|Shares Out. (in M):||83||P/E||0||0|
|Market Cap (in $M):||3,750||P/FCF||0||0|
|Net Debt (in $M):||1,445||EBIT||0||0|
|Borrow Cost:||General Collateral|
Mastec is a specialty E&C company that operates across 4 segments: Oil & Gas, Communications, Electrical Transmission, and Power Generation. Mastec was previously written up on VIC as a long in early 2016. That call has proven well-timed as its Oil & Gas pipeline segment’s explosive organic growth enabled it to generate outsized financial results, leading to substantial stock appreciation. The company anticipates current Oil & Gas pipeline results will prove sustainable at current levels while robust 5G spending will enable Communications to drive the next leg in the company’s EBITDA growth.
Yet a combination of secular enthusiasm, promotional management, and a seemingly reasonable relative valuation have induced thematic investors to substantially underestimate structural defects in Mastec’s business model. This same emphasis on high level drivers has also blinded the analytical community to pervasive accounting and related party abnormalities.
Our analysis will focus on the two most important segments – Oil & Gas and Communications – before diving deeper into the accounting/related party activities/ legal history that make Mastec a compelling short.
We believe that Mastec is trading at an elevated multiple on elevated (and potentially overstated) earnings. As such, we think any potential upside (downside to a short) should be manageable while the downside (upside to short) could prove quite dramatic should the concerns articulated in this right prove accurate.
Oil & Gas (53% of FY ’17 Segment Revenue; 57% of FY ’17 Segment EBITDA)
This division has been the primary driver of Mastec’s Segment EBITDA growth in recent years.
|Oil & Gas||157||304||402|
|Segment EBITDA||$ 319||$ 523||$ 690|
|Oil & Gas||49%||58%||58%|
As shown below, Mastec has spent $650 million (including contingent consideration) on six Oil & Gas acquisitions over the last decade. $370 million was spent on four US acquisitions while $280 million was spent on two Canadian acquisitions. These acquisitions were added to Mastec’s small, pre-existing gas gathering operations.
|Nov-09||Precision Pipeline||132||34||40||206||Large Diameter, Long Haul||Wisconsin|
|Dec-12||Bottom Line Services||68||-||11||79||Midstream||Texas|
|Dec-12||Go Green Inc.||25||-||-||25||Midstream||Texas|
|$ 269||$ 43||$ 51||$ 363|
|$ 232||$ -||$ 49||$ 281|
|Total O&G Acquisitions|
|Total||$ 501||$ 43||$ 100||$ 644|
The Canadian acquisitions proved ill-time as deteriorating results soon forced a material asset write-down. While upside arguably exists from Canadian operations, they currently represent just 5% of O&G revenue, rendering them immaterial to overall O&G results.
In the US, the outcome has been much better. Initially, management’s aim was to build a balanced pipeline business comprised of Mastec’s legacy gathering operations, Pumpco’s Midstream business, and Precision Pipeline’s higher beta, long-haul, large diameter pipeline projects. However, Precision Pipeline has since stolen the show and currently accounts for ~76% of Oil & Gas EBITDA and an even greater % of its EBITDA.
Precision’s focus is in large, higher risk, higher margin projects that typically span several states. In recent years, Precision has been Energy Transfer’s largest contractor on the Dakota Access Pipeline ($3.8 billion project; crude oil from Bakken) and Rover Pipeline ($4.2 billion project; natural gas from Marcellus/Utica). It was recently awarded a large contract on EQT’s Mountain Valley Project ($3.7 billion; natural gas from Marcellus/Utica).
The impact of these large projects can best be seen by the increased % of O&G revenue from Energy Transfer Affiliates. As shown below, Energy Transfer went from ~ 20% of O&G revenue in ’15 to ~ 76% of O&G revenue in ’17.
|Energy Transfer Affiliates||277||295||1,386||2,643|
|Total Oil & Gas Revenues||$ 1,731||$ 1,495||$ 2,024||$ 3,497|
|Energy Transfer/ Total Oil
These large contracts have turbocharged growth. Yet long-haul, large diameter pipelines have long economic lives and minimal recurring revenue opportunities. This leads to violent cyclicality. It is for this reason that its larger competitor Quanta has carefully endeavored to grow its more recurring revenues – “base business” -- within the Oil & Gas division so the company “won’t be so reliant on this large diameter pipe as we move out past 2020 and then beyond, because that (large diameter pipe) is the market that is cyclical.”
Mastec argues that there is tremendous short-term visibility, but this does not change the need for earnings to be materially reset at some point in the future, with the underlying earnings power of this business dramatically lower than current levels. Mastec has fully embraced outsized exposure to an extremely cyclical end-market as it experienced unprecedented strength. This has enabled its pipeline division to dramatically outperform its competitors. However, it also makes it perilous to capitalize current O&G EBITDA at anything close to a normal multiple.
From a larger perspective, Oil & Gas now contributes ~ 60% of Mastec’s Segment EBITDA with 80% of Oil & Gas EBITDA derived from one-off projects in a notoriously cyclical end-market closer to the top than the bottom. This exposure has materially altered Mastec’s overall business risk profile. This change can be seen in the shift in revenues away from more stable Master Service revenue to Installation/ Construction revenues.
The long-haul strength has also not prevented Mastec from experiencing operational challenges on its current work. In particular, its work on Rover has led the Republican head of the Ohio EPA to sue Pre-tec Directional Drilling (Pre-tec is 100% owned by Precision). This was followed by similar lawsuits from the West Virginal EPA. As a result, the timeline for Rover has been pushed back roughly ½ a year from Q4 ’17 to Q2 ’18. Mastec does not discuss individual projects, but the accusations do seem consistent. These operational challenges likely explain some of the deterioration in working capital metrics discussed later.
Communications (37% of FY ’17 Segment Revenue; 35% of FY ’17 Segment EBITDA)
Mastec’s communication division has 3 components; wireless (~ 50% of revenue), wireline (~ 25% of revenue), and “install-to-home” (primarily DirecTV installation and ~ 25% of revenue). AT&T accounts for 70% of Communication revenues. This concentration partly reflects the merger of its two largest customers (AT&T and DirecTV) in 2014.
Mastec’s Communications segment is anticipated to experience outsized growth relative to recent trends on the back of several drivers - 5G spend by wireless operators, FirstNet spend by AT&T, and 1 Gigabit Fiber roll-outs by both telco and cable operators. The company expects growth to be as robust as the multi-year double digit CAGR’s experienced with the roll -out of 4G.
Mastec’s view that 5G will prove as strong as 4G is not universally shared. Morgan Stanley, for one, expects carriers’ 5G infrastructure spend will drive 6 year CAGR of +1% from ’19-’25 vs. a four year CAGR of +3% experienced for 4G from ’10-’14.
But even if we accept Mastec’s premise that 5G will equal 4G, other factors will blunt the impact this would have on Mastec’s results.
First, Communications EBITDA is ~ 36% of Segment EBITDA today vs. 46% of Segment EBITDA in ’10 (prior to 4G upcycle), thereby reducing Mastec’s overall exposure to an upswing in Comunications.
Second, “install-to-home” accounts for ~ 25% of Communications revenue. This will be stagnant (if not decline) over the next few years. During the previous 4G cycle, “install-to-home” did not weigh down Communications because DirecTV was still growing its subscribers. This has now reversed, as shown below.
Third, fiber is slated to play a larger role in 5G than it did in 4G. While Mastec once did have an overweight position in wireline, it has spent the last decade de-emphasizing wireline as it sought to become dominant in wireless (tower crews). This is reflected in the outsized contribution that wireless now makes to its Communications revenue. We think the recent wireline acquisitions are a belated attempt to address this imbalance.
In short, any bullish argument for Mastec predicated on 5G driving enormous organic growth in Communications revenue is mitigated by the smaller size of Communications relative to total company than it was the recent past, the stagnation of 25% of Communications Revenues, and Mastec’s relative underexposure in wireline vs. wireless.
Earnings Quality and Revenue Recognition
As shown below, Mastec experienced a substantial jump in DSO’s at the end of FY ’17 both on an annual and quarterly basis.
However, Mastec’s non-recourse receivable factoring implies an even greater increase. Under these “non-recourse” financing arrangements, certain receivables are settled with a customer’s bank in return for a nominal fee. These arrangements enable Mastec to “improve collection of related receivables.” This boost in operating cash flow has the added benefit of not penalizing “adjusted EBITDA” (because the discount fee paid to banks is recorded below the line).
Mastec began this activity in 2014. In every 10-K from 2014-2016, Mastec disclosed both the amount of receivables factored and the discount fees paid to banks. The 2017 10-K omitted the amount of receivable factoring, but noted that discount fees increased 120% from $2.7 million ’16 to $6 million in ’17. So while the precise amount of receivables factored in ’17 cannot be determined, the outsized jump in discount fees coupled with company’s omissions suggests a material increase from 2016 levels. The table below provides the historical data on receivables from 2014-2016 and the discount fees paid from 2014-2017. From this an estimate of the amount of receivables sold can be made.
The important point here is that already challenged DSO trends were likely much worse than they optically appear.
The composition of the AR jump is also alarming with most of the increase attributable to outsized growth in unbilled AR and change orders.
Unbilled AR reflects revenue recognized by a contractor for work which they deem complete but for which the customer has not yet been billed. The contractor’s revenue estimate is determined by dividing total project costs incurred by its own estimated project costs. A surge in unbilled AR can suggest a difference between internal estimates of completion costs and external realities of the project. A company that underestimates completion costs will overstate revenues (and gross profit). While adjustments are supposed to be made concurrent with cost over-runs, managements have a wide degree of latitude in assessing when such an over-run has occurred. The beauty of POC accounting for unscrupulous (and growing contractors) is that it enables them to both overstate (and frontload) profitability, while enabling them to defer economic reality until a later period (after bonuses paid, shares sold, etc.)
“Change Orders” are revenues recognized for costs that fall outside the immediate scope of the initial customer/contractor agreement but for which the contractor anticipates compensation. The issue here is that contractors might book revenue even if substantial disagreement might exist over responsibility for the incremental costs.
In Mastec’s case, these two potential areas of abuse are not theoretical. Mastec has faced two shareholder lawsuits alleging accounting impropriety in the last fifteen years. The first (2003) focused on the abuse of change orders while the second (2015) focused on the abuse of POC accounting and unbilled AR. Both lawsuits alleged that the company used these accounting conventions to overstate results and inflate profitability.
Moreover, the last time Mastec experienced a similar jump in DSO’s, Unbilled DSOs, and change orders was late 2013/ early 2014. This preceded a protracted period of earnings shortfalls and stock weakness. The initial cause of the decline was weakness in AT&T cap. ex. and Canadian Oil & Gas. However, Mastec soon started booking enormous losses in Electrical Transmission Segment.
This sharp reversal in Electrical Transmission performance occurred after an accounting investigation was launched into that division’s use of percentage-of-completion accounting. The audit resulted in only minor tweaks to historical Electrical transmission results. And Mastec cited end market weakness, “project fade”, and distractions caused by the audit itself to explain the sharp reversal in Electrical Transmission performance.
While the Electrical Transmission division is small, its issues are a microcosm of larger issues that might ultimately reveal themselves at the company’s Oil & Gas division. To that end, we found a 2015 shareholder lawsuit (focusing on the alleged accounting impropriety in the Electrical Transmission division) to be extremely illuminating and worth of discussion.
Electrical Transmission Accounting Allegations
This lawsuit – which contains the testimony of several employees of the electrical transmission division at the time the alleged impropriety occurred – accuses Mastec of “altering cost-to-complete estimates and improperly allocating costs to unrelated projects.” This was done “to defer or avoid recognizing losses on certain projects that were unprofitable when actual costs exceeded “estimated” costs.”
Beyond just underestimating completion costs, Mastec also manipulated timing of revenue recognition by “ordering all of the materials too early when these materials would not have been used until much later phase of the contract. This practice created the appearance, through consideration of purchase orders and invoices, that certain required costs for a certain project had been incurred prior to material use or utility of such materials in the timeline of the project, allowing the company to prematurely recognize revenue.”
Additionally, the lawsuit alleges that Mastec “shifted cost over-runs incurred on certain unprofitable contracts and construction projects, to other unrelated projects that could absorb the costs and mask the losses.”
In short, these employees say that Mastec used POC accounting to both inflate project revenue and project gross profit by underestimating project cost. Furthermore, they say that the inevitable losses from these contracts would be absorbed into larger projects where (presumably) generous profitability assumptions could mask the legacy losses.
Consistent with this interpretation, these employees found Mastec’s explanation for its sudden Electrical Transmission losses absurd, and insisted the culprit was really “unwinding of the financial improprieties” that been built up over the years when ET was reporting record results
Management, on the other hand, still maintains that they did nothing wrong. Even the small historical accounting adjustments, they said, were done only to placate the SEC. They blame “former SEC officials living in $2 million houses looking to cause trouble” for the issue. And they insist that the decline in ET profitability in 2015 was caused, in large part, by the distractions caused by the audit itself. If there were indeed accounting issues, one can be fairly confident that no lesson has been learned by the current management team.
We do not find management’s interpretation of events convincing. We find the specificity of the lawsuit’s allegations coupled with the enormity of the swing in ET’s performance to favor the plaintiff’s explanation.
Mastec massively overstated profitability in its Electrical Transmission division from ’11-’14. Yet when its backlog (and revenues) began to slow, Mastec could no longer use estimated profits to conceal actual losses. Confronted with this dilemma, the company could either come clean about the extent of past financial impropriety or attempt to let past malfeasance run through current earnings. Management could not reveal the magnitude of the ’12-’14 overstatement without seriously damaging credibility (and jobs), so they elected to make minimal retroactive adjustments, and to run the impropriety though current earnings. This decision was made easier by the small relative size of the Electrical Transmission division coupled with the clear visibility the company had into its long-haul mainline backlog at the time.
Any such unwind in a larger segment would me much more difficult to manage.
Related Party Transactions
Adding to our concerns about weak internal controls and/or malfeasance is a substantial increase in the pace of related party activity (both w/respect to vendor purchases and corporate transactions). This began over the last few years, but has accelerated in recent quarters. While any one of these might be explained away, the sheer number of them casts substantial doubt over their economic purpose.
Below are the annual and quarterly purchases from four related party vendors.
The four entities are,
(1) CCPS —Juan Carlos Mas (brother of Mastec Chairman Jorge and Mastec CEO Jose) is Chairman of this company. Jorge, Jose, and an unidentified member of Mastec management own minority stakes in the company. Mastec itself invested $15 million into the company in '13. They supply Oil and Gas Equipment.
(2) "Other Mas plus subsidiary owned entity"- Name of entity is not revealed but it is a subcontractor in which an entity owned by Jorge and Jose Mas, along with two members of management of Mastec subsidiary, own minority interests.
(3) "Related Lease Payments to Subsidiary Management”— these are payments to companies that are owned by or associated with members of subsidiary management for operational facilities and equipment. These companies typically belong to former owner/operators of acquired businesses.
(4) "Related Party Supplies and Services to Subsidiary Management" - Same as #3 above except the payments are not for operational facilities and equipment but for supplies and services.
Note the substantial increase in costs to all these entities (both on an absolute basis and relative to overall Mastec Cash Costs).
Moreover, Mastec made 2 acquisitions last year that were owned in part or in whole by management of subsidiary companies.
In Q3 ’17, Mastec acquired, "an oil and gas pipeline equipment company that was formerly owned by a member of subsidiary management for $40.6 million in cash and an estimated earn-out liability of $57.2 million. Mastec previously leased equipment from this company." In reality, Mastec was the only company that this entity leased equipment to.
In Q4 ‘17, they bought out the minority interest of management in a fully consolidated Oil & Gas subsidiary for $21.4 million and an earn-out liability of $9 million..
Finally, there are two other recent related party transaction that I cannot refrain from mentioning. They are small but revealing. In Q3 ’17, Mastec paid $2 million for 4% of common stock plus warrants to purchase additional shares in a SPAC in which Jose Mas is a director. In February 2018, Mastec “acquired a construction management firm specializing in steel building systems, of which Juan Carlos Mas was a minority owner, for approximately $5 million in cash plus contingent earn-out consideration.”
This rapid increase in related party activity is dizzying and it is impossible to know the motivations behind every single one. Yet we think there are 4 possible scenarios – none of which are positive for Mastec and only one of which is not a material negative.
All above board.
Mastec's management is expropriating wealth from non-management and non-Mas shareholders by paying above market rates for outside services owned by related parties. If, for example, Mas brothers own 40% of an outside entity against 20% of Mastec they will gain more than they lose from such a transaction. For subsidiary managers, the incentive is even more perverse. Any bullish argument that this means the company is under-earning is more than offset by the notion that one owns stock in a company where management is deliberately stealing from shareholders. This usually does not end well.
Mastec is paying below market rents to these entities. They do this to inflate Mastec's results. The bag holder here would be the outside entities? Why would they be willing to accept lower payments from Mastec than they could elsewhere? Mastec overpays them upfront? Mastec agrees to buy them out on the back end at a premium equal to the amount of cumulative underpayments?
Mastec using them as an earnings smoothing device. This would be somewhat similar to #3 above.
An additional consideration is the effects that the use of vendors owned by “members of subsidiary management” might have on that subsidiary’s incentives. The arrangement enables those managers to derive a personal benefit independent and separate from any benefits accruing to Mastec. In effect, there exists an asymmetric risk/reward in contract bidding whereby the manager will always be biased to booking contracts, no matter the longer term costs to Mastec of so doing. This calls into question the quality of Mastec’s record backlog.
We value Mastec on a SOTP basis. For reasons previously articulated, we believe that the % of EBITDA that Mastec currently derives from its long-haul, Mainline work means that the O&G multiple must be aggressively discounted.
While the 5.6X ’18 multiple underpinning out price target appears low in the current market, this partially reflects distortions caused by the outsized contribution of the pipeline business. Beyond that, we also maintain that Mastec’s fundamental business attributes – intense customer concentration, lack of recurring revenue, weak FCFF/ EBITDA conversion – suggest a low multiple is warranted. Finally, its history of past (and potentially future) accounting misstatements and blow-ups makes such a discount even more prudent.
Our timeline is 12-18 months. The immediate risk is that the strength in end-markets coupled with a recent flurry of acquisitions enable them to generate a few earnings “beats” in the short-term. We believe both the probability and extent of the eventual downside more than adequately compensates us for this risk.
(1) Eventual Slowdown in Mainline Pipeline Construction leads to huge airpocket in O&G Revenues, and material hit to EBITDA.
(2) Further accounting shenanigans is revealed, forcing another restatement of numbers. This will likely occur concurrently with #1 as growth is no longer able to paper over faulty POC assumptions.