhan and Van Soest show that factor like ethnicity and unemployment in the region are also important for determining the possibility of participating in the formal sector. Likewise in Bolivia, in India the education level is also an important factor for rising migrants’ possibility of being employed in the formal sector (Banerjee, 1983). The migrants with intermediate and higher level of education have a greater chance of getting employed in the formal sector. As expected, in many countries the return of education is also found to be higher in the formal sector (Pradhan and Van Soest, 1993; Banerjee, 1983; Funkhouser, 1997).
The growth of the informal sector could raise the urban inequality; however, the inequality could be reduced if there is a significant social mobility between the formal and informal sectors. Nevertheless, the social mobility might not be high for many cases. Banerjee (1983) shows that in 1976’s Delhi only 24% of those who entered the informal sector on arrival was switched to the formal sector. He also shows between the years 1965-1975, only 5-15% of the new arrivals were switched to the formal sector in a year. In addition, the potential mobility was found to be low; only 15% of the informal sector wage employees and 12% of the non-wage workers were actively looking a job in the formal sector.
The network of between the agents is important for enabling the mobility between the formal and informal sectors. In India, 60% of migrants who moved from the informal to the formal sector found out their current job with the help of their relatives and friends. The level of education is again found to be an important factor of mobility. The informal sector employees who have a middle school or intermediate college level education have a greater likelihood of shifting to the formal sector.
In summary, along with the other factors the access to education is an important factor for enabling the social mobility and reducing the urban inequality. The migrants or the children in the migrant families should have an access to an education to secondary or tertiary education for the process depicted by Kuznets to happen. One impediment of that is the usage of children as a labor force in the informal sector. Many of the children in the migrants’ families work (Acikalin, 2008), they cannot have an access to higher level of education. Also, the quality of the education that some children get is extremely low even for the primary school level, since they could not have time to focus on their school, while they are working.
Another extension that could be made to the Kuznetsian theory is related with the services sector. As we mentioned previously, Kuznets claims that the growth of services sector would reduce inequality, since the inequality in the services sector is lower compared to the industrial sector. The arguments of Kuznets on services are mostly relevant for the service sectors of 1950’s. However, the relations of production are significantly changed in many of today’s service sectors; employer-employee type of relations improved, whereas the proportion of self-employed is declining. Also, improvements on branding and franchising also created new opportunities for increasing the incomes of the wealthier individuals in the service sector. Therefore, the difference between the levels of inequality within the industrial and service sectors may not be great in today’s world, due to the developments in the services sector.
The property income is another issue that might create a need for the revision of Kuznets’s theory in today’s world. In fact, in his article “Quantitative Aspects of the Economic Growth of Nations”, Kuznets (1963) shows that “the shares of property in total income are lower in the less developed regions… they would tend to widen inequality less in the low income, underdeveloped regions than in the developed regions” in the United States.
The current literature on financialization also signs that the situation depicted by Kuznets might be true for the country wise comparisons. The US economy has the highest GDP per capita compared to other large scale countries1. Following the economic growth, the share of financial income has also risen in the US economy (Krippner, 2005; Crotty, 2009). The share of financial and real estate incomes reached to the level of 23% in 2001 and it continued to rise afterwards. In his empirical study, Frazer (2006) finds a modified inverted U curve with inequality rising after an income level. This result could be partially related with the higher urban inequality caused by the rising share of property incomes in the high income countries.
In his 1955 article, Kuznets does not focus on the changes in rural inequality. He explains the reduction in inequality majorly by the changes in the urban inequality. However, in his 1963 Kuznets develops an argument on the changes in rural inequality. He claims that:
“The rise in productivity within the A sector, indispensable for modern economic growth, may have been associated with technological changes that raised the scale of production on farms and introduced a cleavage between the large commercial farms in the progressive part of agriculture and the small units lagging behind, which would make for wider inequality of income within the A sector, at least until the process of modernization had been introduced throughout the sector.”
The technological improvements could increase the rural inequality for a period of time. As it happened in Philippines during the green revolution (Boyce, 1993), the “technological improvements” in agriculture might even lead to the centralization of land and would make the higher rural inequality permanent. However, in some cases the small scale farms can also compete with the large scale farms that are technologically developed. In fact, many recent studies (Sen, 1996; Griffin, Khan and Ickowitz, 2002) show that the small scale farms could have higher land productivity. It is true that the small scale farms could be technologically behind; however, the small scale farms have advantages on monitoring and effort. The large scale farms owned by absentee owners and cultivated by hired workers under supervision of a manager are often considered as inefficient. These farms have high monitoring costs and the effort shown by the workers is usually lower compared to the family farms. Also, the proportion of uncultivated land is higher in the large scale farms.
In summary, the small scale farms can compete with the large scale farms; the technological improvements followed by growth might or might not lead to a permanent rise in the rural inequality. Lastly, successful land redistribution could play an important role on changing the rural inequality. The rural inequality would be significantly reduced, if land is redistributed during the development process (Griffin, Khan and Ickowitz, 2002)