Description

Short MACOM Technology Solutions (MTSI).  MTSI’s core business is in decline and its newer growth drivers are beginning to roll over.  Management has attempted to hide this deterioration with a promotional campaign touting several unrealistic high growth opportunities that they claim could double the size of the company, while engaging in confusing rhetoric and contradictions.  Furthermore, they have utilized a complicated revolving door of M&A to further obscure the fundamentals by understating the impact of acquisitions and overstating organic growth, in addition to resorting to aggressive and misleading accounting.  MTSI is a poster child for horrible earnings quality and has employed some of the most egregious earnings manipulation tricks of the trade.   MTSI trades at a premium multiple and appears poised to miss expectations.

MTSI is an analog semiconductor company focusing on applications like optical networking and RF.   Its Networks segment contributes 73.5% of revenues, Aerospace and Defense 12.7%, and Multi-market 13.8%.  The history of MTSI goes back to the mid 20^th century when it focused on microwave for aerospace and defense applications.   It has changed hands numerous times and gone through numerous iterations since then.  In 1999 Tyco International acquired MTSI, which was then a part of AMP Inc.   In 2008 Tyco sold this business to Cobham, which in turn sold off what is now MTSI to John Ocampo for $60 million plus a $30 million earn out.  Ocampo acquired MTSI after selling his previous company, Sirenza Microdevices, to RFMD.  Ocampo then took MTSI public in 2012, garnering ~10x the valuation he paid just three years prior, but that was just the beginning.

There are 57.8 million shares outstanding (including options, warrants, and RSUs).  At a price of $40.07 MTSI’s market cap is $2.27 billion.  There is $90.5 million of cash and $345.1 million of debt for and enterprise value of $2.53 billion.

It should perhaps come as no surprise that a company that was once a part of Tyco International uses aggressive accounting (dare we say fraud?) to promote the illusion of growth and profitability.  MTSI first came to our attention as a result of poor earnings quality.  Burgeoning inventory levels and consistently anemic free cash flow belied the strength of the presented (and adjusted) financials. 

Inventories have been steadily increasing.  On a GAAP basis, inventory days on hand were 150 in F3Q16 (6/16), up from 116 for F2014. Ironically, according to the metrics that management prefers to focus on, which exclude acquisition related amortization from COGS to inflate gross profit, inventory days are 175.  Either way, this is much too high for a healthy semiconductor company because bloated inventory increase the risk of obsolescence and write offs.  It can also indicate a lack of end market demand, which could lead to lower production volumes, which on a large fixed cost base, results in significant operating de-leverage.  As a result, free cash flow from F2013-2015 was cumulatively only $20 million.  Working capital was a consistent drain on cash flow during this time, although free cash flow generation has improved recently and for the first nine months of this fiscal year it was $30 million.  

MTSI relies heavily on non-GAAP adjusted financials.  Certainly GAAP accounting has its shortcomings, but it is generally accepted for a reason.   And as analysts we should endeavor to ascertain the true economic normalized cash flows of a business, but we take umbrage at management teams trying to force feed investors with patently disingenuous “adjusted” earnings.  MTSI’s adjusted earnings table is awe inspiring, if not unorthodox, changing from quarter to quarter sometimes approaching 100 lines in length.  The bulk of the adjustments are amortization and compensation expenses, and impairment charges.  Long story, short, through the legerdemain of adjustments in the first nine months of this fiscal year, management transformed operating income of $3.2 million into adjusted operating income of $95 million, or 24.3% of revenues.  That’s a decent margin!  Likewise for the same period, EPS was inflated from $(0.07) to $1.37.  Clearly, constant reorganization of the business is going to create a lot of one time issues that could be excluded to get a better idea of the ongoing normalized trends, but this is over the top.  First of all, compensation is a real ongoing expense and shouldn’t be added back.   Furthermore, amortization related to acquisitions is not an ongoing expense, but if the business is a roll-up that relies on M&A for its growth strategy, then it is fair to recognize a charge that reflects that capital allocation decision.  And again, if a company takes restructuring and impairment charges every quarter, they might be considered an ongoing cost of doing business, like it was for former parent Tyco, or Lucent and a slew of others. 

Management excludes all sorts of costs from their earnings calculation, but for some reason never excludes one-time events that increase earnings.   Management has a history of making questionable decisions, like accounting policy changes that boost results, along with adding non-recurring income into its adjusted earnings, and completing acquisitions to hit targets, while spring loading them to surreptitiously improve profitability.  The most glaring example of this occurred in F1Q15 (12/14) when inexplicably and without prior notice there was a revenue recognition change whereby MTSI switched from recognizing revenue on a sell through basis to distributors to recognizing revenue on a sell in basis. The change resulted in a one-time increase of $15.1 million in revenue at a strong 69% gross margin, which increased its GAAP loss per share that quarter by $0.18.  The change was not disclosed in the guidance management gave on the previous earnings call and the explanation on the F1Q15 call was characteristically vague and misleading.  Without the change of policy F1Q15 revenue would have been $99.8 million, rather than the reported $114.9 million, a fairly substantial miss to say the least.  For the full year the change ended up contributing $17.4 million in revenue.  Inventories on the company’s books were already at elevated levels, but this now raises concern about the level of channel inventories, as well.

Where there is smoke there is often fire and we believe that management is using these various tactics to conceal the declines in its core business and a lack of growth prospects.  Furthermore, adjusted earnings are inflated and unsustainable, and expectations will be severely disappointed. 

Management is very promotional about several areas they claim will be massive new markets for MTSI, like 100G optical, base station power amps with GaN on Si, and active antennas, but they always seem just around the corner and management estimates seem overly optimistic.  Management claims that each of these opportunities could potentially double the size of the company in 3-5 years, sometimes positing they could garner >50% market share of markets they are not even in yet.  Several years ago management was promising that some of these products would be ramping by now, but they continue to be pushed out.  The fact of the matter is that these “opportunities” are basically research projects that that may (or may not in some cases) be commercialized in 5-10 years or more, rather than in the near term as management claims.  There is also no assurance that MTSI will win a material portion of the business.  We are reasonably confident the net present value of these opportunities is close to zero.  The narratives of the three growth stories are almost identical.  Management talks about the potential of markets that do not exist yet and their potential explosive growth and huge market shares they can achieve.   They cite high levels of interest from customers clamoring to get product yet all there is to show for it are vague allusions to trials, work in labs, reference designs, but no qualified products and no clear path to commercialization. 

Specifically, with respect to 100G optical: management and analysts are enthusiastic about continued metro build out and converting a strong market position with service providers into datacenters.   The datacenter market could be orders of magnitude larger than the service provider market.  Management also points out that datacenter products could be replaced every 3 years, unlike service provider products which are designed to last decades.  However, the increased turnover of parts in datacenters leads to one of the biggest obstacles to 100G adoption in datacenters.  Because the products for service providers are designed to be so resilient they are much too expensive for near term adoption in datacenters.  MTSI has been investing to greatly expand the capacity of its Lowell, MA facility, but apparently has not been able to reduce the cost of unit production enough yet to get product qualified by customers.  Management made misrepresentations about making volume shipments of 100G product to datacenter customers earlier this year, but apparently MTSI does not have a commercially viable product.  Despite referring to this as an explosive growth opportunity for several years it has yet to materialize and management continues to push out expectations, most recently on their F3Q16 (6/16) earnings call, saying it is too early to know when the ramp will happen.

GaN on Si is also touted as another growth area that could potentially double the size of the company in 3-5 years.  Current wireless base station RF power amps use LDMOS technology.  Management claims that its IP with GaN will allow them to offer higher performance, but at the cost of current LDMOS products.  The market opportunity is over $1 billion annually and management thinks they could achieve a dominant market share despite there being no market for GaN currently.  >80% of the market supply is produced by NXPI/FSL and Ampleon (the legacy NXPI RF business that was disposed of to complete the FSL acquisition).  On the demand side of the equation three main players (Huawei, ALU/NOK, and ERIC) control >80% of the market, so this is a highly concentrated market with very strong incumbents where economies of scale are of paramount importance that will be extremely difficult for a new player to break into with limited volume.

That said, management claims to have purchase orders in place to deliver prototypes for 4G/LTE.  Management had said this business would be ramping in 2016, but on the F3Q16 (6/16) call they pushed this out to 2017, although this too will probably prove optimistic.  Wireless service providers have already made substantial investments in 4G infrastructure and their end markets remain very price competitive, so it is unlikely that they would change horses in midstream and embark on another large capex cycle when their current base stations works just fine, particularly as they begin to plan for the next generation. 

The GaN opportunity in 5G is also nebulous, since standards have yet to be set and commercialization remains many years off.  GS released an interesting report on the potential of 5G (in April, 2016) where they predict full commercialization of 5G to take place in 2020 and later based on the historical precedents of the roll out of earlier generations and the time line for standards and specifications to be finalized, which jibes with other industry estimates.  However, the GS analyst that covers MTSI for some reason believed it was a 2016 story and now a 2017 story, perhaps because that was the guidance that management provided, but at this point it remains just a story.  It appears likely that compounds like GaN may play an important role in 5G power amps, but there are still many technological and cost hurdles to overcome.  So an opportunity that was supposed to double the size of the company in 3-5 years and was already supposed to have started ramping, may result in the sale of some prototypes in the next few years, but the real market development appears to be five or more years away with standards and specs yet to be finalized and with powerful incumbents unlikely to just sit on their hands.

Active antenna is the third and final growth area that management theorizes could double revenues in less than five years.    MTSI is working on a Multifunction Phased Array Radar (MPAR) and has branded its own solution Scalable Planar Array Tiles (SPAR).     MTSI is attempting to cut the bill of materials in half, while dramatically improving performance compared to radars currently in the field by introducing tiles that contain thousands of transceiver/receiver channels, whereas the current system operates with a single channel.  MTSI has already licensed its technology to a lab at MIT and has an agreement with NOC.  Management claims that they are well positioned to win a sole sourced contract of >$500 million annually for five years for replacing the civilian radar system in the US, which they say could start ramping in 2017.

However, in its 2015 report to Congress, NOAA working in conjunction with the FAA characterized MPAR as an “ambitious concept” that showed promise, but “historically, the development and transition of new radar technology into operations requires 20-25 years.”  So this could possibly be a big home run for MTSI in the 2025-2030 timeframe, assuming the scope of the project does not change much, and no new technology or standards emerge, and the competition decides to ignore this market.  It seems likely that some version of MPAR will be implemented at some point in 10-15 years, but the format is still undecided.  The initial investment decision will not even be finalized until December, 2018, and the final investment decision will not happen until December, 2020.  After that prototypes would be evaluated from 2020 to 2025 and finally a service decision could be made in December, 2025, so management’s guidance on the F3Q16 (6/16) call that Aerospace and Defense could grow >30% in F2017 with active antenna as a meaningful driver seems highly suspect.  At another point in the call though management seemed to contradict this and back off comments earlier in the year that active antenna production volume would ramp in 2017, rather explaining that it is still too early to know for sure.

There were also supposed to be other opportunities with active antenna, like upgrading F-16s, which management said could be worth $300 million, but this has never showed up in reported results.  Once a key talking point of management, they have more recently backed off the hyperbole citing national security issues that prevent them from discussing some of their defense business.

Meanwhile the core Aerospace and Defense segment will have declined at a 5% CAGR by the end of this fiscal year (9/16) since FY2012 (9/12).  In F2012 Aerospace and Defense contributed $95 million of revenues and it is on track to do $72 million for F2016 (9/16).   It had been 31% of total sales, but has become increasingly less relevant as it is currently 13% of sales.  This segment appears to be in secular decline.  Fairing slightly better, the Multi-market segment has been stagnant since the IPO.

The Automotive segment, which drove growth in the first two years as a public company, was sold in F4Q15 (9/15).  The Automotive segment generated about $30 million of sales in F2010, growing to $53.4 million in F2012.  In F2013 it surged 58% to $84.4 million, but topped out there, declining to $80.8 million in F2014.  In F2013 the Networks segment was flat, while Aerospace and Defense and Multi-market both declined 8-10%, so Automotive was the only growth driver that year.  Furthermore, Automotive’s CAGR from F2010-2013 was roughly 30%, compared to the overall growth rate of approximately 5%.  During this period Networks grew at a 4.5% rate and Aerospace and Defense was essentially flat.  Automotive sales were predominantly related to Ford vehicle communication and entertainment systems, which experienced tremendous silicon content growth, but was subject to increasing competition in 2014, which led to the top line rolling over.  Revenues in this segment were in decline from the end of F2013 until the time it was sold.   The sale price was ~$100 million for a business that did $71.7 million in revenue and $20 million in operating income, or 1.4x and 5x, respectively, suggesting its best days were behind it.

Automotive was a great business for a few years, but it was a bit of a one trick pony that was in the right place at the right time.  According to management it was non-core and did not fit their long term vision, so they decided to divest it, which is a fine explanation, but the narrative of that era was that Aerospace and Defense was a strong organic, secular growth story.  Automotive buoyed the overall results and to some extent masked the challenges the core businesses were experiencing.  Also, when Automotive was classified as discontinued operations it reduced the prior year revenue by ~$20 million quarterly, so when acquisitions were used to backfill the lost Automotive revenue the growth rate looked more impressive off of a lower base, further obfuscating the situation.

As the Automotive segment began to fade, MTSI cobbled together a series of optical networking acquisitions to enhance its Networks segment.  These acquisitions have provided almost all of its growth since F2014 and optical now accounts for 54% of total revenue.  Much of the optical portfolio came to MTSI via its acquisition of BinOptics, which was a manufacturer of Passive Optical Networking (PON) components and lasers.  BinOptics had significant exposure to the rapidly growing metro market in China.  The industry was capacity constrained at the time of the acquisition and MTSI set out to quadruple its PON laser production capacity within a year.  While BinOptics continued to contribute strong growth soon after the acquisition in the ~40% range, this implies extreme pricing pressure if they were shipping 4x the volume.  With this in mind, it is important to understand what management means when they talk about exponential or explosive growth because they might be referring to unit growth and not revenue growth.  After only about 18 months since the acquisition closed the market appears to have peaked.  Throughout 2016 there has been evidence of double ordering and softening conditions.  In F3Q15 (6/16) MTSI reported that PON sales fell 16.5% qoq, so the capacity additions in PON lasers may have been another unwise capital allocation decision.  Despite the weakness in PON, overall the Network segment remained strong with 8.6% qoq growth due to market share gains, which they achieved partially from price concessions.  Industry estimates suggest that the 10G PON market, which was driven by the dramatic build in Chinese metro, will peak in 2016 and decline perpetually in 2017 and beyond.  Management was very bullish on the PON business in the months leading up to this slowdown, reassuring analysts that they were in constant contact with their colleagues on the ground in China and that everything was going swimmingly.  Since they have since been caught unawares, one should probably be skeptical of the fact that management also maintains that channel inventories are at normal levels and that this is just a short-term inventory correction.  With optical contributing over half of MTSI’s revenue and most of its growth, this slowdown, which is not contemplated in consensus estimates, could result in a meaningful reduction of expectations.

There is obvious cognitive dissonance between management’s outward ebullience and their actions.   If the core business was doing well and the outlook was so positive there would not be a need to resort to  accounting games with specious earnings adjustments and aggressive revenue recognition policy changes.  Management would not rely on perplexing, ambiguous, and contradictory language.  Management would not feel the need to overpromise and never deliver.   Aside from these red flags, shareholders should be concerned with the allocation of their capital because MTSI has spent ~$650 million on acquisitions since F2013, much of which appears to have been done to create the illusion of growth and hit predetermined guidance.  Management harps on the organic growth story, but there is no organic growth to be found.  It is worth spending a little time on each acquisition since MTSI came public to understand management’s strategy or lack thereof.

During F1Q16 (12/15), MTSI closed two acquisitions in December, 2015, Aeroflex/Metelics and FiBest, for $38 million and $59 million, respectively.   The contribution to sales and earnings was small in the quarter, but given management’s penchant for adjusting earnings it should have been itemized as a non-core boost to earnings.  The following quarter reported revenue growth was 15% qoq, but on an organic basis it would have been approximately 3-4%.   MTSI’s CEO on the F2Q16 call guaranteed that organic growth was double digits when it was clear to anyone who bothered to look at the 10-Q that it was obviously much lower than that.  This is a prime example of how management seemingly tries to deceive analysts about the underlying health of the core business and understate the impact of acquisitions.

MTSI acquired BinOptics in December, 2014 for $223 million.  Interestingly, $15 million of the purchase price was classified as prepaid compensation.  Management evidently decided that rather than paying these employees out of normal operating expenses, they would capitalize it and include it in the cost of the acquisition.   This spring loading further helped MTSI to hit its targets.  Later they would have to reclassify this item from a cash outflow from investing activities to a cash outflow from operating activities.  In F2015 (9/15) BinOptics contributed $61.5 million of revenue.  In addition to the $17.4 million of revenue recognition resulting from the inventory accounting change, and excluding the Automotive segment which was discontinued, organic revenue would have been basically flat for F2015.

MTSI acquired Nitronex in February, 2014.  This acquisition has been heralded as a key reason MTSI will be able to address the billion dollar market of newer generations of RF power amplifiers for base stations due to its GaN intellectual property, which begs the question: why did Nitronex sell out for $26 million with such amazing prospects in the midst of one of the easiest VC fundraising environments ever?   The answer probably has something to do with the fact that MTSI’s chairman was the controlling shareholder of Nitronex and using MTSI shareholder capital was an appealing option for his exit, since the venture was only generating about $1 million of revenue.  

Mindspeed was acquired in December, 2014 for $313 million, which was a troubled analog semi company that went through a difficult sale process that was rumored to not have many interested parties, other than MTSI.   This acquisition provided virtually all of MTSI’s growth in F2014.

With all of the moving pieces here it is no surprise that management and their auditor, Deloitte & Touche, found a material weakness in internal control over financial reporting, which is likely to be an ongoing issue.

In April, 2016, MTSI sued International Rectifier, which had been acquired by Infineon, over GaN related patents that MTSI acquired through Nitronex.  While it is nearly impossible to ascertain the validity of the suit as an outsider it will be interesting to watch.  MTSI seems concerned that Nitronex licensed its GaN IP to a competitor, which is integral to its next generation base station story.  MTSI claims that Infineon is in violation of the agreement and needs to assign the patents back to MTSI.  This is interesting because this raises the possibility that MTSI itself does not have the rights to this IP.  As usual with MTSI’s management’s explanations, they usually lead to more questions than answers.  To wit, management says this will not impact near term results, but is rather an effort to protect the business long-term.  If it has no bearing on near term results, can GaN power amps be on the cusp of an inflection point into volume production?

Perhaps most damning, on slide 27 in the 2016 investor day presentation there is a photo of Tom Brady and Robert Kraft with the caption: franchise players control their destiny.  I think the message they’re trying to send is, “hey, these guys are real winners, just like us!”  But the message I’m getting is, “yeah, you’re a lot alike because you guys are also a bunch of lying, cheaters.”  So the allegory (and unintended humor) is: if you cheat your way to winning you might have a lot of meetings with a commissioner and potentially wind up in court. 

For F2016 MTSI will generate ~$540 million of revenue and on a GAAP basis it will be barely breakeven.  In F2017 the consensus revenue estimate is $635 million, which implies about 18% growth, and 4x EV/revenue.  MTSI organic revenue has been flattish its entire time as a public company and has never approached anything like 18% growth without a lot of help from acquisitions.  Given the declines in the core business, the hype surrounding its growth areas, and the incipient weakness in the Networks segment, barring any other machinations there is a decent chance that organic revenue could actually decline in 2017.

The consensus adjusted EPS estimate is $1.93, implying about 21x what we consider very aggressively adjusted earnings on unrealistic growth expectations, so this appears priced for perfection.  MTSI’s heavily adjusted EBITDA for F2017 based on consensus would be ~$240 million, or ~37.5% of revenue.  And it is trading at 10.5x that.  A more realistic unadjusted EBITDA number for F2017 is ~$110 million, which at 12-14x would be a $20-24 stock, which would also equate to roughly 2x sales.  If any of the accounting irregularities turn out to more nefarious, there could be in excess of 50% downside.

Private equity investor, Summit Partners, sold 4.1 million shares in the past year, reducing their stake 77%.  John Ocampo, former controlling shareholder and current Chairman of the Board, sold two million shares in the past year.  Ocampo still owns over 20 million shares, but he has been a fairly consistent seller since the IPO.  Together Summit and Ocampo have taken several hundred million dollars off the table this year.  The rank and file have been following their lead, as well.  While insiders have been unloading their shares, shareholders have been steadily diluted, as diluted share count rose from 47.1 million in F2013 to 55.3 million last quarter.

What will happen next?  Without more chicanery there will be significant earnings disappointments. Revenue might actually decline in F2017 without more acquisitions.  I would expect Ocampo to try to sell as much stock as he possibly can as fast as he possibly can.  Inventory write downs seem likely.  There might even be issues with filing the 10-K.  If they make it through F4Q16 without blowing up, expect a few more crappy acquisitions in December of companies you have never heard of, which has become a sort of MTSI annual holiday tradition.

Clearly there is a risk to shorting any stock where management talks about potentially tripling the size of the company in the next five years, so if there are any impending signs that they might achieve any modicum of success we will reevaluate our position.  Also, the investment banks and brokers that cover the stock have a vested interest in promoting the story along with management due to all of the deal fees and commissions for selling blocks for insiders.  So even if the growth opportunities are continuously pushed out they will try to keep the hype machine going.  The Aerospace and Defense segment is in secular decline, but if defense budgets improve this could turn into a tailwind.  For instance, management already guided for an improvement in catalog MMIC in Aerospace and Defense next quarter.

Catalyst

Earnings disappointments

Slower growth

Bad acquisitions

Inventory writedowns

Insider selling

Accounting irregularities