Deciles (percentiles) analysis of price data is useful as a way to identify the middle or median price and provide some estimates of the probability of high and low prices occurring, using the historical data. The leap of faith is that the selected historical data is a guide to the future. This article takes a look at this.
Who in the wool industry has not heard of the March 1951 market when wool reached pound a pound? How many wool growers have you met that sold at that level or at the peak level in March 1973, or more recently on October 1 2002? Not many is the answer as these prices were only around for a very short time. This is where percentiles (see more here) are useful as they show the amount of time a market has traded above and below a given price level. The greater amount of time a market has traded at or above a price level, the greater the chance of that price being achieved.
Figure 1 shows the eastern 19 MPG from 2024 onwards in nominal, Australian dollar terms. Overlaid on the graph is the rolling five-year percentile, which shows the ranking of price, compared to the preceding five years. The rolling percentile has ranged from 0% to 100% during the past two decades, showing the variation in the 19 MPG.
While percentiles are on rock-solid ground for analysing historical data, how good is it as a tool for looking forward? Figure 2 takes the rolling five-year percentile for the 19 MPG from Figure 1 and shows how price changed one year into the future (bars – left-hand axis). For percentiles to be useful at forecasting we need to see the price fall when the percentile is high and the price rise when the percentile is low. It does this, sometimes, and sometimes it does not. That is a problem.
Table 1 analyses Figure 2. It shows the range of price changes, across the rows, for a given rolling five-year decile (see the left-hand column). Take the top price decile (91-100%) as an example. Intuitively, we would think this flags a high price level therefore price changes one year forward should generally be a fall to lower price levels. In practice since 2004 it has been a coin toss as to whether the 19 MPG rises or falls. In other words, the top decile for the 19 MPG has no forecasting ability. For the lower price deciles, there is a slight skew to higher prices 12 months forward (which looks to be an inflationary effect).
The floating Australian dollar acts as a shock absorber for the Australian economy, and in doing so also acts as a shock absorber for wool prices. Table 2 repeats the format of Table 1 for the 19 MPG in nominal US dollar terms. In US dollar terms, changes in the 19 MPG according to the rolling five-year percentile level are more closely matched to the intuitive expectation that prices are likely to fall if they are high and likely to rise if they are low, in this instance when looking forward 12 months. There are still significant exceptions, for example, in the top decile of price, the 19 MPG fell over the coming year 60% of the time and rose by 29% or more for 20% of the time.
What does it mean?
Deciles should not be used as a guide to future price risk for wool (or beef/sheepmeat) prices. They look backward, not forwards. Their ability to forecast price levels is nil to poor, especially for Australian dollar prices. There is an old saying among market traders that prices are never too high to buy and never too low to sell.
Key Points
- Deciles are a good tool for analysing historic price data.
- In Australian dollar terms they have little to no forecasting ability, flip a coin instead.
- The forecasting ability for the 19 MPG deciles improves when viewed in US dollar terms.
Click on figure to expand
Click on figure to expand
Click on figure to expand
Click on figure to expand
Data sources: AWEX, RBA, ICS, Mecardo
