Life costs money, and so does running an MLP. To buy or build a new asset, there is an associated one-time cost. The money MLPs spend for this is called growth or expansion capex. Once finished, there’s an amount of money it will cost to maintain it. For MLPs, the money used to maintain their assets (and often their cash flows) is called maintenance capital expenditures, or maintenance capex.
This week, we will examine maintenance capex, and next week we’ll look into how growth capex numbers have and will change.
We looked at the 18 MLPs in the Alerian Large Cap MLP Index (AMLI) that are not general partners. Luckily, despite no legal requirement, the majority of these companies not only disclose their capex numbers and break them out in maintenance and growth, but they also provide future guidance for the coming year’s capex.
Falling Maintenance Capex
Typically, maintenance capex goes up every year. Inflation and the addition of new assets would all drive rising maintenance capex numbers. Now, it can be a little lumpy: bumping up if more work needs to be done this particular year or bumping down if the asset is idle and requiring minimal maintenance. Divestitures and dropdowns will also affect maintenance capex.
Investors have asked several times to substantiate a rumor they’ve heard that MLPs, given the downturn, are no longer maintaining their assets in order to save money. None of these askers have been able to provide me any articles or data to back this up. I don’t know who’s claiming this, but in a week of research, I have found nothing to indicate this would be the case. Not maintaining assets isn’t like skipping happy hour in order to lose weight, it’s like living on cigarettes and diet coke. It won’t be cheaper in the long run, and it’s quite likely to end very badly.
That being said, on a composite basis, the 18 MLPs as a whole did spend 0.9% less on maintenance capex in 2015 than they did in 2014. However, looking at the median maintenance capex numbers compared to 2014, maintenance capex was up over 9%. The reason for the difference: most MLPs increased maintenance spending, but four names decreased y/y spending by more than 20%. Is this necessarily a red flag? Or could this be within the normal range of variance?
Western Gas Partners (WES), a gathering & processing MLP which raised maintenance spending by 1.3% in 2015, indicated that it expects maintenance to be 7%-10% of adjusted EBITDA. Using this as a rough guideline (let’s say 5%-12% to allow for the lumpiness noted above) and keeping in mind that every MLP defines maintenance differently, let’s look at those four companies whose maintenance capex decreased over 20% y/y.
Enterprise Products Partners (EPD) saw maintenance spending fall 24.8% from $369 million to $273 million in 2015. As management explains in the 10-k, “Fluctuations in spending for sustaining capital projects are explained in large part by the timing and cost of pipeline integrity and similar projects.” In 2014, maintenance spending was 6.8% of adjusted EBITDA, and in 2015 it was 5.2% of adjusted EBITDA.
Cheniere Energy Partners (CQP) also saw maintenance spending fall 28.1%. However, CQP does not break out maintenance spending as separate from other operating expenses, which means this figure includes gains from not only derivatives, but also their purchase agreements. According to the 10-k, “Operating and maintenance expense decreased $31.9 million in the year ended December 31, 2015, as compared to the year ended December 31, 2014, due to a $32.2 million increase in fair value for our natural gas purchase agreements recorded during the third quarter of 2015.” Since derivatives and other financial agreements are separate from the physical integrity of the assets, the decline here is actually a sign Cheniere has made money on this. Removing these benefits, CQP’s maintenance capex went up over 10%.
Enbridge Energy Partners (EEP) spent 28.2% less on maintenance in 2015. Midcoast Energy Partners (MEP) made a dropdown acquisition from EEP in mid-2014, which means that MEP will now be responsible for maintaining those assets. While this may not account for the entire difference, maintenance continues to be within the (admittedly arbitrary, but informed) range mentioned above.
Finally, Phillips 66 Partners (PSXP) saw a drop in maintenance capex of 35.3%. In the 10-k, it is noted that this number is lower, and that the general partner makes the determination of what is considered maintenance capex, but it is not directly addressed. I have two potential explanations: the partnership has been remarkably acquisitive, and perhaps they are doing full maintenance as soon as the assets hit their books (in much the same way one might immediately change the oil and top off the fluids in a recently acquired used car), or that since parent company Phillips 66 (PSX) was doing maintenance on its facilities in 2014 resulting in lower volumes on PSXP’s assets, management seized that opportunity to do necessary maintenance on its own assets. In 2015, maintenance did drop to 2.9% of adjusted EBITDA (from 8.4% in 2014). However, PSXP is projecting maintenance costs of $14 million in 2016, nearly double 2015’s $7.7 million, so it is also quite likely that 2015 was also just timing.
The Maintenance Cliff That Doesn’t Exist
Pipelines are not like cellphones. There is no scheduled point at which the whole thing breaks and needs to be replaced. Maintenance includes using PIGs to determine which sections may have corrosion, product build-up, or tiny cracks, and replacing those sections as necessary. Now, as pipelines go, maintenance costs do increase; however, some of the pipelines built during WWII are still in service today.
Often, in the Risk Factors section of a 10-k, an MLP will list that older assets may require more or more expensive maintenance. To quote EPD’s, “Many of our assets have been in service for many years and require significant expenditures to maintain them. As a result, our maintenance or repair costs may increase in the future.”
Shell Midstream Partners (SHLX) has a very detailed explanation of the expected 2016 maintenance work in the 10-k, and I think it’s quite instructive to see how they will be spending the money, “We expect Zydeco’s maintenance capital expenditures, which are asset integrity projects in nature, to be approximately $19.6 million for 2016, of which approximately $12.8 million is to replace a two-mile section of a 22-inch diameter pipe under the Atchafalaya River and Bayou Shaffer . . . We expect Auger and Lockport’s maintenance capital expenditures, which are asset integrity projects in nature, to be approximately $10.0 million for 2016. This includes $4.8 million for electrical and storm water upgrades at Lockport, and $5.2 million for control system upgrade and personnel escape pods at Ship Shoal 28 for Auger.” Zydeco covers over 350 miles, so replacing a two-mile section is about 0.6% of the pipeline. To give you an idea of the lifespan of a pipeline, at a rate of two miles per year, the entire Zydeco pipeline would take 700 years to be fully replaced.
Is There Such a Thing as Too Much Maintenance?
Enable Midstream Partners (ENBL), is spending (and will spend) 15%-20% of its adjusted EBITDA on maintenance capex. According to the 10-k, ENBL defines maintenance capex as, “cash expenditures (including expenditures for the construction or development of new capital assets or the replacement, improvement or expansion of existing capital assets) made to maintain, over the long-term, our operating capacity or operating income.” As a gathering & processing company focused on the Mid-Continent and the Bakken, its exposure to these unconventional resource plays can mean that maintenance expenses relative to volumes may increase over time. As these fields decline faster, in order to “maintain operating income,” more gathering pipelines need to be built to sustain volumes. Some G&P MLPs classify replacing the cash flows as maintenance capex, and some G&P MLPs don’t. In ENBL’s case, we would argue that the high maintenance spending percentage comes down to semantics of the definition of maintenance capex.
Spectra Energy Partners (SEP), a natural gas transportation company, also spends 16%-17% of EBITDA on maintenance capex. The 10-k, unfortunately, does not provide further details. However, their factsheet does list 15,000 miles of transmission and gathering pipeline. If you look closely, you’ll notice that Gulfstream is underwater in the Gulf of Mexico. SEP does not explicitly state this, but it just seems logical that an underwater pipeline would be more expensive to maintain.
However, WES, the G&P MLP mentioned above for its expected 7%-10% maintenance ratio, is even more explicit with its definition of maintenance capex, “those expenditures required to maintain the existing operating capacity and service capability of our assets, such as to replace system components and equipment that have been subject to significant use over time, become obsolete or reached the end of their useful lives, to remain in compliance with regulatory or legal requirements or to complete additional well connections to maintain existing system throughput and related cash flows” (emphasis added).
On the opposite end of the spectrum, Williams Partners (WPZ) does not define maintenance capex, nor does it list maintenance and expansion capex separately. Like with CQP, if I use “operating and maintenance expenses” as listed in the 10-k, WPZ has spent nearly 40% of adjusted EBITDA on these costs. Given that operating expenses are included, this is almost certainly an over estimation, but still worth noting.
What does a high ratio mean for investors? Like ENBL, it may simply mean that the company is more transparent in the costs incurred to maintain not just its assets, but its business. A rising ratio could mean that the cash flows (volumes) are getting more and more expensive to replace. Otherwise, if a high ratio indicates an overly maintained system, I doubt that would cause any operational problems.
In the End
Since many investors use EBITDA as a valuation metric, and many more investors use DCF, it’s important to point out that a DCF calculation typically removes maintenance capex, whereas an adjusted EBITDA calculation does not. An increase in DCF often signals an increase in the distribution. The red flag would be if DCF is increasing at the expense of maintenance capex, absent any explanation. Some years, less maintenance is simply required, or a divestiture changes the algebra. Among the management teams with whom I have spoken, the idea of delaying or canceling maintenance in order to increase the distribution is simply absurd.
For MLP investors, tracking maintenance capex should be part of the basic due diligence. Just like maintenance records should be asked for when buying a car or a house, investors should check that the infrastructure they are buying has also been properly maintained. Too little maintenance puts up a red flag, but too much may indicate a business very difficult to maintain (as might happen with a G&P MLP) or a business that is not as profitable as it might be.