2017年-世界发展银行全球_Food_Price_Seasonality_in_Africa___Measurement_and_Extent_14页_983kb

Summary of "Food Price Seasonality in Africa: Measurement and Extent"

Core Content

This study investigates the extent of food price seasonality in seven Sub-Saharan African countries, analyzing 193 markets for 13 staple and non-staple food commodities. It highlights the challenges in accurately measuring seasonality and emphasizes the need for improved methodologies to avoid overestimation of seasonal price gaps.

Main Points

• Seasonality in African Food Markets: Seasonality is significant, with the seasonal gap—the difference between pre- and post-harvest prices—being notably higher than in international markets. For example, the seasonal gap for tomatoes is 60.8%, and for maize, it is 33.1% on average.

• Key Commodities: Among staple grains, maize shows the highest seasonality (33.1%), while rice has the lowest (16.6%). Vegetables and fruits also exhibit high seasonality, while non-seasonal commodities like eggs and cassava show lower seasonality.

• Malawi's Case: Malawi has the most acute seasonal differences due to maize being its main staple, leading to a double seasonality burden for most households.

• Sustainable Development Goal (SDG) 2: Achieving Zero Hunger requires addressing food price seasonality, as it affects dietary intake and nutrition, especially during the first 1000 days of life.

Key Issues in Seasonality Measurement

• Bias in Estimation Methods: Common methods like dummy variable regression and moving average deviation (MAD) can lead to substantial upward bias in estimating the seasonal gap, particularly with short datasets (5–15 years). This is because the gap is a non-linear function of the dummy variables, which are individually unbiased but collectively lead to overestimation when taking the difference between the maximum and minimum seasonal factors.

• Impact of Sample Length: The upward bias decreases with longer samples. For example, in the no seasonality case, the bias is 21% with 5 years of data, but it drops to 7.5% with 40 years of data.

• Statistical Significance: The MAD method tends to overstate the statistical significance of seasonal patterns, even when seasonality is not present. This is due to autocorrelation in the error terms introduced by the moving average transformation.

• Alternative Models: Trigonometric and saw-tooth models are more parsimonious and reduce the number of parameters to be estimated. These models help mitigate the upward bias and improve the reliability of the seasonal gap estimates.

Methodology and Findings

• Three-Step Procedure: The authors propose a three-step procedure to select the preferred specification and minimize upward bias. This includes:

  1. Estimating the seasonal gap using a stochastic trend model.
  2. Comparing with the MAD approach.
  3. Using more structured models to reduce parameter estimation burden.

• Data Characteristics: The price data used are typically short and have missing observations, which can affect the accuracy of seasonal gap estimates. Most data come from initiatives like FEWSNET and GIEWSNET, which are relatively recent.

• Empirical Results: The study finds that the seasonal gap is consistently overestimated using dummy variable methods, especially in cases of diffuse seasonality. For instance, in the no seasonality case, the gap is estimated at 21% with 5 years of data, but it should be zero.

• Implications for Policy: The findings suggest that policy attention should focus on reducing the impact of seasonality through mechanisms like better storage, access to credit, and market integration. This can help improve food and nutrition security, especially for vulnerable populations.

Conclusion

The study concludes that while seasonality is a significant feature of African food markets, the standard methods of measuring it can lead to substantial overestimation. The stochastic trend model is slightly preferred due to its better statistical inference and parsimony. Future research and policy design should adopt more reliable methods to accurately capture the extent of seasonality and its effects on food security and nutrition.