Methodology to establish the impact of FDI inflows on water pollution in India

The aim of this study is to determine the impact of Foreign Direct Investment (FDI) inflows on environmental pollution in India for the time period 2002-2017. To this effect, ‘environmental pollution’ has been split into two types; water and air pollution. A previous article of this study established that FDI inflows have a significant positive impact on air pollution in the form of greenhouse gas emissions. This article establishes the methodology adopted for empirically testing the impact of FDI inflows on water pollution of India.

For this purpose, various indicators of water pollution were identified, out of which the following have been considered for empirical analysis in this study:

  • pH level
  • Biochemical Oxygen Demand (BOD)
  • Water temperature
  • Electrical conductivity
  • Faecal coliform

The above indicators have been chosen from the initially identified list based on the availability of data from secondary sources.

Data collection procedure

The data for the chosen water pollutant indicators were derived from the official website of National Water Mission and Ministry of Statistics and Programme Implementation. The data in these reports were available for 15 major Indian rivers:

North India

  • Ganga
  • Yamuna
  • Satluj
  • Beas.

West India

  • Mahi
  • Tapi
  • Narmada

South India

  • Godavari
  • Krishna
  • Cauvery

East India

  • Mahanadi
  • Brahamani
  • Baitarni
  • Subarnarekha
  • Brahmaputra

Data was present in the form of ranges showing the minimum and maximum values. Therefore, the average of the range is taken in order to represent the water quality of each river. Finally, the average for all 15 rivers was taken for empirical tests. The averages method has been used by researchers in the past for time series regression analysis (Dangi, Sharma and Uppadhyay, 2017).

In the case of FDI inflows, data was derived from the World Bank database. The table below shows the indicators representing the independent variable (FDI) and dependent variable (water pollution), along with the sources of data.

Variables Indicator Data Source
Water Pollution Water TemperaturepHConductivityDissolved Oxygen (D.O.), Biochemical Oxygen Demand (BOD) Faecal Coliform. Status of Water Quality in India 2011 report from the National Water Mission official website EnviStats India 2018 Report from the official website of the Ministry of statistics and programme.
FDI Net FDI Inflows.World Development Indicators from the World Bank.

Table 1: Variables considered for the study and their data sources

Data analysis procedure

Natural log-transformation (Ln) was performed on the data. It helps in stabilizing the data and establishing a better analysis of the relationship between the variables (Lütkepohl & Xu, 2009). Thus, before analysis FDI inflows and water pollution indicators are transformed into their natural log-form.

STATA was used to test the below hypothesis.

H0: There is no significant impact of FDI inflows on water pollution in India.
HA: There is a significant impact of FDI inflows on water pollution in India.

The procedure followed for testing the hypothesis is presented below.

  1. Linear regression test was run on the data, considering FDI as the independent variable and water pollution indicators as the dependent variable.
  2. Since the null hypothesis had to be rejected, diagnostic tests were applied to determine the nature of variables and the distribution of the data. The test was done at a 5% level of significance.
  3. The first diagnostic test was the stationarity test for all the variables. Based on the results derived from the ADF (Augmented Dickey-Fuller) test, the variables are re-generated into their stationary form.
  4. Furthermore, the Johansen cointegration test and normality test were done. Johansen cointegration test helps in stating the information about the long-run relationship of the dependent and independent variables. Shapiro-Wilk normality test was done to verify the existence of normal distribution in the dataset of the variables. After the results were derived (i.e. cointegrated and normal distributed variables) the regression analysis for transformed variables was done.
  5. Next, residuals for the newly regressed model were predicted and Durbin Watson test of autocorrelation was performed in order to derive the relationship between the residuals.
  6. Lastly, Bartlett’s Periodogram test of heteroscedasticity was performed. A non-autocorrelated and homoscedastic model was tested using the p-value test. The results help in deciding the hypothesis which needs to be accepted or rejected.

Purpose of applying the tests

The purpose of applying these tests has been explained below.

Tests procedure Purpose
Simple Regression To study the original relationship
Diagnostic Tests – Stationarity (ADF), Cointegration (Johansen Cointegration Test), and Normality (Shapiro-Wilk Test) To test the stationarity, long-run relationship, and the existence of normal distribution for the variables and the dataset
Regression To study the relationship between the diagnosed variables
Diagnostic Tests – Autocorrelation (Durbin Watson), and Heteroscedasticity (White-Noise Heteroscedasticity test) To check the presence of relationship among the variables and determine volatility presence over the time period

Table 2: Data Analysis Procedure

Following the above-stated procedure, the analysis for the impact of FDI inflows on each water pollution indicator was performed, independently, after which the overall hypothesis i.e. impact of FDI inflows on water pollution in India was judged.

References

  • Dangi, P., Sharma, BK., and Uppadhyay, B. (2017). “BOD, Total and Faecal coliforms bacterial status of Lake Pichhola, Udaipur, Rajasthan.” International Journal of Fisheries and Aquatic Studies, 5 (3). pp 176-180.
  • Industries, A. (2015). Manufacturing Industries. In India : People and Economy.
  • Kalami, S., Zandi, F., Avazalipour, M. S., & Farahani, S. M. (2013). The Effect of Foreign Direct Investment on Water Pollution. Journal of American Science, 9(4). https://doi.org/10.1057/s41302-019-00140-9
  • Lütkepohl, H., & Xu, F. (2009). The role of the log transformation in forecasting economic variables. Empirical Economics, 42(3), 619–638. https://doi.org/10.1007/s00181-010-0440-1

Riya Jain

Research analyst at Project Guru
Riya is a master in Economics from Amity University. She has a keen interest in econometrics and data analysis. She was a part of the Innovation Project of Daulat Ram College, Delhi University. Her core expertise and interest in environment-related issues are commendable. Apart from academics, she loves music and travelling new places.
Riya Jain

Related articles

Discuss

We are looking for candidates who have completed their master's degree or Ph.D. Click here to know more about our vacancies.