A Member of the Metallicarum International Alliance
BME's mathematical modelling of copper prices
"For subscribers to BME's reports who need to conduct scenario analysis or run Monte Carlo simulations, BME's Copper Price Model is the ideal tool to help. It is fully up-to-date in assessing the interaction between traditional price drivers - such as stock levels, the rate of demand growth and exchange rates - and newly important financial drivers - particularly commodity index funds' long positions in the futures market and hedge funds' net long or short positions. The model is already set up to include the effect of possible future ETF buying of stocks whether on or off the exchanges. The model runs to end-2020 using quarterly figures and can also help focus long term price risk assessments on crucial factors outside the traditional physical market parameters. For instance, it can allow users to explore how a combination of ETF launches and SRB buying might shorten the next copper price trough. Or longer term still it could help companies to assess 'doomsday scenarios', such as financial market dis-investment in copper, combined with excess physical stocks and excess capacity: perhaps leaving prices well down in the production cost structure for a decade." "Many people and companies are uneasy with mathematical price modelling, thinking of it as being rarified and far removed from their own actual experiences of markets. It need not be so. There are various ways of going about mathematical price modelling. BME's approach is not to begin with the mathematics. We begin by using market participants' and analysts' ideas on what drives prices. We then go on to use mathematics to check whether these ideas stand up to testing, and then to quantify the relationships if they prove valid, to test rival theories, and to spot very quickly when market relationships change (new error patterns emerge). To give an example of mathematical testing of rival theories, between 2000 and mid-2005, there were two opposed schools of thought on the main price driver. Single-commodity specialists tended to believe that production-consumption differences and stock trends for 'their' commodity drove its prices. Multi-commodity specialists tended to believe that turning points in the industrial production (IP) growth cycle determined the turning point in the price cycles of all industrial raw materials. When these theories were tested mathematically, some metals (copper and tin) turned out to have price cycles closely related to their stock cycles, whilst others' (especially aluminium) prices showed almost no link to their stock cycles but strong links to the IP cycle. BME went on to test the IP cycle and the stock cycle as joint drivers of price, with dramatic results: combined, these gave much stronger fits to the price cycle. Restated in 'human brain' terms, low stocks and rapid IP growth gave a price spike, but neither low stocks nor rapid IP growth in isolation did so. Easy! But analysts had missed it for years, while a mathematical model sorted it out very quickly. Using simple three-driver models - stocks, IP growth, and strength or weakness of the dollar - BME developed very powerful predictive models of price that worked very well until mid-2005. From that point on, all of our base metals price models began to go wrong the same way, one after the other. All of the base metals' prices were taking off, having lost their traditional relationships to stocks and IP growth cycles. BME was very quick to establish that a drying up of producer hedge shorts and a surge in index fund buying of long positions in the nearby futures market were the twin causes. In late 2005 and early 2006, the nearby futures market was dragging up cash prices. BME quickly adapted its models to the influence of investments in commodities futures. That provides two illustrations of the common-sense usefulness of modelling in general: testing rival theories and adapting quickly when market behaviour changes. As for our specific service today, BME's Mathematical Copper Price Model is designed to fit perfectly with the traditional physical market data set out in such detail in the Quarterly Report on Copper, including its full analysis of stocks. But it also accommodates the price effects in the nearby futures market of commodity index fund long positions and the hedge short positions of exchange stock owners in contango markets. It accommodates hedge fund net long or short positions. And it is fully prepared in advance to deal with ETF stock buying either off or on the exchanges, together with the un-hedging that would accompany purchases of exchange stocks in a contango market. Whether you are or are not already a subscriber to BME's Quarterly Report on Copper, the Model can help you in general forecasting, asset valuation, scenario analysis and risk appraisal. To learn more, please contact us to arrange a demonstration."