Investigating the relationship between independent economic variables and dependent variables of municipal waste generation (case study: Tehran )

Investigating the relationship between macroeconomic variables and variables of waste generation is of great importance in urban management planning with a waste disposal reduction approach. Therefore, in this study, using linear regression method and using data from 56 months (March 2014 to October 2018) related to five independent economic variables and four dependent variables of waste generation in Tehran, four models were developed. In this study, the statistical relationship between independent economic variables and dependent variables of waste generation at the level of 90% confidence has also been investigated. The results showed that 74% of the changes in the tonnage of construction and demolition waste using the economic variable of the price index of goods and consumer services and 69% of the changes of the tonnage of recycable waste separated at the source by the two economic variables of the price index of goods and services and US dollar price announced by Central bank can be described. On the other hand, by using the variable price of the Euro currency announced by the Central Bank, 35.6% of the changes in the tonnage of sweeping & green waste in Tehran can be predicted. It is worth noting that only 21.4% of the changes in the tonnage of municipal mixed waste can be described by the economic variable of the US dollar price in the market.

In this study, relevant data of independent economic variables and dependent waste generation variables during 56 months were analyzed. The five independent economic variables considered in this study were as follow: 1.consumer price index (X1), 2. US Dollar currency price announced by central bank of Iran (CBI) (X2), 3. Euro currency price, announced by CBI (X3), 4. US Dollar currency open market price (X4) and 5. Euro currency price, open market price (X5). Also, four dependent waste generation variables of Tehran included in this study were: amounts of mixed MSW transported to transfer stations (Y1), sweeping & green waste (Y2), and source separated recyclable waste (Y3) and construction and demolition waste tonnage (Y4) collected from Tehran. Then, Shapiro-Wilk test was used to check the normality of data distribution received from Tehran waste management organization (TWMO). According to the results of similar studies in the field of the prediction of MSW generation regarding to different variables impacts, the backward removal method and the following equation used to develop a multiple linear regression model:Yi=β0+β1X1+β2X2+β3X3+β4X4+β5X5+ϵWhere Yi is the dependent variable, β0 is the intercept, X1 to X7 are independent variables, β1 to β7 are regression parameter and ϵ is residuals. On the other hand, by examining the dependent variables data over time using R software, it was determined that the data may also have a significant relationship with time. So the modified model is presented as follows:Yi=β0+β1X1+β2X2+β3X3+β4X4+β5X5+β6T1+β7T2+ϵWhere T1 is auxiliary variable of time and T2 is square of T1.

Regard to the outputs of R and SPSS software, it was found that the data of dependent variables considered in this study follow the normal distribution. Then, using the backward elimination method in multiple linear regression, the developed model for each of the dependent variables was presented as follows:MAE MARE RMSE R2 Adjusted Developed models Dependentvariables156.5 0.023 259.2 0.214 Y1=641.560-2.883 × 10-18 X 44 + 11.473 T1 Y151.3 0.056 85.5 0.356 Y2=-743.398 + 0.093X3 -1.35× e-6 X34 Y236.3 0.033 56.5 0.69 Y3 =1647.268 – 0.0004X1X2 + 32.554T1 Y30.058 0.0053 0.084 0.74 Ln(Y4) = 7.181 + 0.060X1 + 0.00015X12 – 0.001T2 Ln(Y4).

In this study, linear regression method and 56 months data (April 2014 to November 2016) related to five independent economic variables 1- Price Consumer Price Index 2- US dollar price announced by the Central Bank 3- US dollar price in the market 4- Euro currency price announced by the Central Bank and 5- Euro currency price in the market and four dependent variables of waste production 1- Tonnage of mixed urban waste 2- Tonnage of sweeping & green waste 3- Tonnage of recycable waste separated at souce and 4- Tonnage of construction and demolition waste in Tehran were used to fit four new models. Based on the results, 74% of the changes in the tonnage of construction and demolition waste in Tehran can be described using the economic variable of Consumer Price Index. The low values of error criteria in this regard indicate the high power of this model in predicting changes in the dependent variable of tonnage of construction and demolition waste in Tehran using the values of the independent economic variable Consumer Price Index. 69% of the changes in the production tonnage of recycable waste separated at source can also be described by the economic variables of Consumer Price Index and the price of the US dollar announced by the Central Bank. Also, using the Euro currency price variable announced by the Central Bank, 35.6% of the changes in the tonnage of sweeping & green waste production in Tehran can be predicted. However, only 21.4% of the changes in the tonnage of municipal mixed waste production can be described by the economic variable of the US dollar price in the market.Also, at the 90% confidence level, the two response variables of municipal mixed waste tonnage transported to intermediate transfer stations (Y1) and recycable waste tonnage separated at source (Y3) have shown a statistically significant relationship with time (T1). On the other hand, the response variable of sweeping & green waste tonnage in Tehran entering Aradkooh disposal site (Y2) have shown a statistically significant relationship with the independent variable price of Euro announced by the Central Bank (X3) .also the response variable of construction and demolition waste tonnage (LnY4) have shown a statistically significant relationship with time variable T2 and independent economic variable of Consumer Price Index (X1) at the 90% confidence level.