Introduction / 1
Relationships between Economic Growth
and Poverty in Developing Countries:
The Importance of Migration
Christopher David Yamaoka
Submitted to the Department of Economics of AmherstCollege
in partial fulfillment of the requirements for the degree of
Bachelor of Arts with Honors
Assistant Professor Christopher Kingston, Faculty Advisor
Associate Professor Steven Rivkin and Professor Geoffrey Woglom, Readers
April 23, 2004
Acknowledgements
First and foremost, thank you to my advisor, Chris Kingston, for countless meetings, discussions, and emails to help develop and organize this thesis over the course of the year. I would also like to give special thanks to Steve Rivkin, for indispensable help with the econometrics involved in the paper. Finally, many thanks to Sumir Meghani, for taking the time to correspond with me and for referring me to numerous papers and data sources that proved important to the development of the paper.
Table of Contents
I.Introduction / 1II.Growth and Poverty / 5
A.Poverty in India / 5
B.Literature Review: Is Growth Enough? / 7
III.Migration / 13
- Migration in India
- Literature Review: Probabilistic Models and Their Critics
IV.Relationships between Poverty and Migration / 18
- Demographics of Migration
- Motivations of Rural-Urban Migration
- Consequences of Rural-Urban Migration
Rural Consequences / 22
Urban Consequences / 25
V.Data and Econometric Models / 28
A.Data / 28
B.Models and Variables / 29
C.Procedures and Results / 32
Basic Results / 32
Lagged Migration / 37
Instrumental Variables / 40
VI.Conclusion / 43
Appendix / 46
Bibliography / 48
I. Introduction
The links between economic growth and poverty reduction are difficult to measure. To begin with, there is considerable variance in the way growth or contraction of a nation’s GDP translates into increases or decreases in average household income or consumption. Compounding this, there is further variance in the way changes in mean incomes translate into changes in the level of poverty (Ravallion and Datt, 2002). While growth is both intuitively and traditionally viewed as having a poverty-reducing effect, some have recently raised doubt as to the strength of the effect (Wolfensohn and Stiglitz, 1999). The concern is that an exclusive focus on growth, the so-called trickle-down approach, ignores crucial facets of the problem. A more complete understanding of the relationship between growth and poverty would lead to more effective development efforts.
One of the most important aspects of poverty in developing nations is the disparity between rural and urban areas. Wages, literacy rates, and access to services are likely to be greater in urban areas (Yap, 1977). However, while poverty rates are generally higher in rural areas than in urban areas[1], this is not simply a story of rich cities and poor villages. Often, urban poverty rates, even if lower than rural rates, are alarmingly high. Urban centers in less developed countries are commonly overcrowded and underequipped to handle their populations.
In all developing nations, migration between rural and urban areas plays a key role in the way the sectors interact and therefore the way in which economic growth affects poverty in both areas. Growth takes place in both rural and urban areas, but spillover effects between sectors are not always equal; the effect of rural growth on urban poverty can be very different from the effect of urban growth on rural poverty (Ravallion and Datt, 1996). In thinking about strategies to reduce poverty, the question of how and why these relationships differ is of significant interest. Figure 1.1 provides a diagram as a starting point for the discussion of the relationship between growth and poverty. The arrows in the figure should not be read as “leads to”, but rather “affects” or “influences”. It is a vague picture, for now; after some discussion of the relationships in the next chapter, the diagram will be redrawn.
To begin with, the relative conditions in rural and urban areas affect the level of migration from one sector to the other. In turn, this migration has a significant impact on both rural and urban conditions. Indeed, a widely-held view is that urban poverty is an outflow of rural poverty (Arrow C). Some have also argued that migration from rural areas can hurt these areas through the loss of human capital. Therefore, the way in which economic incentives promote or discourage the movement of people from one area to another can have a significant impact on the quality of life in both areas.
It is my argument that migration plays a central role in the way in which economic growth has an impact on poverty. In particular, I argue that it is a key issue in terms of the critique raised by Wolfensohn and Stiglitz, that growth itself is not sufficient for reducing poverty. In this paper, I use data from household surveys for the 16 major states of India over the period 1960-1991 to study the relationship between economic growth and poverty, focusing on the role of migration. I use state-level data because migration is dependent on a lot of factors that may differ from place to place, such as the distance between rural and urban areas; being able to control for state-specific effects is important.
The structure of the paper is as follows: Chapter II begins with a brief overview of the history of poverty in India in order to provide a context for the discussion. I then review some of the vast literature on poverty, focusing on work highlighting the importance of inter-sectoral relationships. Among other things, this work finds that the relationship between urban growth and rural poverty is not as expected (that is, it may not be inverse). In other words, arrow D from Figure 1.1 is not as readily explained as arrows A, B, and E.
In Chapter III, I present a discussion of migration in order to give a picture of the way migration is modeled in economic literature. The structure is similar to that of Chapter II; a brief description of migration in India is followed by a review of the literature. I present two leading models, followed by some of their critiques.
Chapter IV brings together the discussions of poverty and migration more explicitly. In particular, I look at way conditions in rural and urban areas influencethe migration decision and the way in which migration, in turn, can affect conditions in both sectors.
Chapter V presents my attempt at an econometric model of poverty that focuses on the role of migration. It includes descriptions of the data and procedures, as well as a discussion of the many problematic issues associated with such a model. Based on the results, I discuss the suggested relationships between poverty, growth, and migration. Chapter VI concludes with a summary of findings and suggestions for future research.
II. Growth and Poverty
In this chapter I look briefly at the nature of poverty in India in order to provide some context for the discussion of the rest of the paper. I then review some of the literature on poverty, highlighting work done to study intersectoral relationships. That is, I am mainly interested in the way urban conditions affect rural poverty and the way rural conditions affect urban poverty. In each case, we would expect to see a direct effect, through things such as demand for the goods of the other sector. However, there may also be an indirect effect taking place through the mechanism of migration, which may move in the opposite direction. Migration is discussed more explicitly in the following chapter.
A. Poverty in India
Throughout the history of India, numerous projects and papers have attempted to measure, reduce, and better understand the nature of poverty. Early on in the nation’s modern history, a colonial regime that was focused on the interests of the rulers was blamed by most for a poverty rate that was estimated around forty percent (Mann, 1948[2]). When India achieved sovereignty from Britain in 1947, there was a great deal of optimism with respect to the issue of improving the quality of life of the nation’s entire population. Since then, the eradication of poverty has been a primary focus of Indian policy. Each of the five-year plans that have been launched since 1951 (the tenth began in 2002) have contained measures intended to address the issue. Yet despite this clear and consistent policy focus, poverty remains a significant problem in India. Current estimates place the poverty rate at 30-35% of the total population.[3],[4]
That poverty has persisted to such an extent throughout India’s entire history makes it a particularly pressing issue to be studied. Fortunately, compared to most developing nations, India has fairly good data quality and availability. One of the most common criticisms raised against studies that attempt to analyze the way poverty responds to growth in different economies is that these studies often look across countries in order to make comparisons. This raises significant issues of data comparability (Deaton, 2001, p. 127). While the data employed in this paper is certainly far from perfect, the fact that it is primarily gathered from a survey that has remained consistent over an extended period allows for a cross-sectional time series analysis while avoiding the comparability problem that would arise from using data from a number of different countries.[5]
As a response to the data-comparability criticism, many recent studies have turned to state-level analyses. These studies benefit from the fact that statistics can be collected from common sources. At the same time, it is possible to control for state-specific effects. India is an apt place to conduct a state-level analysis of the cross-sectoral interaction between growth and poverty. To begin with, there is a considerable amount of disparity between conditions in rural and urban areas. Over the period 1960-1991, real mean per capita consumption was over 18% higher in urban areas than in rural areas, after accounting for differences in cost of living.[6]
Also, while each state is of course part of the same Indian economy, there is a substantial range across states in terms of measures such as poverty, mean income, and human capital development. For example, at the beginning of the period examined in this paper (1960), poverty rates ranged from 64% in Orissa to 16.5% in Assam.[7] In the last year of the period (1991), Orissa still had the highest incidence of poverty (48.7%) and Assam the lowest (6.5%), reflecting the persistence of poverty within states. Indeed, of the five poorest states in 1960, three were still among the five poorest in 1991, and all were within the poorest seven.[8] It is not entirely surprising that there would be significant variation between states; under the Constitution of India, it is the state and not the national government that is largely responsible for the provision of most social services and the implementation of rural economic policies (Meghani, 2003, p. 5).
B. Literature Review
Is Growth Enough?
As mentioned above, poverty in India has remained a significant problem despite a history of policy intended to address the issue. Over a half-century and ten five-year plans later, poverty rates are still alarmingly high. A review of India’s development strategy shows an emphasis on capital-intensive industrialization concentrated in urban areas (Ravallion and Datt, 1996, p. 20). As a result of the shortcomings of this policy, recent literature has begun to ask the question “is growth enough?” The suggestion of a lot of this research is that economic growth is a necessary but insufficient condition for reducing poverty.[9]
India features both a large rural population as well as several large (and growing) urban centers.[10] As such, it fits the role of a dualistic economy very well. In studying the way economic growth leads to changes in well-being in India and other developing countries, a good deal of work has focused on the sectoral (urban-rural) composition of the growth. A focus on the dualistic nature of the economy implies that asking the question “is growth enough” involves asking two other questions: In what way does urban growth affect rural poverty, and in what way does rural growth affect urban poverty? The study, How Important to India’s Poor is the Urban-Rural Composition of Growth? (Ravallion and Datt, 1996), attempts to address this question. Examining household survey data from the same data set employed in this paper, the authors find rural growth to have an inverse relationship with both rural and urban poverty. Urban growth, however, is found to have a reducing effect on only urban poverty. They find no clear relationship between urban growth and rural poverty.
In the paper, A State-Level Examination of Rural Poverty in India from 1983-1994 (Meghani, 2003, p. 30), the author finds urban growth to be associated with an increase in rural poverty, though the coefficients are small and not significant. Given this perverse (or at best, unclear) relationship between urban growth and rural poverty, it is not surprising that a development strategy focused on urban growth has been less than successful in reducing overall poverty. It is important then to attempt to understand why this effect has not been as expected. A possible explanation is that the urban growth may promote an effect opposite in direction to the one intended. While urban growth may have a positive effect on rural conditions through increased demand for food and other rural goods, it may also have a poverty-increasing effect (through its encouragement of migration) that would mitigate this positive effect.
While urban conditions affect rural well-being in at least these two ways, it is also the case that rural conditions have a significant impact on the urban sector. A commonly expressed view is that rural underdevelopment hampers the growth of industry in urban areas because more rural poverty implies higher costs of food and raw materials (Khan, 2000, p. 7). This can be thought of as the direct effect of rural conditions on urban well-being. Also included in this effect would be the fact that worse rural conditions would be associated with lower demand for urban goods. Likewise, rural growth may benefit urban areas through increased demand and lower costs of raw materials.
A second popular view, focused on population issues, can be considered the indirect effect of rural conditions on urban areas. As stated before, urban poverty is often seen as an outflow of rural poverty. This view holds that a high level of rural poverty perpetuates poverty throughout the nation by leading to overcrowding through increased fertility rates and migration to already crowded cities. This may be largely attributable to the fact that those in poverty are generally very vulnerable to fluctuations in climate, health and market conditions. Possessing few resources with which to cope with such shocks, they are often forced to be relatively short-sighted in their decision making. For example, with few resources outside of human labor, many poor families may have incentive to have more children, who can provide current as well as future income.[11] This strategy can be seen as both a short-term solution and a long-term contributor to the problem of poverty.
Migration is another such strategy employed by families living in poverty in rural areas.[12] In the study, The Migration of Labor (Stark, 1991), portfolio investment theory is extended to migration. Under this theory, the migration decision is a group (family) one, made for the reason of diversifying labor over different markets in order to reduce risk. The strategy is one of remittance, family resource pooling, and consumption smoothing. However, while this strategy may help families avoid risk, it may also lead to a worsening of poverty in urban areas. Regardless of motivation, what seems clear is that rural conditions can have an indirect effect on urban conditions through the process of migration and the expansion of the urban labor supply. Literature on migration is discussed in the following chapter. Chapter IV presents some of the causes and consequences of migration.
While not explicitly modeled, the importance of migration is suggested in a number of studies on poverty. In what has become a very influential paper, Ahulwalia (1978) finds evidence of a strong inverse relationship between rural poverty and agricultural performance at an all-India level. He does not, however, find a consistent effect at the state level. Among his explanations is the importance of migration as a strategy of the rural poor in coping with difficulties associated with poor agricultural performance. This argument is similar to the one made by Khan (Chapter II, p. 8) in that they both view rural-urban migration as a response to worsening conditions in rural areas. That is, migration is primarily a “push” phenomenon. Most of the literature focused specifically on migration, however, views it in terms of urban “pull”.
Before moving on to further discussion of the topic of migration, it may be helpful to redraw the picture of cross-sectoral effects presented in Figure 1.1. Again, arrows in these figures should not be read as “causes” but rather “influences”. Figure 2.3 can be considered a revision of Figure 1.1 based on some of the preceding discussion. Figure 2.1 illustrates the ways the urban sector can impact rural poverty (Arrow D from Figure 1.1) while Figure 2.2 revises the picture of Arrow E, the impact of rural conditions on urban poverty. In Figure 2.2, it may be helpful to think of “Rural Growth” more in terms of lack of rural growth. That is, the worse economic conditions in rural areas are, the more migration to urban areas we would expect to see. This relationship is discussed more thoroughly in the following chapters.