Midstream MLPs: Maintenance CapEx Naughty List

With the 4Q13 results in the books, we’ve updated our Midstream MLP Maintenance CapEx Comp Table. 

First, some color on the issue…

The widely-held opinion among MLP investors, analysts, bankers, and management teams is that net income doesn’t matter because depreciation, depletion, and amortization (DD&A) is a non-cash expense (and only cash expenses matter).  At the simplest level, “Maintenance Capital Expenditures” (M-CapEx) is substituted for DD&A, and MLPs are evaluated and valued based on “Distributable Cash Flow,” rather than net income or earnings.

With respect to that prevailing opinion, our view echoes Warren Buffet’s:

“Trumpeting EBITDA (earnings before interest, taxes, depreciation and amortization) is a particularly pernicious practice.  Doing so implies that depreciation is not truly an expense, given that it is a “non-cash” charge.  That’s nonsense.  In truth, depreciation is a particularly unattractive expense because the cash outlay it represents is paid up front, before the asset acquired has delivered any benefits to the business.  Imagine, if you will, that at the beginning of this year a company paid all of its employees for the next ten years of their service (in the way they would lay out cash for a fixed asset to be useful for ten years).  In the following nine years, compensation would be a “non-cash” expense – a reduction of a prepaid compensation asset established this year.  Would anyone care to argue that the recording of the expense in years two through ten would be simply a bookkeeping formality?” (BRK 2002 Annual Report).

M-CapEx is now a hotly-debated topic in the MLP sector, as it should be.  Because there’s no simple, “right” answer, most default to simply plugging the management’s number into their models, and then projecting future M-CapEx consistent with it (usually as a % of adjusted EBITDA (another dubious management metric)).  And as we have little insight as to how management calculates (or even defines) M-CapEx (no reconciliation back to GAAP Capital Expenditures), it’s difficult to evaluate the appropriateness of what management is giving us.  In short, MLPs can essentially give us any M-CapEx number they want, it’s difficult for investors and analysts to scrutinize the metric given the lack of transparency, clarity, and disclosure (which is why most don’t even bother to), and MLP managements (especially those with Incentive Distribution Rights) are incentivized to give a M-CapEx as low as is palatable to the sell-side (and this seems to be very low), at least in the short-term.

Buffet gives the best definition of M-CapEx that we’ve heard:

“…the average annual amount of capitalized expenditures for plant and equipment, etc. that the business requires to fully maintain its long-term competitive position and its unit volume.”

He goes on:

“Most managers probably will acknowledge that they need to spend something more than [DD&A] on their businesses over the longer term just to hold their ground in terms of both unit volume and competitive position. When this imperative exists - that is, when [Maintenance CapEx] exceeds [DD&A] - GAAP earnings overstate owner earnings. Frequently this overstatement is substantial” (BRK 1986 Annual Report).

Clearly Buffet doesn’t talk to many MLP management teams, because Midstream MLP M-CapEx equaled only 28% of DD&A in 2013!

Bruce Greenwald (Professor of Finance and Asset Management at Columbia Business School) defines M-CapEx as the capital cost of keeping revenues flat, relying upon the ratio of gross-PP&E-to-revenues over the prior seven year period (see these notes for more on his method).  One could perform a similar calculation using a below-the-line figure like EBIT.

Note how neither of these definitions have anything to do with keeping assets "safe" or "compliant"...


In our view, the vast majority of MLPs screen as highly aggressive with their M-CapEx figures.  In 2013 there was not one large midstream MLP with M-CapEx higher than DD&A, and the average was only 28% (down from 34% in 2012). 

We believe that M-CapEx vs. DD&A and PP&E are the best metrics when screening for potential aggressive accounting, because the comparison is going to be apples-to-apples across companies, as well as internally consistent.  M-CapEx as a % of EBITDA is less useful for two reasons: 1) the definitions of EBITDA and/or adjusted EBITDA vary from one MLP to another, and 2) this measure does not control for differences in returns on capital or EBITDA margins among companies; all else being equal, a higher EBITDA margin MLP will screen as more aggressive than a lower EBITDA margin MLP in an M-CapEx vs. EBITDA ranking.

To the argument that DD&A is irrelevant because, “The book useful life is understated,” we say: “Show me.”  We have not seen an instance of a natural gas processing plant or long-haul pipeline installed 25 to 35 years ago that has not received significant capital expenditure (in excess of its depreciation) since, and is still in proper working order.  We fully appreciate that book useful lives can be, and often are, extended, but not without major capital expenditure over time.  While the timing of the capital expenditure will be lumpy and difficult to predict, it will come (or the assets will be abandoned/scrapped).  For instance, Enbridge (ENB, EEP) recently announced that it will replace its Line 3 crude oil pipeline after 46 years in service; the line replacement will cost $7 billion, and Enbridge IR noted to us that, “…the replacement cost of the infrastructure is significantly greater than the sum of its original cost and collective subsequent investments made to the pipeline.”   Given the impact of inflation on the replacement cost of long-lived, fixed assets, this really is not surprising.


Why this matters…

The short-term consequence is that MLP investors, today, are discounting decades of future “distributable” cash flows that are likely to be significantly in excess of future “free” cash flows, lending to widespread and substantial equity overvaluation.

The longer-term consequence is that, at some point, CFFO growth will slow (and possibly cease and/or go negative), while the MLP has a large, externally-financed “growth” CapEx budget.  By this stage, returns on incremental “growth” CapEx are below the cost of capital (and perhaps negative), and the coming dilution is obvious to all.  The more aggressive the MLP is with M-CapEx today, the more acute this risk is.    

Here is our updated 2013 Comp Table, followed by some takeaways below:

Midstream MLPs: Maintenance CapEx Naughty List - kk1

(click for printable PDF)

Comp Table Takeaways

  • The group already screened aggressive in 2012, with M-CapEx / DD&A at only 34%, but got more aggressive in 2013, as the ratio fell to 28%.  In 2013, aggregate DD&A increased 22% YoY while aggregate M-CapEx increased only 2%.  Average PP&E increase 27% YoY in 2013, and the M-CapEx-implied useful life of this PP&E (M-CapEx / Average PP&E) in 2013 was 77 years, a 24% increase from 62 years in 2012.  (The average book useful life (DD&A / Average PP&E) was 22 years in 2013, up from 21 years in 2012.) 
  • The Gatherers and Processors (G&Ps) screen most aggressive, with MWE, CMLP, XTEX, SXE, APL, and RGP all near the top of the list.  In our view, this is likely because many of them have invested heavily in new G&P infrastructure in recent years, assigned all of that CapEx to “expansion” or “growth,” and do not include a reserve for the eventual “replacement,” “refurbishment,” or "turnaround" of these fixed assets in their M-CapEx calculation.  Even if “the assets are new,” they will need to be refurbished or replaced, at which point the cash expenditure is likely to be in excess of the accrued depreciation charge given capital cost inflation.  Just as a refining MLP includes a "replacement" or "turnaround" capital reserve in DCF, we see no reason why a natural gas processing plant shouldn’t have one as well.  (Note: this is what depreciation is for!)
  • Absolute M-CapEx down YoY is a red flag, in our view, unless the MLP was a net seller of assets.  WPZ (-37%), SXE (-35%), OILT (-22%), EPD (-20%), GEL (-19%), and XTEX (-18%) screen as most aggressive on this measure in 2013.
  • M-CapEx as a % of DD&A down YoY is a red flag, in our view.  This implies that PP&E increased while M-CapEx decreased.  This was the case for 74% of the MLPs in this group, and the group was -17% YoY on this metric, on average.  SXE (-62%), TLLP (-57%), SXL (-44%), WPZ (-40%), OILT (-39%), and APL (-38%) screen as most aggressive.  OKS (+42%), PAA (+33%), BPL (+25%), and RGP (+25%) were the big MLPs with this metric screening less aggressive in 2013 than it did in 2012.
  • MWE’s 2013 M-CapEx implies a useful life of 351 years!  And this PP&E is largely processing and frac plants…  Amazing.  MWE is, by far, the most aggressive midstream MLP on this measure.
  • WPZ’s 2013 YoY change in M-CapEx is particularly notable.  PP&E and DD&A were up YoY, yet M-CapEx fell from $403MM in 2012 to $255MM in 2013...  That strikes us as questionable...
  • For the second consecutive year, SEP screens least aggressive by a wide margin. 
  • Among the G&Ps, EQM screens least aggressive by a wide margin.

Kevin Kaiser

Managing Director