So, an outfit called Pandora's Box has jumped on the Muddy Waters bandwagon to hide behind the mask of anonymity and attempt to dent the credibility of Golar LNG (GLNG) and Golar LNG LP Partners (GMLP).
This outfit has (in our opinion) incorrectly asserted that Golar and its affiliated LP, Golar LNG Partners, are a "Ponzi scheme". Worse, they have insinuated that John Fredriksen and Tor Olav Troim have acted in less-than-appropriate behaviour. Their analysis (even though helped by the very good analysis of Pareto of which they had no permission to use) misses the mark on nearly every front. John Fredriksen and Tor Olav Troim are responsible for building companies which have created shareholder value of well over $20bn, not including the significant amount of dividends that have been paid out over time.
While we agree that currently Golar LNG is enjoying a premium valuation when compared to the market-adjusted NAV of its fleet (approx. $32/share in our view) - significantly overvalued it is not, especially in light of the potential for the fleet's NAV to grow significantly in the next 2-4 years. The two analyses used to value GMLP are even more abstract, and we will discuss this in detail below.
The gentlemen at Pandora's Box should sharpen their pencils and dig deeper into MLP accounting. Their explanation of how GMLP generates its distribution is somewhat misguided, as we will demonstrate. In addition, regarding the taxation of MLPs they fail to understand that MLPs are exempt from corporate taxation. GMLP is actually structured as a corporation, not a partnership, and its distributions are therefore subject to dividend taxation, not ordinary income taxation - as with most MLPs.
As a result, US investors will be subjected to a lower tax on distributions than for other MLPs and GMLP will not technically be affected by changes in the current MLP taxation legislation (which the report does not even refer to, making it obvious they are not US-centric investors), but of course until investors are reminded of this fact, GMLP may well trade with the MLP group in a negative fashion if such discussion is brought before Congress.
The crux of GLNG/GMLP's current business model is that it follows a very standardized "mother/daughter" GP/LP dropdown model (see TK/TGP, NS/ NSH, PAA/PXP, CHK/CHKM….the list could go on) whereby the Board receives a fairness opinion from an independent bank regarding the potential valuation of dropdowns as well having a conflicts committee to determine whether the valuation of the dropdown is fair to both parties. However, this is largely irrelevant given the transparency of the valuation metrics used to dropdown the valuation.
The Golar Freeze was recently dropped down for $330mm while on a multi-year contract with DUSUP that generates $39mm in annual vessel EBITDA. This results in valuation of 8.5x and was clearly stated in the press release relating to the dropdown. If you don't like the valuation, sell the shares….but this is a lot more transparency than is usually provided by the GP/MLP community and, in our opinion, represents a very fair multiple for a 10 year contract (with inflation adjusters) with a government agency.
Most importantly, their calculation of the "true" distribution of GMLP's distribution is completely inaccurate, uses double counting, and proves that the authors of this piece are not particularly savvy when it comes to MLP accounting: As can be seen on the table on the bottom of page 11 of their report, the authors begin with net income and from that deduct capital lease obligations and the repayment of long-term debt. Why is non-cash depreciation expense not added back to net income to arrive at potential DCF? Well, because the authors correctly state that depreciation is an expense that investors must pay for. And they do, as is seen in the table on the top of page 12, which is the company's calculation of DCF.
On a quarterly basis, GMLP deducts from EBITDA the line item referred to as Replacement Capital Expenditure Reserve, which represents the annual reserve necessary to replace the income-generating assets that are being used to generate the group EBITDA. This is calculated by assessing the replacement cost of the vessels in the group divided by the life of the vessels to arrive at an annual replacement reserve. Now, here comes the tricky part: So long as Replacement capex equates to annual debt amortization, GMLP has no problem. In an ideal world, these two line items will offset each other. But they may not, and this is where the discussion gets a little more technical.
Most MLPs access the capital markets to acquire incremental assets (yes! The whole point of an MLP is to grow one's income generating assets over time). They do this by issuing equity, bonds and having access to credit facilities. As we know, bonds are non-amortizing. However, shipping companies generally finance vessels through the banking market, which is amortizing. As a result, GMLP is unlike most other MLPs in that it carries a heavier annual amortization profile. The goal of the company over time would be to have a balance of debt which allows the amortization profile of the company to more equally reflect annual replacement capex.
Bank debt repayments and capital lease obligations totalled $51mm in 2011 (we have ignored the repayments to World Shipholding as repayments to John Fredriksen could theoretically be deferred without punitive measures) and the annualised Q4/2011 capital replacement reserve was $33mm ($8.1mm * 4). As a result, debt repayments were $18mm more than capital replacement reserves, hence the incremental revolver at the partnership level. Over time, we agree that GMLP may have to raise some additional non-amortizing debt to improve the balance of amortizing and non-amortizing debt to match these two line items appropriately. But it is important for the correct accounting to be illustrated and to point out that the shortfall of replacement capex is simply not significant, in our view.
To call this structure a Ponzi scheme is therefore absurd, and this label would have to be applied to the entire GP/MLP group. We highlight some further errors:
Valuation of Golar Freeze:
Pandora's Box believes that the cost of the Golar Freeze for GLNG was $180mm (most likely correct: $50mm for the vessel, $130mm for the conversion into an FSRU). Is this necessarily the value at which it should be dropped down to GMLP? Is it not safe to assume that the current 10yr contract with DUSUP that the Golar Freeze is currently operating under is worth more than the cost of the vessel in question? At a current dayrate of approximately $125,000/day, with opex of approximately of $16,000, this equates to annual vessel EBITDA of $38.7mm. Would it not be appropriate that the safety, consistency and longevity of this contract be worth an 11.8% earnings yield (inverse of 8.5x)?
We would assert that this contract could actually have a market value of a 10% yield (10x EBITDA multiple), but given the IDR structure in place, the maths stipulates that the valuation that GLNG drops a vessel down to GMLP is actually a wash. Had Pandora's Box modelled the IDR structure, they would have realised this. In our opinion, GLNG should get rewarded for winning the contract with DUSUP and for undertaking the lengthy and costly (approx. $130mm) conversion from an LNGC to an FSRU, and as a result it is appropriate that GLNG nets a profit from this transaction.
Valuation of GMLP:
A few comments on the DCF and Comp Analysis work on GMLP:
- We believe that the newbuilds and other vessels that exist at GLNG add significant future value to GMLP. Given that they will be dropped down to GMLP once they have been contracted in the next few years, they will significantly increment the GMLP distribution. Why else would GMLP's yield shrink from 7% to the current level of 5%....it is because of the anticipated growth of the annual distribution above and beyond what has already occurred from the annualised MQD of $1.54 to the current run rate of $1.72. We estimate that distributions could surpass $2.50 next year on their way to $3.00+.
- The authors should not have used the Teekay-Maersk transaction to value these vessels. Although not publicly disclosed, we believe that Maersk's fleet had contracts in the region of $55,000pd, which is totally different from the likely long-term rate of $90,000 that GLNG could get for its vessels now (no, no one expects that spot rates of $140,000 could be translated into L-T contracts). Once a 160,000cbm tri-fuel LNGC is contracted at $90,000, it can generate EBITDA of approximately $28mm. Recent transactions would suggest that a multiple of 10x would be appropriate for this type of contract.
- But let's use Pandora's Box's work: if they value the Maersk vessels at $215mm each, then that would imply an EBITDA multiple of 14.3x (215/((55,000-11,000)*355 days=15.6mm)). If we apply that multiple to our EBITDA estimate of $28mm that we believe the tri fuel moderns can conservatively earn, that leads to an implied EV per vessel of $400mm! Thank you Mr Anonymous, we'll take that one. That's one way to blow up a short thesis!
- TGP paid up for these vessels as they needed the contracted vessels to increment their distribution growth - simple as that. Golar were priced out of the negotiations as it is more value-creative to order a newbuild, get it contracted and keep the uplift for the company (sale price of $280mm - newbuild cost of $200 = $80mm profit per vessel).
- But this exercise misses the point! GMLP, like all other LPs is valued using a yield approach (not DCF, not NAV). In a 2% 10-year Treasury world, who wouldn't want a money-good, asset-backed piece of paper yielding 5% with a distribution that is growing at 12%-15% per annum?
It's too easy to look back in hindsight and suggest that the management team acted "inappropriately" in trying to develop alternative sources of value from vessels during the 2006-2009 timeframe when LNG dayrates were very low. But this is exactly what they should have been doing with E-Gas, Torp, Gladstone, etc. And the dollar amounts were tiny at the end of the day. In addition, KOGAS was an exceptional trade. The authors conveniently leave out the fact that GLNG sold the Golar Frost to the Livorno project for an EV of $231mm in 2008 when vessel values were much lower. An incredible deal.
This report goes on and on ad nauseum about a series of small projects which were executed in good faith by an ownership team which will go down in the record books as one of the greatest value creative teams in financial history.
Supply and Demand of the Fleet:
The report states that 15% of the current global fleet of LNG carriers is currently on order (62/364…it's actually 17%, in a report in which so many numbers are inaccurate, or simply wrong, what's one more?). Firstly, one should examine the uncontracted supply of newbuilds as a % of the total global fleet to calculate what is actually available. We estimate that approximately 17 of these 62 vessels on order are currently committed to contracts, leaving 45 uncommitted vessels, or 12% of the global fleet currently on order and available for work. But more importantly, has the author ever heard of the term "demand"?
The demand for LNGCs is likely to ramp up meaningfully over the next few years based on the number of liquefaction projects that are starting up all over the world. The world is currently net short LNGC tonnage - as is evidenced by the fact that dayrates have risen from approximately $19,000pd to the current level of $140,000pd - and we estimate that other than some short periods of a structural balance in the market, the 62 vessels on order will not make the global fleet loose. Of course, if the US exports 60mmtpa by 2016ish, and if a lot of those molecules are destined for the Far East, a further 120-180 vessels will be required due to the ton-mile effect, not to mention demand that will come from exporting terminals being planned in the Middle East and Australia, just to name a few.
We are long both GMLP and GLNG and believe significant upside potential exists for both securities. Core long reasons being:
·Leadership is provided by the John Fredriksen Group ("JFG"), which over time has demonstrated to be a world-class steward of capital as well as a best-in-class operator of energy-related maritime entities (SDRL, FRO, GOGL NO, SFL).
·The financial posture provides flexibility to execute the stated growth strategy. Importantly, very significant assets to be dropped down by the parent company are identified, in place and are owned by the parent company with all covenants in place to accommodate such a transfer.
·The core business of operating contracted LNG tonnage with the highest quality counterparties is a highly cash generative business. This core business combined with the MLP structure provides for a consistent stream of annual adjustments to earnings, and in the case of GMLP, cash distributions.
·Forward looking demand for LNG tonnage will be driven by:
oIncremental demand for natural gas as a major global energy source driven by the fact that it is:
- One of the cleanest sources of energy
- Significantly cheaper than oil
- Abundant in many regions of the world
- Safer than nuclear power and a chief beneficiary of a shift away from nuclear power