An Empirical Analysis of the Environmental Kuznets Curve for River Water Pollution in Uttar Pradesh

Abstract

This article attempts to test the environmental Kuznets curve (EKC) hypothesis for river water pollution for a panel dataset of 15 districts of Uttar Pradesh. There are sharp socio-economic and demographic variations within India; therefore, a regional study can give a better insight into the pollution income relationship of a specific region compared to a national study. Panel unit root tests, Pedroni cointegration test and fully modified ordinary least square (FMOLS) method have been employed to investigate EKC for two water pollutants—biochemcial oxygen demand (BOD) and total coliform (TC). Findings suggest that there is no evidence of an EKC for BOD, but results validate the existence of an EKC for TC. The Swachha Bharat Mission launched in 2014 aimed at eliminating open defecation and increasing toilet access in rural India can be a credited for the reduction in TC levels since 2014. The success of NAMAMI Gange so far seems to be localised to regions where political thrust has expedited the completion of projects.

Keywords

Environmental Kuznets curve (EKC)river water pollution biochemical oxygen demand (BOD)total coliform Ganga basin

JEL Classification: Q56, R11, Q53

I. Introduction

The environmental Kuznets curve (EKC) posits an inverted U-shaped relationship between pollution and economic growth. This implies that as poor countries become richer, pollution problems will be redressed (Thompson, 2012). After the pioneering work of Grossman and Krueger (1992), there has been an exponential increase in the volume of EKC literature. However, there is no concrete consensus regarding the validity of EKC or its shape. Early EKC literature attributed the inverted U-shaped phenomenon to a change in the structure of output as environmental norms became more stringent with economic growth. The shape of the curve, however, is very sensitive to the data, location and pollutant considered in the analysis (Harbaugh et al., 2002).

This article attempts to test the EKC hypothesis for river water pollution for 15 districts of the Indian state of Uttar Pradesh. A substantial part of the EKC literature has focussed on cross-country or single country analysis. Studies examining specific regions of a country are scarce.

There are many Indias within India. The socio-economic, institutional and demographic variations among Indian states are sharp and significant. A national-level study tends to ignore the regional differences within a country. Thus, the pollution income relation (PIR) may not only differ across nations but also across regions of the same nation (Borghesi, 1999). Since water pollution is local, a district- or state-level study would be more insightful than a national or cross-country study (Dean, 2002; Vincent, 1997).

Uttar Pradesh is the most populous Indian state with an average population density of 520 persons per square kilometre in the Ganga basin as against 312 for the entire country as per 2001 census (National Mission for Clean Ganga).

The estimated population of the state in 2019 was 230 million (Aadhar Saturation Report, 2020).

Major cities of Kanpur, Lucknow, Agra, Meerut, Varanasi and Allahabad are situated in the basin. The cities in the basin have large and growing populations and a rapidly expanding industrial base. In terms of Gross State Domestic Product, UP is the fifth largest Indian state (Financial Express, March 2020). Being an economically significant and densely populated state warrants a probe into the pollution income relationship (PIR) at a rudimentary level.

To the best of our knowledge, this is the only study that tests the EKC hypothesis systematically districtwise for one state. A large part of the EKC literature is devoted to air pollution studies. India-specific studies on water pollution are woefully scarce. This article aims to address that gap in literature and probe into the behaviour of water pollutants with evolving output in the region of UP.

In 2018, 12 polluted river stretches were identified in Uttar Pradesh [3]. River water pollution has been a persistent and stubborn problem in the state. Saharanpur, Ghaziabad, Meerut (Gulaothi town), Allahabad, Varanasi, Kannauj and Kanpur have been identified as polluted districts by the CPCB Restoration of Polluted River Stretches Report (September 2018). Among these, Saharanpur, Ghaziabad and Meerut have been accorded the highest priority, signifying the grave state of pollution in the rivers along these regions is almost threatening the existence of the river. Allahabad, Varanasi, Kannauj and Kanpur feature under the ‘third, fourth and fifth’ priority slot, which implies that water pollution levels are severe. In the dataset used in this study, the mean values of TC and BOD of all districts (for the period 2001–2018) exceed the permissible standards. Therefore, none of these districts can be called non-polluted.

Many factors contribute to the current state of river pollution in the state. There are economic, political, sociocultural, ecological and even local reasons. A summary of the reasons for river pollution is given below.

Kannauj, Kanpur, Allahabad to Varanasi is a densely populated stretch. Allahabad is the most populated city of Uttar Pradesh (2011 census). As the river traverses along this stretch, there is over-extraction of water for agricultural, domestic and industrial uses. The river loses its assimilative and self-cleaning capacity. All that returns to it is waste, and the river has little ecological flow.

The growing volume of untreated sewage finds its way into the river.

There is a glaring gap between installed sewage capacity and sewage generation. Sewage generation was 123 per cent higher than estimated wastewater (Sunita, 2014).

Inadequate and dilapidated conveyance network: Large parts of cities such as Kanpur, Allahabad, and Varanasi do not have sewage networks. Drains are outdated and dilapidated; thus, wastewater is not entirely intercepted and conveyed into the sewage treatment plants. Millions of litres of untreated sewage go directly into the river.

The local municipal bodies in all states along the Ganga are in the abysmal financial state. The state of affairs is stark in UP. Municipal authorities do not have enough funds to run Sewage Treatment Plants (STPs) let alone refurbish them. Apart from financial constraints other factors are multiple monitoring and enforcement agencies. The Uttar Pradesh Pollution Control Board (UPPCB) is the agency that monitors and enforces rules but is considered one of the most corrupt institutions in the state (Singh, 2006). The operation and maintenance of STPs is designated to the local bodies like the UP Jal Nigam and municipal corporations. The functioning of these bodies is fraught with apathy, poor governance, resource constraints, corruption and complacence. There is hardly any concept of governance. Thirty per cent of the STP’s are not in a working condition (CPCB, 2013). While multiple agencies are involved in the whole affair, there is no department to co-ordinate or give decisive direction to the attempts to alleviate pollution (Singh, 2006). Also, there needs to be a cadre of technically qualified personnel for maintenance of STPs.

● Weak enforcement against polluting industries

Among the states situated in the Ganga basin, Uttar Pradesh ranks the highest in terms of industrial water consumption as well as industrial wastewater generation. Industrial effluents account for around 25 per cent of the total wastewater but are significant because of their highly toxic and non-biodegradable nature. Besides, the industries are usually concentrated in clusters, which significantly increases the pollution load in these stretches. River stretches along the industrial clusters are highly toxic because of high pollution load and compromised assimilative capacity of the river. There are 687 grossly polluting industries along the banks of Ganga in various cities of UP. Most are found to be non-compliant. The industrial common effluent treatment plants (CETPs) are either obsolete, undercapacity, inadequate or non-functional. For either or all of these reasons, millions of litres of untreated industrial effluents flow directly into the rivers (Pollution Assessment: River Ganga, 2013).

There are some factors peculiar to different cities that augment the pollution load in those regions. Seventy per cent of the industrial wastewater comes from sugar, pulp and paper, chemical industries and distillers. The pulp and paper industry is concentrated in Western UP—Meerut, Muzaffarnagar and Bulandshahr. Slaughterhouses and abattoirs are situated in Saharanpur and Ghaziabad. There is an agglomeration of some 350 tanneries on the outskirts of Kanpur, which contribute about 8 per cent to the industrial wastewater (Pollution assessment: River Ganga).

In addition to the common sources of pollution , around 32,000 bodies are burnt every year at the two cremation ghats of Harishchand ghat and Manikarnika ghat in Varanasi. People come from distant places to perform the last rights of their deceased in Varanasi. Environmentalist B. D. Tripathi pointed out that a staggering 350 tonnes of partially burnt flesh is dumped into the river every year (New Indian Express, 2016). Corpses of infants and animal carcasses are also dumped in the river.

A slew of initiatives have been taken by the Indian government over the decades for river pollution abatement, of which the most visible are the efforts to rejuvenate the Ganga. The Ganga Action Plan (GAP I) was taken up by the Ministry of Environment and Forests in 1986, followed by GAP II in 1993. These were followed by National River Conservation Plan (1995), National River Ganga Basin Authority (NRGBA) in 2009 and the government clean-up campaign in 2010. The ‘Namami Ganga Programme’ was launched in 2014. Despite these projects, there has not been any sustainable improvement in river water quality.

There has been an exponential increase in EKC literature since the pioneering work of Grossman and Kruegner (1992). Shafik and Bandyopadhyay (1992) tested EKC using log-linear, quadratic and cubic models by taking 8 indicators of environmental quality for 149 countries between the years 1960 and 1990. The study concluded that income has the most significant impact on environmental quality. Suspended particulate matter and sulphur dioxide improve as countries approach middle-income levels. Access to clean water and sanitation improve with increase in per capita income. No EKC was found for river water quality indicators. Trade, debt and other macroeconomic policy variables seem to have little effect on environment. Evidence suggests that it is possible to overcome environmental problems but with necessary policies and investments.

Paudel et al. (2005) tested the EKC hypothesis on water pollution using parish-level data aggregated to the watershed level for the state of Louisiana, USA. The parametric model indicated the turning points within the range of US$10,241–12,993, US$6,636–13,877, and US$6,467–12,758 for nitrogen, phosphorus, and dissolved oxygen, respectively.

Lee et al. (2009) studied a sample of 99 countries for 1980–2001. The authors grouped countries into continental subgroups and found evidence of an EKC in America and Europe with turning points of US$13,956 and US$38,221, respectively, but not in Africa, Asia and Oceania. Thus, the findings support regional differences of EKC for water pollution. Thompson (2014) tested EKC evidence for water pollution in countries that share major rivers as their border. The data consisted of a panel of 21 countries and 30 years, of which 7 were border countries. Results indicate the existence of an EKC for both subsets. However, border countries have a lower turning point compared to non-border countries. Liu and Chen tested EKC for river water quality for the Shenzhen city in China and concluded that river water quality follows an inverted N-shaped curve, but air quality and quality of near coastal waters follow a U-shaped curve.

In the Indian context, some works that have explored EKC in water pollution are as follows. Narayanan and Palanivel (2003) analysed the correlation between industrialisation and environmental quality in the analytical framework using EKC. They found evidence to support an inverted U-shaped relationship between industrial value-added per capita and carbon emissions. However, the relationship turned out to be U-shaped in case of water pollution.

Barua and Hubacek (2008) examined the relationship between per capita income and water pollution for 16 states of India over the period of 1981–2000. Out of the 16 Indian states, a significant relationship between water pollution (BOD and COD) and per capita income was found for 12 states; 4 states showed inverted U-shaped curves while 8 states showed an N-shaped curve. The study concludes that water pollution in most states of India is getting worse after initial improvements. Mythili and Mukherjee (2011) used panel data for river pollutants (BOD and PH) for 14 Indian states for the period from 1990–1991 to 2005–2006. The results indicate a ‘tilted S-shaped’ relationship. Most of the regions in the study have crossed the first turning point but are yet to cross the second turning point. The findings did not support the U-shaped EKC for river effluents in India.

Hauff and Mistri (2015) explored EKC concerning safe drinking water access, groundwater resource development and utilisation and the correlation with waterborne diseases in 32 Indian states/UTs within a period of 11 years (2001–2012). The results found no support for EKC for drinking water, groundwater and waterborne diseases. Income had no significant effect on all the indicators. Income growth in lower-income states/UTs improved access to safe drinking water vis-à-vis higher income states/UTs.

II. Data and Methodology

Data

This article has employed a panel dataset of 15 districts of Uttar Pradesh for the period 2001–2018. The districts included are Bulandshahr, Kannauj, Kanpur, Raebareily, Allahabad, Varanasi, Ghazipur, Lucknow, Jaunpur, Saharanpur, Gautam Buddh Nagar, Mathura, Agra, Meerut and Gorakhpur. The main rivers included in this study are Ganga and its tributaries—Yamuna, Gomti, Hindon and Rapti. As of 2011, 21 districts in Uttar Pradesh had water quality monitoring stations. Some districts have multiple stations: upstream, downstream and at some other strategic points. In that case, the average of all stations was computed to ascertain the pollutant value for this study. Out of these 21, we have included 15 districts in our dataset. Reasons for the omission of the remaining cities are as follows: Some monitoring stations like Unnao and Jhansi gauge the river quality of small tributaries like Sai and Betwa, respectively. The data is sporadic for these tributaries, with some years only mentioning ‘dry river’ for Sai. For the same reason, we had to omit Shahpur, Bateshwar, Etawah and Ayodhya—discontinuous and sporadic data of the river pollutants. As the monitoring network expands, new monitoring stations are added every few years. Bijnor, for example, was added in 2018. For the data to be meaningful, we need at least 15 years of observations. Hence, we could not incorporate districts which have been recently added in the monitoring network.

Biochemical oxygen demand (BOD) and total coliform (TC) are taken as proxy variables for water pollution. The proxy variable for income (Y) is Net district domestic product (NDDP). In EKC literature, BOD is the most commonly used pollutant to proxy water pollution (Gassebner et al., 2011; Hettige et al., 2001). Some studies that used BOD in an EKC framework are Grossman and Kruegner (1995), Hettige et al. (2001), Sigman (2002), Cole (2003), Archibald et al. (2009), Taguchi and Yoshida (2010), Gassebner et al. (2011), Thompson (2013), Lee et al. (2010), Barua and Hubacek (2008), Mythili and Mukherjee (2011). A common pollutant allows for comparison of EKC across regions and time periods.

BOD was also chosen because of its relevance in policymaking. In India, BOD is used as one of the criterions to classify grossly polluting industries (GPI). The Environment Protection Act, 1986, defines GPI as Industries discharging effluents into a watercourse and (a) handling hazardous substances, or (b) effluents having BOD load of 100 kg per day or more, or (c) a combination of (a) and (b). TC indicate the presence of pathogens and are essential for assessing the impact of human activity on the river.

Data Sources

Data for NDDP has been taken from Uttar Pradesh Directorate of Economics and Statistics and DistrictsofIndia.com.

Referring to statistical database of central pollution control board. Water pollution data has been taken from the statistical database of the Central Pollution Control Board, Uttar Pradesh Pollution Control Board and Indiastat.

Methodology

Sinha et al. (2019) pointed out that the data used in EKC modelling might not be standardised in terms of base year. Before carrying out the empirical analysis, all variables should be brought to one common base year as it will ensure the nearly similar temporal impact on the variables. Accordingly, Net District Domestic Product (NDDP henceforth), NAS 1999–2000 and NAS 2004–2005, have been rebased to NAS 2011–2012 for all the years in the dataset, using the method followed in Sinha (2016), Sinha and Bhattacharya (2016, 2017) and Sinha and Rastogi (2017).

The pollution data was compiled for 15 districts of Uttar Pradesh. In districts where there was more than one monitoring station, the average of different stations was computed. BOD is a commonly used water pollution index and has been used by Mythili and Mukherjee (2011) and Barua and Hubacek (2008) as one of the water pollutants in the EKC framework. We faced the problem of missing data for BOD and TC for a few years for certain monitoring stations. To address the problem of missing data points, Sinha et al. (2019) advise to fill up data points by linear interpolation and extrapolation or sometimes by simple or moving average. We have filled the missing data point using the moving average.

The Model

To examine the pollution income relationship, we have used the quadratic form of the model, where BOD is taken as the dependent variable and NDDP and NDDP squared are the explanatory variables.

E = f (Y, Y^{2})

(1)

where E, Y and Y² denote environmental pollution, income and income squared, respectively. This model has been adapted as follows:

E_{i t} = β_{0 i} + β_{1 i} Y_{i t} + β_{2 i} Y_{i t}^{2} + ε_{i t}

(2)

where E, Y and Y² denote BOD, income and income squared, respectively.

β₀ denotes the intercept term, β₁ and β₂ denote the parameters to be estimated, i denotes the district and t the time period. ε denotes the error term.

Shahbaz et al. (2012) argue that log-linear models give better results than linear models in terms of efficiency; thus, all variables have been transformed into their first natural logarithm. Equation (2) is rewritten as

\ln E_{i t} = β_{0 i} + β_{1 i} \ln Y_{i t} + β_{2} \ln Y_{i t}^{2} + ε_{i t}

(3)

Equation (3) allows for the testing of following forms of the pollution income relationship (de Bruyn et al., 1998):

β₁ = β₂ = 0 means that pollution and income are unrelated.

β₁ > 0 and β₂ = 0 indicate a monotonically increasing relationship between pollution and income.

β₁ < 0 and β₂ = 0 indicate a monotonically decreasing relationship between pollution and income.

β₁ < 0 and β₂ > 0 show a U-shaped relationship between pollution and income.

β₁ > 0 and β₂ < 0 give an inverted U-shaped relationship or EKC.

Equation (3) has been tested separately for BOD and TC.

Panel Unit Root Tests

The stationarity properties of the series have been tested using the Levin et al (LLC), Im et al (IPS) and ADF–Fisher tests. Out of the several panel unit root tests available, Im et al. test (IPS, 2003) allows for heterogeneous autoregressive coefficients. Im et al. (2003) suggest averaging the augmented Dickey–Fuller (ADF) unit root tests, while allowing for different orders of serial correlation. The null hypothesis is that each series in the panel contains a unit root.

Maddala and Wu (1999) tests offer a strategy to overcome the limitations of both LLC and IPS tests. They suggest a non-parametric test, which is based on a combination of the p-values and t-statistics for a unit root in each cross-sectional unit (the ADF test). More specifically, the testing approach has the advantage of allowing for as much heterogeneity across units as possible. Under the hypothesis that the test statistic is continuous, the significance of p-values is independent in a uniform manner. The advantage of this test is that it does not require an infinite number of groups to be valid (Apergis & Sorros, 2009).

Panel Cointegration Tests

Long-run cointegration among the variables was tested using the Pedroni cointegration test. Pedroni (2000, 2004) proposed a few different statistics, which were based on Engel and Granger (1987) cointegration regression, to test the null hypothesis of no cointegration in the heterogeneous panel. There are two test groups. The first test group is ‘ within dimensions’. This set of tests includes the panel v statistics, panel rho statistics, panel PP statistics, and panel ADF statistics. The second set of tests are ‘between dimensions’, comprising group rho statistics, group PP statistics and group ADF statistics. Both sets test the null hypothesis of no cointegration.

Panel Fully Modified Ordinary Least Square

After conducting the unit root and cointegration tests for the dataset, we have employed the panel fully modified ordinary least square (FMOLS) technique developed by Pedroni to determine the long-run elasticities. This method allows for estimating heterogeneous cointegrated vector for panel members. The main advantage of this method is that it corrects for both serial correlation and simultaneity bias. Another reason OLS is not appropriate is that its estimation produces bias results, since the regressors are endogenously determined in the I(1) case (Kasman & Duman, 2015). To name a few, Bilgilie et al (2016), Mitić et al. (2017), Apergis and Payne (2009) and Kasman and Duman (2015) have applied this technique to ascertain EKC in a panel data framework.

III. Estimation Results

The results of panel unit roots tests are reported in Table 2. LLC, IPS, ADF–Fisher, and PP–Fisher Chi-square tests were done on intercept, intercept and trend and no intercept and trend. All the series are integrated of the first order/All series are I(1).

Table 1.

Minimum Standards for Water Pollutants

Classification	Category	Tolerance Limit
Drinking water source without conventional treatment but after disinfection	A	1. Total coliform organisms MPN/100 ml shall be 50 or less 2. pH between 6.5 and 8.5 3. Dissolved oxygen 6 mg/l or more 4. Biochemical oxygen demand 5 days 20°C 2 mg/l or less
Outdoor bathing (organized)	B	1. Total coliform organisms MPN/100 ml shall be 500 or less 2. pH between 6.5 and 8.5 3. Dissolved oxygen 5 mg/l or more 4. Biochemical oxygen demand 5 days 20°C 3 mg/l or less
Drinking water source after conventional treatment and disinfection	C	1. Total coliform organisms MPN/100 ml shall be 5,000 or less 2. pH between 6.5 and 8.5 3. Dissolved oxygen 4 mg/l or more 4. Biochemical oxygen demand 5 days 20°C 3 mg/l or less
Propagation of wildlife and fisheries	D	1. pH between 6.5 and 8.5 2. Dissolved oxygen 4 mg/l or more 3. Free ammonia (as N)1.2 mg/l or less
Irrigation, industrial cooling, controlled waste disposal	E	1. pH between 6.0 and 8.5 2. Electrical conductivity at 25°C micromhos/cm max. 2,250 3. Sodium absorption ratio max. 26 4. Boron max. 2 mg/l

Source: Uttar Pradesh Pollution Control Board.

Table 2.

Panel Unit Root Test Results for the Panel of 15 Districts of Uttar Pradesh for the Period 2001–2018

Variable	lnBOD	lnTC	lnY	lnY²
LLC intercept	−6.78516 (0.0000)	−5.63002 (0.0000)	−4.93891 (0.0000)	−4.65403 (0.0000)
Intercept and trend	−5.90486 (0.0000)	−3.51600 (0.0000)	−2.72388 (0.0032)	−2.71059 (0.0000)
None	−12.6481 (0.0000)	−11.3426 (0.0000)	−7.69140 (0.0000)	−7.43731 (0.0000)
IPS intercept	−7.71494 (0.0000)	−6.47817 (0.0000)	−7.98422 (0.0000)	−7.73218 (0.0000)
Intercept and trend	−5.79787 (0.0000)	−4.11105 (0.0000)	−5.34038 (0.0000)	−5.24800 (0.0000)
ADF-Fisher intercept	113.407 (0.0000)	96.8152 (0.0000)	117.558 (0.0000)	114.265 (0.0000)
Intercept and trend	85.5548 (0.0000)	66.1610 (0.0000)	79.8590 (0.0000)	78.8988 (0.0000)	27.1667 (0.6145)
None	180.827 (0.0000)	161.295 (0.0000)	111.012 (0.0000)	108.753 (0.0000)	76.8119 (0.0000)
PP-Fisher Chi square intercept	700.240 (0.0000)	253.267 (0.0000)	905.971 (0.0000)	925.883 (0.0000)	56.3231 (0.0025)
Intercept and trend	220.843 (0.0000)	205.944 (0.0000)	225.802 (0.0000)	223.909 (0.0000)	78.2857 (0.0000)
None	286.084 (0.0000)	284.424 (0.0000)	234.265 (0.0000)	224.537 (0.0000)	99.3706 (0.0000)

Source: Authors own calculations estimated using the software EVIEWS 11.

Note: *Values in parenthesis denote the p-value.

The Pedroni panel cointegration results are reported in Table 3. There is a cointegrating relationship between ln BOD, lnY and lnY² according to the panel PP and panel ADF and Group PP and Group ADF statistics in the intercept form of the model. The cointegrating relationship among the variables is also validated by the Panel PP, Panel ADF, Group PP and Group ADF in the intercept and trend form of the model as well as no intercept and no trend form of the model.

Panel FMOLS results are presented in Table 4. The coefficient of lnY is positive and significant while the coefficient for Y² is negative but insignificant. The coefficient for LnY is 0.273009 and significant. Therefore, a 1 per cent increase in Y (or net district domestic product) will increase BOD levels by 27.3009 per cent.

Table 3.

Panel Cointegration Test Results for 15 Districts of Uttar Pradesh for the Period 2001–2018 (ln BOD lnY and lnY²)

Test	Intercept	Intercept and Trend	None
Panel v statistic	0.911836	−0.163149	1.506190
Panel rho statistic	−1.205381	0.249216	−1.225164
Panel PP statistic	−4.100054	−6.479847	−3.005257
Panel ADF statistic	−3.467008	−5.397308	−2.076115
Group rho statistic	−0.214269	1.759332	−0.892535
Group PP statistic	−7.394311	−6.091982	−7.603452
Group ADF statistic	−3.195484	−2.879857	−2.879677

Source: Authors own calculations estimated using the software EVIEWS 11.

Table 4.

Panel FMOLS Test Results for 15 Districts of Uttar Pradesh for the Period 2001–2018 (ln BOD lnY and lnY²)

Variables	lnY	lnY²
Coefficients	0.273009 (4.476489)	−0.006808 (−1.060768)

Source: Authors own calculations estimated using the software EVIEWS 11.

Note: *The value in parentheses shows the value of t-statistic.

There is no evidence of an EKC for BOD for the 15 districts of Uttar Pradesh under study. Instead we find a monotonically increasing relationship between income and BOD where the first turning point yet to be attained. Our findings are partly consistent with the conclusions of Mythili and Muherjee (2011) that, for BOD pollutants, if per capita NSDP is taken as the income variable, then Uttar Pradesh has not crossed the first turning point. The applicability of EKC for developing countries has been widely neglected in EKC literature. Studies have not addressed the question: how long will it take for developing countries to experience an improvement in environmental quality (Cole & Neumay, 2004). A plausible explanation for delayed turning points for the state of UP is weak institutions, weaker enforcements and a vast population. Clearly, larger populations generate higher emissions. Although the same legislations apply to the entire country, some states have exhibited better environmental performances than UP, and that can be attributed to better institutions and stronger enforcement in addition to other factors.

Tables 5 and 6 present the panel cointegration test results and panel FMOLS results, respectively, for the pollutant TC. lnTC, lnY and lnY² are found to be cointegrated by panel rho statistic, panel PP statistic, and group pp statistic when taken at the intercept form. At the intercept and trend forms, the three variables are found to be integrated by the panel pp statistic, panel ADF statistic and group pp statistic. At no intercept and no trend, panel v statistics, panel rho statistic, panel pp statistic , group pp statistic and group ADF statistic suggest that the three variables are cointegrated. Based on these results, we infer that a long-run cointegration exists between the variables TC, income and income squared.

FMOLS results for TC gives a positive and significant coefficient for lnY and a negative and significant coefficient for lnY², thus yielding an inverted U-shaped curve or EKC. The value of coefficient (long run elasticity) for Y is 3.257965, implying that a 1 per cent increase in NDDP will lead to 325 per cent increase in TC levels.

Table 5.

Panel Cointegration Test Results for 15 Districts of Uttar Pradesh for the Period 2001–2018 (ln TC lnY and lnY²)

Test	Intercept	Intercept and Trend	None
Panel v statistic	1.183816	−0.733218	2.610859
Panel rho statistic	−2.021560	−0.917953	−2.700284
Panel PP statistic	−4.038587	−5.283959	−3.703766
Panel ADF statistic	−1.215225	−0.481910	−1.494924
Group rho statistic	0.922641	2.299410	−0.129834
Group PP statistic	−1.936848	−2.983186	−3.165257
Group ADF statistic	−0.890354	−0.196116	−1.840635

Source: Authors own calculations estimated using the software EVIEWS 11.

Table 6.

Panel FMOLS Test Results for 15 Districts of Uttar Pradesh for the Period 2001–2018 (ln TC lnY and lnY²)

Variables	lnY	lnY²
Coefficients	3.257965 (19.97165)	−0.227155 (−13.19700)

Source: Authors own calculations estimated using the software EVIEWS 11.

Note: *The parenthesis value shows t-statistics.

Coliform levels indicate the presence of pathogens, which are essential in assessing the impact of human activity on rivers. Open defecation and TC are therefore directly linked. In our analysis, a declining trend in TC or the downturn of the curve is observable since 2014. This could be because of the Swachh Bharat Abhiyan launched in 2014, which aimed at eliminating/reducing open defecation. According to the Swachh Bharat Mission data, all districts in UP have become open defecation free in 2019. This is a questionable claim, as visuals of open defecation are still common in the state. However, there has been an undeniable increase in toilet access since 2014, the extent of which needs to be determined by independent research.

FMOLS estimation for individual districts gives mixed results (Table 7). Meerut, Kannauj, Jaunpur, Gautam Buddh Nagar and Saharanpur are districts with no pollution income relationship. The mean BOD values are also highest for this set of districts. River Kali and Hindon, tributaries with the highest pollution load, correspond to these districts as well. Kali passes through Meerut and Hindon through Meerut, Gautam Buddh Nagar and Saharanpur. Allahabad, Kanpur, Ghazipur and Mathura show an increasing pollution income relationship. U-shaped curve is found for Lucknow, which implies an initial fall and subsequent rise in BOD levels.

Table 7.

FMOLS Results for Individual Districts

District	lnY	lnY²	PIR
Lucknow	0.551260 (−3.415336)	0.075482 (4.71063)	U-shaped
Allahabad	0.009422 (0.135976)	0.012854 (1.825119)	Increasing
Varanasi	1.644655 (7.734406)	−0.153873 (−6.670547)	EKC
Kanpur	0.225941 (3.193807	−0.006601 (−0.946715)	Increasing
Meerut	0.035482 (0.173057)	0.038008 (0.020859)	No relationship
Bulandshahr	0.465150 (3.368508)	−0.037314 (−2.585788)	EKC
Kannauj	0.156170 (1.047699)	−0.000401 (−0.022762)	No relationship
Ghazipur	0.221048 (4.423178)	−0.007314 (−1.323874)	Increasing
Mathura	0.510050 (3.727671)	−0.028044 (−1.874060)	Increasing
Jaunpur	0.109437 (0.831338)	0.004223 (0.286580)	No relationship
Gautam Buddh Nagar	0.116434 (0.615477)	0.022435 (1.228803)	No relationship
Saharanpur	0.648204 (0.82390)	−0.030601 (−0.368257)	No relationship
Gorakhpur	−0.149551 (−1.123813)	0.028711 (2.050072)	Decreasing
Agra	0.731698 (4.059225)	−0.049168 (−2.691563)	EKC
Raebareily	−0.335497 (−4.83879)	0.054248 (6.859154)	U-shaped

Source: Authors own calculations estimated using the software EVIEWS 11.

Note: *PIR—pollution–income relationship.

Table 8.

Mean of BOD, TC and NDDP for 15 Districts of UP for the Period 2001–2018

District	River	Mean BOD	Mean TC	Mean NDDP
Lucknow	Gomti	8.56	57,778	23,539
Allahabad	Ganga	3.87	16,627	19,016
Varanasi	Ganga	9.24	56,231	10,043
Kanpur	Ganga	4.84	26,970	32,850
Meerut	Kali-Hindon	58	175,833	18,770
Bulandshahr	Ganga	2.86	2,113	14,222
Kannauj	Ganga	3.66	8,697	4,825
Ghazipur	Ganga	4.04	28,356	10,001
Mathura	Yamuna	10.28	2,241,144	9,286
Jaunpur	Gomti	3.81	29,352	7,539
Gautam Buddh Nagar	Hindon	39.96	184,647	35,809
Saharanpur	Hindon	33.3	103,377	12,951
Gorakhpur	Rapti	3.03	1,595	16,366
Agra	Yamuna	11.78	17,588,782	19,567
Raebareily	Ganga	3.37	7,219	6,375

Source: Authors own calculations estimated using the software EVIEWS 11.

Varanasi is one city that exhibits a conventional inverted U shaped EKC. Varanasi is not just the spiritual capital of India, it is also a politically significant city. The ruling Prime Minister of India, Narendra Modi contested the Lok Sabha election of 2014 from Varanasi. One of the poll promises of his 2014 election campaign was to transform Ganga into nirmal and aviral. The NAMAMI Ganga, a colossal pollution abatement and river rejuvenation programme for Ganga and its tributaries, was announced in July 2014 with an outlay of ₹20,000 crore. The projects sanctioned for Varanasi have been executed rather swiftly. Three new STPs were sanctioned—Dinapur, which became operational in October 2018, Goitha and a third at Ramana. Collectively, these have increased Varanasi’s sewage treatment capacity to 412 million litres per day (MLD) within the prime minister’s 5-year tenure. Twenty out of 23 drains were tapped by June 2019. Political will, therefore, is paramount for achieving policy outcomes.

The U-shaped curve found for Lucknow could be explained by the riverfront project started in 2015. The Irrigation and Water Resources Department of Government of UP initiated a riverfront development project from 2015 till 2017 for Gomti river in Lucknow. The project has disturbed the river ecosystem and aggravated the deterioration in river water by hampering the natural flow and, thus, the self-cleansing capacity of the river (Dutta et al., 2018).

IV. Conclusion

This article attempted to investigate the relationship between river water pollutants BOD, TC and net district domestic product for a panel of 15 districts of the Indian state of Uttar Pradesh for the period 2001–2018. We employed the techniques of panel unit root tests, Pedroni panel cointegration method and FMOLS to determine the pollution income relationship for the given dataset. Test results suggest a long-run cointegrated relationship between the variables. FMOLS results found no evidence of an EKC for BOD; instead, there appears to be a monotonically increasing relationship between BOD and NDDP, with the first turning point yet to be attained. FMOLS results, however, validate the existence of an EKC for TC. The Swachh Bharat Mission, launched in 2014, has significantly increased toilet access in rural India; therefore, a reduction in TC levels can be observed since 2014. Another remarkable improvement was seen for the city of Varanasi, which exhibited a conventional inverted U shaped EKC. The NAMAMI Ganga programme has brought improvement in the river water quality in Varanasi where political thrust has expediated the completion of sanctioned projects.

In developing countries, it is notoriously hard to get reliable and continuous data. Since ours was a district-level investigation, we faced serious data constraints, as some indicators may be available at the national or state level but are not available at the district level. Dasgupta et al. (2002), Grossman and Krueger (1995) and Panayotou (1993) suggest that effective environmental policy brings an improvement in environmental quality. Although India has strong environmental legislations such as the Water (Prevention and Control of Pollution) Act, 1974; Cess Act, 1977 and the Environment Protection Act, 1986, it has some of the most polluted rivers in the world. Some measures that can be taken to augment river pollution abatement are as follows:

Rivers without water are drains. Even though the release of additional water reduces the volume available upstream, increasing the water flow is imperative. This additional water can be provided by the state government’s allocation of riparian water. The state government should be encouraged to build storage for collecting rainwater for diluting the rivers (Sunita, 2014).

State government can consider regulating water usage. The agricultural sector, which extracts the maximum amount of water is unregulated. Common effluent treatment plants (CETP) and STPs have a long gestation period. There can be a lag of 3 years from the time of commissioning of the plant until it gets functional. In the interim period, the focus has to shift to building and intercepting drains and refurbishing the drainage network so that wastewater reaches the STPs. In-situ techniques like bioremediation can be used to treat wastewater in the drains, especially when treatment plants are under construction or non-functional.

Treated effluents should be reused either for agricultural use or industrial cooling.

Fund-starved water utilities in cities can consider levying payments, perhaps pay as per usage for domestic use as well.

Corruption within the state PCB and other local bodies paralyses government machinery. Despite investments in capacity, there will be no measurable improvements in river quality unless the functioning of these bodies is radically transformed to eliminate corruption and complacence.

The NGT report of 2019 suggests that levying a fine on state PCB’s does not deter corruption because they pay the penalty out of public funds, but penalising individual officers does deter corruption. Therefore, officers responsible for neglecting duty should be taken to account, and disciplinary action should be taken against them.

For the state of UP, economic growth alone cannot explain the grave state of pollution in some rivers. Several environmental initiatives of the past have failed to deliver sustainable results for cleaning the rivers in the state, even in the country. The Swachh Bharat Mission and NAMAMI Ganga programmes have shown some encouraging results particularly in arresting the rise of TC levels and bringing about an improvement in Varanasi. There is a need to probe deeper and further and include additional variables such as corruption index , population density, quality of institutions and governance in the analysis to get a better insight into the aspects that interfere with policy implementation, so that the success of these programmes can be sustained and augmented.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The authors received no financial support for the research, authorship and/or publication of this article.

ORCID iD

Aisha Sheikh

References

Apergis

Sorros

(2009). The role of disaggregated earnings for stock prices: Evidence from listed shipping firms and panel tests. Research in Applied Economics , 1(1), 1–24.

Archibald

S. O.

Bochniarz

Gemma

Srebotnjak

(2009). Transition and sustainability: Empirical analysis of environmental kuznets curve for water pollution in 25 countries in Central and Eastern Europe and the Commonwealth of Independent States. Environmental Policy and Governance , 19(2), 73–98.

Barua

Hubacek

(2008). Water pollution and economic growth: An environmental Kuznets curve analysis at the watershed and state level. International Journal of Ecological Economics and Statistics , 10(2014), 63–78.

Bilgili

Koçak

Bulut

Ü.

(2016). The dynamic impact of renewable energy consumption on CO2 emissions: A revisited Environmental Kuznets Curve approach. Renewable and Sustainable Energy Reviews , 54, 838–845.

Borghesi

(1999). The environmental Kuznets curve: A survey of the literature . https://dx-doi-org.web.bisu.edu.cn/10.2139/ssrn.200556

Business Standard . (2016, January 26). Floating human bodies polluting Ganga in UP, NGT told. https://www.business-standard.com/article/pti-stories/floating-human-bodies-polluting-ganga-in-up-ngt-told-116012700925_1.html

Central Pollution Control Board. (2013). Pollution Assessment: River Ganga . URL http://www.nationalwatermission.gov.in/sites/default/files/waterwiki/2.pdf

Central Pollution Control Board. (2018). River Stretches for Restoration of Water Quality . URL https://nrcd.nic.in/writereaddata/FileUpload/River_STRETCHES_Sept_2018.pdf

Cole

(2003). Development, trade and the environment: How robust is the Environmental Kuznets Curve. Environment and Development Economics , 8(4), 557–580.

10.

Cole

M. A.

Neumayer

(2004). Examining the impact of demographic factors on Air Pollution. Population and Environment 26, 5–21.

11.

Dasgupta

Laplante

Wang

Wheeler

(2002). Confronting the environmental Kuznets Curve. Journal of Economic Perspectives , 16(1), 147–168.

12.

de Bruyn

van den Bergh

J. C. J. M.

Opschoor

(1998). Economic growth and emissions: Reconsidering the empirical basis of environmental Kuznets curves. Ecological Economics , 25(2), 161–175.

13.

Dean

(2004). Does trade liberalization harm the environment? A new test. Canadian Journal of Economics , 35(4), 819–842.

14.

Districts of India. (n.d.). https://www.districtsofindia.com

15.

Dutta

Sharma

Iqbal

Khan

(2018). Impact of river channelization and riverfront development on fluvial habitat: Evidence from Gomti River, a tributary of Ganges, India. Environmental Sustainability . https://dx-doi-org-s.web.bisu.edu.cn/10.1007/s42398-018-0016-0

16.

ENVIS Resource Partner on Control of Pollution Water, Air and Noise. (n.d.). http://www.cpcbenvis.nic.in

17.

Final report of the Monitoring Commitee—River Ganga Phase 1-B [5], [6]. (2019). https://greentribunal.gov.in/sites/default/files/all_documents/Final%20Report%20of%20The%20Monitoring%20Committee-River%20Ganga%20Phase-1%20B%20in%20OA%20No.%20200%20of%202014.pdf

18.

Financial Express . (2020). India’s $5-trillion GDP journey: Contribution of each state crucial. https://www.financialexpress.com/opinion/indias-5-trillion-gdp-journey-contribution-of-each-state-crucial/1886918/

19.

Gassebner

Lamla

M. J.

Sturm

J. E.

(2011). Determinants of pollution: What do we really know. Oxford Economic Papers , 63(3), 568–595.

20.

Grossman

G. M.

Krueger

(1995). Economic growth and the environment. The Quarterly Journal of Economics , 110(2), 353–377.

21.

Grossman

Krueger

(1991). Environmental impacts of a North American Free Trade Agreement (Working Paper No. 3914). National Bureau of Economic Research.

22.

Harbaugh

W. T.

Levinson

Wilson

D. M.

(2002). Reexamining the empirical evidence for an environmental Kuznets curve. Review of Economics and Statistics , 84(3), 541–551.

23.

Hauff

M. V.

Mistri

(2015). Environmental Kuznets Curve (EKC): Implications on economic growth, access to safe drinking water and ground water utilisation in India. Global Journal of Human-Social Science: E , 15(1, Version 1.0), 1–15.

24.

Hettige

Mani

Wheeler

(2001). Industrial pollution in economic development: The environmental Kuznets curve revisited. Journal of Development Economics , 62(2), 445–476.

25.

K. S.

Pesaran

M. H.

Shin

(2003). Testing for unit roots in heterogeneous panels. Journal of Econometrics , 115(1), 53–74.

26.

Kasman

Duman

Y. S.

(2015). CO2 emissions, economic growth, energy consumption, trade and urbanization in new EU member and candidate countries: A panel data analysis. Economic Modelling , 44, 97–103.

27.

Lee

C.-C.

Chiu

Y.- B.

Sun

C. H.

(2010). The environmental Kuznets curve hypothesis for water pollution: Do regions matter. Energy Policy , 38(1), 12–23.

28.

Levin

Lin

C. F.

Chu

C. S. J.

(2002). Unit root tests in panel data: Asymptotic and finite-sample properties. Journal of Econometrics , 108(1), 1–24.

29.

Maddala

(1999). A comparative study of unit root tests with panel data and a new simple test. Oxford Bulletin of Economics and Statistics , 61(s1), 631–652.

30.

Ministry of Jal Shakti Department of Water Resources , River Development & Ganga Rejuvenation. (n.d.). National Mission for Clean Ganga . URL https://nmcg.nic.in

31.

Mitić

Ivanović

O. M.

Zdravković

(2017). A cointegration analysis of real gdp and CO2 emissions in transitional countries. Sustainability (Switzerland) , 9(4), 568.

32.

Murthy

Gambhir

(2018). Analyzing environmnetal Kuznets curve and pollution haven hypothesis in India in the context of domestic and global policy change. Australian Accounting, Business and Finance Journal , 12(2), 135–156.

33.

Mythili

Mukherjee

(2011). Examining environmental Kuznets curve for river effluents in India. Environment, Development and Sustainability , 13(3), 627–640.

34.

Narayanan

Palanivel

(2003). Industrialisation and the environmental quality under economic reforms: An Indian case study (Working Paper No. 106). UNU/IAS.

35.

National River Conservation Plan. (1995). River conservation programme launched in 1995.

36.

Pal

Mitra

K. S.

(2017). The environmental Kuznets curve for carbon dioxide in India and China: Growth and pollution at crossroad. Journal of Policy Modelling , 39, 371–385.

37.

Panayotou

(1993). Empirical tests and policy analysis of environmental degradation at different stages of economic development . ILO.

38.

Paudel

K. P.

Zapata

Susanto

(2005). An empirical test of Environmental Kuznets Curve for water pollution. Environmental and Resource Economics , 31(3), 325–348.

39.

Pedroni

(1999). Critical values for cointegration tests in heterogeneous panels with multiple of economics and statistics. Oxford Bulletin of Economics and Statistics , 61, 653–670.

40.

Pedroni

(2000). Fully modified OLS for heterogeneous cointegrated panels. In Baltagi

B. H.

Fomby

T. B.

Carter Hill

(Ed.), Nonstationary panels, panel cointegration, and dynamic panels (Advances in Econometrics, Vol. 15, pp. 93–130). Emerald Group Publishing Limited. https://doi.org/10.1016/S0731-9053(00)15004-2

41.

Pedroni

(2004). Panel cointegration: Asymptotic and finite sample properties of pooled time series tests with an application to the PPP hypothesis. Econometric Theory , 20(3), 597–625.

42.

Press Information Bureau. (2019). Jal Shakti Minister Shri Gajendra Singh Shekhawat inspects Namami Gange STP projects at Dinapur and Ramana . https://newschrome.com/jal-shakti-minister-gajendra-singh-shekhawat-inspects-namami-gange-stp-projects-at-dinapur-and-ramana/

43.

Selden

T. M.

Song

(1994). Environmental quality and development: Is there a Kuznets curve for air pollution emissions. Journal of Environmental Economics and Management , 27(2), 147–162.

44.

Shafik

Bandyopadhyay

(1992). Economic growth and environmental quality: Time series and cross-country evidence, background paper for the world development report . Washington, DC: The World Bank, 1992.

45.

Shahbaz

Lean

H. H.

Shabbir

M. S.

(2012). Environmental Kuznets curve hypothesis in Pakistan: Cointegration and granger causality. Renewable and Sustainable Energy Reviews , 16(5), 2947–2953.

46.

Shukla

Parikh

(1992). The environmental consequences of urban growth: Cross-national perspectives on economic development, air pollution, and city size. Urban Geography , 13(5), 422–449.

47.

Sigman

(2002). International spillovers and water quality in rivers: Do countries free ride. American Economic Review , 92(4): 1152–1159.

48.

Singh

(2015). Varanasi too has its share of bodies floating in Ganga . Times of India . https://timesofindia.indiatimes.com/home/environment/pollution/Varanasi-too-has-its-share-of-bodies-floating-in-Ganga/articleshow/45921279.cms#:~:text=VARANASI%3A%20The%20unearthing%20of%20over,in%20a%20year%20in%20just

49.

Singh

(2006). Bridging the ganga action plan: Monitoring failure at Kanpur. Economic and Political Weekly , 41(7), 590–592.

50.

Sinha

(2016). Trilateral association between SO2/NO2 emission, inequality in energy intensity, and economic growth: A case of Indian cities. Atmospheric Pollution Research , 7(4), 647–658.

51.

Sinha

Bhattacharya

(2017). Estimation of environmental Kuznets curve for SO2 emission: A case of Indian cities. Ecological Indicators , 72, 881–894. https://doi.org/10.1016/j.ecolind.2016.09.018

52.

Sinha

Rastogi

S. K.

(2017). Collaboration between central and state government and environmental quality: Evidences from Indian cities. Atmospheric Pollution Research , 8(2), 285–296.

53.

Sinha

Shahbaz

Balsalobre

(2019). Data selection and environmental Kuznets curve models—Environmental Kuznets curve models, data choice, sources, missing data, balanced and unbalanced panels. In Özcan

Öztürk

(Eds.), Environmental Kuznets Curve (EKC) (pp. 65–83). Academic Press.

54.

Sunita

(2014). Ganga the river, its pollution and what we can do to clean it . Centre for Science and Environment.

55.

Swachh Bharat Mission-Grameen. (n.d.). https://sbm.gov.in/sbmReport/Report/Physical/SBM_TargetVsAchievementWithout1314.aspx

56.

Swayamprakasham

(2015). Coliform will count . Down to Earth.

57.

Taguchi

Yoshida

(2010). Environment pollution control: Advantage or disadvantage for latecomer’s economies in East Asia? East Asian Bureau of Economic Research.

58.

Thompson

(2012). Water abundance and an EKC for water pollution. Economics Letters , 117(2), 423–425.

59.

Thompson

(2014). Environmental Kuznets curve for water pollution: The case of border countries. Modern Economy , 5(1), 66–69.

60.

Unique Identification Authority of India. (2020). State/UT wise Aadhar Saturation . URL https://uidai.gov.in/images/state-wise-aadhaar-saturation.pdf

61.

Uttar Pradesh Pollution Control Board. (n.d.). URL http://www.uppcb.com/intro.htm

62.

Vincent

J. R.

(1997). Testing for environmental Kuznets Curves within a developing country.” Environment and Development Economics , 2(4), 417–431.

An Empirical Analysis of the Environmental Kuznets Curve for River Water Pollution in Uttar Pradesh

Abstract

Keywords

I. Introduction

II. Data and Methodology

Data

Data Sources

Methodology

The Model

Panel Unit Root Tests

Panel Cointegration Tests

Panel Fully Modified Ordinary Least Square

Table 1.

Minimum Standards for Water Pollutants

Panel Unit Root Test Results for the Panel of 15 Districts of Uttar Pradesh for the Period 2001–2018

Panel Cointegration Test Results for 15 Districts of Uttar Pradesh for the Period 2001–2018 (ln BOD lnY and lnY2)

Panel FMOLS Test Results for 15 Districts of Uttar Pradesh for the Period 2001–2018 (ln BOD lnY and lnY2)

Panel Cointegration Test Results for 15 Districts of Uttar Pradesh for the Period 2001–2018 (ln TC lnY and lnY2)

Panel FMOLS Test Results for 15 Districts of Uttar Pradesh for the Period 2001–2018 (ln TC lnY and lnY2)

FMOLS Results for Individual Districts

Mean of BOD, TC and NDDP for 15 Districts of UP for the Period 2001–2018

Footnotes

Declaration of Conflicting Interests

Funding

ORCID iD

References

Panel Cointegration Test Results for 15 Districts of Uttar Pradesh for the Period 2001–2018 (ln BOD lnY and lnY²)

Panel FMOLS Test Results for 15 Districts of Uttar Pradesh for the Period 2001–2018 (ln BOD lnY and lnY²)

Panel Cointegration Test Results for 15 Districts of Uttar Pradesh for the Period 2001–2018 (ln TC lnY and lnY²)

Panel FMOLS Test Results for 15 Districts of Uttar Pradesh for the Period 2001–2018 (ln TC lnY and lnY²)