An analysis of the price behaviour of selected vegetables at a New Zealand auction : a thesis presented in partial fulfillment of the requirements for the degree of Master of Business Studies at Massey University
The immediate and short run behaviour of prices and volumes of eighteen vegetables at auction are examined. The objectives were to describe the behaviour of weekly prices and volumes and to investigate various relationships.
The data analysed are weekly volume and turnover for eighteen vegetables at a Palmerston North auction for a three year period. Various climatological variables relating to the same period were also analysed.
Much of the descriptive analysis relies on the techniques of Exploratory Data Analysis; boxplots, letterplots, and a resistant smoother are used extensively. These methods
facilitate the analysis of the behaviour of prices and volumes over time.
The auction marketing system is discussed at length, with particular emphasis on the effect of length of run on supply response.
Various relationships are examined predominantly using stepwise reqression. These include: current price and quantity; current quantity and lagged price; quantity and month of the year; price and month of the year; price and various weather variables; quantity and various weather variables; current price and lagged price.
Some transformations are used to try to get a more linear relationship between price and quantity. This relied on
fitting several resistant lines, another EDA technique.
The interdependence or interrelationship of prices examined using Principal Components Analysis, and principal components were extracted and described.
Some useful insights into the behaviour of the market are gained. Immediate run price variation, that is from week to week, is quite large and this is reflected in low R-squared values for the price-volume relationships. The relationship between current marketed volume and lagged prices was also investigated. The results indicate that in the immediate run, using weekly prices, this relationship is weak.
Arguably the most useful analysis was the monthly price boxplots which give a clear graphic display of the behaviour of prices. These and the other plots give indications as to when prices were highest and lowest. They may be useful to growers in planning production and harvesting.
It was discovered that weekly prices were more variable than weekly volumes for 10 vegetables, less variable for six vegetables, and equally variable for two vegetables. 0ften the highest weekly price did not coincide with the lowest weekly supply, and the lowest price did not correspond to the highest supply. This suggests that bidders under or over estimate the quantity on the floor and each other's requirements.