In a February article, I warned about rich MLP valuations, writing:
Going forward, I expect very long-term returns to be positive, but modest. Along the way, there is likely to be a really bad-ass correction.
I re-iterated those concerns in May. I wasn’t too far from the mark. The Alerian MLP index dropped nearly 20% from its high in May.
Now, arriving late to the party is James Kostohryz, who declares that MLPs are "extremely overvalued." If Kostohryz actually follows the MLP sector, you have to wonder why he wasn’t declaring this back in May when MLPs were 15-20% more expensive than they are now. Putting that aside, it’s worth evaluating his claims against actual data. We’ll do that with a few of his key assertions below.
EV/EBITDA ratios demonstrate MLPs are overvalued
Kostohryz’s table posits an EV/EBITDA average of 16.9 for MLPs (vs. 10.5 for the S&P, ex-Financials). Now I track research from all the major firms that cover MLPs (Credit Suisse, Merrill Lynch, Wells Fargo, Citi, Ladenburg-Thalmann, J.P. Morgan, etc.). Their average EV/EBITDA estimate for the sector is 11.5. There isn’t much variation either, with the lowest estimate at 10.8 and the highest at 11.9. Here are the current EV/EBITDA ratios for the top five in the Alerian index (~40% of index’s weighting):
Generally speaking, the ratios are substantially lower for the smaller MLPs. Out of 80+ MLPs, fewer than a half dozen have EV/EBITDA ratios over 15. Kostohyrz’s 16.9 number is not just a little off. It is dead wrong -- so spectacularly wrong, you really have to wonder where it came from. Did he accidentally grab the “maximum” number instead of the average?
Price to Cash Flow (P/CF) shows MLPs are overvalued
Kostohryz gets a few things wrong on this topic. First, he asserts: “P/CF is fundamentally inaccurate due to the fact that a good chunk of the ‘C’ is actually distributed to the General Partner [GP].” While that is correct for MLPs that actually have GPs, many MLPs don’t. The largest MLP in the sector, Enterprise Products (EPD), doesn’t. The largest upstream MLP, Linn Energy (LINE) doesn’t. And the largest pure-play, refined products MLP, Buckeye Pipeline Partners (BPL), doesn’t. Moreover, at the rate MLPs are re-absorbing their GPs, there will be very few left in a couple of years. This is well known to those who actually follow the MLP sector.
Kostohryz further claims “P/CF does not take into account maintenance capital expenditure, thereby overstating true owner earnings.” While this is absolutely true for regular cash flow, MLPs and analysts in the sector mostly ignore that measure. Instead, they focus on Distributable Cash Flow [DCF]. DCF emphatically does account for maintenance cap-ex (see, for example, the notes in LINE’s and EPD’s investor presentations). The 82.4% CF payout ratio in Kostohryz’s table is suspiciously close to the DCF payout ratio for the MLP sector (83%), which makes me think he confused regular cash flow and DCF. If the payout ratios for the other asset classes were based on regular CF payout ratios, this is an invalid, apples-and-oranges comparison.
In the comments, Kostohryz opines: “I am very concerned that many of these distributions are simply not sustainable.” He provides no data to back this hyperbolic concern, so let’s look at the data. By itself, EPD accounts for 14% of the Alerian index. Its payout ratio, based on last quarter’s Distributable Cash Flow (not regular cash flow), was .73. Averaging analyst estimates, EPD will grow DCF by 6-7% each year for at least the next five years. Even with its conservative philosophy, EPD has managed to increase its distribution for 27 consecutive quarters! Precisely what about these numbers suggests a lack of sustainability? Looking at the rest of the sector, I can only find 5 (out of 80+) MLPs that even a single analyst thinks will fail to cover their distributions with DCF. Most of these are tiny newcomers like Niska Gas Storage (NKA) -- which has a 0.21% weight in the Alerian index. And the analysts aren't even unanimous in projecting these companies will fail to cover distributions.
MLPs do deserve a premium
It sounds like Kostohryz may address this topic in a future article. Until then, I can give a few good reasons why valuation metrics should reflect a premium for MLPs. Most notably, MLPs are pass-through entities, so you avoid the double-taxation that C-corp shareholders are subject to. You also benefit from passed-through deductions like depreciation and depletion. The result is the distributions from MLPs are typically about 80% tax-deferred. You don’t pay the rest of the tax until you sell the MLP. If you hold an MLP for 15 or 20 years, as I have, that represents a fantastic deferred tax advantage. So what should the premium for this be? The market actually did put a price on it years ago when the Lakehead Partners ruling stripped MLP status. Prices immediately dropped 30% and stayed there until the ruling was overturned.
Beyond the tax advantages, MLPs tend to be some of the stablest, widest moat businesses around. Let’s take Buckeye Partners (BPL), for example. Buckeye is no fly-by-night; it has run its business profitably for 125 years. Among other assets, Buckeye owns the pipelines that carry aviation fuel to La Guardia and JFK. On most of its pipelines, BPL is allowed to increase its tariffs by PPI+2.65% every year. So not only do you get fabulously stable revenue, you also get excellent inflation protection. BPL has the sort of revenue model that makes the “dividend aristocrats” look risky by comparison. I, for one, am willing to pay a premium for that.
The MLP structure has fostered a special culture within management and among unitholders. The basis of the culture is that the companies’ profits belong to the unitholders. The burden of proof is on management to justify withholding or spending the profits. This contrasts starkly with most U.S. companies, and particularly tech companies. Companies like Hewlett Packard and Google are content to sit on stupendous piles of cash which idles in banks, earning 0. That’s part of what you pay for when you buy these companies: cash that is stupidly under-utilized. Eventually the empire-building instincts get the better of the executives and they end up buying questionable assets at ridiculously inflated prices. Witness HP’s and Google’s recent purchases, for example. MLP culture doesn’t allow for this sort of nonsense. MLP unitholders insist that every acquisition be completely justified. Every acquisition press release had better contain the words “accretive to earnings” (or better yet, “immediately accretive”), or the stock and management will get trashed. MLP managers are keenly aware of this fact. This cultural advantage justifies an additional premium for MLPs over other companies.
When you look at the real data, you’ll see that MLPs aren’t too far from “fair valuation” at this point. The spread of the Alerian index yield over 10-year Treasuries has averaged 320 basis points since 2000. The current spread is about 318 basis points. Price to DCF and payout ratios are also pretty close to historical averages. MLPs aren't a raving bargain at this point. But they're not "extremely overvalued" either.