("Delta Refines A Fallacy," published at 5:13 a.m. EDT on Tuesday, misstated the time frame that refiners have had about a 25% return on capital in the 10th paragraph. The corrected version follows:)
By Gregory J. Millman
A DOW JONES COLUMN
NEW YORK (Dow Jones Banking Intelligence)--Note to investment bankers: start working on your pitchbook for Delta Air Lines Inc. (DAL) to divest its refinery operations. Before long, that pitch will have plenty of support from Delta shareholders.
Delta Airlines Inc. decided it could minimize jet fuel costs by investing in a failed refinery, so on April 30 announced the purchase (through a newly formed, wholly owned subsidiary called Monroe Energy LLC) of the Phillips 66 Trainer (PSX) refinery complex south of Philadelphia. The Commonwealth of Pennsylvania provided $30 million in assistance, so Delta's out-of-pocket investment to acquire the refinery is $150 million, and it plans to spend another $100 million to maximize jet fuel production there. The $250 million that Delta says it is investing is 19% of its last 12 months earnings (and 29% of 2011's bottom line) so it's a material number for shareholders.
There are three problems here. First, Delta won't manage the capital invested in the refinery for optimal return. Second, even if Delta were to do so, the refinery is one of the least economical in the U.S. so that its potential returns are low at best. Third, the refinery will probably require much more capital than Delta's public presentations suggest.
Once shareholders see the consequences of this deal unfold, they'll demand that Delta divest the refinery operations and redeploy the capital in more productive investments-- or give it back to shareholders so they can redeploy it themselves.
Let's take return on capital first. Refiners usually try to optimize returns by producing more of the more profitable products and less of the less profitable. In the case of refiners, this is difficult because crack spreads, the most important part of their gross margin, are volatile. The crack spread is the difference between the price of crude and the price of such refined products as gasoline, jet fuel, diesel, heating oil, etc.
The product prices vary seasonally and over time, so refiners try to adjust production accordingly, for example by making more gasoline during the summer driving season and more heating oil during the winter.
There are limits to how much production can vary, though, the most important factor being the crude oil the refinery can process. Different products boil off of crudes at different temperatures, and various crudes have various compositions of those products. Some refineries, such as Trainer, can only process light, low-sulfur crudes; some can use heavier, cheaper crudes from Canadian tar sands. Moreover, some refineries have access to high quality crudes from the Mid Continent, which are cheap because there are no pipelines to carry them to world markets, and therefore they are in superabundant supply. Since the product price is a world market price, and refineries have access to export markets, those that can use cheaper crudes enjoy fatter margins and are more profitable than those who cannot.
Delta's plans to boost jet fuel production to over 30% of its slate could lock it in to an even more difficult cost structure, since refining that much jet fuel without significant new capital investment will mandate the use of crudes with relatively higher compositions of middle distillates. The airline's own demand for crudes specifically in this range could move the price up.
Delta won't try to vary Trainer's production slate in the same way that other refineries do, and so far it has announced no plans to make the hefty capital investments necessary to process cheaper, heavier crudes. What it has stated, unambiguously and explicitly, is that it will manage the refinery to maximize jet fuel production. In fact, Delta will devote a higher proportion--32% --of Trainer's production slate to jet fuel than any other refinery in the U.S., more than quadruple the jet fuel proportion of other East Coast refineries.
Delta's planned production slate would only optimize return on capital if jet fuel always provided a better return than any other product. But if that were the case, more refineries would be optimizing their returns on capital by producing a lot more jet fuel than they do. Delta will not be optimizing returns on capital, but instead will be optimizing jet fuel availability for airline operations.
Why is this a problem? Optimizing for refinery returns is better for shareholders than optimizing for airline returns. U.S. refiners produced a return on capital of about 25% over the last 12 months, according to S&P Capital IQ, while U.S. commercial airlines earned only a 11% return on capital. (Delta, by the way, produced a 12% return on capital.)
That brings us to the economics of the Trainer refinery. Delta won't improve its own returns on capital by owning Trainer, notwithstanding Chief Executive Richard Anderson's statement that Delta can create value "by capturing the jet crack spread margin that the refiners produce." The plant will have to refine the entire barrel of oil, and even under its ambitious targets for jet fuel production, 70% of a barrel of oil will be refined into non-jet fuel products.
Delta says it will swap the non-jet fuel production for jet fuel from its counterparties BP PLC (BP.LN) and Phillips 66, according to a price index. Delta has not revealed any details about the specifics of the swap, but essentially Delta will be using its gasoline, diesel and other production as currency to buy jet fuel from these counterparties. We have to suppose that prices will approximate the prevailing market prices, since BP and Phillips 66 are unlikely to sell jet fuel to Delta for less than they could get from another customer. (Even if Phillips 66 included a "sweetener" in its exchange ratio to induce Delta to buy Trainer, it hasn't been made public, and would probably be temporary.) So if jet fuel is selling for $135 a barrel, and gasoline for $130 a barrel, Delta will probably swap at a ratio of about 1.04 barrels of gasoline per barrel of jet fuel. This means that Delta will still be paying the jet fuel crack spread - it will just be using gasoline or other refined products to make the payment.
And Delta will pay dearly to create this "swap currency" of refined products. Trainer, originally built in 1925, is one of the highest-cost refineries in the U.S. That's why two previous owners have shut it down rather than make new capital investments. Technology and geography combine to make Trainer uneconomical: the technology is not complex enough to process lower quality crude oils, and the geography prevents access to cheaper high-quality crudes from the mid-Continent of the North Amerca. Those crudes are cheap precisely because there are transportation bottlenecks that prevent them from getting out. When pipelines are built, they will sell at world market prices. Eventually, shale oil may be available from the Marcellus and Utica shales, but those regions are richer in gas than oil. Transportation is also a problem there, as pipelines have not yet been built to take the newly discovered oil to Philadelphia. Marathon Oil Corp. (MRO) for example, has built truck terminals to haul its Utica shale oil to barges, then to Kentucky for refining.
East coast refineries like Trainer must rely on Brent crude, which has been running about $116 per barrel lately. Add to that cost about $2.00 a barrel to transport the crude to the refinery. Trainer's production slate would give it about $125 per barrel in revenue, for a gross margin of about $7.00 per barrel. But fixed and variable refinery operating costs take about $5.25 per barrel, leaving a net cash margin of only $1.75 per barrel. Moreover, all refineries have to make sustaining capital investments to address ¬environmental and other issues. These run from $1.00-2.00 per barrel for a refinery on the Trainer model. So there's no excess return on capital left. It's easy to see why ConocoPhillips (COP) shut Trainer down. But ConocoPhillips Chief Financial Officer Jeff Sheets made it explicit in the company's third-quarter earnings call last year. "The Trainer Refinery was not producing net income," he explained. "The cash generation was also not very strong. We were at a point where we were having to decide about timing on turnaround costs and future capital expenditures, which had a lot to do with the timing of the decision to shut it down in October."
Finally, Delta shareholders should expect Trainer to be a capital sink requiring more investment than Delta has indicated in its public presentations. Delta does seem to have included in its presentation the "turnaround" investment that ConocoPhillips was unwilling to make. Such "turnarounds" are periodically necessary to refurbish and revamp refineries. Tosco, which re-opened the refinery in 1996 after a shutdown by BP, its previous owner, invested $60 million in a "turnaround". Inflation-adjusted, the $60 million Tosco spent on a turnaround in 1996 would come to about $89 million today. Add $15 million to $20 million for a new catalyst that will probably be necessary to support the higher volumes of jet fuel production planned by Delta, and you have a capital investment only slightly above the $100 million that Delta has announced it will be investing in Trainer beyond the purchase price.
However, refinery experts who spoke to DJBI all pointed to one missing piece of the picture: working capital. A refinery of Trainer's scale, about 185,000 barrels per day, would typically require $100 million to $200 million of working capital. Supporting this estimate is the fact that, after shutting down Trainer, ConocoPhillips told analysts that its refining margins had benefited from the liquidation of $180 million in inventory, mainly from Trainer.
Where is the working capital in Delta's outline of the Trainer adventure? A spokesman for Delta told DJBI: "We anticipate that all of our working capital requirements will be satisfied through payment terms with suppliers." BP will be sourcing and delivering Delta's crude, and Delta has not revealed the details of that contract. However, unless BP has failed to sharpen its pencil on this deal, the payment terms referenced will surely compensate it for financing Delta's inventory. So it's reasonable to expect that the present value of those payment terms will approximate $100 million to $200 million. Using the $180 million inventory figure from ConocoPhillips as a rough approximation for the working capital requirements of Trainer, we can expect working capital will increase Delta's real investment in Trainer by 72%--over and above the airline's announced $250 million investment. That's $430 million, half of Delta's 2011 bottom line.
It won't be long before shareholders start asking what they are getting for that money. The answer won't be satisfactory.
(Gregory J. Millman is a senior columnist with Dow Jones Banking Intelligence. He can be reached at +1 (212) 416-2352 or by email at firstname.lastname@example.org.)