Making sense of the Comex...
EVERY FRIDAY the Commodity Futures Trading Commission releases data that enable analysts to 'take the pulse' of various commodity markets, writes Ben Traynor at BullionVault.
The Commitments of Traders (CoT) report gives the aggregate positions held by traders from the previous Tuesday, including the number of long contracts (that stand to benefit if prices rise) and short contracts (that benefit if they fall).
Included in the CoT is positioning in gold and silver futures and options on the New York Comex. A futures contract is a standardized agreement to buy or sell a particular commodity at a particular date in the future. On the Comex, each gold futures contract is for 100 troy ounces, while each silver contract is an agreement to buy or sell 5,000 ounces. A Comex option meanwhile gives its owner the right, but not the obligation, to buy or sell a futures contract.
The CoT breaks traders down into four categories:
- Swap Dealers
- Managed Money
- Other Reportables
Other smaller traders are also accounted for separately as 'Nonreportables'.
This CFTC document gives brief descriptions for the four categories above. In essence, the first, Producer/Merchant etc., is anyone who is in the relevant industry commercially and using the futures market to hedge the price of their inputs or outputs (e.g. mining companies, refiners, jewelry manufacturers in the case of gold).
A Swap Dealer (usually a division of a major bank, see here for a list) may be dealing in swaps with speculative counterparties or with industry clients looking to hedge; the swap dealer may then be using the futures market to hedge their own book.
Managed Money, as the name suggests, includes hedge funds and the like, while Other Reportables are traders large enough to report their positions but who are judged not to fit into any of the other three categories.
One closely-watched metric from the weekly CoT is the so-called speculative net long, which is calculated by taking the total number of open long contracts held by 'speculators' (we'll get to who they are in a moment) and subtracting the number of open short contracts.
The spec net long is viewed by many as a useful gauge of how bullish or bearish the market is. If the spec net long goes up, the implication is that speculators are growing more bullish. If it goes down, they're getting less so.
There is, however, a problem with the spec net long. There doesn't seem to be agreement on what exactly it is. Different analysts calculate it differently, depending on who they class as speculators and the types of contracts they look at.
Who are the speculators?
Until September 2009, the CoT used to break large traders down into two main camps: commercial and noncommercial (there was, as now, also a Nonreportables category to account for smaller players).
Commercials comprised the first two categories mentioned above, Producer/Merchant etc. and Swap Dealers. The noncommercials were regarded as speculative money, and hence it was their positioning that was used to calculate the spec net long.
Since September 2009, when the CFTC started publishing its disaggregated CoT, the picture has become more nuanced. Many analysts still lump Managed Money and Other Reportables together as the noncommercial, speculative end of the market. An example is Japanese trading house Mitsui, whose weekly report quotes the net long in broken down in terms of 'Large Specs' (i.e. Managed Money and Other reportable together) and 'Small Specs' (the Nonreportables), in terms of futures only, options, and the futures and options combined.
South Africa's Standard Bank also lumps both types of noncommercial player together, calculating a net long figure based on the aggregate futures and options positioning of Managed Money and Other Reportables together.
Others analysts prefer to look at Managed Money in isolation, viewing it as a purer measure of speculative sentiment. One example is Commerzbank, whose research notes quote the spec net long in terms of futures contracts held by Managed Money. Brokerage INTL FCStone also looks just at the managed Money category, but it differs from Commerzbank in that it uses futures and options combined to calculate the figures it quotes as the spec net long.
Which classifications of traders you count as speculative can make a difference. As an illustration, here's what happened in the week ended Tuesday 16 April, a week in which gold saw its steepest price drop in three decades:
Managed Money responded to the price drop by cutting aggregate short positions and increasing long ones. Other Reportables did the exact opposite (as did Nonreportables). And this was far from the only week when the two camps of traders moved in different directions. It seems the positioning decisions of Managed Money and Other Reportables are driven by different motivations.
There is no right or wrong answer when it comes to who to include as 'speculative' traders.In many cases it is a judgment call for the regulator, not an immutable fact.
An important point to stress is that it is the trader that is categorized, not trades themselves. This means that an entity classified as 'speculative' (Managed Money or, for some analysts, Other Reportable) may still hold positions that have a nonspeculative motive (e.g. hedging a swap position) and vice versa. Yet those positions will be counted towards the grand total of such positions held by speculators.
What contracts should be considered?
For each commodity market the CFTC publishes two versions of the CoT each Friday. One is based only on positioning in futures contracts while the other also includes options. Most weeks, both reports tell a similar story in terms of changes in the spec net long. But this is not always so.
The table below shows how the spec net long changed in the week ended April 16 2013, according to four different ways of calculating it:
By the end of Tuesday 16 April gold was down more than 10% from a week earlier. As you can see from the table, this price drop was met by a large increase in the futures only net long position of Managed Money, whose futures and options net long also increased, though less dramatically.
If you include Other Reportables, however, then for the week in question it makes a difference whether or not you include options. On a futures only basis the net long of all noncommercials went up, but if you include options it fell. Why?
Our guess is that this is in part explained by what's known as delta-weighting. When calculating traders' positions for its weekly report, the CFTC weights long and short option positions according to what's known as the 'delta', the sensitivity of the option's price to movements in the underlying, in this case the price of a gold futures contract. So if a $1 move in the price of the futures contract results in a $0.50 move in the price of the option, the option is said to have a delta factor of 0.5.
Let's say a trader holds 100 identical call options (the right to buy at a given strike price). He is considered to be long since he stands to benefit from a rising price in the underlying. The CFTC converts the trader's option position into a futures contract equivalent by multiplying the number of contracts by the delta factor. If the option has a delta of 0.5 then the trader's option position appears in the futures and options CoT report as a long position of 50 futures contracts.
By April 16, with gold having fallen so hard, many of the long options held a week earlier were now well out of the money. Those that had not been closed out will have seen their deltas fall dramatically. A call option to buy at $1800 an ounce isn't worth much once gold's fallen below $1350 – and it doesn't get much more valuable even if gold climbs $50 or $100 from there.
The same dynamic works the other way with short positions – any well out-of-the-money puts on Tuesday April 9 that were still open a week later will have seen their deltas rise.
This, we suspect, goes some way to explaining why the changes in futures and options spec net long cited above tell a different story from the futures only. On Tuesday 16 April, the biggest open interest for May call options was at the $1650 strike price, at 12,261 contracts. Yet these calls would have been given a lower weighting in the CoT for that day than they had been a week earlier.
Of course, market-driven changes to delta weighting are not the whole story – no doubt a lot of long option positions were closed and short ones opened when the price started to fall. This aspect of the CoT is however worth bearing in mind when analyzing the numbers, especially at times when the price has moved a long way.
Why are traders long or short?
When trying to make sense of Comex positioning, the fundamental question is why traders are long or short. Are they hedging a commercial activity? A position in another market? Or are they taking an out-and-out speculative bet on price direction?
The CoT does not provide enough information to answer these questions definitively. Classifying traders according to their typical trading activity, as the CFTC does, allows us to inferences, but always bear in mind that even professional analysts don't agree on who should be regarded as a speculator. Remember also that just because a trader may be classed in a 'speculative' category does not mean all their trading should be considered as such.
As we have seen, it can make a significant difference who you regard as a speculator, as well as the types of contracts you take into consideration. These decisions are judgment calls. Often it makes little difference to the overall story whether you look at futures only or futures and options, or whether or not you split out Managed Money. But occasionally it does – and as we saw in April 2013 this can be at times when sharp price moves have already created a confusing picture.