Cheniere Energy Partners (CQP), majority-owned by Cheniere Energy (LNG), will soon, briefly monopolize one of the prime bottlenecks of global energy: Gulf Coast LNG export capacity via its Sabine Pass facility. LNG export capacity is a call option on a locational price spread, and right now the spread between Henry Hub and the East Asian bid is very wide. After a string of long-term deals, the future capacity from Sabine Pass is largely subscribed. Should Cheniere Energy maximize this opportunity by selling all of its optionality to third parties, or should it keep more of the option open for its marketing and trading unit, CMI? This decision has big implications for unit holders of CQP. After modeling the uncertainty of future price spreads, we will see that the terms of the deal to transfer capacity from CQP to CMI should make unit holders of CQP wary. They should rebalance to hold more shares of the parent, Cheniere Energy.
The Sabine Pass facility in Louisiana was originally built to receive LNG imports at a time when global gas fundamentals were markedly different. Lately, it's being fitted with the capex-intensive condensers needed to liquefy and export gas. The site can accommodate up to six trains with 4.5 mtpa capacity (231 Tbtu/y) each, or 64% of Qatargas's current world-leading capacity. The capacity from the first four trains is 90% sold via 20 year take-or-pay deals to external shippers BG, Gas Natural Fenosa, KOGAS, and GAIL. All the financing and DOE export authorizations have been secured for these first four, and construction has commenced. Here's a tentative timeline:
- Q3 2012: commenced construction (trains one and two)
- May 2013: commenced construction (trains three and four)
- September 2013: filed FERC application (trains five and six)
- Q4 2015: first production from train one
- Mid-2016: train two comes online
- 2017: first operations (trains three and four)
- 2018: Train five starts
- 2019: Train six starts
The last two trains still lack financing, and only 40% of the capacity has been inked into long-term deals. FERC hasn't yet allowed export to non-free trade agreement countries from these trains, only FTA countries, which comprise some countries in North and Latin America, Israel, Jordan, and a few others -- in short, not the major importers you think about for LNG (though shippers could make a pit-stop to unload and reload in Mexico or Colombia on their way to Japan. Someone could make a tidy profit on this FTA/non-FTA conversion).
The Deal With CMI
Cheniere Marketing (CMI), wholly owned by Cheniere Energy, will be staffed to trade and ship LNG with several leased vessels, sourcing likely all of its LNG from CQP's facility. In fact, in Cheniere Energy's October 2013 Corporate Presentation (slide 15) we see that up to 104 Tbtu per year from the first four trains has been designated for CMI at a price of $3/MMBtu for the first 36 Tbtu and 20% of the annual profit for the extra volume. Unlike the other customers, CMI isn't locked into a take-or-pay with its volume. If the arb closes, CMI will not pay $3/MMBtu to CQP. CMI has a free option, and this is the point.
The Calculation and Its Big Assumptions
I set up discounted cash-flow valuations for CQP and CMI with a 25-year horizon followed by one discounted perpetuity cash flow. Besides the expected revenues from the long-term capacity sales and anticipated revenues from CMI, this DCF includes the revenue for the remaining 15 years of gasification capacity ($248M/y), the revenue from the Creole Trail Pipeline ($80M/y), opex, the payments for the current financing deals, and the projected payments for the financing of trains five and six.
The first big assumption is that the current valuation of CQP is correct. At the time of writing, the market cap of CQP is $9.4bn. This allows me to calibrate my factors, which is important for the next step. Since we're discussing the tradeoff of optionality between CQP and CMI, we need to see how the DCFs change when prices are volatile and random. To do this, I simulated 10,000 different 25-year price paths for Henry Hub and a generic international LNG price (which grossly oversimplifies the LNG market, but it represents the max bid out there, currently held by Japanese buyers). I assume, and this is important, that Hub price will slowly revert to within $2/MMBtu of the international spot price after 20 years on average. Slow reversion seems entirely appropriate considering that 1) the U.S. has ample supply of unconventional plays to bring online, 2) at a certain Henry Hub level, we'll see gas-to-coal switching in power generation, but 3) arbitrages find a way to close.
Each simulation produces a distribution of DCFs. I want to compare the distributions for four cases: four different amounts of optionality set aside "freely" to CMI:
- Case 1: CQP fully subscribes Trains 5 & 6 and also completely finishes selling the capacity from trains three and four. This leaves CMI with just 61 Tbtu/y with which to trade.
- Case 2: CMI keeps the current allotment of 104 Tbtu/y. The rest of the capacity of trains five and six is sold. This is my base case.
- Case 3: CQP is able to sell 183 Tbtu of train six, leaving 173 Tbtu in total for CMI. (We'll assume that the tranche 1 volume in the CMI deal increases from 36 to 54 Tbtu.)
- Case 4: CQP is only able to sell 91 Tbtu of train six (the same size as the Centrica deal for train five), leaving 265 Tbtu for CMI.
The chart below shows that the theoretical value of CQP decreases when it sells less of its capacity for a fixed option premium. Whatever optionality is not sold externally lands in CMI's plate and does not sufficiently benefit unit holders of CQP. Recall that Case 2 in red is our base case, calibrated to the current market cap.
Click to enlarge images.
When capacity is left unsubscribed and the arb closes, no one in the Cheniere family benefits. When capacity is left unsubscribed and the arb stays open or widens, CQP does not sufficiently benefit from 20% of profit sharing.
The next chart should come as no surprise:
More free option = more money for CMI.
Lastly, let's see how this adds up for the parent, Cheniere Energy, which accrues 55% of CQP's DCF and 100% of CMI's DCF. (Note that this is only the DCF coming from the Sabine Pass import/export facility; it excludes any value from Corpus Christi.)
Now that's interesting. For shareholders of Cheniere Energy, there is very little incentive to intentionally set aside export capacity for the trading unit. On average they would be almost indifferent. In fact, if we sought to maximize risk-adjusted returns using the Sharpe Ratio (expected returns/standard deviation of returns), we could argue that Case 1 is optimal to Case 4. Another take on the chart above is that Cheniere Energy shareholders appear to be self-hedged against undersubscription at Sabine Pass.
Unit holders of CQP should be very keen to see that as much capacity as possible is sold externally rather than freely (i.e., without a long-term take-or-pay obligation) set aside to Cheniere Marketing. LNG shareholders, on the other hand, can sleep easy knowing that whatever capacity is left unsold -- either intentionally or unintentionally -- can still be monetized by its trading company. This risk factor alone should convince buyers of the Cheniere story, like myself, to re-weight in favor of CQP's majority owner, LNG.