Cointegrated movement of food grains production and agricultural inputs: a time series assessment

Abstract

Introduction of the green revolution, modernization of agriculture, encouragement to research and extension in agriculture are some of the factors that contributed to the growth in agriculture. Increasing crop production and productivity are not just about the new technologies or crop management. Environmental sustainability is also of vital importance. The complexity of these issues now faced make improving crop production and productivity a more challenging task. Water, fertilisers, crop protection-inputs and professional advice all need to be managed in the most efficient manner. Fertiliser use has seen a tremendous increase in India and in other parts of the world with the spread of green revolution. Fertiliser was identified as one of the three most important factors, along with seed and irrigation for raising agricultural production and sustaining food self-sufficiency in India. In Kerala, farmers mostly depend on agriculture as a means to earn more money and concentrate more on cash crops other than crops those belong to staple food grains category which is one of the most important factors for human existence. The study intends to scrutinize the movement of food grains production and agricultural inputs through a time series assessment in India and three selected states viz., Kerala, Andhra Pradesh and Tamil Nadu using secondary information collected from various official sources. To identify the trend in production of food grains and agricultural inputs in India for the period 1950-2020 and the states (1980-2020), the linear, quadratic and cubic functional forms were selected with high values of adjusted R 2 . Trend analysis for India depicted an overall growth in an upward direction for the variables under study realizing almost linear trend. Whereas the trend analysis for Kerala, AP and TN with respect to total cropped area, fertilizer consumption and pesticide consumption showed a declining trend. In the case of food grain production, a slow increase was noted in very recent years for all the three states. CAGR was computed to observe the growth rate of the variables and for India, overall growth rate in the variables under study was positive. For total cropped area it was +0.006, +0.089 for fertiliser consumption and +0.048 for pesticide consumption and +0.026 for food grains production. However, in Kerala, the total cropped area (+0.001) and fertiliser consumption (+0.01) showed positive CAGR whereas negative growth rate for pesticide consumption (-0.01) and for food grains production (-0.002). In Andhra Pradesh, CAGR was -0.02 showing a negative growth rate in the case of total cropped area and 0.03 for fertiliser consumption, -0.03 for pesticide consumption and 0.02 for food grain production. In the case of Tamil Nadu, for total cropped area and fertiliser consumption CAGR was 0.004 and 0.02 respectively. Whereas for pesticide consumption it was -0.002 and for food grain production it was 0.02. Overall pesticide use had a negative CAGR in the states of Kerala, AP and TN. Also, the negative growth rate of food grain production in Kerala needs serious attention and it is also worth to identify the factors which discriminates Kerala from AP and TN. Time series model building was used to determine the best fit model and forecast future values of the variables under consideration. In India, Holts’ model was identified as the best to forecast total cropped area, fertiliser consumption and food grains production with adjusted R2 values as 0.96, 0.99 and 0.98 respectively. Regarding pesticide consumption Simple exponential smoothing model was the best with adjusted R 2 = 0.95. For Kerala, Simple exponential smoothing model, ARIMA (1,0,0) and Holts’ model were obtained for total cropped area (adj. R2=0.76), fertiliser consumption (adj. R2=0.66) and food grains production (adj. R2=0.85) respectively. For Andhra Pradesh, ARIMA (0,1,0) model was identified for total cropped area with adj. R2= 0.80, Simple exponential smoothing model for fertiliser consumption with adj. R2=0.93, for pesticide consumption with adj. R2=0.82 and for food grains production with adj. R2=0.82. When coming to Tamil Nadu, ARIMA (0,1,0) was the best for modeling total cropped area with adj. R2=0.76, ARIMA (0,1,6) for fertiliser consumption with adj. R2=0.74, Simple exponential smoothing model for pesticide consumption with adj R2= 0.84 as well as for food grains production with adj. R2=0.43. It is well known that Kerala imports food grains mainly cereals and vegetables from Andhra Pradesh and Tamil Nadu. To examine the pattern and dispersion of variables viz; total cropped area, fertiliser consumption, pesticide consumption and food grains production in Kerala, AP and TN, Box plot analysis was done and found that AP had highest dispersion and Kerala showed lowest dispersion with respect to variables under study. Since variability was found among the states, Mahalanobis D2 was used to estimate the pairwise distance between the states with respect to variables under study. The distance between Kerala - TN (1.94) was more when compared with Kerala - AP (1.93) and the distance between AP - TN (1.74) was the lowest. Discriminant analysis paves a way to pinpoint the casual factors which contribute to the discrepancy between the states and it identifies the root cause for the distance obtained by Mahalanobis D2 among states. Food grain production followed by fertiliser consumption was found to be the discriminating factors in Kerala - AP analysis. The distinguishing factors in Kerala - TN analysis was fertiliser consumption followed by total cropped area. Consumption pattern of fertiliser nutrients such as N, P and K in Kerala was entirely different from the recommended dose. On all Kerala basis, the average use of N, P and K were significantly lower than that of the recommended quantity depicting imbalanced use of fertilisers during the period 1995 - 2020 and for the period 1993 - 2009 for all districts in Kerala. Kerala showed highest imbalance index of 0.24 during the study period. None of the years showed perfect balance or extreme imbalance in Kerala. For district wise study it could be observed that the district Wayanad was having the highest imbalance index (0.212) followed by Kozhikode (0.205) and Idukki (0.202). The Palakkad district was having the least value of imbalance index which was equal to 0.099. To assess the co integrated movement of food grains production and agricultural inputs in India and the states under study, Vector Auto Regression was used by modeling each variable as a linear combination of past values of itself and past values of other variables in the system. The VAR models resulted in an adjusted R2 ranging from 0.95 - 0.99 for India with respect to different variables and for all the states also with significantly high values of adjusted R 2 showing the potential of the VAR approach to quantify the co integrated movement of the variables under study