I think that ACMP’s cash flows will decline due to:
- Less gas being carried on their gathering pipelines. Throughput will decline as shale gas wells naturally decline. Throughput on ACMP’s existing assets cannot grow unless more shale gas wells are drilled.
- Cash flows from minimum volume commitments ending.
*Disclosure: I am short ACMP common. This is not one of my better short ideas and I may cover this position in the future.
ACMP’s assets largely consist of gathering pipelines. The company does not generate much revenue from processing facilities, which have different economic characteristics.
Well and pipeline economics
The economics of gathering pipelines are largely tied to two things:
- Existing wells, their economics, and their production volumes.
- Future wells.
A shale gas well will have strong cash flow in the very beginning. Over time, production will drop as the well depletes. This will cause cash flow to drop. Ideally, the payments to the gathering pipeline company should be high in the beginning and much lower as the well enters its decline phase. If not, there could be a situation where the gathering costs are so high that the well operator would have to stop production on its well even though the well can still generate meaningful amounts of cash. Such a scenario ultimately doesn’t make sense for either party.
There is a large degree of uncertainty as to how much money a shale gas well will actually make. Because the underlying shale gas wells utilize new technology (e.g. horizontal drilling with very long laterals, multi-stage fracturing, etc.), we don’t have past experience that we can use to make good predictions of the future. We don’t know exactly how much gas the well will produce. More gas would be beneficial for both parties.
There is also uncertainty about future wells that might be drilled in the same field. If more wells are drilled, then another segment of gathering pipeline will be built to connect the new wells to the existing network of gathering pipelines. Because most of the pipeline infrastructure is in place, the cost of the new segment is very low. The economics of that segment of pipeline is probably very good. Contracts set out how the pipeline company is compensated in this scenario and how much of the pie they get. If other companies want to join the gathering pipeline network, then the gathering company can potentially make an unusual profit. I believe that the pipeline companies can only make unusual profits if their pipeline network expands beyond initial expectations.
In practice, the well operator will have a lot more technical data than the gathering pipeline company. They will be more informed about their existing wells’ decline rates (or type curves) and how many future wells they plan on drilling. In an ideal world, the gathering company would enter into a contract that ensures that they get a decent return on capital without the risk of being on the wrong side of this informational disadvantage.
ACMP’s economics are very much tied to the contracts it has struck with producers. Unfortunately, while ACMP files their contracts on EDGAR most of the contract details are redacted. This applies to its contracts struck with CHK, even though both companies are publicly listed. This lack of transparency makes ACMP difficult to evaluate.
In any case, this is what I’ve been able to figure out.
Minimum volume commitments
Some of the contracts are based on a minimum volume commitment. The gathering company is basically guaranteed a set minimum amount of revenue. Both parties split the upside above the minimum volume commitment. (I’ve oversimplified things as the details are slightly more complicated. You might want to read the redacted contracts for yourself.) The gathering company largely gets paid a fixed fee per MMBTU. The more gas, the more it gets paid. The contact stipulates minor adjustments to the fee per MMBTU. This is to try to cover a scenario where cost inflation causes the gathering company to lose money on the gathering pipeline. Such a scenario would push the gathering company to try to break the contract or to simply default on the contract. Both parties are better off if they can avoid conflict and legal fees.
Cost of service
Under this style of contract, the gathering company is guaranteed a particular internal rate of return. There may be conflicts as to how much the gathering company is or isn’t making. The gas producer wants to pay as little as possible while the gathering company wants the opposite. This ACMP contract sets out a dispute resolution process, with the final arbiter being a consultant selected by the AAA office in Dallas. If the gathering pipeline network expands, then the gathering company will have to spend the capex to connect new wells to the network.
This will allow (?or force?) it to invest capital at a set rate of return. I didn’t figure out if this contracted rate of return exceeds ACMP’s average cost of capital. (If the number was less than ACMP’s cost of capital then this would be a bizarre contract. ACMP could lose money if the gas producer drilled more wells in the field. Because this isn’t listed as a risk factor in ACMP’s 10-K, this may not be the case.) My guess is that the rate of return exceeds ACMP’s cost of capital. ACMP wants to see the gas producers drill more wells so that it can invest more capital at rates that exceed its cost of capital.
In this style of contract, the gathering company also gets paid a fee per MMBTU. This fee is adjusted up or down to guarantee the rate of return. If the gas producer makes less gas than they said they would make, the fee per MMBTU goes up (and vice versa). Some of the contracts have a cap on how high the fee per MMBTU may be adjusted upwards.
I have no idea why ACMP does this, but it actually doesn’t own its own compressors. It rents them from Chesapeake. Chesapeake also happens to account for most of ACMP’s revenues. I could not figure out how the rates for these compressors are determined and what the rates will be in the future after they are adjusted to “market” rates.
This is a layer of financial engineering that doesn’t make sense to me.
What will cash flow be several years from now?
I think that ACMP’s cash flows will go down. The 10-K contains diagrams showing all of the minimum volume commitments. Several years from now, all of these minimum volume commitments will end. I believe that the last minimum volume commitment is for the Barnett Shale Region and that it will end in 2019 (see page 42 of the 10-K):
ACMP will largely be paid a fee per MMBTU. Because natural gas wells naturally decline, ACMP will likely see less throughput on their system than today (ignoring any acquisitions). It is possible that ACMP’s clients will drill more wells. However, Chesapeake is ACMP’s largest customer and has announced that it will be largely focusing on NGL plays and will be drilling less wells than before.
In particular, I’m paying attention to the Haynesville Shale Springridge gathering system. The Haynesville Shale produces dry gas and producers are leaving the area due to low natural gas prices. From YE2011 to YE2012, ACMP saw throughput on its Springridge system decline from 197.5 to 131.4 Bcf, a drop of 33.5%. This is a massive decline. I’d be curious to see the type curves of Haynesville wells.
Going forward, the decline will likely slow down (like other shale wells) but it seems to me that ACMP will likely lose money on its $500M purchase of Springridge (the deal closed at the end of 2010) from Chesapeake Midstream Development. Here’s how I see the numbers working out.
- In YE2010, Springridge generated $2.082M in revenue. (Revenue is low because the acquisition closed at the end of the year.)
- In YE2011, Springridge generated $93.107M in revenue.
- In YE2012, Springridge generated $65.144M in revenue.
- I could not figure out the revenue for YE2013 as the acquired Mansfield Gathering System was also included in Haynesville numbers.
Suppose we assume a 100% EBITDA margin, 10% decline in EBITDA, 20 years of cash flow, and a 5% discount rate (ACMP borrows money at 4-6%+). The net present value would be $430M, less than the $500M paid.
If we assume a more realistic 56.5% EBITDA margin, then the net present value is $253M as shown below:
My assumptions are likely off but the big picture is that this deal didn’t make much sense for ACMP shareholders. For there to be a return of capital, Chesapeake would need to drill a lot of wells in the Haynesville to grow Springridge volumes. However, the ACMP YE2013 10-K states:
We continue to see a trend by our producer customers of shifting drilling activity from dry gas shale plays, such as those in the Barnett Shale and Haynesville Shale regions, to NGL-rich plays, such as the Eagle Ford region, Marcellus region, Niobrara region, Utica region and Mid-Continent region. We believe this trend is likely to continue for the foreseeable future.
The other thing I don’t like about the deal is that the minimum volume commitments aren’t protecting ACMP on the downside (it was not triggered in YE2011 or YE2012). It could be that ACMP is taking similar risks on its other contracts.
The company paints a very different picture about its Springridge gathering system in this presentation slide (PDF):
The presentation date was Feb 18, 2014. At that time, Springridge was out of its minimum volume commitment period while Manfield’s minimum volume commitments end in 2017. I find it a little misleading to describe the Haynesville assets as having “escalating annual minimum volume protection”. Springridge’s minimum volume protection seems to have done nothing and right now there is no such “protection”.
Does ACMP make good deals?
In my previous post on ACMP, I argued that the GP/LP structure was extremely one-sided in favour of Chesapeake. ACMP itself is a bad deal. (At the time, the company was called Chesapeake Midstream Partners and traded under the symbol CHKM.)
Since ACMP went public, it has bought assets from Chesapeake. Springridge was one of them. At the time, Chesapeake controlled the General Partner of ACMP and therefore Chesapeake had conflicts of interest. The GP/LP structure allows for this and Chesapeake is allowed to direct ACMP to buy Chesapeake’s assets. I think that the self-dealing was beneficial to Chesapeake to the detriment of ACMP shareholders. This self-dealing will not continue in the future as Chesapeake sold its GP and LP interests in ACMP to funds affiliated with Global Infrastructure Partners in June 2012.
Is ACMP well managed?
Obviously I don’t think that it is. For example, G&A as a percentage of revenue continues to climb.
It would be nice if I could model out ACMP’s future cash flows based on their contracts. Unfortunately, the contracts filed on SEC EDGAR are heavily redacted as mentioned earlier. Estimated type curves of the serviced wells are not published. I do not have a good model for ACMP’s future cash flows. I’m simply taking a guess that they will be much lower than what they are today.
One quick and dirty way of valuing ACMP is to take the company’s adjusted EBITDA figure, assume some decline rate, and to model out a net present value of the future cash flows. Then I subtract debt and ignore the little cash on the company’s balance sheet. I assume that the company won’t pay any taxes (it pays close to zero tax anyways) so that adjusted EBITDA and cash flow are interchangeable. I add the “expense for non-cash equity awards” back into adjusted EBITDA. This yields a valuation of $2.2B as shown below:
I am probably undervaluing ACMP here. ACMP may very well sell more stock to make more acquisitions like it has done in the past. This will increase ACMP’s value. Currently, GIP II (a fund managed by General Infrastructure Partners and part-owner of ACMP’s General Partner) is selling its ACMP limited partner shares in secondary offerings. Underwriters do not like to flood the market with too much supply of a stock so GIP’s selling will temporarily impact ACMP’s ability to sell shares. The future could change and I expect ACMP to sell shares if they remain so high.
Another scenario is if I am underestimating ACMP’s growth in the short term. More shale drilling will increase ACMP’s cash flows. If I simply double the starting EBITDA figure (and erroneously assume zero taxes), then the NPV rises to $7.6B. I feel that a reasonable valuation for ACMP would be between $2.2B and $7.6B. (You could make an argument for a valuation below $2.2B but from a short selling perspective it’s not that important.) The current market cap is $10.79B. I wouldn’t take my spreadsheet that seriously because I am not comfortable valuing ACMP given that their contracts are heavily redacted.