FORECASTING AGRICULTURAL COMMODITY PRICES WITH LIMITED DATA SETS

Many developing countries wish to, but do not publish forecasts of commodity prices. One difficulty often cited is the length (shortness) of the data series. This study tried to determine whether this was a valid reason. Price data for two commodities in a specific country, where the problem is acknowledged were utilized in a case study approach. Existing models were evaluated for applicability to the data series. The identified models were estimated with estimation data and then while keeping the model parameters constant, prices were forecasted in the out-of-sample period. The error associated with the forecast of each model was calculated and the relationship between quantity of data and model accuracy evaluated. The error values of all the models were high. The results of the analyses were inconclusive in terms of the study objectives.


Issue Date:
1991
Publication Type:
Thesis/ Dissertation
PURL Identifier:
http://purl.umn.edu/10998
Total Pages:
157
Series Statement:
Graduate Research Master's Degree Plan B Papers




 Record created 2017-04-01, last modified 2017-08-23

Fulltext:
Download fulltext
PDF

Rate this document:

Rate this document:
1
2
3
 
(Not yet reviewed)