The United States Natural Gas Fund (UNG) exchange-traded continues to mystify, this time by cutting positions over the last few days just when you would expect it not t0: that is, the day that natural gas futures soared by 8.43 per cent (see below chart).

Since about May the fund has grown exponentially — from occupying some 40,000 natural-gas derivative contracts on May 20 to about 90,000 contracts just last week.
The fund has also become one of the most actively traded ETFs in the commodity sphere. This comes despite some erratic rather than inspiring price movements in the underlying contract, as can be seen below:

What’s more, all this growth has come amidst a contango in the natural gas curve, a market situation which sees the fund register regular losses due to negative rolls month on month, irrespective of price gains in the underlying asset.
All these developments, meanwhile, eerily echo those seen in the UNG’s sister fund the United States Oil ETF earlier this year — just before the SEC announced it would be investigating the fund’s influence on the market.
Mysterious cutback
Due to the UNG’s inflated size this time round, the market had been expecting some degree of disruption at its latest rollover from July to August contracts.
But, just like with the USO, an interesting phenomenon has occured. Rather than continuing to build positions, the UNG has suddenly began to decrease them by a sizeable level. What’s more, just like with the USO, the turnaround has come at a point when many thought the UNG had reached almost a critical mass and when most attention was upon it.
Olivier Jakob of Petromatrix comments as follows regarding its sudden and counter-intuitive position decrease (our emphasis):
NatGas stands out as the commodity that did not follow the trend and yesterday’s change in the UNG positions are intriguing but not in an obvious correlation to the price action. The UNG positions in NatGas Futures were cut sharply lower to move them very close to the Nymex accountability limit.
The Swaps positions were unchanged and the UNG which is supposed to be invested primarily in NatGas Futures is now invested up to 84% in Swaps.
We have pointed out before how it was only when positions in the USO started to be cut back that crude oil prices staged a rally. So does this UNG cutback mark a similar turning point for natural gas? What’s more if it does, why would it be that the ETF’s growing size was holding back the market in the first place? Jakob stresses:
Crude oil prices managed to rally only after the positions in the USO started to be cut and we will need to observe closely in coming days the evolution in the UNG positions.
And with regard to the cut in positions, Jakob states:
In a contango roll it is normal for the overall position to be cut but yesterday’s cut was far above that required by the rolling economics. It seems counter-intuitive to have prices rising when a long ETF stops buying but that is because the ETFs on commodity futures are market disruptive instruments and it is only when the ETF-fear subside that Future markets can go back to their normal function of pricing the balance between supply and demand.
We couldn’t agree more.
Charting your losses
The following chart attempts to explain why it is that positions in a contango market would naturally be cutback as the fund rolls from month to month. (For the sake of the explanation please presume natgas prices move as inferred by the futures curve.)

The chart presumes a $5 contango between June and August at the time of the roll (eg. August is $5 dearer than July).
If you started your investment with $1bn in July at $5 per mmBtu, that would equal a holding of 20,000 contracts (1 trading unit=10,000 mmBtu). To maintain your $1bn investment without seeing your contracts expire, you would be forced to buy August contracts at a $5 premium. Accordingly to maintain the $1bn investment you have to cut back your overall holdings by 10,000 contracts. If not, you have to pay up an additional $1bn investment just to retain your 20,000 contract position. Your loss either way is equal to 10,000 contracts on the day.
And it goes on, and on like this until the contango finally abates.
But there’s a further burden with funds. Fund methodology dictates that positions are rolled days ahead of expiry. Yet as contracts approach expiry they naturally converge in price on account of market arbitrage forces. The funds, however, miss out on this convergence.
A professional investor (especially one who can play both the physical and paper market) is always going to to be able to “track” the front-month price of natgas (as depicted by the blue line in the example) better than a fund because of this reason. In our extreme case the shaded red area shows just how much a fund stands to lose due to this methodology.
Of course funds do try to avoid this by rolling their positions over a number of days, easing out the differentials. However, while this might limit some losses, it doesn’t come close to eliminating them all.
A counter-intuitive UNG, or is it?
When we look at the UNG we realise the fund wasn’t just trying to maintain its initial investment, however. UNG was actually adding positions at a furious rate.
Remember, one ETF unit is supposed to reflect one investment in the underlying market. But as an ETF is an open ended instrument, units can fluctuate according to their own supply and demand fundamentals irrespective of the movements in natural gas. To prevent units from falling out of sync too much, however, ETF managers depend on intermediary players who are authorised to buy or redeem units from the provider on a daily basis at prices equal to the net asset value of the fund (eg. total value of investments divisable by units issued).
The incentive to play this arbitrage keeps the ETF unit prices in check.
If the provider feels demand for units is soaring, he can always apply for permission from the SEC to issue further tranches — something the UNG has been doing in the last month.
Tying it all together
If you consider our above example, the UNG’s NAV would naturally have been falling in value due to the contango. According to the example above if 20,000 units equaled $1bn in July, come August 20,000 units would have been worth $500m at the NAV price. At such a discount, those intermediary authorised participants would have piled in for new units to bridge the arbitrage opportunity. This might account for the unusually large deviations in the NAV/unit prices of the UNG in the past month.
It might also explain the furious rate of the UNG’s growth in the last month just as the contango has been getting wider.
Which leads us to the following story from Dow Jones on Monday (our emphasis):
NEW YORK (Dow Jones)–One of the hottest investments on Wall Street may have gotten too big for its own good. With investors betting on rising gas prices, assets in the United States Natural Gas Fund (UNG) recently swelled to almost $3.7 billion from about $670 million in February, even sparking fears it could be disrupting the futures market.
Now the popular exchange-traded fund is days away from another potential problem: Funds that hold commodities typically face stiff restrictions on the number of shares they can issue to meet investor demand, and United States Natural Gas is running out fast.
Securities and Exchange Commission filings show managers want to increase the number of shares available nearly 10-fold. But such requests can take weeks and there is no telling when the SEC will act. If the fund can’t issue enough shares to meet investor demand, its shares could begin trading at prices higher than the underlying value of their holdings, breaking a key promise ETFs make to investors and possibly influencing prices in the natural-gas futures markets.
Which could neatly explain the drastic and sudden sell-off in UNG positions referred to by Jakob. If the UNG managers can’t issue more units to keep the price/NAV together, they have no alternative but to sell positions to compensate for the contango losses instead.
The UNG simply must be liquidating positions to maintain the NAV valuations, as in our example above.
What’s more, one can presume a similar thing must have happened with the USO, especially after the SEC cottoned on to the troubles its increasing size was causing in the market. It’s unlikely, after all, that the SEC would have sanctioned any further unit issues.
Related links:
The problem with commodity ETFs - FT Alphaville
Strange things still afoot in natural gas - FT Alphaville
Commodity ETF investors move significantly into natural gas - FT Alphaville
Super-natural gas - FT Alphaville
Anything but therm in the US - FT Alphaville
A self propelled pyramid? -FT Alphaville
How contango affects oil ETFs - FT Alphaville