Determinants of price transmission
von Carolin Simone Mengel
Datum der mündl. Prüfung:2014-07-14
Betreuer:Prof. Dr. Stephan von Cramon-Taubadel
Gutachter:Prof. Dr. Bernhard Brümmer
Gutachter:Dr. Sebastian Heß
EnglischThis dissertation aims to improve our understanding on how prices changes are transmitted between markets and what determines these dynamics. We undertake three studies to shed light on price transmission from different perspectives. Based on price transmission estimates from literature and own data, we seek to identify regional- and product-specific differences and political, cultural and geographical influences. One central question is whether distance and international borders have an impact on price transmission between spatially separated markets. The analyses combine in innovative ways large-scale market analyses with up-to-date methods such as meta-analysis and threshold cointegration. The findings contribute with empirical evidence to the theoretical considerations about the determinants of price transmission. The field of price transmission has spawned numerous econometric approaches to depict how prices in different locations or on different value chain levels affect each other. One standard concept is testing if price series move together over time (cointegration). Common parameters of interest are the strength of such a long-run price equilibrium (elasticity) and the speed of price adjustment to short-term deviations. In more recent applications, trade costs are incorporated as threshold effects in the price adjustment. The presented work deals with these aspects and seeks to isolate patterns in the results. The first chapter on price transmission from international to domestic markets aims to improve our understanding of the extent and speed of the transmission of international cereal prices to local markets in developing countries. We analyze two samples of price transmission estimates, one extracted from a comprehensive literature sample of 31 published papers and studies on cereal price transmission and one containing of own estimates of cereal price transmission using the FAO’s GIEWS dataset. We also present the results of a non-parametric analysis of price transmission in which we analyze the share of periods in which domestic and international prices have jointly increased or decreased. We find a higher share of cointegrated commodity market pairs in the literature sample (79% compared to 43%). This may be due to publication bias. Cointegration is more prevalent for rice market pairs and less prevalent for maize market pairs. Both the literature and the GIEWS-based estimates point to average long-run price transmission coefficients (elasticities) of roughly 0.75 and average short-run adjustment parameters of roughly 0.09-0.11. In most cases domestic prices adjust to deviations from the long-run price relationship, but international prices do not. The only notable exception to this rule is rice, which suggests that the determination of international rice prices differs fundamentally from the determination of international wheat and maize prices. In a subsequent meta-regression analysis we measure how much of the variation in the samples of price transmission estimates can be explained by country- or product-specific factors. However, this analysis fails to generate compelling results. An analysis of domestic price volatility reveals that median volatility has increased since July 2007. In the second chapter, we conduct a comprehensive meta-analysis on the effect of distance and border effects on spatial price transmission. We use price transmission estimates for 1189 grain market pairs extracted from 57 studies and seek to explain them by airline distance and existence of an international border. The findings indicate distance and border effects on both price cointegration and price transmission. A border separating two markets reduces the probability of cointegration of price series by 23% compared with markets located in the same country. 1000 kilometers of airline distance reduces the probability of cointegration by 7%. The speed of price adjustment is on average 13% slower in international than in intra-national market pairs. 1000 kilometers of distance within a country yields on average 6-20% slower price adjustment. Distance effects are economically insignificant for international market pairs. Maize price pairs are less often cointegrated compared to rice prices and cointegration is most prevalent for barley. Price transmission is slowest in wheat markets. In peer reviewed studies price transmission is faster. However, the explanation need not be a publication bias but can also result from higher quality methodologies and consequently fewer misspecification errors. Moreover, we identify a set of model specifications that significantly affect price transmission estimates. The study contributes to the literature by presenting a first meta-analysis of spatial price transmission literature and providing insights into distance and border effects on spatial price transmission. The third chapter explores the link between proximity and price cointegration in West African rice markets. Proximity is captured with variables for geographical, political and cultural distance. Linear and threshold cointegration is tested for a set of 756 rice market pairs in 6 West African countries, with threshold specifications accounting for transaction costs. Whether proximity matters for the results is determined in a second step with a multinomial logistic regression. The estimation produces robust and statistically significant evidence for a link between price transmission with air-line and road distance, international borders, contiguity and a common language. We conclude that proximity matters for market integration processes in West African rice markets. Overall, the findings from both literature and own estimations indicate that cointegration is more present between local and international markets, compared with local-local price pairs. In comparison to wheat and maize, the international rice market appears to be more dynamic with regard to price shocks. Furthermore, the evidence confirms the hypothesis of distance and border effects on spatial price transmission. Both decrease the likelihood of cointegration and the speed of price transmission, based on literature estimates and own data. We conclude that price shocks are less likely to spread out over longer distances and across borders. Moreover, more remote markets may not be able to absorb supply or demand shocks. This should be taken into account when targeting policies at food prices. Deeper market integration could potentially be achieved by lowering the trade and communication costs associated with distance and borders, for example via infrastructure measures, more efficient border processing or less restrictive trade barriers.
Keywords: price transmission; agricultural trade; ECM; cointegration; food markets; market integration