While refiners as variable distributions MLPs is a relatively new phenomena (Northern Tier (NTI), CVR Refining (CVRR) and Alon USA Partners (ALDW) are all less than a year old), performance of all three since their IPOs has been exceptional. Through March 13, NTI is up 126% since its July 2012 IPO. ALDW is up 63% since its November 2012 IPO. And CVRR is up 30% since its January 2013 flotation. And that's just their capital appreciation and doesn't include their substantial dividends.
The benefits of the MLP structure are straightforward and material. Instead of paying taxes at both the corporate and investor level on profits, taxes are just paid at the investor level - in effect avoiding double taxation of profits.
Given the success that variable distribution refining MLPs are having in the markets, it is worth walking through the potential value that would be unlocked in HollyFrontier (HFC) stock were to convert to a variable distribution MLP (how it would do this - either as an exchange of MLP units for stock or as an IPO of a refining MLP - we'll leave up to the bankers).
Part 1 - Valuing The Distributable Cash Flow
Assuming that HFC can deliver the same pretax earnings in 2013 it delivered in 2012 (First Call consensus calls for a 53% increase in Q1 Y-over-Y, but a decline for full year 2013 EPS of 18% - which may prove overly conservative), that's $2.8bn in pretax income. Add to that ~$240mm of D&A and subtract the company guidance for capex in 2013 of $450mm and you arrive at ~$2.6bn of distributable income to an MLP investor in 2013.
The question is what value would investors pay for that annual distribution? What yield would investors require? One could, and should arguably, bestow on HFC a lower yield than any of the other similar MLP's due to its size and diversification. HFC is not a one refinery MLP (like Alon USA Partners or Northern Tier) and HFC is not a two refinery MLP (like CVRR). HFC has five refinery complexes located all across the mid-continent and is double CVRR's size in terms of refining capacity. Its refineries as a whole have the highest nelson complexity ratings of any independent refiner (rating over 12 across its system) which arguably allows HFC to take in the most inexpensive crudes and refine them to the most highly valued products.
NTI's yield based on its most recently announced distribution annualized = 16%
CVRR yield based on the low end of their distribution guidance for 2013 = 17%
ALDW yield based on their most recent dividend is = 8.7%
Let's conservatively assume that HFC would be bestowed a 15% yield on its distributable earnings from above (lower yields would only imply higher valuations).
Value of those earnings to HFC shareholders at a 15% yield = $17.3bn of market value or $85 per share
Part 2 - Valuing HFC's Interest In Holly Energy Partners (HEP)
This is more straightforward. HEP is HFC's former logistics business. HFC still owns 42% of the public company and the GP interest. In HFC's investor presentation (investor.hollyfrontier.com/events.cfm) the Company takes the market value of HEP, ascribes a value of the GP interest and comes up with a valuation with the mid-point of the range being ~$1.35BN.
Add $1.35 to $17.3bn for the value of the distributable earnings and you arrive at $18.65bn of market value or $92 per share.
Part 3 - Valuing HFC's Net Cash
HFC is in a truly unique position in the refining industry, no independent refiner of any size has as much net cash (cash less debt) as HFC does. In fact many refiners have a significant net debt position. HFC has $2.4bn of cash on the balance sheet and only $470mm of recourse debt (there's $900mm of HEP debt that's just recourse to HEP that you see when looking at HFC's balance sheet - disregard it). That leaves HFC shareholders with $1.9bn of net cash. Presumably in the context of a conversion to an MLP, virtually all of that excess cash could be distributed
Add that $1.9bn to the $18.65bn and it totals to $20.55bn or $101 per share - 94% above yesterday's closing price.
Beyond the potentially dramatic uplift in the value for HFC shareholders, why would HFC be motivated to convert to an MLP? HFC's management has clearly stated (as recently as their most recent earnings call) that they plan to return excess cash to shareholder through increasing dividends and not share repurchase. Despite the lowest EV/EBITDA multiple in the industry (2.8x TTM EBITDA vs. a peer median of 4.2x) and the lowest P/E net of net cash in the industry (4.7x for 2012 vs. 6.9x for the peer median) which would arguably call for material share repurchases, management has elected not to be aggressive on share repurchase and focus on increasing dividends. They have tripled their permanent dividend since March of last year (now 30 cents vs. 10 cents) and they did five special 50 cent dividends in 2012 and already one in 2013. In addition, HFC is highly familiar with unlocking significant economics through the MLP structure with their association with and structuring of Holly Energy Partners.
Why hasn't HFC issued shares in a variable distribution refining MLP? HFC's management is conservative and variable distribution MLPs (as highlighted above) are a relatively new phenomena. HFC's management is presumably watching to see how these MLPs behave and the market's acceptance of them. Further, there's a broad question in the market about how much appetite there is for variable distribution MLPs. This author is unaware of any other variable distribution MLPs that have $20bn of market value or a multi-billion dollar float. But if CVR Energy (CVI) and CVR Refining's recent success and acceptance in the market are any indication, HFC's IPOing a ~15-20% stake in a a theoretical named "Holly Frontier Refining LP" would be very well received (and potentially unlock significant value for HFC). What would be even more interesting for shareholders is if an offer to exchange MLP units for common stock could be made.
In sum, the question isn't why should HFC convert to an MLP. Given management's pre-disposition towards dividends, the dramatic uplift in value it should gain through conversion and its familiarity with MLPs, the question is why shouldn't it?