Maize Parity Correlations

Introduction

The aim of this report is to investigate the correlations between the pairs (Safex YW, Zar) and (Safex YW, Cbot) as a function of the distance in Zar from Export Parity.

The plot below shows the evolution of Safex YW since 2007. The blue ribbon is the price range defined by export (bottom) and import (top) parity. When prices are at or below export parity we expect Safex YW to be highly correlated to Cbot. Here South African farmers can compete on the global markets. Similarly, when prices are at or above import parity we expect Safex YW to be highly correlated to Cbot. In this case South African farmers cannot compete with global markets since it is cheaper to import maize from abroad. For them to compete prices will have to follow Cbot. The correlations within the range of prices defined by the blue ribbon are less clear.

In a similar vein we can explore the correlations between Zar and Safex YW. If the current price is at or above import parity it is cheaper to import maize from abroad. If the Zar strenghtens we have more buying power and can import more maize from abroad. In order for Safex YW to compete on the international markets the price has to decrease. If the Zar weakens we have less buying power and more incentive to make use of our local crop. This should drive Safex YW prices up. This gives a strong positive correlation for prices above import parity.

Suppose now that prices are at or below export parity. If the Zar weakens we can compete better in the export market, this should drive Safex YW prices up. If Zar strengthens we become less competitive on the international markets and Safex YW should decrease in order to compete. This again gives a strong positive correlation.

The plot below shows what we expect to see given the obove reasoning. For prices at or below export parity and prices at or above import parity we expect to see high correlations. As we move within the blue ribbon above correlations decrease.

The dataset consists of monthly price, currency and parity data. We divide the From Export Parity data into bins of width 100 Zar. Within each of these bins we determine the correlations between the pairs (Safex YW, Cbot) and (Safex YW, Zar). The results are shown in the following sections.

Safex YW vs Cbot

The plot below shows the Safex YW van Cbot correlations as a function of the price difference with respect to export parity. Notice the near perfect correlations at the low and high ends of the distance from export parity. The data shows a negative correlation for the intervals (-100, 200] and (1100, 1200]. This may be a result of lack of data in those intervals. The remainder of the plot agrees with our initial intuition.

Safex YW vs Zar

The plot below shows the correlations between Safex YW and Zar for the same buckets as the plot in the previous section. Notice that the left most data pint shows strong negative correlation. This point seems suspicious. It implies that when prices are deep under export parity a strengthening Zar gives rise to increasing Safex YW prices. This can be due to a lack of data in those buckets and should be taken with a grain of scepticism. The remainder of the plot agrees with our initial intuition.

Avatar
Mauritz van den Worm
Portfolio Manager and Quantitative Researcher

My research interests include the use of artificial intelligence in managing commodity portfolios