Short-run food price prediction methods

The starting point for short-term prediction of food prices is an analysis of the factors underlying their development. This is based on a picture of the medium-term stably-acting macroeconomic variables that affect food prices, including knowledge of their intensity and the timing of their transmission. This analysis is formalised by a partial single-equation model of food price inflation, and the model approach is further supplemented and verified using other information and expert judgement.

The short-term prediction of food prices is based on a paradigm that in the short run (up to 1 year) food prices show, in addition to high seasonal volatility, a strong link to cost factors - particularly input prices in the form of agricultural producer prices and food industry prices - and also the exchange rate. These factors reflect short-term production and trading conditions on the market, which in turn reflect the quality and quantity of harvests both at home and abroad as well as current administrative interventions in the form of barriers to exports/imports of agricultural or food products (export and import licences, tariff protection, hygiene and phytosanitary regulations and suchlike). Demand (both in the wider sense of the output gap and in the narrower sense of household consumption and retail sales) does affect food prices, but its influence within this time horizon is much smaller than the potential magnitude of changes in cost and administrative factors.

Agricultural producer prices are one of the crucial explanatory variables for food price developments. The first reason is that they are a significant cost input at the start of the "product vertical" (agricultural producer prices - industrial producer prices in the food industry - consumer prices of food). The second reason is that agricultural commodities in some cases enter the consumer food market directly (fresh fruit and vegetables, potatoes). The forecast for agricultural producer prices is expertly estimated from prices of key agricultural commodities, primarily on the basis of information from commodity exchanges and the commodity outlooks of the Czech Ministry of Agriculture, together with the exogenous exchange rate scenario. The prediction of aggregate agricultural producer price inflation is then constructed by composition of the price predictions for the individual main agricultural commodities. Other explanatory variables of food prices include the koruna-euro and koruna-dollar exchange rates, foreign inflation and retail sales. The inclusion of these explanatory variables in the single-equation food price model explains about 60% of the variability of the food price index.

Given that the intensity of transmission of inflation factors in the case of food prices is stable in the medium run, the main risks of the factor approach consist primarily in the relatively strong uncertainty of the prediction of exogenous variables (agricultural producer prices in particular) together with the relatively high volatility of food prices in individual months. For these reasons, the food price prediction is reliable with regard to the overall change at the annual horizon, but less reliable with regard to month-on-month changes. Efforts to make the prediction more accurate for the very short run have led to simultaneous application of other complementary methods (an analysis of the CZSO's survey of average consumer prices of selected food products, and monitoring of price changes announced by major producers and vendors). The applicability of these methods is limited, however, and the risks of inaccuracy of the prediction for the nearest month or two are relatively high by comparison with the other price categories in the consumption basket.