Working paper

Urbanisation,
demographic
transition, and
the growth of
cities in India,
1870-2020

Chinmay Tumbe
September 2016

When citing this paper, please

use the title and the following
reference number:
C-35205-INC-1
 Urbanization, Demographic Transition and the Growth of Cities in India,
1870-2020

Chinmay Tumbe*

September 2016

Abstract
This paper examines the nature of urbanization and urban growth in India since the late 19th
century against the backdrop of the unfolding demographic transition. It argues that (a)
Urbanization within India exhibits a tight relationship with economic growth at the regional
level (b) The demographic divergence between rural and urban natural growth rates since
the 1970s that is attributed to stagnant agricultural productivity and rural literacy levels in
large parts of India has slowed down the pace of urbanization, especially in the Northern
hinterlands (c) City population growth rates peaked in the 1980s and will see a marked
deceleration in the coming decades with substantial variations driven by investments in
specific sectors such as Information & Technology (d) Human capital externalities in a period
of demographic transition draw in migrants but also reduce fertility such that its relationship
with city growth rates, as observed in Indian data is mixed (e) Labour mobility in India is high
but is mostly male-dominated, semi-permanent, and remittance-based in nature leading to
masculine urbanization with important implications for urban growth and urbanization. By
integrating the demographic transition with urban processes, this paper explains India’s
relatively slow pace of urbanization, the inter-city variation in population growth rates and
the paradox of faster urban growth combined with slower urbanization in the North relative
to the South.

* Indian Institute of Management Ahmedabad. Email: [email protected]. This research

project was financially supported by the International Growth Centre (India-Central),
London School of Economics, and hosted at the Tata Institute of Social Sciences (TISS)
Hyderabad. The author thanks Uma Gayatri Yarrapothu for excellent research assistance
and mapping work, students of the Urban Economics course at TISS Hyderabad for research
assistance and participants at the following conferences/seminars for valuable comments:
International Conference on Sustainable Urban Development (Delhi, 2015), World Economic
History Congress (Kyoto, 2015), CEPT University (Ahmedabad, 2015), IIM-Ahmedabad
(2016), 7th IGC-ISI India Development Policy Conference (Delhi, 2016), ISEC and IIHS
(Bangalore, 2016).

1
“Why has Indian urbanization been so slow?” (Davis, 1962)

1. Introduction
When the demographer Kingsley Davis asked a question on India’s relatively slow pace of
urbanization over five decades ago, he proceeded to answer that question on the basis of
the “relative slowness of economic development in India.” Since then and especially in the
past two decades, India’s economic growth rates have risen but the question asked by Davis
continues to be relevant. For its level of income, urbanization appears to be low in India
(Deshingkar & Anderson, 2004), even though the size of its urban population is significant
and larger than the population of the United States of America.

It is in this context that this paper examines the nature of urbanization and urban growth in
India since the late 19th century.1 Crucially, it locates urbanization and urban growth against
the backdrop of the unfolding demographic transition in India. The word – Urbanization- is
used in this paper both as a statistic and a process. As a statistic, it denotes the share of
urban population in total population and as a process it indicates the upward movement of
the statistic. Urbanization depends not only on urban growth but also on rural growth, a key
point emphasized in this paper. The demographic transition, on the other hand, refers to
the transition from high birth and death rates to low birth and death rates. Despite the
empirical regularity of both these processes, the interaction between them is a very nascent
field of research (Dyson, 2011). This paper studies these interactions and argues that they
explain several features of India’s urban trajectory. Thus, it contributes to the growing
research enterprise on urban processes outside the Western domain (Chauvin, Glaeser, Ma,
& Tobio, 2016).2

For instance, we find that India has been de-urbanizing in the past four decades in the
absence of migration and reclassification effects due to a demographic divergence between
rural and urban natural growth rates, a phenomenon that is rarely studied in urban studies.
Another example is the role of human capital in explaining cross-city growth variation. The
urban economics literature developed in the West over the past five decades at a point of
time when the societies were at the end of the demographic transition. As a result, human
capital accumulation had a positive impact on city population growth primarily through the
migration channel. Yet, for societies undergoing a demographic transition, human capital
accumulation also has a potential negative effect on city growth rates as fertility rates are
dampened. The urban economics literature has little to say about fertility because there is
meagre variation in cross-city fertility rates at the end of the demographic transition. As this
paper documents, this is not the case in a country like India which is undergoing the
demographic transition.

1
Figure 1 shows the political boundaries within India in 2011 and the location of the ten largest Indian cities.
2 th
For instance, Sub-Saharan Africa faced urbanization without economic growth in the late 20 century posing
a challenge to standard urban theories (Fox, 2012). Urban growth forecasts in developing countries are also
problematic (Cohen, 2004).

2
Understanding long-duration processes such as urbanization and the demographic
transition require a historical analysis and so the time period for our analysis begins in the
1870s with the first Indian Census.3 It ends in 2020 with a comment on future urban
scenarios. The analysis itself is based on a unique database of demographic, economic,
geographic and social indicators covering regions and cities over time. It assembles a
database on the population of nearly 400 urban agglomerations (1871-2011) that
collectively comprise of over 60% of the Indian urban population. Data on 4,000 urban units
in 2001, district level metrics and a State level panel between 1971 and 2011 are also
analysed. The Appendix to this paper provides details on the compilation of the database
along with summary statistics of key variables.

The rest of the paper is arranged as follows. Section 2 documents four distinct features of
Indian urbanization based on a historical and comparative analysis. Section 3 analyses the
relationship between urbanization and the demographic transition. Section 4 studies the
growth of cities in India over the short and long run and section 5 concludes.

3
The Census of India in 1872 was asynchronous. Since 1881, it has been synchronous and has been held
th
regularly every ten years. The 15 All-India Census was held in 2011.

3
2. Urbanization in India
India’s rate of urbanization rose gradually from around 10% at the end of the 19th century to
31% in 2011 (Table 1) with wide regional variations (Figure 2).4 We document four
important features of this process below:

a) Masculine Urbanization:

India is one of the few countries of the world where urban sex ratios are relatively more
masculine than rural sex ratios (Figure 3A). While most parts of India have female deficits in
their aggregate sex ratios, this deficit is more pronounced in urban areas than rural areas , is
widespread across districts (Figure 3B) and has been a persistent feature for over a century
(Table 1, Col. 10 and 11). It occurs primarily because migration to cities for work is
overwhelmingly male-dominated in nature (Tumbe, 2015b). Major cities such as Mumbai
and Kolkata were male ghettos in the early 20th century with sex ratios hovering around 500
females per 1,000 males and rising over time mainly on account of natural growth within
the cities (Figure 4). Delhi and Surat’s phenomenal growth in the first and second halves of
the 20th century respectively were closely associated with worsening of sex ratios due to
large inflows of male migrants. In these male-dominated migration streams, younger
cohorts of migrants often replace older cohorts of migrants who leave the cities to retire in
the native regions with their left-behind families (Tumbe, 2015a). Families are separated for
long periods of time because of restrictions on female mobility either due to cultural norms
or adverse housing conditions. These migration streams are also not seasonal but semi-
permanent in nature yielding large remittance flows to the source regions (Tumbe, 2012).
This phenomenon has an important implication that net rural-urban migration flows are
considerably smaller than gross rural-urban migration flows due to return migration, leading
to slower urban growth and urbanization.

b) Service Sector Urbanization:

Between 1901 and 2016, the share of the manufacturing sector in Gross Domestic Product
(GDP) in India rose from less than 5% to around 15% while the share of the service sector
rose substantially from 25% to nearly 70% (RBI, 2016; Sivasubramonian, 2000). The skipping
of the manufacturing revolution in India is in stark contrast to the experience of East Asian
countries, and is reflected in its cities where most people earn a living from the services
sector. Virtually all Indian districts have over 50% of the main urban workforce engaged in
the services sector with the average crossing over 80% (Figure 5). While major cities such as
Mumbai and Kolkata did have a sizable share of the workforce in the manufacturing sector
in the early 20th century, by the early 21st century, the figure was less than 30%. Barring
4
The urban definition has broadly remained unchanged between 1961 and 2011. A substantial change
occurred between 1951 and 1961 leading to a reduction in the number of towns. Currently, the Indian Census
defines urban areas as (a) All settlements with a local urban body and (b) Settlements with a population of at
least 5,000 persons, density of at least 400 persons per sq. km. and at least 75% of the male main working
population engaged in non-agricultural activities.

4
Surat, all the Top-10 cities of India are service-dominated cities (Table 2) and formal sector
manufacturing plants appear to be moving to rural locations (Ghani, Goswami, & Kerr,
2012).

c) Strong Linkages between Economic Growth and Urbanization:

Globally, there exists a tight positive relationship between per capita income levels and
urbanization rates but limited evidence on the level of urbanization affecting economic
growth (Bloom, Canning, & Fink, 2008). Within India, there is a strong relationship between
per capita income levels and urbanization rates at the State level with the correlation
coefficient as high as +0.9 (Figure 6A). Relatively richer states such as Maharashtra, Kerala,
Tamil Nadu and Gujarat have urbanization rates above 40% while poorer states such as
Odisha and Bihar have urbanization rates less than 20%.

In addition, there also appears to be evidence that urbanization rates affect economic
growth rates at the sub-national level. The correlation coefficient between growth rates of
per capita incomes between 1971 and 2011 and urbanization rates of 1971 at the State level
is +0.8 (Figure 6B). It is also possible to replicate a Granger-Causality test on urbanization
and income levels conducted by Bloom et al. (2008) at the cross-country level, using panel
data on Indian states for the period 1971-2011 at ten-year intervals. Urbanization is said to
‘Granger-cause’ income if on controlling for past income levels, lagged urbanization rates
have predictive power for future levels of income. Table 3 reports the results of such a test
whereby bi-directionality of the Granger-causality between variables is established. This
demonstrates the intimate relationship between urbanization and economic growth within
India over the past four decades. This relationship can be attributed to the productivity
advantages of industry and urban agglomerations that have been documented in the Indian
context (Shukla, 1996; Lall, Shalizi, & Deichmann, 2004; Mohanty, 2014).

d) Relatively low level and slow pace of urbanization in India

At a global level, India’s urbanization rate is observed to be more than 10 percentage points
lower than the rate predicted for its level of per capita income (Deshingkar & Anderson,
2004). Further, the pace of urbanization has been slower in India than in many other
countries. Between 1971 and 2011, India’s urbanization rate increased from around 20% to
31%. In the same time period, many Asian and African countries more than doubled their
urbanization rates from similar starting points. China’s rate of urbanization has crossed the
50% threshold. India thus poses two challenges to urban trends: Why is the level of
urbanization so low relative to its per capita income level and why has the pace of
urbanization been slow in the last four decades?

Two explanations often discussed pertain to urban definitions and migration restrictions.
India uses a conservative urban definition that assigns many settlements to be ‘rural’ when
they would have been classified as ‘urban’ in other countries. A more liberal definition

5
would raise the urbanization rate and explain part of the difference between the actual
urbanization rate and the rate predicted at the corresponding per capita income level. As
shown in Figure 7, the urbanization rate in India in 2011 could vary between 31% based on
the official definition and 47% if rural settlements with population exceeding 5,000 were
classified as urban areas, a practice followed in many countries. However, Figure 7 also
shows that irrespective of the definition used, the change in urbanization rate between
2001 and 2011 remains constant at 3 percentage points. That is, a definition-based
explanation addresses the issue of low urbanization but not slow urbanization.

Another explanation on the relatively low level and slow pace of Indian urbanization focuses
on migration restrictions. Census statistics on in-migration show low rates of mobility
especially for non-marriage related reasons.5 As per one argument, caste networks provides
an important source of informal insurance in rural settings and thus acts as a barrier to
migration (Munshi & Rosenzweig, 2016). There are, however, other perspectives on Indian
migration. Census based statistics on work-related migration have been shown to be
seriously deficient (Deshingkar & Akter, 2009) and the average Indian district’s male out-
migration rate (for work) computed from the National Sample Surveys is above 5% (Tumbe,
2015b), a threshold denoting ‘mass migration’ (Hatton & Williamson, 1998, p. 9-10). As
pointed out earlier, the gendered nature of rural-urban migration for work and eventual
return migration considerably slows down the pace of urban growth and urbanization.

Definition and migration-based explanations have, however, overlooked another potential

reason for India’s relatively low level and slow pace of urbanization. This is related with the
fact that urbanization also depends on rural-urban differences in natural growth rates which
correspond to their respective demographic transitions and this reason, as the next section
will elaborate, is vital in resolving a major urban puzzle.

5
Female migration rates are extremely high on account of marriage due to village exogamy.

6
3. Urbanization and the Demographic Transition
The population of India grew at an annual rate of less than 1% between 1872 and 1921 in a
period marked by intense famines in the 1870s and 1890s, plague in the first decade of the
twentieth century and the influenza epidemic of 1918-19. Since 1921, the annual population
growth rate steadily climbed upwards as disease and climate-linked mortality rates plunged.
The fertility transition began only in the 1960s as birth rates gradually declined from 45
births per 1,000 to 24 births per thousand in 2011. These trends in birth and death rates
over the long run are depicted in Figure 8 which also shows the inverted U-shape path of
natural growth rates that peaked in the 1970s.6

The transition of natural growth rates has however varied substantially at the sub-national
level. For instance, in 2011, Bihar in the north experienced a natural growth rate of 21 per
1,000 or 2.1% per annum as compared to the relatively richer state of Tamil Nadu in the
south that experienced a natural growth rate of only 9 per 1,000 or 0.9% per annum (Figure
9). These wide variations in natural growth rates were driven primarily by differences in
birth rates as death rates have converged across states in the past four decades.7 The
differing pace of the regional fertility transition has been attributed to differences in human
capital accumulation (Murthi, Guio, & Dreze, 1995) and slow diffusion of low-fertility norms
from the coastal south (Guilmoto & Rajan, 2001). As a result, there exists a broad North-
South fertility divide with the poorer states of the northern hinterland exhibiting higher
birth rates than other regions. This is a well-documented feature of Indian demography.

In addition to this North-South fertility divide, there exists another remarkable feature of
India’s demographic transition. This is the demographic divergence between rural and urban
natural growth rates since the 1970s, as depicted in Figure 10. In the 1970s, birth and death
rates were both lower in urban than rural areas but their difference or the natural growth
rates was identical. Since then, rural and urban natural growth rates have diverged such
that rural natural growth rates are now substantially higher than urban natural growth
rates. This is a significant phenomenon as it indicates that in the absence of migration and
reclassification effects, India has been de-urbanizing for four decades. In other words, the
demographic divergence has dragged down the pace of India’s urbanization and needs
careful scrutiny.

A clearer picture emerges in the state-level analysis. As shown in Figure 11, the four
southern states did not undergo the demographic divergence observed at the All-India level.
In these states, the decline of rural and urban natural growth rates occurred at the same
pace. This trend is also observed in the relatively richer states of Maharashtra and Gujarat in
the west and Punjab in the north. The states which did undergo the demographic

6
Natural growth rate is the difference between Birth Rate and Death Rate.
7
Across-state birth rates ranged between 25 and 45 births per 1,000 in 1971 and between 14 and 30 births per
1,000 in 2011. In contrast, death rates have converged across states ranging from 5 and 20 per 1,000 in 1971
to 5 and 8 per 1,000 in 2011.

7
divergence are shown in Figure 12. The poorer northern states - Bihar, Madhya Pradesh,
Rajasthan and Uttar Pradesh, often referred to as the BIMARU states - exhibit a clear
demographic divergence from the 1970s. As these states have a large demographic weight
in India’s population, the divergence within these states translates into the overall
divergence noted at the All-India level. Assam, a relatively poor north-eastern state also
exhibits the demographic divergence. In West Bengal, there is a large difference between
rural and urban natural growth rates but no evidence of divergence from a common starting
point. The difference owes itself to the fact that Kolkata, a city with very low fertility levels,
comprises a large share of the state’s urban population.

The data on birth and death rates make it clear that the proximate cause of the
demographic divergence has been the stagnant rural-urban fertility differential in
conjunction with a narrowing of the rural-urban mortality differential.8 At the sub-national
scale, the demographic divergence occurs in those regions where the rural-urban fertility
differential does not exhibit narrowing tendencies.

In order to further understand the drivers of the demographic divergence, we analyse rural-
urban differences at the district level. Standard fertility metrics such as Total Fertility Rate
(TFR) and Birth Rate are not available separately for rural and urban areas at the district
level. Instead, we construct a Fertility Index from Census data based on the child-woman
ratio- Children aged 0-6/ Females aged 7+, that is available for rural and urban areas. This
index correlates extremely well with standard measures of fertility at the state level. The
correlation coefficient is above +0.9 between the Census based fertility index and TFR
derived from the National Family Health Survey and Crude Birth Rate derived from the
Sample Registration System statistics.

The rural-urban differences in the Census-based fertility index at the district level are
illustrated in Figure 13. In line with the earlier analysis, the map shows districts within the
BIMARU states and Assam with high values and most other districts especially in the south
with low values.9 Across-district variation in rural-urban fertility differences occurs primarily
on account of large variation in rural fertility rather than urban fertility.10 The question
therefore is why some regions have lower rural fertility than others?

The vast literature on fertility stresses the significance of the negative relationship between
fertility on the one hand and education and income on the other. Education expands the
knowledge on contraceptive methods and education of females in particular increases the
opportunity cost of child-rearing activities (Dreze & Murthi, 2001). The income effect on
fertility has been theorized in the quantity-quality trade-off framework whereby higher
8
A narrowing of the rural-urban fertility differential in India based on child-woman ratios was observed before
the 1960s (Robinson, 1961).
9
This provides additional validity to the accuracy of rural-urban differences in natural growth rates revealed
through SRS data.
10
The correlation coefficient between the rural-urban fertility gap and rural and urban fertility is 0.8 and 0.26
respectively.

8
incomes are invested in increasing the ‘quality’ of fewer children (Becker, 1960).
Accordingly, we take rural literacy rates and agricultural productivity per hectare as proxies
for rural education and income in our regression framework to understand the drivers of
cross-district rural-urban fertility differences.11

Table 4 reports the regression results. Both education and income variables are highly
significant variables with the expected signs and together explain about 26% of the variation
in rural-urban fertility differences. These variables are statistically significant even after
controlling for state level conditions that absorb variation in policies and governance
conditions (Column 2, Table 4). To guard against the concern that most of the variations are
essentially North-South differences in demographic conditions, a separate regression is run
on a sample of 90 districts within the southern states (Column 3, Table 4). Agricultural
productivity continues to have strong explanatory power at the within-state level for
southern states whereas rural literacy ceases to be an important factor, presumably due to
limited variation in the variable in the south. The stability of the regression coefficients
shows the robust relationship between rural education and income against rural-urban
fertility differences at the district level.

Based on these empirical results, it appears that regions with relatively lower agricultural
productivity and rural literacy rates are associated with higher rural-urban fertility
differences. Conversely, boosting agricultural productivity and rural literacy rates would
speed up urbanization by narrowing the rural-urban fertility differential. This would not
necessarily increase or reduce out-migration as the observed link is weak in India.12 Thus
rural prosperity is demographically linked with the process of urbanization through a
channel beyond migration.

Finally, it is possible to simulate the impact of the demographic divergence on urbanization

rates. Table 5 reports results from a state level panel regression of urbanization rates on
lagged values of rural-urban differences in natural growth rates and other variables. It
shows the robust negative impact of the demographic divergence variable on urbanization
rates. The results suggest that closing the rural-urban divide between natural growth rates
would lead to an increase in urbanization rates by over 4 percentage points. This would
explain nearly 50% of the observed gap between India’s official rate of urbanization and that
predicted at its level of income.

India’s low and slow pace of urbanization thus hinges on three important factors. The
lowness occurs partly because of the nature of India’s urban definition and the slowness
occurs because of the demographic divergence and the highly gendered nature of work-
related migration that leads to considerable return migration from cities to villages.

11
Multicollinearity issues are avoided as rural literacy and agricultural productivity per hectare have weak
correlation.
12
The correlation between the out-migration variable depicted in Figure 21 and rural literacy or agricultural
productivity across districts is close to zero.

9
4. The Growth of Cities in India
By the end of the 19th century, the British colonial state directly ruled 60% of the area of the
Indian subcontinent and indirectly ruled the remaining area through princely states (Roy,
2006). The colonial state operated chiefly from the three port cities of Kolkata, Mumbai and
Chennai13 that reflected the significance of international commerce. By 1901, they were the
three largest cities of India.14

Figure 14A shows the dominance of these three cities by mapping major urban
agglomerations of 2001, scaled by the population in 1901.15 New cities had sprung up in the
late 19th century as transportation hubs due to the railways, hill stations and cantonments
for the ruling elite, and market towns for exporting agricultural products in the first wave of
globalization (Heitzman, 2008). Small industrial clusters for specialized products also
emerged, notably in Tamil Nadu along the irregular quadrilateral of Coimbatore, Salem,
Tiruchirapalli and Madurai (Ghurye, 1974). Other prominent cities in the early 20th century
included cities in modern day Uttar Pradesh, Ahmedabad and Hyderabad (capital of a
princely state).

After Independence in 1947, the Indian state built selected new cities for political
administration at the state level (Chandigarh, Bhubaneshwar, Gandhinagar) and industrial
townships but planning policies carried an anti-urban bias and a lopsided funding focus on
rural development projects (Ramachandran, 1989). The liberalization of the economy in
1991 marked another milestone as India opened up to the world economy inviting foreign
investment that was directed to specific sectors and cities (Shaw, 2012).

Figure 14B maps the population of major urban agglomerations in 2011 and a comparison
with Figure 14A reveals important similarities and differences. Both maps have large vacant
spots in central and north-eastern India reflecting the absence of major urban
agglomerations in those spaces spanning a century. The most significant difference is the
appearance of Delhi as a large city as it grew rapidly in the 20th century after it was chosen
as the new capital of British India in 1911.16 The triangle between Bangalore, Hyderabad and
Chennai is also more prominent in 2011 than 1901. The rise of Pune, Surat and Ahmedabad
relative to major cities in Uttar Pradesh is also visible.

13
They were called Calcutta, Bombay and Madras respectively in colonial documents. In this paper, current
names are used to describe all locations.
14
This was in contrast to the Mughal empire that operated out of six primate cities - Lahore, Delhi, Agra,
Patna, Burhanpur and Ahmedabad- dominating internal trade. Surat, the port city, served an economic rather
than a political function (Chaudhuri, 1978).
15
62 of the top 100 urban agglomerations in 2001 were also among the top 100 list of cities in 1901.
16
Delhi is the second largest urban agglomeration after Mumbai as per the Indian Census of 2011 but the
largest city of India as per the United Nations. Unlike the Census, the UN considers urban units outside the
National Capital Territory to be a part of the Delhi agglomeration.

10
This brief overview of urban growth patterns follows the work of social scientists amidst a
large body of research on the subject17 and yet curiously enough, urban growth has not
adequately been analysed in conjunction with the demographic transition. Even though
natural increase contributed to over 50% of urban growth in the past four decades (Bhagat,
2011), most studies on urban growth in India focus on migration and not fertility or
mortality.

Figure 15 shows how closely urban growth in India has followed the logic of the
demographic transition. The annual population growth rate distributions of large urban
agglomerations comprising over 60% of the Indian urban population shift in accordance with
the national demographic transition outlined in Figure 8.18 Between 1881 and 1921 the
average annual growth rate was 0.8% with a large proportion of cities experiencing negative
population growth rates. Plague and other epidemics of this era led to high death rates and
the contribution of natural increase to urban growth would have been minimal. Average
annual growth rates rose to 2.6% in the 1921-1951 period and 3.4% in the 1951-1991
period. In these periods, virtually no major city experienced population declines due to
significant declines in mortality rates. Average annual growth rates then fell in the most
recent 1991-2011 period to 2.6% mimicking the decline in natural growth rates observed at
the national level. A few cities also observe population declines in the recent period.

While the demographic transition is the principal reason behind these distributional shifts,
Figure 15 also shows the considerable variation in city growth rates within each selected
period. Why did some Indian cities grow faster than others within particular time periods?
We begin to address this question by considering the proximate determinants of urban
growth for a time period where sufficient data is available i.e. 2001-2011. The three
proximate determinants are natural increase as proxied by fertility19, migration as proxied
by the share of recent migrants in total population and area reclassification. The last factor
is important as boundaries of municipalities and urban agglomerations expand or contract
upon local considerations rather than clearly defined criteria.

Table 6 reports the results of a regression of annual city population growth rates (2001-
2011) on the three factors and other geographical and political variables noted in the
literature.20 All the three core factors are highly significant variables and collectively explain

17
See Bose (1980), Kundu & Gupta (1996), Sivaramakrishnan et al. (2005) and Denis et al. (2012) and the
references therein.
18
This holds true even for the full sample of 4,000 odd urban units. Results not shown for the sake of brevity.
19
City level death rates are unavailable but at the State level, there is little variation in urban death rates
suggesting that differences in urban natural growth are due to differences in urban birth rates.
20
The literature on the growth of cities (da Mata, Deichmann, Henderson, Lall, & Wang, 2007; Duranton &
Puga, 2014; Henderson, 2003) outline many variables that are not easy to operationalize in the Indian context.
Three studies on India (C. M. Becker, Mills, & Williamson, 1986; Mitra & Mehta, 2011; Sridhar, 2010) note the
significance of growth-pole and market size effects linked with agglomeration.

11
nearly 60% of the observed inter-city variation in growth rates.21 In addition, there is strong
evidence of growth-pole or market size effects whereby cities further away from the leading
eight cities experience significantly lower growth rates.

The ‘area’ variable is also highly significant and adds considerable explanatory power to the
regression model. The coefficient size is substantially less than one suggesting that city
expansion is linked with rising densities. This point is elaborated further in Figure 16 which
shows the distribution of population densities of over 4,000 urban units in 2001 across
Census city class sizes. Average population densities nearly double from 2,500 persons per
sq. km. in tiny towns to 5,000 persons per sq. km. in large cities and the rise in densities is
gradual across class sizes. Thus, Indian cities tend to grow by area and density as they grow
larger.

The other two proximate determinants – fertility and migration- are studied in greater detail
below. Figure 17 provides a schematic of city growth that combines combinations of high
and low fertility with high and low rates of net migration. Thus, cities could be classified into
four categories – internal growth (with minimal migration), external growth (with minimal
fertility), explosion (with high fertility and migration) and stagnation (with low fertility and
migration). The demographic transition is mostly a one-way transition from high to low
fertility rates whereas cities can switch from low to high migration rates and vice versa. Such
a schematic is useful to understand city growth in societies undergoing the demographic
transition as there is considerable variation in fertility rates. Figure 18 fits Indian data to the
schematic and shows the wide variation in fertility and migration rates across cities.
Between 2001 and 2011, a city like Kolkata faced ‘stagnation’, Sambhal in Uttar Pradesh had
‘internal growth’, Bhubaneswar witnessed ‘external growth’ and Surat experienced
‘explosion’. Thus the interplay between fertility and migration has a major bearing on the
variation of inter-city growth rates.

Figure 19 maps the fertility index in major urban agglomerations and shows that the broad
north-south divide in general fertility levels extend into the urban realm. The correlation
coefficient between urban fertility and urban literacy rates was below -0.6 in 2011, revealing
the significance of human capital formation in pushing down city growth rates via the
fertility channel. In the absence of migration and area reclassification, northern cities are
growing much faster than southern cities due to lower literacy rates, higher fertility rates
and higher internal growth. This has important implications for urban planning as
infrastructure requirements depend on the population base of the city. It also shows that
reduction of urban population growth rates can be achieved by investing in urban education
rather than restricting migration.

21
The unexplained part of the model is due to the effects of out-migration, mortality and the lagged nature of
the in-migration variable due to data constraints. Migration data for Census 2011 was not available at the time
of writing.

12
Migration to cities itself is the function of many variables including wages, employment
opportunities and amenities and these are analysed in a regression framework. Wages are
proxied by literacy rates to reflect the human capital-productivity links that is a standard
feature of the urban economics literature. Employment opportunities are proxied by
workforce shares in particular sectors that experienced high economic growth rates and
temperature range is taken to be an important physical amenity in the Indian context. Table
7 reports the regression results with in-migration to cities in the 1990s as a share of city’s
population in 2001 as the dependent variable. After controlling for state level variables that
absorb institutional factors, city level literacy rates are a strong predictor of migration to
cities, confirming a stylized fact of urban economics. Two sectors – public administration
and the Information and Technology (IT) sector – that saw considerable expansion in the
1990s also help explain inter-city variations in migration rates.22 Temperature range is not a
significant variable in the model that collectively explains over 40% of the variation in
migration rates. Migration networks are also important in determining migration corridors
to particular cities, though they are hard to operationalize at the cross-city level in a
regression framework. Figures 20, 21, 22A and 22B show the clusters within India that
receive migrants and send out migrants for work and towards particular cities. Networks
from clusters in the Gangetic and coastal plains and other parts of India have sustained for
long periods of time (Tumbe, 2012) and explains why some cities have had larger access to
migrant pools than others.

Taken together, the results above show the mixed impact of human capital on city growth in
a country undergoing the demographic transition. Human capital has a positive impact on
city growth through the migration channel but a negative impact through the fertility
channel. Unsurprisingly, it emerges as an insignificant variable in explaining the cross-city
growth variations between 2001 and 2011.23

A long run regression analysis of city growth rates is constrained by data limitations.
Working within these limitations, Table 8 shows some of the political variables associated
with urban growth over the long run. Until 1921, cities associated with direct British rule
experienced higher population growth rates after controlling for various geographic
indicators. Since mortality was a major factor influencing urban growth in this period, it is
likely that British ruled cities had better investments in health systems due to the residence
of the ruling elite. Further, major political headquarters have had a statistically significant
impact on city growth rates across most periods. The decline in the coefficient after 1991
suggests a weakening impact of political centralization on city growth. By using more
historical data on Indian cities, future research can expand the range of variables that affect
city growth over the long run.

22
The choice of sectors to be included in the regression was based on the level of statistical significance.
Results do not change on the inclusion of non-significant sectors.
23
That is, running the Table 6 regression by replacing fertility and migration with literacy rates. Results not
shown.

13
5. Conclusion
Demographic transitions matter for urban transitions. As the evidence for India shows, the
differing pace of demographic transition along the rural and urban axes can affect the pace
of urbanization. The differing pace of demographic transition across cities can affect the
inter-city variation in population growth rates. The regional dynamics within India lead to
the apparent North-South paradox of rapid urban growth and slow urbanization in the
North versus slower urban growth and rapid urbanization in the South. That is, northern
cities are growing faster than southern cities on account of higher fertility rates but
urbanizing slower than the south because northern villages are also growing much faster
than southern villages. The dynamics of urbanization therefore requires a careful
examination of urban and rural conditions as rural prosperity is demographically linked with
greater urbanization. Boosting agricultural productivity and rural literacy levels in the
northern hinterlands would arguably lead to greater urbanization, not less, by narrowing the
rural-urban fertility differential.

City growth rates peaked in the 1980s and will see a marked deceleration in the coming
decades with substantial variations driven by investments in specific sectors such as IT. It is
worth re-emphasizing these two points of the paper – on slow urbanization and slower
urban growth than before- because the discourse in media outlets is diametrically opposite,
centred as it is on rapid urbanization and booming cities. The rural-urban fertility differential
shows no signs of narrowing in recent data and fertility continues to decline in major cities.
In Kolkata, some wards experienced population declines between 2001 and 2011.

The regional variations have significant policy implications as more urban infrastructure
would be needed in northern India relative to the south due to existing fertility differentials
and projected urban growth rates. Investment in urban education would reduce internal
growth but also lead to external growth via the channel of human capital formation and
migration. Finally, male-dominated migration streams for work leads to masculine and
slower urbanization due to high rates of return migration. Investments in cities that support
gender-balanced migration streams would go a long way in encouraging structural
transformation and urbanization. It is ironic that better agricultural productivity and
women’s hostels in cities appear to be pre-conditions for faster urbanization in India.

14
Appendix: Data Sources
Most of the data used in his paper come from three sources – Sample Registration System
or SRS (Government of India, 2009), Central Statistical Organization (CSO), and the Census.
Census data, where unspecified, refers to the Primary Census Abstract. Other datasets
include the 64th round National Sample Survey (2007-08), United Nations Demographic
Yearbook of 2013 and geographical variables based on satellite imagery.

A dataset on population of nearly 400 urban agglomerations in the broad jurisdiction

boundaries of 2001, for all the Census years between 1872 and 2011 was compiled by
linking data from Mitra (1980), Government of India (2007) and Census 2011. A sample of
this series was cross-checked with Census reports of the early 20th century.

In 2011, there were 35 states and union territories, 640 districts, 468 urban agglomerations
and 7,935 urban units or towns and cities in India.

Summary Statistics: State panel database, 1971-2011, in 5 intervals of 10 years each

Variable Obs. Mean Std. Dev. Min Max Source

Urbanization % 164 30.29 20.84 0.0 97.5 Census
Log Number of Towns 162 4.2 1.9 0.0 9.0 Census

Rural-Urban Natural Growth Difference,

Rate per 1,000 people 138 2.98 3.60 -4.7 13.6 SRS

Log Per Capita Real Income, 2004-05 series

(Net State Domestic Product Rs.) 128 9.9 0.68 8.6 12.0 CSO

Summary Statistics: District Database

Std.
Variable Obs. Mean Dev. Min Max Source
Rural Fertility Index %, 2011 584 34.8 9.0 16.5 61.8 Census
Urban Fertility Index %, 2011 590 28.8 5.9 16.5 57.2 Census
Rural Urban Difference in Fertility
Index %, 2011 581 5.94 5.39 -10.5 31.7 Census
Log Agricultural Yield per Hectare, Chand et al.
2003-04 541 10.2 0.6 8.0 11.9 (2009)
Rural Literacy Rate, 2011, % 584 69.5 10.5 34.7 97.6 Census

15
Summary Statistics: Urban Agglomeration Database
Std.
Variable Obs. Mean Dev. Min Max Source
Annual City Population Growth Rates
%, 1881-1921 301 0.8 1.1 -1.7 4.5 Census
Annual City Population Growth Rates -
%, 1921-1951 349 2.6 1.3 10.5 8.4 Census
Annual City Population Growth Rates
%, 1951-1991 373 3.4 1.4 0.0 11.2 Census
Annual City Population Growth Rates
%, 1991-2011 383 2.6 1.6 -0.5 14.3 Census
Annual City Population Growth Rates
%, 1881-2011 299 2.2 0.6 0.8 4.3 Census
Annual City Population Growth Rates
%, 2001-2011 385 2.6 2.6 -1.6 25.8 Census
Annual City Area Growth Rates %, -
2001-2011 360 2.6 4.9 11.4 27.2 Census
Urban Fertility Index %, 2001 394 32.5 6.5 19.8 50.5 Census
Literacy Rate %, 2001 394 78.6 7.2 52.2 97.2 Census
Recent Migrants (<10 Years) as % of
Population, 2001 393 16.9 7.4 2.9 62.1 Census, D-3
Temperature Range (degree celsius) 393 23.7 11.4 3.0 46.7 Census
% of Workforce in Public
Administration (NIC 75) 327 9.8 5.1 0.9 35.9 Census, B-22
% of Workforce in IT sector (NIC 72) 326 0.2 0.3 0.0 3.2 Census, B-22
Map
British rule=1 394 0.7 0.5 0.0 1.0 inspection

16
References
Becker, C. M., Mills, E. S., & Williamson, J. G. (1986). Modeling Indian Migration and City
Growth, 1960-2000. Economic Development and Cultural Change, 35(1), 1–33.

Becker, G. (1960). An Economic Analysis of Fertility. In NBER (Ed.), Demographic and

Economic Change in Developed Countries. Columbia University Press.

Bhagat, R. B. (2011). Emerging Pattern of Urbanisation in India. Economic and Political

Weekly, 46(34), 10–12.

Bloom, D. E., Canning, D., & Fink, G. (2008). Urbanization and the Wealth of Nations.
Science, 319(5864), 772–775.

Bose, A. (1980). India’s Urbanisation, 1901-2001. New Delhi: Tata Mcgraw Hill.

Chand, R., Garg, S., & Pandey, L. (2009). Regional Variations in Agricultural Productivity: A
District Level Study (No. National Centre for Agriculture for Agricultural Economics and
Policy Research Discussion Paper 01/2009). New Delhi.

Chaudhuri, K. N. (1978). Some Reflections on the Town and Country in Mughal India.
Modern Asian Studies, 12(1), 77–96.

Chauvin, J. P., Glaeser, E., Ma, Y., & Tobio, K. (2016). What Is Different About Urbanization in
Rich and Poor Countries? Cities in Brazil, China, India and the United States (National
Bureau of Economic Research Working Paper Series No. 22002).

Cohen, B. (2004). Urban growth in developing countries: A review of current trends and a
caution regarding existing forecasts. World Development, 32(1), 23–51.

da Mata, D., Deichmann, U., Henderson, J. V., Lall, S. V., & Wang, H. G. (2007). Determinants
of city growth in Brazil. Journal of Urban Economics, 62(2), 252–272.

Davis, K. (1962). Urbanisation in India: Past and Future. In R. Turner (Ed.), India’s Urban
Future (pp. 3–26). Berkeley: University of California Press.

Denis, E., Mukhopadhyay, P., & Zérah, M.-H. (2012). Subaltern Urbanisation in India.
Economic and Political Weekly, xlviI(30), 52–62.

Deshingkar, P., & Akter, S. (2009). Migration and Human Development in India. New York.

Deshingkar, P., & Anderson, E. (2004). People on the Move: New Policy Challenges for
Increasingly Mobile Populations. Natural Resources Perspectives, (92), 1–4.

Dreze, J., & Murthi, M. (2001). Fertility, education and development: Evidence from India.
Population and Development Review, 27(1), 33–63.

Duranton, G., & Puga, D. (2014). The Growth of Cities. In P. Aghion & S. N. Durlauf (Eds.),
Handbook of Economic Growth (1, Vol. 2, pp. 781–853). Elsevier.

Dyson, T. (2004). India’s Population: The Past. In T. Dyson, R. Cassen, & L. Visaria (Eds.),
Twenty-First Century India: Population, Economy, Human Development and the

17
 Environment (pp. 15–31). Oxford University Press.

Dyson, T. (2011). The role of the demographic transition in the process of urbanization.
Population and Development Review, 37(Supplement), 34–54.

Fox, S. (2012). Urbanization as a Global Historical Process: Theory and Evidence from sub-
Saharan Africa. Population and Development Review, 38(2), 285–310.

Ghani, E., Goswami, A. G., & Kerr, W. R. (2012). Is India’s Manufacturing Sector Moving
Away From Cities? (No. NBER Working Paper No. 17992).

Ghurye, G. S. (1974). Cities of India: Their Growth and Location. In M. S. A. Rao (Ed.), Urban
Sociology in India (pp. 178–216). Hyderabad: Orient Longman.

Government of India. (2005). General Population Tables, Part 1. Office of the RGI, Delhi.

Government of India. (2007). General Population Tables, Table A-4, Part II. Office of the RGI,
Delhi.

Government of India. (2009). Compendium of India’s Fertility and Mortality Indicators, 1971-
2007. Office of the RGI, Delhi.

Guilmoto, C. Z., & Rajan, I. Z. (2001). Spatial Patterns of Fertility Transition in Indian Districts.
Population and Development Review, 27(4), 713–738.

Hatton, T. J., & Williamson, J. G. (1998). The Age of Mass Migration: Causes and Economic
Impact. New York: OUP.

Heitzman, J. (2008). Middle Towns to Middle Cities in South Asia, 1800-2007. Journal of
Urban History, 35(1), 15–38.

Henderson, V. (2003). The Urbanization Process and Economic Growth: The So-What
Question. Journal of Economic Growth, 8(1), 47–71.

Kundu, A., & Gupta, S. (1996). Migration, Urbanisation and Regional Inequality. Economic &
Political Weekly, 31(52), 3391.

Lall, S. V, Shalizi, Z., & Deichmann, U. (2004). Agglomeration economies and productivity in
Indian industry. Journal of Development Economics, 73, 643–673.

Mitra, A. (1980). Population and area of cities, towns and urban agglomerations, 1872-1971.
Bombay: Allied Publishers.

Mitra, A., & Mehta, B. (2011). Cities as the engine of growth: Evidence from India. Journal of
Urban Planning and Development, 137(2), 171–183.

Mohanty, P. K. (2014). Cities and Public Policy: An Urban Agenda for India. New Delhi: Sage
Publications.

Munshi, K., & Rosenzweig, M. (2016). Networks and Misallocation: Insurance, Migration,
and the Rural-Urban Wage Gap. American Economic Review, 106(1), 46–98.

Murthi, M., Guio, A., & Dreze, J. (1995). Mortality, Fertility, and Gender Bias in India: A

18
 District-Level Analysis. Population and Development Review, 21(4), 745–782.

Ramachandran, R. (1989). Urbanization and Urban Systems in India. New Delhi: Oxford
University Press.

RBI. (2016). Handbook of Statistics on the Indian Economy, 2015-16. Mumbai.

Robinson, W. (1961). Urban-Rural Differences in Indian Fertility. Population Studies, 14(3),

218–234.

Roy, T. (2006). The Economic History of India, 1857-1947. New Delhi: OUP, 2nd Edition.

Shaw, A. (2012). Indian Cities. Delhi: Oxford University Press.

Shukla, V. (1996). Urbanization and Economic Growth. Delhi: Oxford University Press.

Sivaramakrishnan, K. C., Kundu, A., & Singh, B. N. (2005). A Handbook of Urbanization in

India. New Delhi: Oxford University Press.

Sivasubramonian, S. (2000). The National Income of India in the Twentieth Century. New
Delhi: Oxford University Press.

Sridhar, K. S. (2010). Determinants Of City Growth And Output In India. Review of Urban and
Regional Development Studies, 22(1), 22–38.

Tumbe, C. (2012). Migration persistence across twentieth century India. Migration and
Development, 1(1), 87–112.

Tumbe, C. (2015a). Missing Men, Migration and Labour Markets: Evidence from India. Indian
Journal of Labour Economics, 58(2), 245–267.

Tumbe, C. (2015b). The Missing Men: Sex Ratios and Migration. In S. I. Rajan (Ed.), India
Migration Report 2015: Gender and Migration. Delhi: Routledge.

19
Table 1: Urbanization and Urban Growth in India, 1871-2011

Annual
Annual Inter-
Number Number of Inter-Censal Censal Share of Share of
of Cities/UAs Urban Total Major 10 Major 10
Towns with Urban Total Population Population cities as % cities as % Urban Total
and Million+ Population Population Urbanization Growth Growth of Urban of Total Sex Sex
Year Cities Population (Million) (Million) % Rate % Rate % Population Population Ratio Ratio
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)
1871 1 18 212 8.7 944
1881 1 20 214 9.3 0.75 0.08 962
1891 1 22 234 9.4 1.03 0.92 962
1901 1,916 1 26 238 10.8 1.63 0.19 17.3 1.9 910 972
1911 1,908 2 26 252 10.3 0.04 0.56 19.4 2.0 872 964
1921 2,048 2 28 251 11.2 0.80 -0.03 19.8 2.2 846 955
1931 2,220 2 33 279 12.0 1.77 1.05 19.0 2.3 838 950
1941 2,427 2 44 319 13.9 2.81 1.34 22.0 3.0 831 945
1951 3,060 5 62 361 17.3 3.52 1.26 24.4 4.2 860 946
1961 2,700 7 79 439 18.0 2.37 1.98 26.0 4.7 845 941
1971 3,126 9 109 548 19.9 3.29 2.24 26.5 5.3 858 930
1981 4,029 12 159 683 23.3 3.87 2.23 25.7 6.0 880 934
1991 4,689 23 218 846 25.7 3.16 2.16 25.4 6.5 894 926
2001 5,161 35 286 1,029 27.8 2.78 1.97 26.0 7.2 900 933
2011 7,935 53 377 1,211 31.2 2.80 1.64 24.6 7.7 929 943

Notes: Census in 1871 was asynchronous. Major ten cities refer to top ten cities in 2001 by population. Sex Ratio is females per 1,000 males. Source: Data before 1901 from
Dyson (2004) and after 1901 from various Census volumes, especially Census 2001, Table A-2 and A-4.

20
Table 2: Top Ten Indian Cities and Migrant Characteristics

General Migrants
%
Households Migrant % of
Annual Population with main % workers Migrant
Population Area Density occupation from as % of Workers
Population Growth in Sq. per Sq. Sex % Slum in mfg. % of Sex Rural total in Mfg.
City (Million) Rate % Km. Km. Ratio Population sector Population Ratio Areas workforce Sector
2011 2001-11 2001 2001 2001 2001 2007-08 2001 2001 2001 2001 2001
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
Mumbai 18.4 1.1 1,135 14,478 822 54 27 43 726 66 63 31
Delhi 16.3 2.4 889 14,490 822 19 27 43 784 62 62 27
Kolkata 14.1 0.7 1,034 12,776 869 32 27 28 873 57 44 27
Chennai 8.7 2.9 702 9,346 955 19 25 25 881 38 39 17
Bengaluru 8.5 4.1 540 10,561 908 10 22 37 829 40 50 30
Hyderabad 7.7 3.0 822 6,985 931 17 18 25 860 47 53 20
Ahmedabad 6.4 3.5 438 10,322 884 13 32 34 918 55 47 34
Pune 5.0 3.0 669 5,619 899 19 27 45 912 54 59 30
Surat 4.6 5.0 237 11,868 760 21 64 58 681 77 72 62
Jaipur 3.1 2.8 485 4,792 876 16 27 27 915 48 36 26

Notes: Census Data for all cities refer to the urban agglomerations as defined by Census authorities except for Jaipur, which refers to its Municipal Corporation. Figures for
percentages are rounded to nearest whole numbers. Slum population data refers to only the principal municipal corporation limits of each urban agglomeration.
Manufacturing sector in Col. 7 refers to NIC 2004 codes 15-37 and in Col. 12 refers to Category D Census classification. Migration data is by ‘place of last residence.’
Sources: Col. 1 & 2 from Census 2011, Col. 7 is computed from unit level data of NSS 2007-08. Data for all other columns are computed from Census 2001.

21
Table 3: Granger-Causality Tests between Urbanization and Income Levels at the State
Level, 1971-2011

Dependent Variable

Log Per Capita Real

Income Urbanization %
(1) (2)
Log Per Capita Real Income [T-10] 0.75 7.21
(0.134)*** (2.092)***
Urbanization % [T-10] 0.01 0.50
(0.006)** (0.153)***

State Fixed Effects Yes Yes

Time (Decade) Fixed Effects Yes Yes
Observations 92 92

Notes: Panel data of Indian States at 5-time period intervals: 1971, 1981, 1991, 2001, 2011. Source: Various
Census volumes and Central Statistical Organization. Income is Net State Domestic Product (Rs.), 2004-05
series, constant prices. Constant term not shown in table. Robust standard errors reported in parentheses.
***Significant at 1% level and **Significant at 5% level.

Table 4: Determinants of Rural-Urban Fertility Differences at the District Level

Dependent Variable: Rural-Urban Difference in Fertility Index %, 2011

(1) (2) (3)

Log agricultural yield per hectare -1.16 -0.94 -1.30

(0.347)*** (0.43)** (0.51)**
Rural literacy rate, % -0.24 -0.24 -0.04
(0.021)*** (0.022)*** (0.04)
Constant 34.32 32.40 18.04
(3.25)*** (4.31)*** (4.77)***
State Control Variables No Yes Yes
Restricted sample of Southern States No No Yes
Observations 539 539 90
R. Sq. 0.26 0.61 0.23
Notes: District level database. Source: Census and Chand, Garg, & Pandey (2009). Robust standard errors
reported in parentheses. ***Significant at 1% level and **Significant at 5% level.

22
Table 5: Impact of the Demographic Divergence on Urbanization Rates

Dependent Variable: Urbanization %

(1) (2) (3)

Rural-Urban Natural Growth Rate

Difference (10 year lag) -0.68 -0.71 -0.56
(0.213)*** (0.214)*** (0.202)***

Urbanization % (10 year Lag) -0.19 -0.12

(0.164) (0.152)
Log Number of Towns 6.54
(1.783)***
Constant 30.07 34.62 5.59
(1.264)*** (4.208)*** (8.79)
Observations 105 105 105
Time (Decade) Fixed Effects Yes Yes Yes
State Fixed Effects Yes Yes Yes
Notes: Panel data of Indian States at 5-time period intervals: 1971, 1981, 1991, 2001, 2011. Source: Various
Census volumes and SRS statistics. Standard errors reported in parentheses. ***Significant at 1% level and
**Significant at 5% level.

Table 6: Proximate Determinants of City Growth Rates, 2001-2011

Dependent Variable: Annual City Population Growth Rate, 2001-2011, %

(1) (2)
Urban Fertility Index, 2001 0.10 ** 0.08 ***
Recent Migrants (<10 Years) as % of Population 0.07 *** 0.08 ***
Distance from Major Top-8 Metro City:
(within 100 Km as base category)
Between 100-300 Km -1.23 *** -0.76 **
Between 300-500 Km -1.58 *** -1.00 **
More than 500 Km -1.87 *** -1.30 ***
Annual Growth Rate of Area, % 0.31 ***
Capital City Dummy -0.13 0.48
Hilly Region -0.07 -0.06
Coastal City Dummy 0.19 0.55
River Dummy -0.01 0.07
State Dummy Variables Yes Yes
Number of Observations 384 360
R. Sq. 0.31 0.59
Notes: Database of Urban Agglomerations. Source: Census. Standard errors not reported. ***Significant at 1%
level and **Significant at 5% level. Constant term not shown in table.

23
Table 7: Determinants of in-migration to cities, 1992-2001

Dependent Variable: Recent Migrants (<10 years) as % of population, 2001

(1) (2)
Literacy Rate %, 2001 0.21 *** 0.15 **
% of workforce in Public Administration, 2001 0.21 ** 0.26 **
% of workforce in IT sector, 2001 8.00 *** 8.39 **
Temperature Range 0.08 ** 0.01
State Dummy Variables No Yes
Number of Observations 324 324
R. Sq. 0.20 0.43
Notes: Database of Urban Agglomerations. Source: Census Table B-22 for workforce data and D-3 for migration
data. Robust standard errors not reported. ***Significant at 1% level and **Significant at 5% level. Constant
term not shown in table.

Table 8: Political Variables and City Growth Rates over the Long Run, 1881-2011

Dependent Variable: Annual City Population Growth Rates

1881 1921 1951 1991 1881
to to to to to
1921 1951 1991 2011 2011

Direct British Rule=1 0.39 *** 0.01 -0.15 -0.11 0.01

Capital City=1 1.00 *** 0.43 1.94 *** 0.65 ** 0.96 ***
Control Variables Yes Yes Yes Yes Yes
R Sq. 0.34 0.08 0.28 0.04 0.36
N 279 349 370 382 277

Notes: Database of Urban Agglomerations. Source: Census. Direct British rule=1 if city was not in a princely
state. Capital City=1 refers to important political capitals within each time period. Robust standard errors not
reported. ***Significant at 1% level and **Significant at 5% level. Constant term not shown in table. Control
variables include dummy variables for Bihar, Uttar Pradesh, South India, coast, river and variables on
temperature range and log of base population.

24
Figure 1: States and Top Ten Cities of India, 2011 Figure 2: Change in State Urbanization %, 1901-2011

Source: Government of India (2005), Statement 4 & 8 and Census 2011.

25
Figure 3 (A): Rural-Urban Sex Ratios across Countries Figure 3 (B): Rural-Urban Sex Ratios across Indian Districts

Sex Ratio is females per 1,000 males.

N=120. Source: Table 7, United Nations Demographic Yearbook (2013) N=572. Source: Census 2011, Primary Census Abstract.

26
Figure 4: Selected City Sex Ratios, 1901-2011 Figure 5: District Level Distribution of Urban Workers, 2001

Source: Census 1901-2011. Data on Urban Agglomerations. Source: Census 2001, Table B-4.

27
 Figure 6 (A): Income Levels and Urbanization within India Figure 6 (B): Income Growth (1971-2011) and Urbanization in 1971

N =18 Large States, Correlation Coefficient=+0.9 N= 15 Large States, Correlation Coefficient= +0.8

Notes: CAGR is Compounded Annual Growth Rate. Income is Net State Domestic Product (Rs.) in 2004-05 constant prices. Two-Letter State codes formed by abbreviations
of State names.

Source: Census volumes and Central Statistical Organization.

28
 Figure 7: Urbanization in India in 2001 and 2011 across definitions

50 47

45 44

40 37
34
Urbanization %

35
31
30 28

25 2001

20 2011

15

10

0
Official Census Defintion Plus 10K+ in rural areas Plus 5K+ in rural areas

Notes: ‘Plus 10K+’ and ‘Plus 5K+’ refers to the inclusion of rural settlements with population more than 10,000
and 5,000 respectively in the definition of urban population. Source: Census 2001 and 2011.

Figure 8: The Demographic Transition of India

45
Birth Rate
35
Death Rate
Rate per 1,000

25

15

Natural Growth
5 Rate

1891 1901 1911 1921 1931 1941 1951 1961 1971 1981 1991 2001 2011
-5
Year

Source: Dyson (2004) and SRS statistics.

29
Figure 9: State Level Natural Growth Rates in Rural and Urban Areas, 2011

Source: SRS Bulletin Oct 2012, Vol. 47, No. 2.

Figure 10: Rural-Urban Demographic Divergence of Natural Growth Rates, 1975-2011

Source: Government of India (2009) and bulletins of the Sample Registration System, 2007-2012.

30
Figure 11: Demographic Divergence in the Southern States, 1975-2011

Source: Government of India (2009) and bulletins of the Sample Registration System, 2007-2012.

Figure 12: Demographic Divergence in Selected Northern States, 1975-2011

Source: Government of India (2009) and bulletins of the Sample Registration System, 2007-2012.

31
Figure 13: Rural-Urban Differences in Fertility at the District Level

Source: Census 2011.

32
Figure 14 (A): Population of large UA’s in 1901 (in 2011 boundaries) Figure 14 (B): Population of Large UA’s in 2011

Source: Various Census Volumes. Top 100 Urban Agglomerations in 2001. City bubbles scaled to population size.

33
Figure 15: City Population Growth Rate Distributions across Time Periods, 1881-2011

Source: Census 1881-2011. N=350+ agglomerations. Time Period I: 1881-1921, II: 1921-1951, III: 1951-1991, IV:
1991-2011.

Figure 16: Town & City Level Population Density Distribution across Class Sizes, 2001

Source: Census 2001. Class Sizes: 1 (Above 100,000 population), 2 (50,000 - 99,999), 3 (20,000-49,999), 4
(10,000-19,999), 5 (5,000- 9,999), 6 (Below 5,000 population). 89 urban units out of 4,307 urban units with
population densities exceeding 20,000 have been excluded.

34
 Figure 17: City Growth Typology based on Fertility and Migration

Figure 18: Fertility vs. Migration across Indian Urban Agglomerations

Source: Census 2011. N= 350+ urban agglomerations. Figures next to marked cities reflect annual growth rate
between 2001 and 2011. Fertility Index = % 0-6 Age Children / Women age 7+.

35
Figure 19: Fertility Index in Urban Agglomerations, 2011

Source: Census 2011. 468 Urban Agglomerations. Map bubbles are scaled by population. Fertility Index = 0-6
Age Children / Women age 7+.

36
Figure 20: Urbanization in India, 2011 Figure 21: Out-Migration for Work in India, 2007-08

Source: Census 2011 Source: Tumbe (2015), based on NSS 2007-08 data

37
Figure 22 (A): Source Region of Migrants in Major Cities, 1992-2001

Source: Census 2001, Table D-13. Data refers to internal migrants who migrated between 1992 and 2001 to urban areas of district/s spanning the major city-urban
agglomeration.

38
Figure 22 (B): Source Region of Migrants in Major Cities, 1992-2001: Kolkata, Chennai, Hyderabad, Bangalore, Pune, Surat

Source: Same as for Figure 22 (A).

39
The International Growth Centre
(IGC) aims to promote sustainable
growth in developing countries
by providing demand-led policy
advice based on frontier research.

Find out more about

our work on our website
www.theigc.org

For media or communications

enquiries, please contact
[email protected]

Subscribe to our newsletter

and topic updates
www.theigc.org/newsletter

Follow us on Twitter
@the_igc

Contact us
International Growth Centre,
London School of Economic
and Political Science,
Houghton Street,
London WC2A 2AE

Designed by soapbox.co.uk