Author: Sachin Sudhakar More
Author Address: Associate Professor, Department of Agricultural Economics, College of Agriculture, Vasantrao Naik Marathwada Krishi Vidyapeeth, Parbhani-431 402 (Maharashtra)
Keywords: ADF test, ARIMA, cotton, forecast, PP test, stationarity
JEL Codes: C32, C53, Q13, Q17, Q22.
Reliable and timely estimates of cotton production are important providing useful inputs to policymakers for proper foresighted and informed planning. So an attempt was made to forecast the production of cotton at all India level using a time series model. The annual data on production of cotton for the period 1951-52 to 2018-19 was processed. The data were transformed into logarithmic series to stabilize the variance of the series. The stationarity of the data was checked with the help of the Augmented Dickey-Fuller and Phillips-Perron tests. The results of ADF and PP tests confirmed the cotton production series was non-stationary at level, so stationarity in the data was brought by differencing the data series at a first lag. The pattern present in ACF and PACF and results of SCAN and ESACF provided guidelines to select the order of non-seasonal ARIMA model. The best fit ARIMA model (ARIMA: 3 1 1) was selected based on AIC criteria and residual diagnostic. The performance of the model was judged based on the MAPE value. The out of sample forecast of cotton production at all India level was carried out for the period 2019-20 to 2021-22. The forecasted values indicated a slight increase in the production of cotton compared to 2018-19.
Indian Journal of Economics and Development
Volume 16 No. 4, 2020, 583-590
Indexed in Clarivate Analytics (ESCI) of WoS
Scopus: Title Accepted
NAAS Score: 4.82