I focus separately on the effects of climate change on non-primary and primary
output. Based on the past literature, the expectation is that climate change is most
likely to have important impacts on cities in coastal or riverine locations, in
resource-dependent regions, and in locations at risk at risk from extreme weather
events, especially those undergoing rapid urbanization or those whose economies
are closely linked with climate-sensitive resources (Hunt and Watkiss 2007).
Equation (1) was estimated at the district-level using a basic model. In this, we
examined simply the district’s net non-primary output (per capita for 2003–04) as
being dependent on various exogenous factors indicated in Eq. (1). In order to do
this, DDP estimates published by state directorates of Economics and Statistics in
the various states were obtained, with the result that we obtained information
regarding 500 districts in all states except a few which did not publish this.
2
Using the district instead of the town as the unit of observation has a distinct
advantage which is that only growing places are designated as towns, whereas even
areas that are declining in terms of population are designated as districts, with the
result that the selection bias that could arise with the choice of towns as the unit of
observation, does not arise with respect to districts. The non-primary portion of the
net DDP per capita at constant (1999–00 prices) was used as the measure of
city-level output. While gross and net DDP estimates at current and constant prices
were available for districts in all states except a few, the net DDP estimates for
2003–04 were chosen since for a few states such as Assam that was the most recent
year for which the data were available. The net estimates were chosen to exclude
any depreciation. The net DDP was computed in per capita terms to account for any
scale effects in the case of large or highly urban or metropolitan districts such as
Mumbai.
The exogenous variables summarized in Eq. (1) were constructed with utmost
care. The literacy rate for 2001 for the districts was obtained from the Census of
India. The remaining variables were constructed from the Census of India’s town
directories. The town directories contain data on maximum and minimum temperatures for all the 5000 towns in the country. The temperature difference was
calculated between the maximum and minimum temperatures for every town and
aggregated at the level of the district, the unit of observation here for various
reasons discussed earlier; this was used as an exogenous factor influencing city
economic growth, following the literature (for instance Haurin (1980) examines the
influence of climate on population and migration). The difference between the
maximum and minimum temperature for towns was aggregated (averaged) across
districts containing them and in this way, district-specific estimates of these temperature differences were obtained for all districts.
2
The states that did not publish DDP estimates are—Goa, Gujarat, Nagaland, Puducherry, and
Tripura.
286
K.S. Sridhar
output. Based on the past literature, the expectation is that climate change is most
likely to have important impacts on cities in coastal or riverine locations, in
resource-dependent regions, and in locations at risk at risk from extreme weather
events, especially those undergoing rapid urbanization or those whose economies
are closely linked with climate-sensitive resources (Hunt and Watkiss 2007).
Equation (1) was estimated at the district-level using a basic model. In this, we
examined simply the district’s net non-primary output (per capita for 2003–04) as
being dependent on various exogenous factors indicated in Eq. (1). In order to do
this, DDP estimates published by state directorates of Economics and Statistics in
the various states were obtained, with the result that we obtained information
regarding 500 districts in all states except a few which did not publish this.
2
Using the district instead of the town as the unit of observation has a distinct
advantage which is that only growing places are designated as towns, whereas even
areas that are declining in terms of population are designated as districts, with the
result that the selection bias that could arise with the choice of towns as the unit of
observation, does not arise with respect to districts. The non-primary portion of the
net DDP per capita at constant (1999–00 prices) was used as the measure of
city-level output. While gross and net DDP estimates at current and constant prices
were available for districts in all states except a few, the net DDP estimates for
2003–04 were chosen since for a few states such as Assam that was the most recent
year for which the data were available. The net estimates were chosen to exclude
any depreciation. The net DDP was computed in per capita terms to account for any
scale effects in the case of large or highly urban or metropolitan districts such as
Mumbai.
The exogenous variables summarized in Eq. (1) were constructed with utmost
care. The literacy rate for 2001 for the districts was obtained from the Census of
India. The remaining variables were constructed from the Census of India’s town
directories. The town directories contain data on maximum and minimum temperatures for all the 5000 towns in the country. The temperature difference was
calculated between the maximum and minimum temperatures for every town and
aggregated at the level of the district, the unit of observation here for various
reasons discussed earlier; this was used as an exogenous factor influencing city
economic growth, following the literature (for instance Haurin (1980) examines the
influence of climate on population and migration). The difference between the
maximum and minimum temperature for towns was aggregated (averaged) across
districts containing them and in this way, district-specific estimates of these temperature differences were obtained for all districts.
2
The states that did not publish DDP estimates are—Goa, Gujarat, Nagaland, Puducherry, and
Tripura.
286
K.S. Sridhar