tiểu luận kinh tế lượng ECONOMETRICS REPORT AIR QUALITY AND ITS DETERMINANTS

FOREIGN TRADE UNIVERSITY FACULTY OF INTERNATIONAL ECONOMICS ECONOMETRICS REPORT: AIR QUALITY AND ITS DETERMINANTS Group Class : : 25 Japanese-style International Business – K57 Members : Nguyễn Khánh Linh – 1815520193 Phùng Bích Loan – 1815520198 Đỗ Quỳnh Trang – 1810520228 Instructor : Ph.D Từ Thúy Anh Hanoi – October, 2019 TABLE OF CONTENTS ABSTRACT INTRODUCTION I/ LITERATURE REVIEW II/ THEORETICAL BACKGROUND Air quality: Determinants of air quality: III/ DESCRIPTIVE STATISTIC OF DATA 12 IV/ ECONOMETRIC MODEL 14 V/ ROBUSTNESS CHECK 18 Multicollinearity 18 Normality of residual: 20 White’s test for heteroskedasticity 21 T - test of the hypothesis relating to a regression coefficient 22 VI/ ADJUSTED MULTIPLE REGRESSION MODEL: 24 VII/ FINDING AND DISCUSSION: 25 CONCLUSION 27 REFERENCES 28 ABSTRACT Our topic is “Air quality and its determinants” We collected our data from GRETL (data Ramanathan 7-10: 30 Standard Metropolitan Statistical Areas (SMSAs) in California, compiled by Susan Wong for 1970 – 1972.) This report includes eight main parts (Literature review, Theoretical background, Descriptive statistic of data, Econometric model, Robustness check, Finding and Discussion, Conclusion) Our report aims at: • Providing a clear sight about theoretical and method that we use to analyze information through Literature review and Theoretical background About data and objectives that we focus on, we explain those data in descriptive statistic of data • Running and checking the model if there are problems about variables or results through Econometric model and Robustness check • Discussing about model and recommend solution to have clear model and efficient in reality by Finding and Discussion • Summing up our report by Conclusion INTRODUCTION Pollution is now a commonplace term, that our ears are attuned to We hear about the various forms of pollution and read about it through the mass media Air pollution is one such form that refers to the contamination of the air, irrespective of indoors or outside Air is essential for life itself; without it we could survive only a few minutes It constitutes immediate physical environment of living organisms It is a mixture of various gases like nitrogen, oxygen and carbon dioxide, and others in traces; along with water vapour perceptible as humidity and suspended solids in particulate form A physical, biological or chemical alteration to the air in the atmosphere can be termed as pollution It occurs when any harmful gases, dust, smoke enters into the atmosphere and makes it difficult for plants, animals, and humans to survive as the air becomes dirty Air pollution is a mix of particles and gases that can reach harmful concentrations both outside and indoors Its effects can range from higher disease risks to rising temperatures Soot, smoke, mould, pollen, methane, and carbon dioxide are a just few examples of common pollutants In societies that are rapidly developing sufficient resources may not be invested in air pollution control because of other economic and social priorities The rapid expansion of the industry in these countries has occurred at the same time as increasing traffic from automobiles and trucks, increasing demands for power for the home, and concentration of the population in large urban areas called mega cities The result has been some of the worst air pollution problem in the world In many traditional societies, and societies where crude household energy sources are widely available, air pollution is a serious problem because of inefficient and smoky fuels used to heat buildings and cook This causes air pollution both out door and indoors The result can be lung disease, eye problems, and increased risk of cancer Worldwide, bad outdoor air caused an estimated 4.2 million premature deaths in 2016, about 90 percent of them in low- and middle-income countries, according to the World Health Organization Indoor smoke is an ongoing health threat to the billion people who cook and heat their homes by burning biomass, kerosene, and coal Air pollution has been linked to higher rates of cancer, heart disease, stroke, and respiratory diseases such as asthma In the U.S nearly 134 million people—over 40 percent of the population—are at risk of disease and premature death because of air pollution, according to American Lung Association estimates Air pollutants cause less-direct health effects when they contribute to climate change Heat waves, extreme weather, food supply disruptions, and other effects related to increased greenhouse gases can have negative impacts on human health The U.S Fourth National Climate Assessment released in 2018 noted, for example, that a changing climate "could expose more people in North America to ticks that carry Lyme disease and mosquitoes that transmit viruses such as West Nile, chikungunya, dengue, and Zika." In Vietnam, according to incomplete statistics, in recent years, the number of patients relating to air pollution is increasing The number of children hospitalised for asthma treatment, respiratory bacteria, and cough in some hospitals in Hanoi and Ho Chi Minh City has increased many times According to medicinal specialists’ forecast, the number of cancer cases including lung cancer in Vietnam will increase significantly in the next years In some big urban areas in Viet Nam, the continuous automatic monitoring system shows that the PM2.5pollution is the most serious problem for air pollution It is a very difficult issue not only for Viet Nam but also of many Asian countries The pollution assessment can be carried out through the comparison of average daily monitoring data and average annual monitoring data to the allowed levels It could be seen through years that, average daily value (24 hours) of many days within a year in some locations in big cities is 1.5 - times higher than the allowed levels); the number of days having average daily value higher than the allowed level is increasing through years; average hourly value within a day changes, mostly highest in peak hours when the density of cars and motorbikes is high in the streets In some days, the value is even three to four times higher than the allowed level It could be seen that, air pollution issue in some urban areas of Viet Nam is very problematic when the average value and the annual value higher than the allowed level increase As air pollution is a very concerning problem not only in Vietnam but also on a global scale, our team decided to choose this topic I/ LITERATURE REVIEW Our main purpose in this report is the explore the correlation between air quality and other factors There are numerous studies and articles that examine the relationship between air quality and its determinants, yet there are not many articles that can adequately illustrate the positive or negative correlation between them In addition, we are going to demonstrate it in this report through practicing econometric knowledge Before going deeply into the analyzing process, we looked at numerous studies and articles to understand the definition of air quality, and how the good & poor state of air can affect the surrounding According to British Columbia (BC), the term “air quality” means the state of the air around us Good air quality refers to clean, clear, unpolluted air Clean air is essential to maintaining the delicate balance of life on this planet — not just for humans, but wildlife, vegetation, water and soil Poor air quality is a result of a number of factors, including emissions from various sources, both natural and “human-caused.” Additionally, we examined the determinants that can directly impact the quality of the air Our first reading was “Clean Air in the UK” on air-quality.org.uk, which mentioned Topography, Weather conditions, Exposure to pollution, Time and Pollution emissions as the main factors affecting the air quality The article proposed that there are many factors that affect air quality, making the search for clean air quite a complicated issue This is because air pollution enters the atmosphere by different amounts at different times and in various places Air pollutants may also change in the atmosphere to make other pollutants In addition, people can have differing sensitivities to air pollutants For example, the elderly, the very young and those with asthma or other breathing difficulties are likely to be more sensitive to high levels of pollutants The article demonstrated quite adequately the relationship as well as the correlation, however, the structure of the article made it hard for the readers to follow There is no classification of factors, and the factors have yet fully listed According to Matt Williams in “What Causes Air Pollution?” (Universe Today, 2016), air pollution can be divided into Primary and Secondary types of pollutants Whereas primary pollutants are caused by primary sources – i.e the direct result of processes (such as industrial emissions or volcanic eruptions) – secondary pollutants are the results of intermingling and reactions by primary pollutants (such as carbon emissions and water vapor, which creates smog) In other words, we can divide the factors into two types: natural causes and anthropogenic causes The natural causes include Dust and wildfires, Animal and vegetation, Volcanic activity; the anthropogenic factors include Fossil-fuel emissions, Waste and Agriculture and animal husbandry However, the anthropogenic factors are still missing a lot of human-caused factors, which made us not fully satisfied with the determinants listed in the article Rainald Borck, Philipp Scharauth, Munih Society for the Promotion of Economic Research, in “Population density and urban air quality” went deep into one particular matter – population and its density They find that increasing population density by one standard deviation can increase PM10 (particulate matter with diameter less than 10PM) by about percent and NO2 by around 12 percent The study thus contributes to knowledge about the economic costs of agglomeration However, there is much less robust evidence on the costs of agglomeration In “The relationship between air pollution emission and income: US Data” by Richard T.Carson, Yongil Jeon and Donald R.McCubbin, they showed another interesting aspect of how the quality of the economy can affect the air quality To be specific, the individual income can be of great contribution to the quality of the air The article used data from the 50 US states to find that emissions per capita decrease with increasing per capita income for all seven major classes of air pollutants… Without exception, the high-income states have low per capita emissions while emissions in the lower-income states are highly variable Questions still remain as to why this relationship exits “Air pollution and poverty: Does the sword cut both ways?” (2003) by F.W.Lipfert, J Epidemiol Community Health also demonstrated similar perspective In many industrialized nations (including the United Sates) poverty leads to substandard medical care, substandard nutrition, substandard housing, and reliance on inefficient and excessively polluting vehicles and heating and cooking appliance As mentioned in World Energy Outlook Special report “Energy (Electric) and Air Pollution”, International Energy Agency, the energy usage can also be listed as one determinants of the air quality The energy sector is the largest contributor to emissions of air pollutants in the United States and ongoing efforts to combat air pollution are in the contect of its rapid transformation After extensive reading, our list of factors that can affect the air quality: Population & its density, precipitation, coastal locations, individual income, poverty, electricity usage, and industrial factors eg Fuel oil, value added in manufacturing and industrial establishments We divided it into two types: natural determinants (precipitation and coastal locations), and human-caused determinants (population, density, income, poverty, electricity usage, and industrial factors) We came to realization that population, density, and precipitation can have a positive correlation with air quality, while the rest of the determinants have the negative correlation The detailed of our findings would be illustrated in detailed below II/ THEORETICAL BACKGROUND Air quality: “Air quality” refers to the condition of the air within our surrounding Good air quality pertains to the degree which the air is clean, clear and free from pollutants such as smoke, dust and smog among other gaseous impurities in the air Air quality is determined by assessing a variety of pollution indicators Good air quality is a requirement for preserving the exquisite balance of life on earth for humans, plants, animals and natural resources As such, human health, plants, animals and natural resources are threatened when pollution in the air reach high concentrations Poor air quality can affect or harm human health and/or the environment Air quality can be degraded by natural or man-made sources Natural sources include volcanic eruption, windstorm dust Man-made source include pollution from moving vehicles, toxic gases from industries, coal powered plants, burning wood or other material in open air, landfills Both these sources can seriously affect the overall air quality and can lead to severe health problems for humans Determinants of air quality: 2.1 Population: The larger population means bigger weight of air quality Many people cook and heat by using fuels that dramatically pollutes air quality, which increases the level of household air pollution Moreover, vehicles also strongly impact on the air qualiy Pollution from automobiles is clearly visible in every city of the world Vehicles run on fossil fuels such as petroleum and gasoline that emit soot and harmful gases such as CO and NOx which are among the major air pollutants in the environment 2.2 Density: Larger number of density means bigger weight of air quality The dense population results in bad air quality because of emissions from vehicles, factories, low-tech heating tools, etc The cities often suffer from bad air quality than rural areas where there are less people living, which means less pollutants emitted to the air 2.3 Precipitation: Higher rainfall rate results in bigger weight of air quality Rainfall is recognized as one of the main mechanisms to reduce atmospheric particulate pollution, which typically results in less pollution since it washes away particulate matter and can also wash out pollutants that are dissolvable 2.4 Coastal locations: Standard Metropolitan Statistical Areas near the coast have lighter weight of air quality Air pollution of urban atmospheres arises pollutants arising from the high density of human activity found in cities Coastal locations have more winds; therefore those places are likely to experience better air quality 2.5 Income: The higher medium per capita income results in the lighter weight of air quality People having higher income are likely to have lower emissions than people having lower income due to better living standards They normally use equipment that is eco-friendly 2.6 Poverty: The higher percentage of poor families leads to lighter weight of air quality Nearly 92% of pollution-related deaths occur in low- and middle-income countries They regularly use vehicles, equipment which negatively affect air quality This accelerates the level of air pollution 2.7 Electricity usage: The more electricity consumed by industrial manufacturers, the lighter the weight of air quality Almost all forms of electricity generate waste For example, natural gas releases carbon dioxide and nitrogen oxide Earth's atmosphere traps these gases, leading to air pollution and smog 10 IV/ ECONOMETRIC MODEL As listed on the table, we can see that there are multiple elements, which affect the air quality; therefore, we must use the Multiple Regression Analysis In this research, there are 10 independent variables, namely popln, valadd, rain, coast, density, medincm, poverty, electr, fuelloil, indestab, const To analyze and test the effect of these variables on the level of air quality, we find out the Coefficients using Ordinary Least Square (OLS) method with the Best Linear Unbiased Estimator (BLUE) assumption The table below shows the multiple regression result among the variables: From the table above, in the first place, we have: Model 1: airquality = 98,863 + 0,09 popln – 0,024 valadd + 0,24 rain – 34,6 coast + 0,00028 density - 0,014 medincm – 0,0116 poverty – 0,023 electr – 0,0016 fueloil – 0,01 indestab • According to the result, the Adjusted R- squared = 0,42201, it means that nearly 42,2% of the variance of airquality are explained by the OLS regression line, this 14 percentage is quite stable even when we add more variables Then, the remaining 58% is explained by other factors In general, we divide the parameters into groups and check whether the results fit theories or not • Positive correlation with airquality: popln, rain, density o The larger population means bigger weight of suspended particular matter: Fits the theory of Household factor on Air quality: “Around billion people cook and heat their homes using polluting fuels (i.e wood, coal, dung, kerosene) and inefficient technologies Cooking and heating with polluting fuels and technologies produces high levels of household air pollution, which includes a range of health damaging pollutants such as fine particles and carbon monoxide.” o Higher rainfall rate results in bigger weight of suspended particular matter: Fits the theory of Weather conditions factor on Air quality as “rain washes out or dispersed pollutants in the atmosphere” o Larger number of density means bigger weight of suspended particular matter: Fits the theory of Population density and urban air quality: “They find that increasing population density by one standard deviation increase PM10 (particulate matter with diameter less than 10PM) by about percent and NO2 by around 12 percent” • Negative correlation with airquality: valadd, coast, medincm, poverty, electr, fueloil, indestab o The more value added by industrial manufacture, the lighter weight of suspended particulate matter: Not fit the theory of Industrial Emission: “Industries contribute significantly to the air pollution problems.” o Standard Metropolitan Statistical Areas near the coast have lighter weight of suspended particulate matter: Fits the theory of Weather conditions factor 15 on Air quality: “Coastal locations and open areas often experience more windy weather and are therefore likely to experience better air quality.” o The higher medium per capita income results in the lighter weight of suspended particulate matter: Fits the theory of “The relationship between air pollution emission and income: US Data” “Using data from the 50 US states, we find that emissions per capita decrease with increasing per capita income for all seven major classes of air pollutants… Without exception, the high-income states have low per capita emissions while emissions in the lower-income states are highly variable.” o The higher percentage of poor families leads to lighter weight of suspended particulate matter: Not fit the theory “Air pollution and poverty”, “In many industrialized nations (including the United Sates) poverty leads to substandard medical care, substandard nutrition, substandard housing, and reliance on inefficient and excessively polluting vehicles and heating and cooking appliance” o The more electricity consumed by industrial manufacturers, the lighter the weight of suspended particulate matter: Not fit the theory “The energy sector is the largest contributor to emissions of air pollutants in the United States” by International Energy Agency o The more barrels of fuel oil consumed in industrial, the lighter weight of suspended particulate matter: Not fit the theory of Industrial Emission o The higher number of industrial establishments, the lighter weight of suspended particulate matter: Not fit the theory of Industrial Emission: “Industries contribute significantly to the air pollution problems.” Therefore, we will exclude valadd, poverty, electr, fuelloil, indestab in the adjusted multiple regression model • F-test of the overall significance of the model: 16 The F-statistic shows that at 5% of significance, we have enough evidence to reject the Joint Null hypothesis: All the coefficients on the independent variables are zero Because: P – value = 0,0159 < 0,05 That means we can reject the hypothesis that: None of the independent variables have an effect on airquality In other words, this model might work However, it is clear that there are a lot of insignificant parameters in the model without any “*”; therefore; we need to test the model to see whether it follows BLUE assumption and come up with the most efficient model 17 V/ ROBUSTNESS CHECK Basically, we need to function types of test: Multicollinearity, Auto – correlation, Normality of residual, Heteroskedasticity However, because type of date in this assignment is cross-sectional data, without data related to time factor, we not have to test Auto – correlation Multicollinearity Multicollinearity test is the test to see whether one of the regressors is a linear combination of the other regressors We need to test this because the goal of multiple regression function is to see the effect of one independent variable on a dependent variable, holding other variables constant • Collinearity of independent variables: 18 • Correlation among independent variables: As a result of the two table of Collinearity and Correlation, we have to take the link between popln, indestab & medincm into account: Cor(popln,medincm)= 0,9957 > 0,8 Cor(popln, indestab)= 0,98 > 0,8 There is a strong linear relationship There is a strong linear relationship between the two variables of Population between the two variables of Population and Medium income and Number of individual manufaturers Vif (popln) = 400,274 >> 10 Vif (popln) = 400,274 >>10 Vif (medincm) = 399,713 >>10 Vif (indestab) = 78,68 >>10 These two variables show a problem of These two variable show a problem of collinearity collinearity Cor (indestab, valadd) = 0,9205 > 0,8 19 There is a linear relationship between the two variables of Number of individual manufacturers: Vif (valadd) = 14,528 > 10 Vif (indestab) = 78,68 >10 These two variable show a problem of collinearity To fix the problem, we must exclude medincm and indestab out of the model, just leave popln Normality of residual: In order to make valid inferences from the regression, the residuals of the regression should follow a normal distribution If they are, they will conform to the diagonal normality line indicated in the plot 20 As presented in the graph and table: Null hypothesis: error is normally distributed Test statistic: Chi-square(2) = 6,2377 with p-value = 0,0442077 < 0,05 Thus, at 5% level of significance, we have enough evidence to reject the Null hypothesis In other words, error is not normally distributed However, at 1% level of significance, we have enough evidence not to reject the Null hypothesis Reason behind this controversial p-value for the Null hypothesis is that the number of observations is limited: 30 observations and there are 10 independent variables, meaning that the degree of freedom is 19, which is insufficient for the u i distribution to reach Normal distribution To fix this problem, we need to add more observations to the data or exclude insignificant from the model to increase the number of degree of freedom White’s test for heteroskedasticity This test is based on the assumption about the distribution of u i conditional on Xi is that it has a mean of zero, thus, the variance of this conditional distribution does not depend on Xi.: 21 Null hypothesis: heteroskedasticity not present Test statistic: TR^2 = 17,6377 with p-value = P(Chi-Square(19) > 17,6377) = 0,54675 > 0,05 At the 5% level of significance, we have enough evidence not to reject the Null hypothesis In other words, the errors are homoskedastic T - test of the hypothesis relating to a regression coefficient This test uses t-statistic to test the hypothesis about the slope (coefficients) and see whether the independent variable is significant to the model or not 0: 1: =0 ≠0 22 • It is clear from the result that at 5% level of significance, popln, coast, medincm are significant with the p-values are all smaller than 0,05 • Meanwhile, excluding constant, electr & density have the highest p - value (0,89 & 0,86, respectively), meaning that at % level of significance, we have enough evidence not to reject Null hypothesis In other words, electr & density are highly insignificant As a result, we will exclude all the independent variables that are not significant Therefore, we just have popln, coast, medincm as significant variables 23 VI/ ADJUSTED MULTIPLE REGRESSION MODEL: After Robustness check and T-test, we decide to just include popln and coast in the model as independent variables We have the adjusted OLS regression: airqual = 118,627 + 0,0047 popln – 33,4 coast It can be seen from the result of Gretl that: At 5% level of significance, we have enough evidence not to reject the Null hypothesis that: error is normally distributed At 5% level of significance, we have enough evidence not to reject the Null hypothesis that: Heteroskedasticity is not present Meaning that the errors are homoscedastic 24 VII/ FINDING AND DISCUSSION: Finding • The OLS estimate of the intercept is 118,627 Meaning that if the area is not on the coast and population is zero, the weight of suspended particulate matter is 118,627 This estimation is meaningless and not real • The OLS estimate of the coefficient on the population in thousands is 0,0047 The number let us know if the population increases by thousand, the weight of suspended particulate matter increases by 0,0047 (unit of measurement) • The OLS estimate of the coefficient on SMSA’s coastal typography is -33,4 Through this, we can see if the area is on the coast, the weight of suspended particulate matter decrease by 33,4 (unit of measurement) • The adjusted = 0,335518 : It means that the two variables explain 33,55 % of the variance of airqual Because of limitation of observation, the percentage of 33,55 % meets our objective set in advance 2 Discussion Accordingly, T-test shows that in this model have variables, which not have meaning in statistic The three left that have an effect on are quality are popln, coat, medincm However, multicollinearity test shows that popln and midenicm have a linear relationship, then we must exclude medincm Test of normality of residuals of both the Model and Adjusted model show with the p – value for Null hypothesis: The errors are normally distributed is around 0,05; shows that we need to have more observation for the distribution of residuals to reach the Normal distribution Another way is to exclude variables to increase the degree of freedom, which is the way we choose to come up with the adjusted regression model According to survey, it shows that population have an effect on weight of suspended particulate matter Factor of geography: coastal also significantly affects air quality 25 Even though, coastal position is unchangeable, which belongs to the nature Authority of each metropolitan areas should take population into consideration, avoid the problem of air pollution to spiral out of control 26 CONCLUSION The objectives of this research was to investigate the underlying factor that determine the air quality of Standard Metropolitan Areas by measuring weight of suspended particulate matter According to the data, there are 10 factors affecting the air quality And we established the model for calculating which factor affect quality of the air with ten factors (price, brand name, phone features) and check the model have any problem about normality of residuals, heteroskedasticity, etc To use T-test and F-test for checking variable statistic meaning The result used multiple regressions analysis to test the effects of two significant independent variables (price, brand name, product feature) on the weight of suspended particulate matter However just 33, 55% of the variance explained Therefore, we need to keep working on the data, adding time series data if possible to update the new factor of industrial era to the regression model 27 REFERENCES Conserve Energy Future: “What is Air Quality?” https://www.conserve-energy-future.com/what-is-air-quality.php Air quality Organization: “Source of indoor pollution” http://www.airquality.org.uk/05.php?fbclid=IwAR1FB2itSmBo742Hporc800UCRAjnIurGUocdLBTpgfxCoclSnciZYXzwI Department for Environment Food & Rural Affairs: “Causes of air pollution” https://uk-air.defra.gov.uk/air-pollution/causes? fbclid=IwAR2WtcrNiIVErUvYtB6V7SLoUfSliKZDBDRuWSg pL6zsNjiJx9szUqZM-A Universe Today: “What causes air pollution?” https://www.universetoday.com/81977/causes-of-airpollution/?fbclid=IwAR086zdGNr0POOdYjHIZAmdGaH9tfwULMz9tyT8l8DTyg3frDp1HVtWeWc Air Pollution: “Clean Air in the UK” http://www.air-quality.org.uk/06.php? fbclid=IwAR1wg06YZD3Mn1jDRJT9exkn34OrqZhExQk7LFNnoqMQCGthD4LVQzMkFk “Population density and urban air quality”, Rainald Borck, Philipp Scharauth, Munih Society for the Promotion of Economic Research – CESifo GmbH “The relationship between air pollution emission and income: US Data”, Richard T.Carson, Yongil Jeon and Donald R.McCubbin, Department of economics, University of California, Sandiego, LaJolla, CA 92093, USA “Air pollution and poverty: Does the sword cut both ways?” F.W.Lipfert, J Epidemiol Community Health: first published as 10.1136/jech.58.1.3 on 18 December 2003 World Energy Outlook Special report “Energy and Air Pollution”, International Energy Agency 28 ... in this report is the explore the correlation between air quality and other factors There are numerous studies and articles that examine the relationship between air quality and its determinants, ... THEORETICAL BACKGROUND Air quality: Air quality refers to the condition of the air within our surrounding Good air quality pertains to the degree which the air is clean, clear and free from pollutants... “What is Air Quality? ” https://www.conserve-energy-future.com/what-is -air- quality. php Air quality Organization: “Source of indoor pollution” http://www.airquality.org.uk/05.php?fbclid=IwAR1FB2itSmBo742Hporc800UCRAjnIurGUocdLBTpgfxCoclSnciZYXzwI