Journal of Social and Development Sciences

Vol. 5, No. 2, pp. 111-122, Jun 2014 (ISSN 2221-1152)

Climate Change and Sugarcane Productivity in India: An Econometric Analysis

*Ajay Kumar, Pritee Sharma

Indian Institute of Technology Indore, India

*[email protected]

Abstract: This study provides an understanding for the relationship between climatic factors and
sugarcane productivity in India. The main objective of this paper is to estimates the impact of climatic and
non-climatic factors on sugarcane productivity. To check the consistency of empirical results, simple
linear regression model, Ricardian productivity regression (non-linear) model and Cobb-Douglas
production function models are employed. The data set incorporates 390 observations corresponding to
thirteen states with panel data for 30 years during 1980 to 2009. These all models include sugarcane
productivity as dependent variable. Irrigated area, agriculture labour, consumption of fertilizers, literacy
rate, tractors and farm harvest price (at constant level) are considered as explanatory variables. Average
rainfall, average maximum and average minimum temperature include as climatic factors to capture the
effect of climatic conditions on cane productivity. These climatic factors are incorporate for three weather
seasons such as rainy, winter and summer. Empirical results based on Prais Winsten models with panels
corrected standard errors (PCSEs) estimation shows that climatic factors i.e. actual rainfall, average
maximum and average minimum temperature have a statistically significant impact on sugarcane
productivity. The climatic effect for various factors on cane productivity are varies within different
seasons. Average maximum temperature in summer and average minimum temperature in rainy season
have a negative and statistically significant effect on sugarcane productivity. While, sugarcane
productivity positively get affect with increasing average maximum temperature in rainy season and
winter seasons. The study concluded that there is non-linear relationship between climatic factors and
sugarcane productivity in India.

Keywords: Climate change, sugarcane productivity, India

1. Introduction

Sugarcane is the most important cash crop and it has an important position in the agrarian economy of
India. Almost six million farmers grow sugarcane and also large numbers of agricultural laborers are
engaged in cane cultivation.1 It provides the employment opportunities to more than half a million
people, either skilled or semi skilled workers, mostly from rural areas. 2 Approximately 7.5% rural
population gets their basic livelihood resources, directly or indirectly, from sugar industries. In addition
to this it also gives some fuel as a byproduct along with a large number of high-cost, value-added products
(Shrivastava et al., 2011). It has a lion’s share in accelerating industrialization process and bringing socio-
economic changes in rural areas (Pandey, 2007). It is the second largest agro-processing industry, costs
almost Rs. 30000 crore, after textiles in India. 3 The area which occupied with sugarcane is around 4.4
million hectares and an average productivity is 68 tonnes/ha.4 India is one of the major producer as well
as consumer of sugar. It produces 18.9 million tonnes of sugar which is nearly 11.8% of the total sugar
production of the world.5 The sugarcane crops grow in tropical and subtropical zone in India covers 40-
60% of the total cane area of the country. Preceding section is dealing with importance of sugarcane
crops in India. Despite that there is limited research available in case of sugarcane productivity and
climate change. Based on extensive literature review of earlier studies the authors are confirmed that
productivity of major food grain and cash crops are likely to decrease with climate change (see literature
review section). But in case of sugarcane crops, this is unclear whether sugarcane productivity would be
increase or decrease in presence of climate change (Srivastava & Rai, 2012).

1 http://www.iisr.nic.in/iisrvision2030.pdf
2 http://www.iisr.nic.in/iisrvision2030.pdf
3 http://www.iisr.nic.in/iisrvision2030.pdf
4 http://www.iisr.nic.in/iisrvision2030.pdf
5 http://www.iisr.nic.in/iisrvision2030.pdf

111
In this regards there are many another crucial questions which sugarcane industries and agriculture
scientists have to answer like: What is influence of climatic and non-climatic variables on sugarcane
productivity in India? What is the annual variation in cane productivity due to climate change? How
sugarcane cane industries get affect with annual variation in cane productivity due to climate change.
Thus there is a need a research for climate change and sugarcane productivity to facilitate development of
appropriate farm policies. Due to this major drawback of earlier studies the present study tries to provide
an understanding about climatic, non-climatic factors and sugarcane productivity in India. The main
objective of this study is to assess the impact of climatic and non-climatic factors on sugarcane
productivity in different weather seasons such as rainy, winter and summer seasons in India. Since
sugarcane is annual crop and grows 12-18 months in the agriculture field in India. Hence this objective
would provide the scientific information that what are the influences of various climatic factors in
different growing time of sugarcane crop. The importance of this study tries to assess the seasonal wise
impact of climatic factors on sugarcane productivity.

2. Literature Review

This section of study is dealing with brief review of literature regarding climatic change and it impact of
crop productivity at global and national level. Onyeji & Fischer (1993) observed that climate change has
decreased the agricultural productivity, raised food prices and declined consumer incomes. This study
also reveals that climate change is declined the per capita food consumption in Egypt. Gbetibouo and
Hassan (2005) study identify that agriculture is a more vulnerable sector, physically and economically
due to climate change compared to other sectors of the economy. Further this study observed that climate
change has drastic negative impacts in agricultural production. Bosello and Zhang (2005) study estimated
the relationship between climate change and agriculture. This research shows that climate change is
complex and higher temperature will influence the production patterns. Deressa et al. (2005) applied a
Ricardian cross section regression model and found that sugarcane production is highly sensitive to
climate change. It has a negative impact on sugar production in South Africa (Deressa et al., 2005).
Masters et al. (2010) mentioned that climate change has a significant negative effect on agriculture
production that occupies around 40% of the land globally.

Ramulu (1996) analyzes the impact of annual rainfall and other socio-economic factors on cane
production in Andhra Pradesh. This study sown that rainfall does not have any significant impact on
sugarcane crop. Ramulu (1996)’s study employed a Cobb-Douglas production function with time series
data during 1973-1990. But this study did not include any other climatic variables like maximum and
minimum temperature to capture the temperature effect on cane production. Srivastava & Rai (2012)
also mentioned in their review article that there is need a research to identify the climatic effect on cane
productivity in India. In case of food grain crops, several studies provide the evidence that productivity of
food grain crop negatively affect due to climate change such as Saseendran et al. (2000) found negative
effects of temperature on rice productivity. They also mentioned that change in temperature up to 50C
can lead to continuous decline in rice yield and every one degree increment of temperature will lead to a
6% decline in rice yield in Kerala (India). Study by Nandhini et al. (2006) observed that cultivable land of
rice is lead to decrease due to the scarcity of inputs and low rainfall in Tamil Nadu (India). Hundal &
Prabhjyot (2007) employed a simulation model and mentioned that the increase in temperature by 10C
lead to decrease rice and wheat yield by 3% and 10% respectively in Punjab. Kar & Kar (2008) employed
a Cobb-Douglas production function to assess the rainfall effect on jowar production in Orissa (India).
This study included annual rainfall as climatic factors and concluded that low rainfall has negative impact
on jowar production as well as income of the poor farmers. This study also suggested that more
investment in irrigation would be useful to improved farm income in Orissa (India).

Kalra et al. (2008) undertook a state wise analysis for four states of India, namely Punjab, Haryana,
Rajasthan and Uttar Pradesh. The study also concluded that wheat, mustard, barley and chickpea
production has decreased due increase in seasonal temperature. Study by Kapur et al. (2009) mentioned
that rainfall may decrease crops yields by 30% by the mid 21 st century. This study also justified that there
would be reduction in arable land that could be results in more pressures on agriculture production In
India. Kumar & Parikh (2001) argued that the projected large-scale changes in the climate would lead to
significant reductions in rice and wheat crop yields by 2060. It may affect the food security of more than
one billion people in India. Haris et al. (2010) used a simulation model, mentioned that rice production
may lead to decrease by 31% in 2080 due to climate change in Bihar (India). Hari et al. (2010) analyzed
the climatic effects on paddy and corn crops. The analysis observed that climate change adversely affect

112
the paddy and corn crop since last decade in most of the districts of Uttar Pradesh (India). Kumar et al.
(2011a) concluded that climate change has shifted the weather condition which affected the seasonal
crops and reduced the available growing time of rice and sugarcane crops in Uttarakhand and Uttar
Pradesh (India). Geethalakshmi et al. (2011) also showed that the productivity of rice has declined by
41% with 40C increase in temperature in Tamil Nadu (India).

Kumar et al. (2011b) argued that irrigated area for the production of maize, wheat and mustard in
northeastern and coastal regions; rice, sorghum, and maize production in Western Ghats may decline due
to climate change. Gupta et al. (2012) analyzed the climatic impact on crop productivity of rice, sorghum
and millet at macro level. The authors included average temperature and actual rainfall in growing time
of these crops. The empirical findings of this study showed that climate change is likely to reduce the
yields of rice, sorghum and millet crop in 16 major agriculture intensive states of India. Kumar (2009)
employed a Ricardian cross sectional regression model to investigate the effect of climate sensitivity on
farm net revenue in India. Kumar (2009)’s study undertaken maximum temperature, minimum
temperature and actual rainfall in three weather seasons like autumn, summer, rainfall and winter. The
study concluded that climate change is results in 9% reduction in agricultural farm net revenues in India.
Kumar et al. (2014) investigated the impact of climatic and non-climatic factors on productivity of major
food grain crops in India using a Cobb-Douglas production at state level panel data in India. In this
analysis, the authors include average minimum temperature, average maximum temperature and actual
rainfall as climatic factors in growing time of each crops (sowing time to harvesting time). Empirical
result of the study reveals that productivity of wheat, barley, gram and rice crops are declined due to
increase in actual average minimum temperature. The productivity of rice, maize, sorghum, and ragi
crops are lead to decrease with increase in actual average maximum temperature in growing time of
corresponding crops.

Above discussion represents that climate change is very harmful for agricultural sector. Many studies give
the clear evidence that it decreased the agriculture productivity in different regions of India and other
countries of the world. In India, productivity of wheat, rice, maize and other food grains crops are likely to
be affected due to climate change. But in case of cane production, there are few study has been done such
as Srivastava & Rai (2012), Kumar et al. (2011a), Deressa et al. (2005), and Ramulu (1996). Hence the
present study tries to fill this research gap and provide a scientific understanding about climate change
and its impact on cane productivity in India. The present paper describes the empirical evidence about
impact of climate change on sugarcane production. The authors employed an average maximum
temperature, average minimum temperature and actual rainfall in rainy, winter and summer seasons.
This investigation would be helpful to identify that which climatic factor has negative effect on sugarcane
productivity in India. How various climatic factors have a seasonal influence on cane productivity in
different weather seasons?

3. Methodology

Source and Data Description: The data set for present study is covering 30 years at state level panel
data during 1980-2009. Thirteen major agricultural intensive states are taken from different agro-
ecological zones. The states which included from tropical zone are Bihar, Orissa, Uttar Pradesh, Punjab,
Haryana, Gujarat, Madhya Pradesh, West Bengal, Maharashtra and Rajasthan. Andhra Pradesh, Tamil
Nadu and Karnataka are included from subtropical zone. All these states are the major food grain and
cash crops (commercial crops) producers which contributes more than 75%, of the total production for
each crop of the country. To identify the missing values in the data set, the interpolation and graphical
methods are used. The data for agricultural, socio-economic and climatic variables are taken from
following sources:
1. Agricultural Data: Sugarcane production, area sown under cane cultivation, irrigated area, number of
tractor, consumption of fertilizer and farm harvest price of sugarcane (at constant level 1993-94 prices)
are taken from the Centre Monitoring Indian Economy (CMIE), Directorate of Economics, Statistics
Ministry of Agriculture (Government of India) and Agricultural Informatics Division National Informatics
Centre Ministry of Communications and IT (Government of India). Number of agricultural labors and
cultivators are taken from various publication of Census (Government of India).
2. Demographic Data: Agriculture labour and state wise number of literate population in rural area is
taken from various publication of Planning Commission (Government of India).
3. Climatic Variables: Minimum and maximum temperatures were obtained from the Indian
Meteorological Department (IMD) (Government of India) database. This data was available on daily

113
intervals with latitude and longitude (0.5 0C×0.5 0C grid scale) information of monitoring stations. Due to
the unavailability of city wise data of temperature the stations pertaining to specific latitude and
longitude information were identified. Based on this information the geographical regions were
identified. From the groups of such stations different geographical region were linked to arrive at the
state level data points. Monthly district wise rainfall information was taken from Hydromet Division,
Indian Meteorological Department (IMD, Government of India). These data were converted into city wise
monthly averages and then the data transformed in to state wise. Monthly maximum temperature,
minimum temperatures and rainfall for selected specific cities were collected from the 354
meteorological stations in thirteen states of India. The C++ software was used to process basic information
on climatic factors like rainfall, minimum temperature and maximum temperature data. The SPSS
software was used to extract and bring data to excel format. Average minimum temperature and average
maximum temperature and average rainfall in three weather seasons like rainy, winter and summer are
incorporated for the regression analysis. Regression analysis is run by Minitab, STATA and SPSS
softwares to fit the various proposed models.

Econometric Model: There are many methods are available in the existing literature to investigate the
impact of climate change on agriculture productivity. These are Production function model, Ricardian
cross sectional regression model, Agronomic-economic model, Agro-ecological zone model and integrated
assessment model (United Nations Report, 2011). To assess the impact of climatic and socio-economic
factors on sugarcane productivity, simple linear regression model, Ricardian type productivity regression
(non-linear) model and Cobb-Douglas production function model are employed in the present study.

1. Simple linear regression model (LR Model): This model assumes that climatic variables are similar
to other agriculture and socio-economic inputs for agriculture crop growth. This model is used by Mongi
et al. (2010) to investigate the impact of climate change on agriculture production in Tanzania. Another
study based on this model is undertaken by Haim et al. (2008) in Israel. Let (tp)st is sugarcane production
in a certain time period for particular state and this is a function of several socio-economic and climatic
factors. Simple linear regression model will be as-
(tp)st = f{(as)st, (ia)st, (tf)st, (al)st, (tt)st, (lr)st, (fhp)st, (arfrs)st, (arfws)st, (arfss)st, (amaxtrs)st, (amaxtws)st,
(amaxtss)st, (amintrs)st, (amintws)st, (amintss)st} (1)
Where, tp is total production for sugarcane; s is cross sectional groups of states 1 to 13; and t is the time
period for 1980-2009; as is area sown under sugarcane crop (in hectare); ia is irrigated area under
sugarcane crops (in hectare); tf is consumption of fertilizers at planted land under sugarcane crops (in
000 tones); al is utilization of agricultural labour for cane cultivation (in numbers); tt is use for tractor
for cane cultivation (in numbers); lr is literacy rate (in numbers) that is defined as ratio of literate rural
population with gross sown area multiply by area sown under sugarcane crop; fhp is the farm harvest
price for sugarcane crops (in rupees at constant level 1993-94 prices); arfrs is average rainfall in rainy
season (in millimeter); arfws is average rainfall in winter season (in millimeter); arfss is average rainfall
in summer season (in millimeter); amaxtrs is average maximum temperature in rainy season (in degree
Celsius); amaxtws is average maximum temperature in winter season (in degree Celsius); amaxtss is
average maximum temperature in summer season (in degree Celsius); amintrs is average minimum
temperature in rainy season (in degree Celsius); amintws is average minimum temperature in winter
season (in degree Celsius); amintss is average minimum temperature in summer season (in degree
Celsius). Now divide by tp to as (for production per hectare land or sugarcane productivity) and after
apply simple linear regression model than equation (1) will be as-
(tp/as)st = β0 +β1 (ia)st +β2 (tf)st + β3 (al)st + β4 (tt)st + β5 (lr)st + β6 (fhp)st + β7 (arfrs)st + β8 (arfws)st + β9
(arfss)st + β10 (amaxtrs)st + β11 (amaxtws)st + β12 (amaxtss)st + β13 (amintrs)st + β14(amintws)st+β15
(amintss)st+µst (2)
Where, (tp/as) is cane productivity that is output per hectare land; β0 is estimated constant coefficient and
β1, β2, β3, β4, β5, β6, β7, β8, β9, β10, β11, β12, β13, β14 and β15 are estimated regression coefficient for
corresponding explanatory variables and µst is an error term. Since this study is a panel data analysis and
that includes sugarcane productivity as output factors and other independent variables for thirteen states
for a time period, 1980-2009. So there are needed to estimate another test like state-level fixed effects
that is quite beneficial in capturing unobserved heterogeneity across states. Year-specific effects model is
useful to control for annual difference in sugarcane productivity for common to all states. State-by-year
fixed effects model is quit beneficial to capture the unobserved heterogeneity and to control annual
difference in sugarcane productivity (Gupta et al., 2012). After incorporating these variables, equation (2)
will be as-

114
(tp/as) = β0 +β1 (ia) +β2 (tf) + β3 (al) + β4 (tt) + β5 (lr) + β6 (fhp) + β7 (arfrs) + β8 (arfws) + β9 (arfss) + β10
(amaxtrs) + β11 (amaxtws) + β12 (amaxtss) + β13 (amintrs) + β14 (amintws) + β15 (amintss)+ξ1(s-1)SD(s-1)+€1(t-
1)TD(t-1)+ ψ1(s-1)+ (t-1) SD(s-1)×TD(t-1)+µst (3)
Where, SD(s-1) is the vector for states dummies; TD(t-1) is the vector for time dummies; ξ1(s-1) is the
estimated regression coefficient for state dummies; €1(t-1) is the vector of estimated regression coefficients
for time dummies. In the equation (3) state dummies and time dummies are used to capture the state-
level fixed effects and to control for annual difference in sugarcane productivity to all states. SD(s-1)×TD(t-1)
is the vector of combine states and time dummies; and ψ1(s-1)+(t-1) is vector of estimated regression
coefficients for combine states and time dummies to state-by-year fixed effects to capture the unobserved
heterogeneity and to control annual difference.

Ricardian productivity regression (non-linear regression) model (RP Model): In the present study
Ricardian productivity regression (non-linear regression) model is used to estimate that whether climatic
factors have a linear or non-linear relationship with sugarcane productivity. For this square root terms of
each climatic factor with original terms are also added in regression analysis. Ricardian regression
approach was used by many researchers to investigate the climatic impact on agricultural productivity in
different region of the world like Zhai & Zhuang (2009) in Southeast Asia; Kurukulasuriya & Mendelsohn
(2008) in Africa; Gbetibouo & Hassan (2005) in South Africa; Mano & Nhemachena (2007) in Zimbabwe;
Seo et al. (2005) in Sri Lanka. In present study sugarcane production on per hectare land is taken as a
proxy instead of net revenue on per hectare land and other explanatory variables are taken in the similar
ways. Specifically Ricardian cross sectional regression approach is a quadratic formulation of climatic
factors and a linear function of all other socio-economic factors (Mendelsohn et al., 1994; Kumar, 2009).
We added square root terms for all climatic factors instead of quadratic term and after apply a Ricardian
productivity regression model equation (3) can be written as-
(tp/as)st = §0 +§1 (ia)st +§2 (tf)st + §3 (al)st + §4 (tt)st + §5 (lr)st + §6 (fhp)st + §7 (arfrs)st + §8 (arfrs^1/2)st + §9
(arfws)st + §10 (arfws^1/2)st + §11 (arfss)st + §12 (arfss^1/2)st + §13 (amaxtrs)st + §14 (amaxtrs^1/2)st + §15
(amaxtws)st + §16 (amaxtws^1/2)st + §17 (amaxtss)st + §18 (amaxtss^1/2)st + §19 (amintrs)st + §20 (amintrs^1/2)st
+ §21 (amintws)st + §22 (amintws^1/2)st + §23 (amintss)st + §24 (amintss^1/2)st + ξ2(s-1)SD(s-1) + €2(t-1)TD(t-1) + ψ2(s-
1)+(t-1)SD(s-1)×TD(t-1)+ϕst (4)
Where, dependent and all explanatory variables are described in equation (1); §0 is estimated constant
coefficient. While, §1, §2, §3, §4, §5, §6, §7, §8, §9, §10, §11, §12, §13, §14, §15, §16, §17, §18, §19, §20, §21, §22, §23 and §24
are the estimated regression coefficients for corresponding explanatory variables and ϕst is an error term
in the econometric model. Equation (4) is similar to Ricardian model which was applied by Mendelsohn et
al. (1994) and includes climatic factors as a non-linear from and other variables as a linear form.

Cobb-Douglas production function model (C-D Model): Cobb-Douglas production function proposed
by Knut Wicksell (1851-1926) and tested against statistical evidence by Cobb & Paul (1928) in 1928 is
used.6 The authors also employed a Cobb-Douglas production model to assess the impact of climate
change on sugarcane productivity. This model was used by Kar and Kar (2008), Gupta et al. (2012),
Kumar et al. (2014) in India; and Oduol et al. (2011) in Sub-Saharan Africa. After apply a Cobb-Douglas
production function model equation (3) will take the following specification-
ln(tp/as)st = λ0 + λ1 ln(ia)st + λ2 ln(tf)st + λ3 ln(al)st + λ4 ln(tt)st + λ5 ln(lr)st + λ6 ln(fhp)st + λ7 (arfrs)st + λ8
ln(arfws)st + λ9 (arfaj)st + λ10 (amaxtrs)st + λ11 (amaxtws)st + λ12 (amaxtss)st + λ13 (amintrs)st + λ14 (amintws)st
+ λ15 (amintss)st + ξ3(s-1)SD(s-1)+ €3(t-1)TD(t-1)+ ψ3(s-1)+ (t-1) SD(s-1)×TD(t-1) + ξst (5)
Where, λ0 is estimated constant coefficient that is also known as total factor productivity (TFP) or ‘Solow
Residual’; ln is natural logarithms; λ1, λ2, λ3, λ4, λ5, λ6, λ7, λ8, λ9, λ10, λ11, λ12, λ13, λ14 and λ15 are the estimated
regression coefficients for corresponding explanatory variables. These estimates are also known as
estimated elasticities for respective variables under Cobb-Douglas production function model. ξst is the
error term in the model. This function assumes that the production function is constant returns to scale.
This is a liner production function and homogeneous degree one. Equation (4) implies the real functional
form of Cobb-Douglas production function model. Similar model was used by Nastis et al. (2012) to
analysis the climatic impact on agricultural productivity in Greek. Gupta et al. (2012) employed a Cobb-
Douglas production function model to investigate the effect of climatic and non-climatic factors on rice,
sorghum and millet productivity utilizing panel in India. Kumar et al. (2014) investigate the impact of
climatic and non-climatic variables on productivity of major food grain crops using a Cobb-Douglas
production function model in India.

6http://docentes.fe.unl.pt/~jamador/Macro/cobb-douglas.pdf

115
Hypothesis Testing and Selection of Appropriate Model: In this study several regressions models are
done for each proposed models to select an appropriate model. Random effects model is applied to
assuming that the variation across states is to be random and uncorrelated with cane productivity. After
that to capture the unobserved heterogeneity in states and to control annual difference in sugarcane
productivity, fixed effects and time fixed models are used that is described in equation (3), (4) and (5)
(Gupta et al. (2012), Kumar et al. (2014).

Testing for random effects: To decide either random effects or an ordinary least square regression
model is appropriate or not. Breusch-Pagan Lagrange multiplier (LM) test is used for all three proposed
model (Kumar et al., 2014). Null hypothesis is that variance across states is zero and it means that there is
no significant difference among all states and there is no panel effect. Here we are fail to reject the null
hypothesis and concluded that random effects model can be used to estimate the regression coefficient
(see table 1 in appendix).

Testing for fixed or random effect: Hausman specification test is used to check the quandary of fixed
and random effects model (Kumar et al., 2014). The authors tested the null hypothesis is that the
preferred model is random effects and the unique error (ui) terms are not correlated with regressors. But
Hausman Chi2 values were statistically significant at 1% level for all proposed model and null
hypothesizes are rejected (see table 1 in appendix). It implies that unique error (ui) term significantly
correlated with regressors in all proposed models. Hence fixed effects model can be considered for
further regression analysis.

Testing for cross-sectional dependence/contemporaneous correlation: Cross sectional dependence

is the major problem of macro level panel (over 20 years) data. If outcomes are correlated with across
state then there is presence of cross sectional dependence in fixed effects model. To identify the cross
sectional dependence, Breusch-Pagan Lagrange multiplier (LM) and Pesaran's (CD) test are applied for
three proposed models (Kumar et al., 2014). The null hypothesis under B-P (LM) test is that the residual
across states are not correlated. In this case null hypothesis is rejected. It reveals that residual across
states are correlated and panel data set have a cross sectional dependency. Similar hypothesis test is used
for Pesaran's (CD) test and this also shows the presence of cross sectional dependency in panel data set
(see table 1 in appendix).

Testing for heteroskedasticity: Modified Wald test is applied to identify that whether
heteroskedasticity exist or not (Kumar et al., 2014). The null hypothesis is that there is homoskedasticity
(or constant variance). Here Modified Wald test is rejected the null hypothesis since chi2 values are found
statistically significance at 1% level for linear, Ricardian and Cobb-Douglas production model (see table 1
in appendix). We can be concluded that there is presence of homoskedasticity in the panel data.

Testing for serial correlation: If serial auto-correlation exists in fixed effects model then outcomes are
correlated across years for a given state. To address the presence the autocorrelation, Wooldridge test are
used (Kumar et al., 2014). Null hypothesis is that sugarcane productivity is correlated with across year.
Here null hypothesis are rejected for all proposed model since sugarcane productivity is statistically and
significantly correlated with across year at 1% level (see table 1 in appendix). It implies that there is
presence of serial correlation and can be concluded that panel data set have a first order auto-correlation.

Final Estimation: Finally Prais Winsten models with panels corrected standard errors (PCSEs)
estimation are used for all the proposed regression models to avoid the problems of heteroskedasticity,
serial correlation, auto-correlation and serial auto-correlation in fixed effects regression model (Gupta et
al. (2012), Kumar et al. (2014).

4. Results and Discussion

The table 1 shows the empirical results for three different models; simple linear regression model (L-R
model), Ricardian productivity regression model (R-P model) and Cobb-Douglas production function
model (C-D model). Regression coefficients are estimated by Prais Winsten models with panels corrected
standard errors (PCSEs) estimation. Empirical results show that average maximum temperature in
summer and average minimum temperature in rainy season appears negative and statistically significant
impact on sugarcane productivity at 1% significance level for L-R and C-D model. Sugarcane crop will get
benefits with increasing maximum temperature in rainy and winter seasons. The regression coefficients

116
and elasticity’s of maximum temperature have a positive and statistically significant effect on cane
productivity. Increasing average minimum temperature in winter and summer seasons are also found
positive and statistically significant. The estimate implies a positive impact on sugarcane productivity for
LR and C-D models. Average rainfall in winter seasons have a positive and statistically significant impact
on sugarcane productivity. Regression coefficient of rainfall is not consistent with earlier study by
Ramulu (1996). There could be one basic reason that Ramulu (1996)’s study was based only one states.
Finally based on empirical findings here can be concluded that climate change through an increase in
average maximum, average minimum temperature and average rainfall have significant impact on
sugarcane productivity. This effect is varying within seasons for various climatic factors.

Table 1: Regression results for L-R model, R-P Model and C-D model Prais Winsten models with
panels corrected standard errors (PCSEs) estimation
Applied Simple Linear Regression Ricardian Productivity C-D Production Function
Model Model Regression Model Model
Variables Reg. Coef. Std. Error Reg. Coef. Std. Error Reg. Coef. Std. Error
ia 0.047910* 0.0080828 0.0240739* 0.00748 0.0198025 0.02143
tf -0.16130** 0.0642938 -0.1790186* 0.05431 -0.0329882 0.02550
al -7.55E-08 2.69E-06 -7.41e-06* 2.01E-06 -0.069963* 0.01760
tt -0.000102 0.0002279 -0.0000746 0.000204 -0.0160224 0.01835
lr 4.752133* 1.177407 2.310179*** 1.323545 0.1410651* 0.0398
fhp 0.017637** 0.0080331 0.0239894* 0.007585 0.0752636* 0.01966
arfrs -0.00253 0.0026178 0.0044096 0.007997 -0.0264597 0.03340
arfrs^1/2 NA NA -0.18179 0.470289 NA NA
arfws 0.00994*** 0.00522 0.032026** 0.013675 0.0342118 0.01594
arfws^1/2 NA NA -0.886456** 0.378393 NA NA
arfss 0.00500 0.003272 -0.023564** 0.009823 0.0446389 0.03693
arfss^1/2 NA NA 1.401276* 0.523348 NA NA
amaxtrs 8.129719* 0.8173265 133.8864* 37.68776 3.59126* 0.77953
amaxtrs^1/2 NA NA -1452.561* 433.2398 NA NA
amaxtws 1.67827** 0.8066474 118.474* 24.41869 2.496673* 0.66712
amaxtws^1/2 NA NA -1248.472* 269.5023 NA NA
amaxtss -7.28427* 1.408118 119.9152** 48.89536 -4.494785* 1.11148
amaxtss^1/2 NA NA -1539.73* 592.8962 NA NA
amintrs -4.09978* 0.5222698 -0.738061 10.03223 -0.384118* 0.09489
amintrs^1/2 NA NA 11.77637 106.0642 NA NA
amintws 1.81255* 0.6318183 14.82338** 5.755262 0.039280 0.22209
amintws^1/2 NA NA -101.4968** 47.51834 NA NA
amintss 1.37561* 0.377451 16.25332 15.59702 0.5075401* 0.18114
amintss^1/2 NA NA 175.4554 144.979 NA NA
con. coef. 36.2764 46.64806 11949.32* 2048.546 -0.3297855 1.56667
Source: Estimated by Authors and *, ** and *** indicates the 1%, 5% and 10% significance level for
respective variables. NA=No data.

In case of non-climatic variables irrigation is found an important to increase the cane productivity.
Because irrigated area has a positive and statistically significant influence on cane productivity under LR
model. The elasticity of irrigated area with sugarcane productivity is also positive but statistically in-
significant under C-D model. Here can be justified that additional irrigation facilities could be a crucial
option to mitigate the adverse effect of climate change on cane productivity. Increasing application of
fertilizers in cane cultivation would be harmful. The estimated regression coefficient of fertilizers has a
negative and statistically significant impact on cane productivity under simple linear regression model.
Another harmful effect of more fertilizers application would be caused in greater climatic and
environmental damage (Ranuzzi & Srivastava, 2012). The elasticity of agriculture labour with cane
productivity shows that the negative and statistically significant association. Estimate implies that more
utilization of agriculture labour in cane cultivation would be caused in decline cane productivity. More
specifically, more utilization of human power may not be useful to improve the marginal productivity of
land.

117
The participation of literate population in cane cultivation is found a significant variable to increase cane
productivity. Our estimates i.e. the elasticity and regression coefficient of literate population has a
positive and statistically significant impact on cane productivity. There could be one major reason that
literate farmer has a more understanding compared to illiterate. Literate farmers can chose an
appropriate crop for cultivation. They are more aware for suitable sowing time of crops, irrigation time.
They have more understanding that how much fertilizer and pesticides have to use in cane cultivation.
Another important thing is that they are able to select adaptation methods to mitigate the adverse effects
of climate change. Cane productivity negatively affect with the use of tractor. Although the regression
coefficient and elasticity of tractor with cane productivity are statistically in-significant. Therefore
negative sign for both show the negative effect on cane productivity. The estimates imply that
mechanization may not be useful to increase the cane productivity. There could be many scientific
reasons: 1) in India most of farmers are small and marginal, 2) the average size of smallholder farming
system is (<2.0 hectare per farm) that accounts for 78% of the total operational holdings, 3) this
operational holdings occupies 32% of total agricultural area and the average size of 84 million small farm
holdings is (<1 hectare per farm). Due to small sizes of farm holding, marginal productivity of land would
lead to decrease with increase in mechanization. Appropriate price of crop could be caused to increase
the productivity of sugarcane. Our estimated regression coefficient and elasticity of farm harvest price
with cane productivity are found positive and statistically significant. There could be one crucial reason is
that farmer give the preference to those crops which may provide the more financial benefits. It would
increase the decision power of farmers to select an appropriate crop for cultivation. More specifically, it
would provide the motivation to farmers for sugarcane cultivation. They will give more preference to
crop cultivation and productivity would be increased. All regression coefficients for original terms of
climatic factors and square root terms of climatic factors have an opposite sign. These estimates reveal
show that climatic factors have a non-linear relationship with sugarcane productivity.

5. Conclusion and Recommendations

The main aim of this study is to assess the impact of climatic and non-climatic on sugarcane productivity
in different weather season. Sugarcane productivity as dependent variables and other climatic and non-
climatic factors as explanatory variables are compiled a panel data set for thirteen states during 1980-
2009. To check the consistency of empirical results, simple linear regression model (L-R model),
Ricardian productivity regression (non-linear) model (R-P model) and Cobb-Douglas production function
model (C-D model) are employed. Regression results based on Prais Winsten models with panels
corrected standard errors (PCSEs) estimations show that average maximum temperature and average
minimum temperature in summer and rainy season are appears negative and statistically significant
impact on sugarcane productivity. While, sugarcane crop will get benefits with increasing average
maximum temperature in rainy and winter seasons. Rising average minimum temperature in winter and
summer seasons are found positive and statistically significant influence on sugarcane productivity.
Sugarcane productivity would be improved with increase in average rainfall in winter seasons. These
estimates imply that climate change, through increase in average maximum temperature and average
minimum temperature and changing rainfall pattern have a statistically significant effect on sugarcane
productivity in different weather seasons in India. Finally this study provides the empirical evidence that
climatic factors have a non-linear relationship with sugarcane productivity.

In brief our estimation shows that climatic factors have negative influence on sugarcane productivity.
Thus it would be a serious threat for most of the marginal and substantial farmers those are engaged in
sugarcane cultivation for their livelihood and agricultural labour and industry based on sugarcane in
India. Hence the study tries to provide an understanding to Indian policy makers toward the sugarcane
farm policies in presence of climate change. This study also provides a several policy suggestions like
Indian policy maker is need to provide more irrigation facilities for cane cultivation. Our results show that
irrigation has positive and statistical significant influence on sugarcane productivity. Another important
suggestion is that policy makers should provide appropriate price to farmers for their agriculture
production. Our empirical findings also shows that farm harvest price of sugarcane crops are very crucial
thing to improve the productivity of sugarcane crop. There would be one significant reason that farmers
give preference to those crops which will provide more financial benefits. Appropriate price of crops will
also increase the decision power of farmers to select a more financially beneficial crop for cultivation.

Acknowledgements- This paper has been presented at 7th Biennial Conference of Indian Society for
Ecological Economics (INSEE) on Global Change, Ecosystems, Sustainability, December 04-08, 2013,

118
Tezpur University, Tezpur (Assam). The authors are grateful to Professor Jagdish Krishnaswami (Session
Chair in my paper presentation in this Conference), and Dr. Amarnath Tripathi (Assistant Professor,
Institute of Economic Growth, Delhi) for giving their valuable suggestions and comments in this paper.
Authors also would like to give thanks to conference participants to giving their helpful comments in this
paper.

References

Bosello, F. & Zhang, J. (2005). Assessing Climate Change Impacts: Agriculture. CIP–Climate Impacts and
Policy Division Working Paper 02. Available at:
http://www.eclac.cl/ilpes/noticias/paginas/1/35691/Bosello_assesingclimatechangeimpacts.pd
f.
Cobb, C. W. & Douglas, P. H. (1928). A Theory of Production. American Economic Review, 18(1), 139-156.
Deressa, T., Hassan, R. & Poonyth, D. (2005). Measuring the Impact of Climate Change on South African
Agriculture: The Case of Sugarcane Growing Regions. Agrekon, 44(4), 524-542.
Gbetibouo, G. A. & Hassan, R. M. (2005). Measuring the Economic Impact of Climate Change on Major
South African Field Crops: A Ricardian Approach. Global and Planetary Change, 47, 143–152.
Gupta, S., Sen, P. & Srinivasan, S. (2012). Impact of Climate Change on Indian Economy: Evidence from
Food Grain Yields. Centre for Development Economics Working Paper 218, Delhi.
Geethalakshmi, V., Lakshmanan, A., Rajalakshmi, D., Jagannathan, R., Sridhar, G., Ramara, Bhuvaneswari, A.
P., Gurusamy, K. L. & Anbhazhagan, R. (2011). Climate Change Impact Assessment and Adaptation
Strategies to Sustain Rice Production in Cauvery Basin of Tamil Nadu. Current Science, 101(03).
342-347.
Haris, A. A., Biswas, S. & Chhabra, V. (2010). Climate Change Impacts on Productivity of Rice (Oryza
Sativa) in Bihar. Indian Journal of Agronomy, 55(4), 295-298.
Haim, D., Shechter, M. & Berliner, P. (2008). Assessing the Impact of Climate Change on Representative
Field Crops in Israeli Agriculture: A Cause Study of Wheat and Cotton. Climatic Change, 86,425-
440.
Hari, B. D., Gernot, B., Charity, G. O. & Hemanta, K. T. (2010). Capitalizing on Assets: Vulnerability and
Adaption to Climate Change in Nepal. Social Development Papers Social Dimensions of Climate
Change Paper, 121.
Hundal, S. S. & Prabhjyot, K. (2007). Climate Variability and its Impact on Cereal Productivity in Punjab
India. Current Science, 92(04), 506-512.
Kumar, A., Sharma, P. & Ambrammal, S. K. (2014). Climatic Effects on Food Grain Productivity in India: A
Crop Wise Analysis. Journal of Studies in Dynamics and Change, 1(1), 38-48.
Kumar, K. S. K. & Parikh, J. (2001). Socio-economic Impacts of Climate Change on Indian Agriculture.
International Review for Environmental Strategies, 2(2), 277-293.
Kumar, K. S. K. (2009). Climate Sensitivity of Indian Agriculture does Spatial Effects Matter. SANDEE
Working Paper 45-09, Kathmandu.
Kumar, V., Sharma, Y. & Chauhan, S. (2011a). Impact of Climate Change on the Growth and Production of
Saccharum Offcinarum and Magnifera Indica. International Journal of Science Technology and
Management, 2(1), 42-47.
Kumar, S. N., Aggarwal, P. K., Rani, S., Jain, S., Saxena, R. & Chauhan, N. (2011b). Impact of Climate Change
on Crop Productivity in Western Ghats, Coastal and Northeastern Regions of India (Special
Section: Climate Change: Projections and Impact for India). Current Science, 101, 332-341.
Kar, J. & Kar, M. (2008). Environment and Changing Agricultural Practices: Evidence from Orissa, India.
Indus Journal of Management and Social Sciences, 2(2), 119-128.
Kurukulasuriya, P. & Mendelsohn, R. (2008). A Ricardian Analysis of the Impact of Climate Change on
African Cropland. AFJARE, 2(1), 1-23.
Kalra, N., Chakraborty, D., Sharma, A., Rai, J., Monica, H. K., Subhash, C., Kumar, P., Ramesh, B. S., Barman,
D., Mittal, R. B., Lal, M. & Sehgal, M. (2008). Effect of Increasing Temperature on Yield of Some
Winter Crops in Northwest India. Current Science, 94(1), 82-88.
Kapur, D., Khosla, R. & Mehta, P. B. (2009). Climate Change: India’s Options. Economic and Political Weekly,
36(31), 34-42.
Mendelsohn, R., Nordhaus, W. D. & Shaw, D. (1994). The Impact of Global Warming: A Ricardian Analysis.
American Economic Review, 84(4), 753-771.
Mano, R. & Nhemachena, C. (2007). Assessment of the Economic Impacts of Climate Change on
Agriculture in Zimbabwe: A Ricardian Approach. Policy Research Working Paper, World Bank.
Available at: http://elibrary.worldbank.org/doi/book/10.1596/1813-9450-4292.

119
Mongi, H., Majule, A. E. & Lyimo, J. G. (2010). Vulnerability and Adaption of Rain Fed Agriculture to
Climate Change and Variability in Semi-arid Tanzania. African Journal of Environmental Science
and Technology, 4(6), 371-381.
Masters, G., Baker, P. & Flood, J. (2010). Climate Change and Agricultural Commodities. CABI Working
Paper, 02.
Nastis, S. A., Michailidis, A. & Chatzitheodoridis, F. (2012). Climate Change and Agricultural Productivity.
African Journal of Agricultural Research, 7(35), 4885-4893.
Nandhini, U. S., Alagumani, T. & Shibi, S. (2006). Economic Analysis of Agriculture in Southern Parts of
Coastal India. Agricultura Tropica et Subtropica, 39(4), 279-284.
Oduol, J. B., Binam, A., Olarinde, J. N. B., Diagne, L. & Adekunle, A. (2011). Impact of Adoption of Soil and
Water Conservation Technologies on Technical Efficiency: Insight from Smallholder Farmers in
Sub-Saharan Africa. Journal of Development and Agricultural Economics, 3(14), 655-669.
Onyeji, S. C. & Fischer, G. (1993). An Economic Analysis of Potential Impacts of Climate Change in Egypt.
IIASA Working Paper, 93(12), Laxenburg.
Pandey, A. P. (2007). Indian Sugar Industry -A Strong Industrial Base for Rural India. MPRA Paper, 6065.
Ranuzzi, A. & Srivastava, R. (2012). Impact of Climate Change on Agriculture and Food Security. ICRIER
Policy Series 16.
Ramulu, M. (1996). Supply Response of Sugarcane in Andhra Pradesh. Finance India, 10(1), 116-122.
Saseendran, S. A., Singh, K. K., Rathore, L. S., Singh, S. V. & Sinha, S. K. (2000). Effects of Climate Change on
Rice Production in the Tropical Humid Climate of Kerala, India. Climate Change, 44, 495-514.
Seo, S. N., Mendelshon, R. & Munasinghe, M. (2005). Climate Change and Agriculture in Sri Lanka: A
Ricardian Valuation. Environment and Development Economics, Cambridge University Press, 10(5),
581-596.
Srivastava, A. K. & Rai, M. K. (2012). Sugarcane Production: Impacts of Climate Change and its Mitigation.
Biodiversitas, 13(4), 214-227.
Shrivastava, A. K., Shrivastava, A. K. & Solomon, S. (2011). Sustaining Sugarcane Productivity under
Depleting Water Resources. Current Sciences, 10(06), 748-754.
United Nations Report. (2011). An Assessment of the Economic Impact of Climate Change on the
Agriculture Sector in Jamaica. Economic Commission for Latin America and the Caribbean, Sub-
regional Headquarters for the Caribbean.
Zhai, F. & Zhuang, J. (2009). Agriculture Impact of Climate Change: A General Equilibrium Analysis with
Special Reference to Southeast Asia. ADBI Working Paper Series, 131.

APPENDIX

Table 1: Results for hypothesis testing

Applied Test/Model Specification Simple Linear Ricardian C-D Production
Regression Model Productivity Function model
Regression model
Breusch-Pagan Lagrange multiplier (LM) test 0.00 0.00 0.00
for random effects [Chibar2(01)]
Hausman test for fixed or random effects 311.20* 171.29* 131.62*
[Chi2(13)]
Breusch-Pagan Lagrange multiplier (LM) test 123.221* 127.61* 157.20*
for cross-sectional dependence correlation
[Chi2(78)]
Pesaran's (CD) test for cross-sectional 1.781*** 1.781*** 2.822*
dependence correlation
Modified Wald test for heteroskedasticity 185.55* 119.84* 145.09*
[Chi2(13)]
Wooldridge test for serial correlation 23.869* 33.756* 40.293*
[F(1, 12)]

Source: Estimated by Authors; and *, ** and *** indicates the 1%, 5% and 10% significance level.

120
Table 2: Regression results for linear model by Prais Winsten models with panels corrected
standard errors (PCSEs) estimation
No. of Observation 390 R-squared 0.7716
No. of States 13 Wald Chi2(15) 1024.73
No. of Obs./States 30 Prob> Chi2 0.0000
(tp/as) Regression Panel z P > |z| 95% Confidence Interval
Coefficient Corr. Std. Errors

ia 0.0479096* 0.008083 5.93 0.000 0.032068 0.063752

Tf -0.1613009** 0.064294 -2.51 0.012 -0.28731 -0.03529
al -7.55e-08 2.69e-06 -0.03 0.978 -5.34e-06 5.19e-06
Tt -0.0001015 0.000228 -0.45 0.656 -0.00055 0.000345
Lr 4.752133* 1.177407 4.04 0.000 2.444458 7.059809
fhp 0.0176374** 0.008033 2.20 0.028 0.001893 0.033382
arfrs -0.002536 0.002618 -0.97 0.333 -0.00767 0.002595
arfws 0.0099418*** 0.00522 1.90 0.057 -0.00029 0.020173
arfss 0.0050032 0.003272 1.53 0.126 -0.00141 0.011416
amaxtrs 8.129719* 0.817327 9.95 0.000 6.527788 9.731649
amaxtws 1.678273** 0.806647 2.08 0.037 0.097273 3.259273
amaxtrs -7.284266* 1.408118 -5.17 0.000 -10.0441 -4.52441
amintrs -4.09978* 0.522270 -7.85 0.000 -5.12341 -3.07615
amintws 1.812548* 0.631818 2.87 0.004 0.574207 3.05089
amintss 1.375609* 0.377451 3.64 0.000 0.635819 2.1154
con. coef. 36.2764 46.64806 0.78 0.437 -55.1521 127.7049
Source: Estimated by Authors; and *, ** and *** indicates the 1%, 5% and 10% significance level for
respective variables.

Table 3: Regression results for Cobb-Douglas production function model by Prais Winsten models
with panels corrected standard errors (PCSEs) estimation
No. of Observation 390 R-squared 0.6872
No. of States 13 Wald Chi2(15) 809.20
No. of Obs./States 30 Prob> Chi2 0.0000
(tp/as) Regression Panel z P > |z| 95% Confidence Interval
Coefficient Corr. Std. Errors

Ia 0.0198025 0.021426 0.92 0.355 -0.022191 0.061796

Tf -0.0329882 0.025491 1.29 0.196 -0.016972 0.082949
Al -0.0699627* 0.017592 -3.98 0.000 -0.104441 -0.03548
Tt -0.0160224 0.018346 -0.87 0.382 -0.051979 0.019935
Lr 0.1410651* 0.039802 3.54 0.000 0.063054 0.219075
fhp 0.0752636* 0.019659 3.83 0.000 0.036733 0.113794
arfrs -0.0264597 0.033394 -0.79 0.428 -0.09191 0.038991
arfws 0.0342118 0.015937 2.15 0.032 0.002977 0.065447
arfss 0.0446389 0.036932 1.21 0.227 -0.02775 0.117024
amaxtrs 3.59126* 0.779529 4.61 0.000 2.063411 5.119109
amaxtws 2.496673* 0.667123 3.74 0.000 1.189137 3.804209
amaxtrs -4.494785* 1.11148 -4.04 0.000 -6.67325 -2.31632
amintrs -0.3841186* 0.094889 -4.05 0.000 -0.57010 -0.19814
amintws -0.0392798 0.222088 -0.18 0.860 -0.47456 0.396004
amintss 0.5075401* 0.181138 2.80 0.005 0.152517 0.862563
con. coef. -0.3297855 1.566667 -0.21 0.833 -3.40040 2.740826
Source: Estimated by Authors; and *, ** and *** indicates the 1%, 5% and 10% significance level for
respective variables.

121
Table 4: Regression results for Ricardian productivity model by Prais Winsten models with panels
corrected standard errors (PCSEs) estimation
No. of Observation 390 R-squared 0.8379
No. of States 13 Wald Chi2(24) 2196.78
No. of Obs./States 30 Prob> Chi2(24) 0.0000
(TP/AS) Regression Panel z P > |z| 95% Confidence Interval
Coefficient Corr. Std. Errors

ia 0.0240739* 0.00748 3.21 0.001 0.009 0.039

tf -0.1790186* 0.05431 -3.30 0.001 -0.285 -0.073
al -7.41e-06* 2.01e-06 3.69 0.000 3.48e-06 0.00001
tt -0.0000746 0.000204 0.37 0.715 -0.0003 0.0005
lr 2.310179*** 1.323545 1.75 0.081 -0.284 4.904
fhp 0.0239894* 0.007585 3.16 0.002 0.009 0.039
arfrs 0.0044096 0.007997 0.55 0.581 -0.011 0.0201
arfrs^1/2 -0.181790 0.470289 -0.39 0.699 -1.103 0.739
arfws 0.032026** 0.013675 2.34 0.019 0.005 0.0588
arfws^1/2 -0.8864564** 0.378393 -2.34 0.019 -1.628 -0.145
arfss -0.0235639** 0.009823 -2.40 0.016 -0.043 -0.004
arfss^1/2 1.401276* 0.523348 2.68 0.007 0.376 2.427
amaxtrs 133.8864* 37.68776 3.55 0.000 60.020 207.753
amaxtrs^1/2 -1452.561* 433.2398 -3.35 0.001 -2301.695 -603.426
amaxtws 118.474* 24.41869 4.85 0.000 70.614 166.334
amaxtws^1/2 -1248.472* 269.5023 -4.63 0.000 -1776.687 -720.257
amaxtss 119.9152** 48.89536 2.45 0.014 24.082 215.748
amaxtss^1/2 -1539.73* 592.8962 -2.60 0.009 -2701.786 -377.675
amintrs -0.738061 10.03223 -0.07 0.941 -20.401 18.924
amintrs^1/2 -11.77637 106.0642 -0.11 0.912 -219.658 196.105
amintws 14.82338** 5.755262 2.58 0.010 3.543 26.103
amintws^1/2 -101.4968** 47.51834 -2.14 0.033 -194.631 -8.362
amintss -16.25332 15.59702 -1.04 0.297 -46.822 14.316
amintss^1/2 175.4554 144.979 1.21 0.226 -108.698 459.609
con. coef. 11949.32* 2048.546 5.83 0.000 7934.242 15964.39
Source: Estimated by Authors; and *, ** and *** indicates the 1%, 5% and 10% significance level for
respective variables.

122