Rice is one of the most important cereal in the world, its cultivation occupies high crop area. The annual area harvested in 2003 reached 153 million hectares with a production of 589 million tons worldwide. In this production almost 200 million harvested in China and India, similar amounts were obtained in other Asian countries (^{FAO, 2004}). In Venezuela production is 700 000 tons, thus the number of people who are economically dependent crop is high. It is estimated that the world population in the next 50 years will have a 40% increase therefore will generate increased consumption and production 70% by 2050 (^{FAO, 2012}) should be increased. The world rice market is large producers and large consumers. Latin America holds about 5% of world rice area harvested. Brazil, Colombia, Peru and Venezuela have the largest acreage in the region. In terms of unitary yield in Latin America: Colombia, Peru, Venezuela and Uruguay hold the top positions, while Brazil yields are low. The current values of per capita consumption in Venezuela are between 15 and 20 kg of rice per person per year, while in Colombia stands at 57 kg per person (^{Montilla, 2012}).

Similar to the behaviour of other cereals, rice prices in the international market is affected by supply, demand and population growth, as well as biological and climatic phenomena that affect productivity variables. These factors confer high volatility of prices, so it is necessary to rationally plan decisions at both producer and government agencies, to determine the area to cultivate and financing planting.

One method that can be used for forecasting with time series is integrated moving average ARIMA (^{Gujarati, 2003}) autoregressive model, commonly known as Box Jenkins methodology (^{Box-Jenkins, 1978}). The objective of this research was to evaluate the suitability of the model to predict the behaviour ARIMA time series of international rice prices, taking the time interval in June 2002 to November 2012, for estimating over a period of 16 months, from September 2013 to December 2014 for reliable forecasts.

Based on information from the Organization for Food and Agriculture (^{FAO, 2012}) and rice cereal for the significant contribution was selected to the world economy, the time series data used corresponded to reflect the international price of rice for sale the period June 2002 to November 2012. First of all, the series was transformed using operations such as natural logarithm, first order differences and cycle order to obtain a stationary series; then the ARIMA (1,0,1) (Figure 1) model was used to develop the forecasts. Free software R was used for statistical analysis (^{R Development Core Team, 2008}) and Excel for graphics.

Figure 1 Basic Flowchart of the methodology of ^{Box-Jenkins (1978)}. 

In the 10 years of records, notes that in 2007 and 2008 showed particularly significant price increases (Figure 2a). According to ^{Abbassia, (2012)} that crisis was mainly due to natural disasters such as droughts and floods due to climate change, simultaneously there were other factors such as the increase in oil prices which caused increases in the costs of production and transportation. The number of consumers also increased in emerging economies like China, India and Brazil. The Price of rice in the coming years 2013 and 2014 tend to rise (Figure 2b) remaining in the line of mean values with little variation.

Figure 2 (a) Trend rice prices; and (b) Estimated rice prices in December 2014. 

The estimated price for the first month clinches US $648 a tonne for December 2014 and may reach US $665 a tonne. The estimated model (1) is integrated moving the average period of 12 months of autoregressive components. Equation (2) is the result of applying the natural logarithm, first order differences of order one cycle and the econometric variable price. It was observed that the model is a good predictor adequately reflects the stochastic process generated by the data series (^{Brockweel, 2002}).

The real price of a tonne of rice in June, 2002, was US $202 in November 2012 was US $590.73 and in December 2014 the graphical estimate has a value of US $665 (Figure 2b), in both series the rise reflects a small gap in the cycle from 2008; i.e., the stochastic process interpreted the price movement in all years considered in the study.

The inter-monthly prices presented in graphic form and its respective band estimate 95% of trust in December, 2014 tend to increase (Figure 3) . Any phenomenon that occurs will be within this range, since the model has the ability to reflect what happens in the future at market prices.

Figure 3 Tendency in prices of rice and estimation in December 2014 with their respective lower and upper bands at 95% prediction. 

The inter-monthly prices in absolute terms between the estimates lower and upper band until December 2014 tend to rise. (Table 1). The price band in figures reflected well the process, which means from the viewpoint of stochastic processes modeled the process generated by the series (^{Karlin, 1975}). The prediction by the adjusted econometric model indicated prices US $ 648 per tonne in the first month and $665 per ton at 16 months; i.e. December 2014. For this period, the estimated prices show a variation between the lower bands of US $191.7 to top $2 309.0 per tonne respectively.

Table 1 Estimated prices and prediction bands 95% of rice in December 2014 

The estimation of the first month filed a maximum value in the range of US $1 220.70 and a low of US $344.6 per ton in the lower band, whereas at 16 months estimation, the model showed a maximum value of US $2 309.0 in the top band and a minimum of US $191.7 per ton in the lower band. The lower band is maintained closer to the upper band estimated, and the latter average spread rapidly increased as the average time estimation.

The price increase may be associated with unfavourable climatic kind in Asia and South America. However, some experts suggested that, this may be offset by the fact that global stocks of the product, particularly exporters are quite abundant (^{Méndez del Villar, 2012}).