Examples In MLP Due Diligence - Magellan Midstream Partners

| About: Magellan Midstream (MMP)

Summary

This is the seventh in a series of "due diligence example" articles on MLPs.

MMP is a large-cap investment grade MLP with superior safety attributes.

MMP has relatively superior distribution coverage - which indicates distribution growth will continue.

MMP has declining DCF growth - which indicates distribution growth should slow.

This is the seventh in a series where I provide my example of "What is good due diligence for MLP (energy Master Limited Partnership) investors?" For many of you, this will appear to be hyper due diligence. You can opt to weed out some of the hyper diligence to arrive at the level that feels right to you. Prune prudently. For those who have read multiple articles in this series, it is time for you to be asking what - if any - part of my due diligence process should you omit in yours.

Before I begin with the Magellan specific numbers - let's review some sector data. MMP answers the question "What does good look like?" with most of its attributes. I want to provide more color on that.

For example, you may be asking what is a good DEBT/EBITDA ratio? To some degree, it is all relative. A good ratio is "well below the credit covenants". A more utilitarian answer would parse the 'mature' MLPs something like this: Great - under 4.1x; Good - 4.2x to 4.4x; OK - 4.5x - 4.7x; Trouble - over 4.8x; Big trouble - over 5.0x. But, those numbers are true when EBITDA expectations are stable. A falling EBITDA expectation can turn a good ratio into a troubled ratio. And when a newly constructed asset hits the balance sheet with earnings generation that is well below expectations, the same bad thing can happen.

What is "Interest Coverage" and why is it important? The ratio is EBITDA dividend by interest expense. It is a second way to answer the question of how manageable is the debt load. Good is over 6x. Bad is under 4x.

What is RRR? It is "Required Rate of Return". It is a 'dividend discount model' variable this is - for most brokerages - a part of the calculation of a target price.
To this investor, it is a valuation concept you should pick up via observation over the years. Why do "yield + CAGRs (analyst projections of forward Compound Annual Growth Rates of the distribution)" vary? Because of different levels of risk or 'regular oscillations' in a company's earnings. For REITs (Real Estate Investment Trusts) - the RRR is the portfolio cap rate. Why do MOB (Medical Office Building) heavy health care REITs sell at higher valuations (or lower yield + CAGRs) than REITs owning senior housing? MOBs have lower cap rates.
Why do senior housing REITs sell at higher valuations (or lower yield + CAGRs) than REITs owning skilled nursing facilities? Senior housing properties have lower cap rates.
Why do BDCs (Business Development Companies) with lower 'portfolio weighted average yields' sell at higher valuations (or lower yield + CAGRs) than BDCs with higher portfolio weighted average yield'? Lower yields exist due to lower default risks.
Why do MLPs with higher percentages of fee based income sell at higher valuations (or lower yield + CAGRs) than MLPs with higher commodity sensitive income? There is lower risk of extreme income oscillations.
What can one do to account for the expected variation in the "yield + CAGR" numbers? You set variable "Required Rates of Return" based on the risk level.

A really new investor might be asking if there is any evidence that concepts like CAGRs and RRRs have any influence on MLP valuations. Let's look at that MLP 'sector wide' data (meaning the coverage universe includes large cap, small cap, gathering and processing and refinery logistic MLPs) by parsing the data by the RRR related data point of credit rating:

For MLPs with corporate credit ratings of "BBB+ or BBB" - their average yield is 7.00%; CAGR is 5.65%; Price/DCF is 14.11; Analyst Rating is 2.39; and Accuracy Rating is 1.80.
For MLPs with corporate credit ratings of "BBB-" - their average yield is 7.18%; CAGR is 5.97%; Price/DCF is 11.46; Analyst Rating is 2.27; and Accuracy Rating is 2.17.
For MLPs with corporate credit ratings of "BB+, BB or BB-" - their average yield is 7.74%; CAGR is 5.10%; Price/DCF is 11.41; Analyst Rating is 2.36; and Accuracy Rating is 2.26.
For MLPs with corporate credit ratings of "B, B+ or B-" - their average yield is 10.53%; CAGR is 4.00%; Price/DCF is 8.09; Analyst Rating is 2.45; and Accuracy Rating is 3.38.

The "analyst rating" comes from Yahoo Finance - where the lower the rating, the better the rating. The "accuracy rating" is my "historical DCF projection accuracy rating" - which has been fully explained in prior articles. On average, MLPs with better credit ratings sell at lower yield and higher Price/DCF valuations. There appears to be a high correlation between the accuracy rating and the credit rating. This is evidence that credit ratings are based on more attributes that credit metrics alone. If one closely reads the updates by the rating agencies, then you are already aware of that. I also see a correlation between credit rating and analyst rating.

Let's look the at sector wide data by parsing the data by my CAGR projection:

For MLPs with CAGR projections over 8.5% - their average yield is 5.11%; Price/DCF is 15.13; Analyst Rating is 2.19; and Accuracy Rating is 1.94.
For MLPs with CAGR projections under 8.5% but over 5.9% - their average yield is 5.78%; Price/DCF is 12.40; Analyst Rating is 2.26; and Accuracy Rating is 1.56.
For MLPs with CAGR projections under 6.0% but over 3.5% - their average yield is 7.59%; Price/DCF is 10.45; Analyst Rating is 2.28; and Accuracy Rating is 2.73.
For MLPs with CAGR projections under 3.5% - their average yield is 10.81%; Price/DCF is 9.49; Analyst Rating is 2.62; and Accuracy Rating is 3.14.

On average, MLPs with higher CAGR projections sell at significantly lower yields and significantly high Price/DCF valuations. I also see some correlation between the CAGR projection and the analyst rating.

One last question: What is the difference between an RRR assessment of 10.0 and 12.5? Or put differently, does such an assessment difference matter? The answer - it matters a lot. Let's say you are comparing two MLPs with roughly equal CAGR projections with two different risk assessments of 10.2 and 12.5. By my penning a specific number on my risk assessment, I have effectively said that the lower risk option merits selling at 250 bps lower of a yield. I also generate different constants in my Price-implied CAGR formula (a metric discusses in prior articles) for those different risk levels.

Let's look at the year to date numbers and pry loose their message on MMP:

MLP Midstream 6-17-16
The consensus Distributable Cash Flow or DCF projections were last updated on 5-30-16. The distribution Compound Annual Growth Rate or CAGR projections were updated 5-30-16. Yields are based on the Q2-16 distribution. Under the 'year to date' header, the change in the distribution is the change since Q2-15 - or the change over the last twelve months. The change in the target, EPS and DCF is the percentage change in the consensus 2016 projection that has happened since the beginning of 2016. The Dist/DCF number is the ratio of the Q2-16 distribution to the 2016 DCF projection. The 2016 DCF projection is an adjusted average of eleven DCF projections from the major brokerages covering MLPs. The CAGR is intended to be the percentage change in my CAGR projection since the beginning of the year - but I lack confidence that I have the timing of this stat right. The target prices and EPS projections are from Yahoo Finance.

Large Cap Midstream

Current Distrib/ Q2 Dist Dist/dcf Dist/dcf Year-to-Date Percent Change
Company name Price Quarter Yield 2016 2017 Price Pr+Dist EPS Target DCF Dist* CAGR
Buckeye Partners, L.P. (NYSE:BPL) 69.12 1.1875 6.87 89.79 86.84 4.79 8.39 0.24 -2.19 0.38 3.26 0.00
Enable Midstream Partners (NYSE:ENBL) 14.34 0.3180 8.87 102.58 100.16 55.87 62.78 -4.88 6.55 -10.79 1.76 -97.50
Enbridge Energy Partners, L.P. (NYSE:EEP) 21.99 0.5830 10.60 105.52 100.95 -4.68 0.37 -40.34 -28.99 -3.07 2.28 -86.84
Enterprise Products Partners L.P. (NYSE:EPD) 27.72 0.3950 5.70 77.83 74.18 8.37 11.45 -6.85 -5.09 -2.87 5.33 -3.64
Energy Transfer Partners, L.P. (NYSE:ETP) 38.56 1.0550 10.94 107.93 91.94 14.32 20.58 -62.76 -19.96 -18.71 3.94 -66.10
Kinder Morgan, Inc (NYSE:KMI) 18.04 0.1250 2.77 23.26 21.83 20.91 22.59 -6.67 -3.75 -5.70 -73.96 0.00
Magellan Midstream Partners LP (NYSE:MMP) 74.43 0.8025 4.31 80.05 72.46 9.55 11.91 -5.69 -0.48 1.01 11.85 -10.11
Oneok Partners, L.P. (NYSE:OKS) 37.92 0.7900 8.33 100.32 96.93 25.85 31.10 13.20 13.05 3.28 0.00 50.00
Plains All American Pipeline, L.P. (NYSE:PAA) 26.75 0.7000 10.47 137.93 112.45 15.80 21.86 -31.67 -24.07 -24.54 2.19 -96.67
Spectra Energy Partners, LP (NYSE:SEP) 46.27 0.6338 5.48 81.00 74.56 -3.00 -0.34 6.73 1.53 10.99 3.85 7.69
Sunoco Logistics Partners L.P. (NYSE:SXL) 27.22 0.4890 7.19 87.32 72.99 5.91 9.72 -23.26 37.98 -3.03 16.71 -2.11
Williams Partners L.P. (NYSE:WPZ) 32.54 0.8500 10.45 90.67 85.00 16.84 22.94 -48.65 0.00 -2.34 0.00 -33.33
Large Cap Average 7.67 90.35 82.52 14.21 18.61 -17.55 -2.12 -4.62
The Alerian MLP index ETN AMJ is 6.21% and with dividends is 8.06%.
The S&P500 index ETF SPY is 1.30% and with dividends is 1.81%.
The Russell 2000 index ETF IWM is 1.28% and with dividend is 1.57%.
Click to enlarge

MMP has a year to date "Price plus Distribution" of 11.91% - which is well below average for this MLP sub-sector. The current yield is 4.31% - which is significantly below this grouping's average. MMP has under performed while having smaller than average declines in its EPS projection, a smaller decline in the price target, an increase in its DCF projection, and a moderate fall in its five year forward distribution CAGR projection as a result of company guidance.
MMP is one of seven MLPs in this grouping that does have a 2016 DCF projection that is higher than their current distribution. One of those seven have current yields that are higher than 10% - which is WPZ, a stock facing increased uncertainty due to WMB being in parent or general Partner.
With the WTI rising, 2016 has been a "risk on" year. MMP is built to outperform in risk neutral and risk off years. While MMP is losing to the average large cap midstream stock, it is till beating the Alerian MLP index.

The spreadsheet that follows summarizes key metrics in the income statement and balance sheet since Q1-13:

Q1-2016 Q4-2015 Q3-2015 Q2-2015 Q1-2015 Q4-2014 Q3-2014 Q2-2014 Q1-2014 Q4-2013 Q3-2013 Q2-2013 Q1-2013
MMP Operating Margin
Refined Products 170.863 203.840 241.570 148.190 183.421 251.975 200.493 162.704 255.033 209.183 146.805 177.802 160.195
Crude 100.697 95.190 94.667 106.858 84.650 88.766 70.325 73.479 63.260 62.661 50.624 40.480 22.655
Marine Storage 27.355 29.243 31.418 29.695 27.928 31.307 27.018 26.365 28.438 23.593 24.461 32.861 25.283
Click to enlarge

Magellan Midstream Partners metrics: 57% Refined Products; 34% Crude; 9% Marine Storage. 85% fee based.

Q1-2016 Q4-2015 Q3-2015 Q2-2015 Q1-2015 Q4-2014 Q3-2014 Q2-2014 Q1-2014 Q4-2013 Q3-2013 Q2-2013 Q1-2013
Revenues 519,816 573,048 577,232 487,543 522,090 667,070 521,601 496,446 618,606 577,438 443,835 443,912 432,421
Costs 321,466 343,502 289,670 304,322 311,869 387,331 297,409 320,183 343,983 355,081 291,586 260,191 291,967
EBITDA 270,132 315,529 292,017 281,613 284,934 303,559 231,438 247,989 297,894 289,791 192,051 217,342 166,264
DCF dollars 205,337 256,897 230,004 222,829 233,124 248,055 183,386 195,846 253,182 236,570 141,057 168,187 123,906
Units 227,849 227,781 227,427 227,631 227,525 227,426 227,068 227,288 227,141 227,068 226,679 226,679 226,679
DCF/unit $0.9012 $1.1278 $1.0113 $0.9789 $1.0246 $1.0907 $0.8076 $0.8617 $1.1146 $1.0418 $0.6223 $0.7420 $0.5466
Upcoming Distrib. $0.8025 $0.7850 $0.7625 $0.7400 $0.7175 $0.695 $0.6675 $0.6400 $0.6125 $0.5850 $0.5575 $0.5325 $0.5075
Coverage 1.1230x 1.437x 1.326x 1.323x 1.428x 1.569x 1.210x 1.346x 1.820x 1.781x 1.116x 1.393x 1.077x
2016 Guidance: EBITDA of $1.166 billion and DCF of $910 million
Broker1 DCF/unit $1.13 $1.01 $0.98 $1.02 $1.09 $0.81 $0.86 $1.12
Broker2 DCF/unit $0.901 $1.128 $1.01 $0.980 $1.025 $1.091 $0.808 $0.862 $1.115
Broker3 DCF/unit $0.90 $1.13 $1.01 $0.98 $1.02 $1.09 $0.81 $0.86 $1.12
LTM EBITDA 1,172.935 1,172.003 1,106.619 1,101,544 1,067,920 1,080.880 1,067,112 1,027,725 997,078 865.488
Long Term Debt 3,800,000 3,400,000 3,407,114 3,326,936 3,183,750 2,982,895 3,303,707 2,910,496 2,435,316 2,435,316 2,486,715 2,390,021 2,391,718
Debt/EBITDA 3.24x 2.90x 3.08x 3.08x 2.98x 2.76x 3.10x 2.83x 2.44x 2.81x
For MMP, even the credit facility expense is included in the long term debt numbers
Cash 210.000 28,700 9,007 29,030 52,813 17,063 14,853 731 196,630 25,235 14,228 119,491 221,003
Net Debt 3,590,000 3,371,300 3,398,107 3,297,906 3,130,937 2,965,832 3,288,854 2,909,765 2,238,686 2,410,081
Net Debt/EBITDA 3.06x 2.85x 3.07x 2.99x 2.93x 2.74x 3.08x 2.83x 2.24x 2.78x
Interest Expense 34.151 35,421 34,993 37,265 34,100 34,855 39,779 39,756 36,416 35,168 31,852 31,720 31,723
Interest Coverage 8.31x 8.91x 8.35x 7.56x 8.36x 8.71x 5.82x 6.24x 8.18x 8.24x 6.03x 6.85x 5.24x
In March of 2015 MMP issued $250 million of 3.20% notes due 2025 and $250 million of 4.20% notes due 2045
In March of 2014 MMP issued $250 million 5.15% notes due 2043
In October of 2013 MMP issued $300 million of 5.15% notes due 2043
Click to enlarge

Since the beginning of 2014, LTM (last twelve month) EBITDA numbers are up (1,172.935/865.488) 35% while the unit (MLPs call their "shares" units) count has grown 0.3% and interest payments on a growing amount of debt has fallen. If you are reading this series of article and have not read the ones on ETP and PAA - you should pause and do so. That last set of metrics was amazing.
DCF/unit growth and EBITDA growth depends on your start date. DCF is basically flat when compared to the quarters since Q4-14. The DCF is up big over the last three years. Distribution growth has been consistently good.

What can we learn from these numbers?
1. I have verified the DCF/unit (or Distributable Cash Flow per unit) calculations with the DCF numbers from three different brokerage analysts. The DCF calculation is simple - there are no adjustment to the DCF dollars. Adjustments for 'normalization' have not been needed.
2. I have complete in confidence in my DCF calculation because my numbers always match those done by the professional analysts.
3. DCF/unit change since 2013 has been positive. Most MLPs have faced headwinds due to commodity price falls.
Let's move on to some attributes I know due to producing similar spreadsheets for many other MLPs. As you absorb the data from future articles in this series, you will have the data in which to arrive at similar assessments.
4. DCF/unit numbers by quarter are moderately volatile.
5. The 'annual' distribution coverage has been as strong as it gets. Coverage is both a safety and growth attribute.
6. The debt metrics for MMP are the role model of what good is. That produces a very low cost of debt capital.
7. In a low inertia environment, a good forward DCF projection will be close to the LTM (last twelve month) DCF number. For MMP, LTM DCF is (.90 + 1.13 + 1.01 + .98) $4.02/unit. 2016 EBITDA guidance of $1.166 billion is slightly below - but almost equal to - the LTM EBITDA of $1.173 - which tells us with certainty and clarity that earnings will be flat in 2016. The 2016 DCF guidance of $910 million results in a DCF/unit projection of (910,000/227,849) $3.99 - which almost replicates the LTM number.

I have 8 2016 DCF projections from the 12 brokerages I use to create a consensus number. The projections range from $3.95 to $4.14. The average for the projections is $4.01 - and 7 of the 8 projections are within ten cents of the average. The very small spread in the analyst projections comforts me. MMP has the highest grade I give on historical DCF projection accuracy. Once again - readers should compare this text just created on MMP to the text in same section for ETP and PAA. The difference - using popular phraseology - is huge.

Now that I have a very good DCF - let's move on the CAGRs - or a five year forward "Compound Annual Growth Rate" projection of the distribution.

Long term metric trends
The average calculation for growth is for ten years - for 2006 through 2015
The first average is the sum of changes for each individual year over ten year period - with that result divided by 10
The 2nd average is the difference between the current and beginning number, divided by the beginning number - with that result divided by 10
The distributions shown are annualized first quarter distributions

Company 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 average average CAGRs PI-CAGRs
MMP DCF 1.33 1.56 1.78 1.51 1.83 2.03 2.23 2.87 3.87 4.13 4.01 4.43 My 8.00% RRR-yield 5.99%
growth 17.3% 14.1% -15.2% 21.2% 10.9% 9.9% 28.7% 34.8% 6.7% -2.9% 10.5% 13.60% 20.15% Last5 15.57% P/DCF Ratio 12.69%
Dist. 1.10 1.20 1.31 1.42 1.42 1.51 1.63 2.00 2.34 2.78 3.14 Broker1 9.10% Broker2 10.60%
growth 9.1% 9.1% 8.0% 0.0% 6.7% 7.6% 22.7% 17.0% 18.8% 12.9% 11.19% 18.42% Broker3 10.30% Broker4 11.50%
Dist/DCF 83% 77% 74% 94% 78% 75% 73% 70% 60% 67% 78% 74.62%
Click to enlarge

My median 2018 DCF projection for MMP is $4.68 and the median 2019 projection is $5.04. Presuming a 76% payout of the 2019 DCF of $5.04, the distribution would be $3.83 - and that would average (383/314 . . divided by 3 years) 7.32% growth per year. I show the high brokerage projection at 11.5% and the low one at 9.1%. I want conservative CAGRs. My 8% CAGR is pretty conservative. Given the current low yield, the market is putting a valuation (averaging my two PI-CAGRs of 5.99 and 12.69 to get 9.34%) on MMP that is over that projection.

Now that I have arrived at a growth projection with which I can have a fair amount of confidence - it is time to make a numeric risk assessment.

When it comes to historical DCF projection accuracy - MMP is as good as it gets. When it comes to credit ratings - MMP is as good as it gets. When it comes to credit metrics - MMP is as good as it gets. When it comes to distribution coverage - it is temporarily going from great towards average. The relatively low yield tells us a large percentage of the market's valuation of MMP is based on growth. I want to slightly increment the risk of owning MMP due to that attribute. Because of the sum of all attributes, I assess MMP with a 10.3% Required Rate of Return ("RRR" or risk assessment). Low RRRs are in the 10s and mainly go to BBB+ rated companies. Average RRRs for covered distributions are in the 11s and mainly go to BBB rated companies. Uncovered distributions result in RRRs in the 12s and mainly go to BBB- rated companies.

And with that - you should be sufficiently prepped for the ending spreadsheet that compares valuations:

Yield + CAGR Total Return Expectations

Company Q2-16 Consensus Total Bonds DCF My Total Rtn Consensus Price Implied CAGR Distrib Price
Yield CAGR Return Ratings Accr RRRs - RRR Ratings RRR-Yld P/DCF / DCF / DCF
Large Cap Midstream
Buckeye Partners, L.P. BPL 6.87% 4.40% 11.27% BBB- 2.50 11.00 0.27 2.1 4.13 3.82% 89.79 13.07
Enable Midstream Partners ENBL 8.87% 0.10% 8.97% BB+ 3.00 12.50 -3.53 2.6 3.63 1.97% 102.58 11.56
Enbridge Energy Partners, L.P. EEP 10.60% 0.50% 11.10% BBB 2.00 12.00 -0.90 2.9 1.40 -0.09% 105.52 9.95
Enterprise Products Partners L.P. EPD 5.70% 5.30% 11.00% BBB+ 1.30 10.00 1.00 1.7 4.30 3.06% 77.83 13.66
Energy Transfer Partners, L.P. ETP 10.94% 2.00% 12.94% BBB- 3.00 12.50 0.44 2.2 1.56 0.67% 107.93 9.86
Kinder Morgan, Inc KMI 2.77% 6.00% 8.77% BBB- 1.30 11.00 -2.23 2.4 8.23 -4.83% 23.26 8.39
Magellan Midstream Partners LP MMP 4.31% 8.00% 12.31% BBB+ 1.00 10.30 2.01 2.1 5.99 12.69% 80.05 18.56
Oneok Partners, L.P. OKS 8.33% 1.50% 9.83% BBB 3.00 11.50 -1.67 2.9 3.17 1.00% 100.32 12.04
Plains All American Pipeline, L.P. PAA 10.47% 0.10% 10.57% BBB 3.00 12.50 -1.93 2.7 2.03 6.80% 137.93 13.18
Spectra Energy Partners, LP SEP 5.48% 7.00% 12.48% BBB 1.00 10.30 2.18 2.2 4.82 5.70% 81.00 14.78
Sunoco Logistics Partners L.P. SXL 7.19% 9.30% 16.49% BBB 1.10 12.00 4.49 2.1 4.81 3.98% 87.32 12.15
Average 7.41% 4.02% 11.43% 2.35 4.01% 3.16% 12.47

Click to enlarge

"Total Return minus RRR" is my the buy, hold or sell number. Positive is buy - the stock is selling below what the valuation assessments suggest. Negative is sell - the stock is selling above what the valuation assessment suggest. Close to zero means the stock is correctly priced. In most sectors there is superior metric transparency compared to MLPs - and I produce TR - RRR numbers that are close to zero. Projected Total Return or "TR" = Yield + CAGR.

What does all of this math mean?

MMP is one of six MLPs where my "Yield + CAGR - RRR" numbers are higher than zero. Six out of six of those MLPs have Yahoo Finance sourced analyst "consensus ratings" that are lower (or better) than average. I may have low balled my MMP CAGR projection at 8% while my formulas are split on whether MMP is currently priced for a lower or higher projection. Based on the numbers, MMP is a buy.

This is a time where MLPs with superior safety attributes have relatively high valuations. One should expect some short to mid term under performance due to those valuations.

Conservative investors should want to invest predominantly in stocks where they have "faith and confidence" in the CAGR projections. That label fits MMP. That attribute should influence any weighting decision you would make on an MMP purchase.

One last comment: In my write-up on EPD, I provided the hyperbole that if you are retired or nearing retirement (my target audience) - and "if EPD, MMP and SEP (the large-cap MLPs with great balance sheets) are not your #1, #2 and #3 MLP holdings - you are probably doing something wrong. Or - if more than 30% of your MLP holdings have current yields over 7% - you are probably doing something wrong."

That comment was reader tested in a poll of comment likes at the end of the article. 82 readers agreed with that comment. 8 disagreed. Thus this site generated 90 reader poll participants out of the 10,313 that read the article. It may be the case that an internet site on investing skews towards a younger audience - thus phrasing the question for retirees assisted in a low response rate. Article like this could be too brain taxing - and reader fatigue contributed to a low response rate. But... I have the perception that my readers would have greater faith and confidence if the reader polls generated a response rate closer to 500. (And I think that most Seeking Alpha content should have polls that 'reader tests' the articles main thesis.) Can you make that higher participation rate happen?

Disclosure: I am/we are long EPD, MMP.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.