October 3, 2017 | Mr. G R Manjunatha (Scientist B, Central Sericultural Research & Training Institute, Berhampore)
In this study the major food grains considered to examine the food price volatility are paddy, wheat, bajra, tur and bengalgram, i.e. considering two major cereals, one major millet and two major pulses. Paddy and wheat are the major food grains produced and paddy forms the single largest food grain produced forming around 40 percent of the total. Monthly data on whole sale price index with 2012 base year was obtained from 2012 to 2017 from Ministry of Commerce, Government of India for analysis to estimate the food price volatility.
The graphical representation of the wholesale price index at constant prices indicates the non-stationary nature of the data which does not necessarily imply the presence of food price volatility in all these crops. Therefore there is a need to make the wholesale price index series stationary by first differencing which makes the series stationary. The statistical tool to find whether there is stationary or not is attempted by using the Augmented Dickey Fuller (ADF) test with the null hypothesis that the price series has unit root (i.ethe series is non-stationary/ the data has variable mean and variance) against the alternative hypothesis of no unit root (ie the series is stationary/ the data has constant mean and variance).
The ADF performed on paddy, wheat, bajra, tur and bengalgram indicated that the wholesale price index at constant prices is non- stationary in all these five crops. In the second step, the first differencing is done to make the whole sale price index series stationary, which is a necessary condition to check for food price volatility. Upon first differencing, all the data on wholesale price index at constant prices were made stationary (Table 1). The differenced series was regressed with intercept and the obtained residual was subjected to ARCH-LM Test (Auto Regressive Conditional Heteroscadasticty -Lagrangian Multiplier) using Eviews software to identify the presence of volatility. In the ARCH test, the null hypothesis is the presence of ARCH effect i.e., the presence of food price volatility (presence of conditional heteroskedasticity or autocorrelation in the squared residuals). It was found that the ARCH effect was found to be significant at 5 percent level in all the five crops except that of wheat (Table 2). Therefore, it is concluded that there is food price volatility in paddy, bajra, tur and bengalgram and not in wheat. This implies that in wheat crop, there is no price volatility, i.e there is no ARCH effect and there is no serial correlation and conditional heteroskedasticity. But in all the crops paddy, bajra, tur and bengalgram, there is price volatility and there is serial correlation.
Thus, despite offering minimum support price and procurement of paddy, there is price volatility. Paddy being the single largest food crop forming 40 percent of the food grains produced as also the major crop which is procured, faces the food price volatility. Next, we have attempted to forecast the prices using the ARCH family models such as ARCH, GARCH, TARCH and EGARCH (Table 4 to 7). In order to choose the most suitable model for forecasting, the AIC (Akaike information criterion) and SIC (Schwarz information criterion) are used (Table 3). The most suitable model will be that with low values of AIC and SIC. Accordingly the wholesale price index has been forecasted and presented below (Fig. 1 to 4).
Therefore, the study concludes that there is food price volatility in paddy, bajra, tur and bengalgram as observable in the price series of paddy from 2012 to 2017. The reasons are that despite minimum price support and procurement operation, farmers are not insulated against the price fluctuations. Therefore it is crucial for the policy to consider developing measures towards removing market imperfections, providing market infrastructure, road infrastructure, storage infrastructure and providing market information and other relevant measures. Wheat is the only crop which was found to be having no price volatility.
Achal Lama, Girish K Jha, Ranjit K Paul and BishalGurung, Modelling and Forecasting of price volatility: An application of GARCH and EGARCH models. Agricultural Economics Research Review.Vol 28 (No. 1): pp: 73-82
Bhardwaj S. P., Ranjit Kumar Paul, Singh D. R. and Singh K. N. (2014) An Empirical Investigation of Arima and Garch Models in Agricultural Price Forecasting, Economic Affairs, 59(3) : 415-428
PiotrFryzlewicz, Lecture notes: Financial time series, ARCH and GARCH models (http://stats.lse.ac.uk/fryzlewicz/lec_notes/garch.pdf)
Ranjit Kumar Paul, Bhardwaj S. P., Singh D. R., Anil Kumar, PrawinArya and Singh (2015). Price K. N. volatility in food commodities in india - an empirical investigation, Int. J. Agricult. Stat. Sci. Vol. 11, No. 2, pp. 395-401.
Robert Engle, An Introduction to the Use of ARCH/GARCH models in Applied Econometrics (http://www.stern.nyu.edu/rengle/GARCH101.PDF)