The discussion of the recent short report on Brookfield Infrastructure Partners (NYSE:BIP) has largely subsided and the price of BIP units has recovered. The drop and the subsequent rise in unit price should, however, mostly be attributed to the move in long-term treasury yields, as is evidenced by price movement for other yieldcos — so it seems the short report, though more credible than previous such reports, did not make a dent in unitholder acceptance of BIP. I don’t want to rehash all the previous discussions, but briefly put together all the points I consider important. Judging from SA comments, many are not aware of how strong the case against investing in BIP is. (This does not suggest shorting it. I don’t know how to successfully deal with potentially unlimited losses from shorting, and BIP has the support of BN which can turn the tables on investors being short, so it’s probably way better to find something else to short if you must.)
1. Brookfield tends to be very promotional in their talk and presentations. BIP is being sold to the public as a stable infrastructure company with tons of cash flowing upstream from the operating subsidiaries. Certain disclosures that might suggest otherwise have been gradually removed from supplemental information. At present, it is impossible to determine how much cash flows to BIP from the subsidiaries, but it seems way less than indicated by FFO or AFFO. According to the short report, there are subsidiaries like Arteris (toll roads) that actually consume cash. (I’ve not verified it because it is hard work and I don’t need to know it for my decisions.)
2. Brookfield avoids talking about mistakes and failures, and presents their operations in such a way that it is hard to track individual investments: subsidiaries are often renamed on acquisition, not referred to by a name but by a description instead (“our UK ports”), and dissolved in aggregate reporting or swept under the rug if failing and never mentioned again. IRRs and MOCs are reported by top examples or aggregately, without mentioning instances with poor results. It is well known that in the investing business, even the best fail a lot — as per a Peter Lynch quote, “In this business, if you’re good, you’re right 6 times out of 10. You’re never going to be right 9 times out of 10.” Where are the 4 out of 10 when BIP was not right? Buffett / Berkshire have no trouble mentioning theirs, and if anyone is as good as Brookfield claims they are, I’d expect them to admit mistakes without much hesitation. Such asymmetry in reporting successes and failures is not necessary: for instance, BDCs regularly report investments with non-accrual and payment-in-kind status by company name and by amounts of capital involved, and discuss them in some detail on conference calls, often without being asked about.
3. Let’s take a look at the data from 4Q21 supplemental information (credit for pointing out reconciliations of “adjusted earnings” goes to A. Steinberg). On p. 43, AFFO is $1412M, but $1093M of that is net income. On p. 44, net income is shown to consist of 1093M of realized gains (from sales of subsidiaries). In other words, 77% of AFFO comes from capital recycling, not actual infrastructure operations! They also mention some 739M of items that could be considered “non-expenses”, so possibly the net income is understated which would weaken this argument, but still — a major part of FFO and AFFO is realized gains. The situation is not much different today, as confirmed by IR in responses to e-mail questions.
4. While BIP tells SEC that FFO is unsuitable as a liquidity measure because it does not fully capture actual cash flows and is partially accrual-based, they are perfectly willing to tell unitholders that “FFO is the best representation of the operating cash flow our businesses generate” (check the comment by LasVegasInvestor) and freely involve FFO whenever they are discussing cash flows (with unitholders and analysts, not SEC).
FFO payout ratio is an accounting construct that is disconnected from business reality even if one believed FFO measures cash from operations (will the tooth fairy pay for maintenance of assets?). Only the AFFO payout ratio makes some sense, but there are important caveats even there. I’ll illustrate it on a railroad BIP built years ago in Australia. The railroad connects a mine to a port. If the mine is closed, the railroad is worthless. So its life is finite, perhaps 20 years, and thus any cash from that railroad is partially a return of capital, and AFFO does not reflect that. When the railroad is new, there is barely any maintenance cash outlay, so AFFO overstates true economic earnings — there will be much more maintenance needed later (Buffett said once that BNSF requires capital expenditures above depreciation to remain competitive). This problem pertains to all businesses to some extent, but we see that payout ratios derived like this might be deceptive.
5. The business model employed by BIP (and private equity in general) works well in environments where debt is cheap and prices mostly go up. Just what we had over the last 15 years, i.e. pretty much all the time BIP existed. Past results might be shiny, but don’t get fooled into naively projecting them into the future. Heed the warnings from the more experienced like H. Marks. “Capital recycling” is not magic, it requires availability of capital for debt financing and willing buyers.
6. BIP raised quite a lot of capital by issuing expensive BIPC units in recent years. That might not be viable in the future on such good terms, which might put a lot of pressure either on distributions or on future growth (if there is no cash to retain).
7. On the subsidiary level, BIP uses as much debt as the rating agencies permit (for the entities to remain investment grade). This worked well while rates were going down; and BIP also extended maturities where it could, so is not affected by rising rates in the near future. But getting cash from operating subsidiaries by raising debt against increased asset valuations / underlying cash flows is likely to be a lot harder over the following years.
8. The hefty fees BIP pays to Brookfield Asset Management (BAM) take away most (or even all) of the upside. BIP was created at times when 50% IDRs for master limited partnerships were common, and BIP’s 25% incentive distribution looked relatively attractive. But such exorbitant fees proved unsustainable and those times are long past. BIP fees take away about 1/3 of the generated cash flows. I do believe that BAM adds considerable value with its capabilities, but one cannot rely on getting stellar returns all the time for decades (the cyclical business environment won’t permit that). If BAM manages to hit the upper bound of their target range (15% returns), only 10% goes to the unitholders. Thus you get the historical average return of U.S. equities at best, and it’s imprudent to count on more.
9. BIP is, in principle, a collection of separate businesses, a fund. It is reasonable to value it on a sum of the parts basis to determine some kind of net asset value (NAV). However, there is no disclosure that would allow an investor to do so. Aggregate numbers can hide quite a lot and make discounted cash flow valuation very unreliable. The recent short report suggests that units trade at 2-3x NAV. While I think that is far-fetched, what exactly are reasons for BIP to trade at a premium to NAV vs. at a discount as most other such vehicles and yieldcos trade? After considering the fees paid to BAM, the discount should be massive (the fees paid are not reflected in SOTP / NAV calculations, so their present value should be subtracted, giving rise to a discount). And what kind of discount should one demand for a fund that is not even willing to disclose its NAV?
10. A similar collection of separate businesses managed by Brookfield is Brookfield Business Partners (BBU),”a global business services and industrial company focused on owning and operating high-quality providers of essential products and services”, according to the website. They forgot to mention they also aim to flip the businesses all the time, but they post impressive IRRs and MOCs for their deals (just as BIP does). They also report FFO (renamed to EFO) and ignore depreciation. Unlike BIP, BBU distributes very little, and fees are paid only on capital appreciation. Unlike BIP, which typically trades at > 10x FFO, BBU trades for < 5x FFO. Is there really such a big difference between the two, considering that most of BIP cash flow / FFO comes from gains on asset sales, similar to BBU? BIP has a longer history than BBU, esp. the history of distributions, but we don’t truly know how it will perform in adverse environments when it is not possible to sell assets for a good price. BBU can at least hold onto them, but from what is BIP supposed to pay the promised distributions?
11. Another subsidiary of Brookfield, similar to BIP, was BPY, focused on real estate. I’ve never been able to figure out if their distributions were covered by cash flow (and it seemed like they were not covered by cash from rents). When hard times arrived with covid, BPY traded down to a distribution yield above 15% and price less than 0.5x NAV (based on somewhat questionable cap and discount rates, but those were at least clearly published — that discount vs. NAV persuaded me to buy some units). After that, BPY was acquired by BN at a slight premium to market price, but way below NAV (it was possible to elect to be paid with undervalued BN shares at least partially, so one cannot say BN stole it from the unitholders). BIP is a somewhat different story because it is more diversified across sectors, but BPY distributions and NAV got separated from the business reality. How can we be sure this is not happening / will not happen to BIP? Will a unitholder be able to notice such a change before a meaningful unit price drop?
12. Fiduciary duties (and personal interests) of people managing BIP are with BN, not BIP. They are interested in growing fees BIP pays, not enriching unitholders. (BN holds a significant stake in BIP, which mitigates the issue somewhat.) An example of this risk materializing is how the Dalrymple coal terminal was IPOed — while BAM’s institutional fund was allowed to completely sell its interest in the terminal, BIP only sold it partially and kept the shares no one wanted at the IPO. Also, when new deals are split between various entities BAM manages and third-party co-investors, I’d say that BIP is the lowest in the pecking order — I can hardly imagine that a co-investor would be getting worse conditions than BIP, or that BIP’s availability of capital would be the primary determining factor for deciding how big a chunk of a juicy deal BIP gets.
Alternatives to BIP
Investors are not compensated well for all the risks they undertake by owning BIP units (it’s even worse for BIPC shares that are more expensive). At the prevailing prices, you get 5-6% yield plus perhaps 6-7% growth (assuming no deterioration in business conditions that helped BIP achieve such growth in the past, which might be a tall order). Quick capital appreciation is speculative at best, you are relying on sustained growth for 10+ years to get a decent total return. There are various alternatives against which BIP seems overvalued:
- Brookfield Corporation (BN), the top Brookfield entity (apart from semi-private entities like Partners Value Limited that mostly just holds BN shares), offers a ~10% distributable earnings yield while DE is projected by the management to grow by 17-25% a year for the next 5 years. The yield and projected growth are way higher than BIP’s. If you trust the Brookfield people, why not go with BN instead of BIP? They own > 20% of BN and almost no BIP, after all. And if there is a massive infrastructure boom, BIP might participate, but the winner would be BAM/BN which can freely raise funds and earn fees, unconcerned whether BIP has capital to invest or the ability to issue BIPC shares.
- Enterprise Products Partners (EPD), a high-quality midstream business that uses conservative accounting and has well-covered distributions, yields ~8% with expected growth of perhaps 5%. Expected total return is higher than BIP from similar types of assets and the risk that high future growth will not materialize is lower. (Credit to A. Steinberg on insisting on this comparison — no one suggested a counterargument in the SA discussions, though plenty of people favor BIP.)
- BN has several preferreds with yields much higher than BIP’s (traded only in Canada). For instance, BN.PR.K is floating rate (i.e. offers a degree of protection against rising rates / inflation), yields 11.3%, and has a minimum expected yield of 3.2% (based on the lowest historical Canadian Prime Rate of 2% — when bought at today’s price, the preferred’s yield is 1.6x Prime Rate). BN.PF.J yields 8.6%, will have the distribution reset in 2027 to Canada govt. bond yield + 3.1%, and the minimum yield is 6.5% (contractual). These instruments are way higher in the capital structure than BIP — the payments on preferreds are rather negligible when compared against BN’s distributable earnings (which is much closer to actual regular repeatable cash flow compared to BIP’s AFFO). Other companies also offer promising preferreds if one is seeking yield.
- There are many decent BDCs paying yields much higher than BIP, while exposed primarily to credit risk (I don’t think one can easily argue that equity in BIP’s investments, many of them in emerging markets, is safer than debt of U.S. middle-market companies; it is more likely to be the opposite). For instance, FSK yields ~14%, and BXSL earns similarly but distributes less, so it only has a yield of 11%. If one does not like BDCs’ internal ~2:1 leverage, there are private credit funds like CCLFX that borrow less.
- A decent way of obtaining yield is to put 1/3 of the portfolio in cash (or short-term treasuries) and 2/3 in stocks with high capital appreciation, say, Berkshire Hathaway (BRK.B), Apollo Global (APO), and Blue Owl (OWL). Even if interest rates were zero, you could draw 4% of your portfolio from the cash pile for 7+ years (with rates at 5%, it is 10+). In scenarios where this approach fails, BIP won’t perform great either and is unlikely to raise distributions much.
I have to admit that expected returns from S&P 500 at current valuations are significantly lower than from BIP, and there are yieldcos and utilities with little if any growth trading at yields not much higher than BIP. So it’s not true that BIP is mispriced against every other option.
Most of the mentioned concerns also apply to BEP. For solar and wind, some degree of depreciation is very real, so confusing FFO with “owner earnings” is likely to lead to disappointment. Water reservoir levels seem to be at risk due to climate changes, so the normalized hydro power generation figures that BEP publishes are nice, but not reliable. BEP operates in a single industry, so it is susceptible to eroding returns from new renewable power generation projects — they can’t do much if competitors are satisfied with lower returns and outbid them. In my opinion, these concerns are not reflected enough in the unit price and thus BEP is not a viable alternative to BIP.
Conclusion
The points listed above contain a couple of red flags. It cannot be said that BIP distributions are in immediate danger, but there are risks that are real, and Brookfield is doing very little to assuage well-founded worries of unitholders: a bit of cash flow disclosure would go a long way, and they are purposefully not providing it (though they were willing to do so years ago).
Thinking about this, I’ve discovered the story of Allied Capital. A neat summary or a full book is also available. A takeaway from this and other similar stories is that fraud or deception can go on for a very long time, and so can questionable business models. In particular, issuing shares at excessive valuation can mask quite a lot of poor business developments. I’ve not studied it yet, but an article on Main Street Corporation (MAIN) contains more food for thought on this topic.
A notable difference from these stories is that the whole Brookfield’s reputation is at stake, so I don’t think they would run some fraud specifically at BIP that would scare institutional investors away from BAM. But they might operate on the belief that those big investors don’t care much about what small guys invested in BIP get. On the other hand, BN shares have many of those risks mentioned above removed (and another set of risks added), and they are cheap enough to warrant consideration, but the stake has better be limited because of the tail risk. We should also not forget the story of Edper, a Brookfield predecessor employing a similar pyramidal structure, which met adverse business conditions and collapsed.
The risk that companies are not as great as they appear is real, and it’s extra high for businesses that liberally use non-standard accounting measures. I’d prefer to be compensated for risks, and from the alternatives I’ve suggested above, it seems quite clear that BIP unitholders are not. I am now in a wait-and-see mode on Brookfield, keeping a limited BN stake only, and awaiting Buffett’s proverbial tide to go out, or if more cockroaches will appear.
Read the full article here