Efficiency gains in commodity forecasting using disaggregated levels versus more aggregated predictions

This study evaluates the efficiency gains in forecasting of using disaggregated data in time series modeling compared to high levels of aggregation. This represents an important topic for agribusinessmen and farmers because it could provide them with insights on how to obtain more accurate predictions. This information then can be use to improve their hedging and negotiation strategies. In our research, we simulated commodity prices and evaluated them under different leves of temporal aggregations are tested (weekly, monthly and annually). The objective is to test whether models based on disaggregated data can produce better price forecasting than the corresponding model using a higher level of aggregation. For example, we test if weekly models can predict better monthly prices than monthly models. Then, we use time series methods to model the prices and select the best estimators at each aggregation level and commodity. For the commodity prices under different sample sizes and long time series, models based on disaggregated levels effectively provided an efficiency gain in forecasting. Among these levels, the best models were the weekly models. The same behavior was consistent across all possible levels of aggregations.


Issue Date:
2016
Publication Type:
Conference Paper/ Presentation
PURL Identifier:
http://purl.umn.edu/235792
JEL Codes:
C53; E17; C10
Series Statement:
Paper
9413




 Record created 2017-04-01, last modified 2017-04-26

Fulltext:
Download fulltext
PDF

Rate this document:

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