Copper Inventories versus Supply/Demand Forecasts

Fundamental analysis of commodity markets is based on statistics for supply and consumption. Estimates of the balance between supply and demand are then used for price forecasts. The International Copper Study Group (ICSG) as well as many analysts and investment banks and independent research institutions forecast that the copper market will be in a supply surplus this year and that this surplus would increase next year. Thus, it is not surprising that copper traded down from almost $8,300 per ton to less than $6,700 per ton. The increasing Chinese GDP growth rate prevented a further slide of the copper price and led to a sideways trading range.

However, this approach has some major drawbacks. The first one is that statistics of historical and actual global supply and demand differ among the institutions providing official statistics. This makes forecasts difficult to compare. Furthermore, historical data is revised frequently for several years in the past. However, commodity markets react on data when it is released first. Therefore, a quantitative analysis of the relationship between commodity prices and the balance of supply and consumption could lead to false estimations. The frequent revisions are also a problem if supply and demand have to be estimated by methods of univariate time series analysis.

The second problem is the data compilation. Surveys among copper mining companies and official production statistics could be used for the supply side. However, the difficulty is to obtain reliable data for copper consumption. Thus, a proxy is used, which is called apparent consumption. The ISCG writes in its recent forecast for copper that “…, ICSG uses an apparent demand calculation for China, the leading global consumer of copper, accounting for about 40% of world demand. Apparent copper demand for China is based only on reported data (production + net trade +/- SHFE stock changes) and does not take into account changes in unreported stocks [State Reserve Bureau (SRB), producer, consumer and merchant/trader], which may be significant during periods of stocking or de-stocking and which could significantly alter supply-demand balances.”

The third disadvantage is that the data is not easily available for the broader public. The ICSG publishes monthly data by press releases, but time series of supply and consumption are only available for subscribers.  

Based on the expectation of higher production from new and existing mines, the ICSG predicts that the copper market will be in a surplus of 387,000 tons this year after a supply deficit of 421,000 tons last year. For 2014, the supply surplus is predicted to increase to 632,000 tons.

In our quantitative fair value model for the copper price, we have not included data for supply and consumption for two reasons. The first is already mentioned above, the frequent revisions make the regression coefficients less reliable. The second reason is that even monthly data is published with a delay.

But fortunately, there is other data available, which is a good proxy for the supply and demand situation and this data is published on a daily basis. The three major exchanges trading copper futures and forwards provide data on inventories held in licensed warehouses. Our analysis showed that there is a high correlation between the supply/demand balance and the development of the warehouse inventories.

Thus, the conclusion is that if there is an increasing oversupply of copper, then also the warehouse inventories as reported by the exchanges should show a rising trend. A supply deficit should be accompanied by falling inventories.

As the first chart shows, total inventories held in licensed warehouses of the three exchanges reached a high in the early second quarter, which was slightly exceeded at the end of June this year with 928,093 tons. However, since the start of the third quarter the warehouse inventories declined steadily and are down by around one third at 611,335 tons. At all three exchanges, the reported warehouse inventories declined. It is thus a broad based trend. This development does not indicate that the trend of increasing supply surplus continued.

More important for the price development of commodities are the free available inventories. Unfortunately, our data sources provide only figures for the LME. Here, the free inventories (on warrant) already peaked in the first quarter and then declined as the second chart shows.

The inventory data should send at least an alert that consumption of copper might be stronger than the ICSG and some analysts currently predict. Of course, copper prices do not only depend on inventor levels or changes. Other fundamental factors also play a role, especially the development of leading indicators for the Chinese economy. Nevertheless, the drop of warehouse inventories should provide some support for copper. If inventories decline further, copper might even rebound.