Price multiples

Having investigated general boom and bust characteristics of commodity cycles, it is now time to investigate the prevalent price increases that occur in the course of going from a cycle minimum to their subsequent peak. When an-alyzing these price multiples as defined in eq.

(7.7), some strong differences emerge again be-tween CW and CCW cycle movements. While it was already observed earlier that CCW cy-cles are about 60% shorter than CW cycy-cles (cf.

section 7.2.2), they also have much lower price multiples than CW cycles (cf. figure 7.26). In addition we can note that the population of CCW data points is found in a narrow band along the cycle length axis with a price multi-ple Φp of ≈1.6. It therefore appears that no matter how long a CCW cycle lasts, the price multiples change only slightly¹².

Figure 7.26: Relationship between cycle lengths and price multiples

CCW lengths and price multiples are thereby with 0.48 moderately and significantly correlated (cf. table 7.12).

The CW population is much more dispersed in

12The extreme value of Φp = 4.53 is to be taken with caution, since data exists only for one CCW cycle of Olympic Dam

terms of cycle length and price multiples. The correlation of the latter parameters is thereby much stronger and more significant than that of the CCW motion.

When looking at it from a perspective on how price multiples of the two movements compare to each other directly, we can observe yet again distinctive patterns (cf. figure 7.27).

T_bb vs. Φp

CW CCW

t 5.6311 3.1074

p-value 3.17E-06 0.0039 correlation 0.7055 0.4815

Table 7.12: Pearson product-moment correla-tion results for Tbbvs. Φp

To start with, we can see that the majority of metals exhibit a Φppratio of greater than one (i.e. most of the data points right of iso-ratio line Φpp=1). As before in 7.2.2, we can observe a clear separation of metal zones that corre-spond to A - base metals, B - precious metals, C - minor metals and D - energy metals. We now describe these groups more thoroughly.

A - Base metals

The boomerang shaped base metals zone A is closest to the Φpp = 1. We can discern three metal subgroups with their related ore types. These are copper (a2) to the upper left, followed next to the lower right by zinc (a1), and with furthest to the lower right, nickel (a3).

Figure 7.27: Price multiple relationship of CW and CCW cycle motions

The copper ore zone a2 show a clear lin-eation that is the most parallel to Φpp=1. Most of the aligned assets maintain a Φpp between 1.27 and 1.32, while we can clearly separate different ore types the further we move away from the coordinate origin. The lower left of the copper subgroup contains gold bearing copper porphyries such as POR.OYT. Further to the upper right, ores follow where silver becomes more important. POR.GRS and IOCG.GUE ores follow with the copper-silver RED.KUP ores further to the upper right. The pure

metal copper is next in line and at the cen-ter of the subzone with a Φ^CW_p of 2.89 and Φ^CCW_p of 2.19. Further to the upper right, cop-per deposits follow that exhibit an increasing amount of molybdenum. These deposits are all porphyries with the Cu-Mo POR.HVM at the upmost left and poly-metallic POR.BIM and POR.BAG ores sitting closer to copper.

This alignment suggests that molybdenum has an important influence on the potential price multiple within the copper deposit group. This effect holds, no matter what cycle motion

pre-vails. On the other hand, contained silver and gold seem to diminish the price multiple up-side.

There are two exceptions to this copper con-taining group. The first is once again the Olympic Dam deposit located at a^′₂, being the only ore type that sits to the left of the Φpp = 1 line. As mentioned earlier, this can be the re-sult of the only one existing CCW cycle in the data set (cf. table A.25). It must be recognized however that its unique geochemical composi-tion of copper, gold, silver and especially ura-nium, might also play into this unique position.

The other exception is the copper-cobalt RED.TNK ore that is located in the nickel bear-ing zone a3. The location of it might thereby be related to the fact that cobalt exhibits a very high CW price multiple, and is ”pulling”

this copper ore closer to itself. This is because cobalt can act as a by- or co-product depending on the pricing regime.

The central band of the boomerang shaped base metals zone is occupied by zinc and lead with their related deposits (a1). As with the deposits in zone a2, an alignment of deposits and met-als exist. This alignment however is less steep than that for the copper deposits. Discounting the presence for tin we can clearly see that zinc containing poly-metallic deposits sit on the left side of this subzone while the constituting met-als sit to their right. As we have already seen in section 4.2.5, the deposits with the highest number of contained metals sit to the left with some of the lowest price multiples (VMS.KID Φ^CW_p = 2.4 and Φ^CCW_p = 1.81). The fewer by-products the ores thereby contain, the higher the price multiples are, and the further to the right they sit. This becomes apparent when looking at MVT.TRI ores (Φ^CW_p = 3.13 and Φ^CCW_p = 1.84) in the middle of this alignment, close to zinc, with lead at its right hand side (Φ^CW_p = 3.73 and Φ^CCW_p = 1.88). The implicit increase of variance that is going along with this transition from left to right is yet again an indirect sign that MPT rules apply.

Discounting aluminum and the RED.TNK de-posit group that was discussed earlier, the lower right subzone a3 of the base metal area A is dominated by nickel and its associated ore types. A linear alignment of the deposits within this group is less pronounced than for the other metals. The nickel group shows lower Φ^CCW_p values than both the copper and zinc zones, but exhibits a higher range for its CW price

multiples (cf. table A.25). As in the previ-ous subzones, we can distinguish the different ore types according to their by-product com-position. The three different metals that seem to mostly influence the location of those de-posits relative to the main product nickel are gold, cobalt and platinum. The presence of gold seems to thereby push the deposit towards lower price multiples, while platinum does the opposite. Cobalt does not seem to have a heightened influence such that these deposits more or less plot around nickel, that sits at the center of this subzone. This might be a result of the comparable low value of cobalt in the total ore assemblage of the respective ores.

B - Precious metals

To the right of the base metals zone in fig-ure 7.27 sits the precious metals zone. While its CCW multiples are comparable to those of the base metals, their CW multiples are clearly elevated. They range from 5.16 to 8.27 and do present quite a phenomenal price upside for poly-metallic deposits (cf. table A.25) over base metals. In addition, gold poses a note-worthy exception of this group, showing a CW multiplier of 13.1. To recap, this implies that in clockwise cycles gold has the potential of a 13 fold price increase, while a 1.5 fold increase is more likely during a CCW cycle. Knowing which cycle motion gold is in becomes there-fore very important for the savvy miner and investor.

C - Minor metals

The two minor metals cobalt and molybdenum show some of the highest CW multipliers in the data set, while their Φ^CCW_p s are more in line with the rest of the groups (cf. table A.25, pg. 202). It is interesting to note that the asso-ciated poly-metallic deposit groups RED.TNK (Cu-Co) and LAT.MOA (Ni-Co) plot closer to their main metals copper and nickel, respec-tively. This might be related to the very fact that in both cases, cobalt is a by-product that helps the economics of those ore types, but does not determine their economic viability as main products do. As we have seen with molybde-num related deposits earlier, the mere presence of molybdenum in a poly-metallic porphyry de-posit seems to increase not only the price multi-ples of their CCW cycles, but in fact their CW cycles, as well.

D - Energy metals

Uranium exhibits with 13.2 the highest CW multiplier of the data set (cf. table A.25). This, with a slightly above average CCW price mul-tiplier of 1.74 (vs. 1.6 avg.) makes it one of the most exceptional metals. This fact might be related to the fact that was discussed ear-lier already, namely the long lead times for its main end users, nuclear power plants. When supply is insufficient for the apparent demand and prices increase dramatically due to it. This is even more so the case for uranium than for gold, since uranium is not as easily stored and transferred as gold is from bank vaults.

Summary & Discussion

The analysis of price multiples revealed the ex-istence of distinct metal group differences be-tween clockwise and counterclockwise cycles.

While clockwise cycles exhibit multiples that range from 2.39 to 13.2, counterclockwise ones only range from 1.2 to 4.53. Compared to base metals, the groups of precious metals, energy metals and minor metals exhibit higher Φps during clockwise cycles than counterclockwise cycles.

The question arises what the root causes for these are. We suggest that similar to earlier finding their respective metal end usage might play a role in this.

However, there might be other, but related rea-sons for these distinctions, as well. They re-volve around their respective abundance in the Earth’s crust and the ability to extract them economically. The latter feature is thereby closely intertwined with the former, since it is generally more economically to extract some-thing that is more common, than what is not.

Hence we produce and use more of what is more present in Earth’s crust. We can illustrate this in figure 7.28, using 2017 production figures by the USGS [33] and crustal composition data of planet Earth¹³ [34]. As we can see, a relation-ship seems to exist between industrial metal production volumes and natural abundance on planet Earth. When we perform a Kendall rank correlation test on this data we can show that indeed a very significant positive relationship exists between the two variables (p= 0.006 <

α= 0.05 cf. table 7.13).

Figure 7.28: Metal abundance vs. production Taking this understanding further and com-paring above production values with the CW price multipliers per metal, it becomes obvious that a relationship between the two exists, also.

parameter value

Table 7.13: Kendall rank correlation results for metal abundance vs. production 2017

As we can see in figure 7.29, it appears that the rarer a metal is, and the less we produce of it, the more prone it is to price spikes signified by higher price multipliers.

Figure 7.29: Metal CW price multiples vs.

world production 2017

We can test how significant this relation-ship is in performing another rank correlation test on the two variables. The results shown in table 7.14 show that such a relationship is indeed significant (p= 0.0049 < α= 0.05).

13Would human societies have evolved on a different planet with a different crustal composition, this general relationship would probably still hold and the proportion of production values could be quite different

parameter value

Table 7.14: Kendall rank correlation results for CW price multiplier vs. production volume

However, price spikes do not happen sim-ply for the lack of abundant resources, but are based on underlying economic fundamentals.

Price spikes usually do happen when there is a sudden shift either in supply or demand. While supply shifts causing price spikes are common in the metals industry (e.g. mine shut downs due to accidents or labor strikes), they are usu-ally only short term in nature (i.e. usually weeks to a few months).

Abundance Supply Industry

Table 7.15: Metal abundance, production and industrial usage

Since this study is looking at long term pricing spike effects, it is the shifts in demand that have a much more profound and long last-ing influence that interest us here. Sustained price spikes such as these usually have a more substantive change in outlook for a particular commodity as a reason. A sequence of such events can be visualized in a series of price vs.

quantity diagrams as shown in figure 7.30.

Figure 7.30: Conceptual representation of a price spike mechanism

Before a sustained price spike happens the market is in balance (cf. chart I in figure 7.30).

The demand curve intersects with the supply curve and the marginal cost of the highest cost producer is setting the price p1 with quantity q1 being delivered to the market. When de-mand increases as shown in chart II, suppliers initially struggle to provide additional quanti-ties of material, since the supply curve is very inelastic. The reason for this is that because miners run their operations at designed capac-ity, very little can be done to satisfy this extra demand¹⁴.

The new equilibrium of the market settles therefore not where the supply curve S1meets the new demand curve D2, but rather on the extended vertical line of q₁. This implies that while the same quantity is provided to the mar-ket, prices move from p1to p2, implying a price spike. When miners finally find the capital and either expand the capacity of current op-erations or build new mines, the supply curve shifts outward (S2). This process usually goes along with a drop in prices from p2 to p3 (cf.

chart III in figure 7.30). The market finds a new balance where quantity q2 is supplied to the market at price p3 (chart IV).

This supply-demand scenario however plays out differently for different metals, and depends on their relative size of production and by that reasoning, relative abundance in Earth’s crust.

14We forgo any discussion on operational improvement, operations that can run above specified capacity, or high-grading to increase productivity and metal supply to the market.

Because of this, it appears that the rarer a metal is, the more difficult it is to add extra capacity to satisfy increased demand and there-fore observed price spikes are higher.