Responsible solver:

Mgr. Jiří Pokorný

Aim of the project:

An improvement of the methodology of predicting prices of agricultural commodities made in TÚ 40/4221.

Summary of the project:

The project develops the prediction software used at IAEI, on the basis of internal projects from previous 3 years.
The first part is focused on cycles and finding their lengths with the analysis of periodogram. Also the informational value of the cycles for the prediction was evaluated. Particularly, the time series of monthly farming prices were analysed, and it was found that the lengths of period of cycles 11, 12, 13, and 14 are chosen wisely, and there was not enough cases found in favour of adding another length of period in the prediction software.
In the second part, the price transmission in the vertical of broiler meat was analysed – from the prices of feed cereals, through the price of compound feed, to the farming price of broilers. The analysis was based on regression models, as well as testing for cointegration and the Granger causality. From the temporal point of view, the cases with no lag, 1 lag, and 2 lags were analysed. The prices concerned were Czech and to some extent also Polish. The cointegration relationship exists between each pair of the vertical for both countries, with the exception of the pairs feed wheat – compound feed in the Czech case, and compound feed – price of broilers in the case of Poland. A change in price of feed wheat causes a change in price of compound feed, and a change in price of compound feed then causes a change in farming price of broilers. There was one exception in the Czech case, and that was a change in price of compound feed causing a change in price of feed corn. Other relationships concerning the feed corn are insignificant. In spite of the previous, the regression analysis did not find any significant relationship in the development of prices in the vertical of broiler meat.
In the third part, upgrades of the prediction software are presented.