Food price trend analysis: Lessons for strengthening food security policy in Tanzania

  • F Maro
  • F Mwaijande
Keywords: Co-integration, Vector Error Correction Model


Increase in global prices for most key cereal crops has had an unprecented effect on local markets prices for maize (Zea mays L.) and rice (Oryza sativa), raising policy concerns especially in eastern and southern Africa. The objective of this study was to analyse maize and rice price transmission within Tanzania domestic markets. The study used monthly wholesale prices from nine local markets in Arusha, Dar es Salaam, Iringa, Lindi, Mwanza, Rukwa, Dodoma and Morogoro from January 2004 to August 2013. The Vector Error correction model was used. Markets were categorised into leading and follower markets. Results showed that 88 percent of maize prices in selected markets were stationary, while for rice it was 100 percent. Further analysis using Johansen test indicates 63 percent of selected maize market pairs and 75 percent for rice markets pairs were co-integrated. Leading markets were found to transmit relatively small percentages (20 percent) compared to more than 70 percent of prices transmitted by follower markets. It took relatively longer for smaller markets to transmit prices to their larger counterparts. This was also supported by granger causality analysis, where larger markets prices
failed to be transmitted to small markets. Very few pairs of markets (5%) had bi-directional movement of prices, indicating limited flow or market rigidity in sharing price information. The speed of price adjustment was also very slow, especially when higher prices originate from smaller markets. This trend implies presence of many layers of markets and the prices were largely controlled by fewer traders rather than marketing forces or other actors like farmers who were down to the value chain. This kind of monopoly leads to price volatility and consumers are forced to pay more, hence, affecting affordability of majority net buyer consumers.

Key Words: Co-integration, Vector Error Correction Model


Journal Identifiers

eISSN: 2072-6589
print ISSN: 1021-9730