Semiconductors

Feature Today’s Insight is a Special Report written by BCA’s Senior Technology Strategist, Brian Piccioni. Brian discusses the reasons for ongoing M&A in the semiconductor industry, and the investment implications. We trust you will find this report insightful and informative. Semiconductor Consolidation Makes Sense But Changes Little We have written extensively about our stance against financial engineering through M&A in the high tech sector.1 However, we view the semiconductor space as somewhat of an anomaly. Unlike most tech goods, a large portion of semiconductor products generate revenues for many years, even decades, after they are first released. Although most of the development costs of these devices are depreciated in the first few years after introduction, price deflation continues. This means that for most such devices, margins do not rise to a very high level. In addition, incremental costs may be associated with "die shrinks" (making smaller devices with the same function), and changes in IC packaging, as the cost of the package can be more than the semiconductor itself. In addition to the inherent benefits of buying a company that makes a product line with long duration revenues, most semiconductors are sold through the same channels and have similar, if not identical, customers. This can allow for the rationalization of sales and marketing efforts. On the surface it might appear there is an opportunity for economies of scale in manufacturing, but these can prove elusive. It is often not worth the effort to consolidate manufacturing for an acquired company due to large differences in manufacturing processes. This is especially true since it would require an investment in R&D for a catalog of mostly dated products. These products would have to be "re-qualified" by customers, as there is no guarantee parts produced in a different factory will function the same as the old one. While most high tech M&A destroys shareholder value, that is less likely to be the case when two mature semiconductor companies combine (Chart 1). However, industry consolidation is not likely to lead to pricing power or unusual profitability post consolidation because: Semiconductor buyers are reluctant to adopt a product made by only one vendor; There is a powerful push for the adoption of technologies based upon Open Standards in order to avoid semiconductor vendors having too much power over customers; and For the most part, with the exception of leading edge process technology used in CPU and commodity memory devices, semiconductor expertise is well understood and widely available, as are the tools for the development of new devices. Total Intangible Asset Writedowns And Restructuring Charges As A Percent ##br##Of Assets By S&P 500 Tech Sub-Sector, 2000 - 2016 Feeding Frenzy: Semiconductor M&A Feeding Frenzy: Semiconductor M&A Semiconductor Buyers Are Reluctant To Adopt A Product Made By Only One Vendor Large device manufacturers always had an aversion to single sourced semiconductors but exceptions were made when there was a critical need or simply no other choice. For example, if you are going to design a PC you are either going to use a single-sourced device from Intel or AMD. Things changed after the "Dot Com" bubble when equipment manufacturers found themselves unable to ship finished products because a single-source vendor had declared bankruptcy and the parts were no longer being made. Even when a single source part is specified, an effort is made to ensure there are substitutes available. This hedges against the possibility the part may no longer be available and also reduces vendor pricing power. Powerful Push For The Adoption Of Technologies Based Upon Open Standards Open standards are standards where form, fit, and function, are both defined and easily referenced. The standard itself is typically inexpensive to license and any related Intellectual Property is available for license on "fair and equitable" terms, meaning that the price is reasonable and the same for all licensees. Open standards have a long history in the semiconductor industry. The market for certain devices such as memory chips would likely have never developed if every vendor had a different way of doing things. Nevertheless, companies such as Intel were able to establish a proprietary standard CPU architecture and profited handsomely as a result. Similarly, purported abuses by companies such as Rambus and Qualcomm have resulted in all players being leery of patent suits. It is now very difficult to get manufacturers to accept a new standard unless it is open. Difficult To Benefit From Competitive Advantage, Even For Largest Players The components sold by most small semiconductor companies do not require cutting edge process technology or expertise. The largest companies such as Intel, Samsung, and TSMC may have an advantage due to their process R&D, but competition among themselves limits returns. In addition, there are very few "must have" products nowadays, and consumers and businesses can typically decide to simply not purchase a new PC, video game, etc., if prices get out of hand. Industry Consolidation Will Not Fuel Growth As we have frequently noted, semiconductor industry growth has slowed to GDP plus or minus a few points (Chart 2). The industry operates within the context of chronic high price deflation, meaning many more units have to be sold each year just to keep revenues flat. Some end markets allowed for the sale of higher value-add components with increased functionality, offsetting some of the deflation. However, the era of hyper growth in PCs, networking gear and smartphones is in the past. This places downward pressure on pricing through the value chain. Chart 2Semiconductor Industry Growth Has Slowed, ##br##Now Near GDP Growth Rate Semiconductor Industry Growth Has Slowed, Now Near GDP Growth Rate Semiconductor Industry Growth Has Slowed, Now Near GDP Growth Rate Loosely speaking the industry can be separated into commodity semiconductors and proprietary ones. Commodity devices are exact functional equivalents to devices sold by multiple vendors. Examples might be discrete devices such as transistors and diodes, memory chips, logic devices, and so on. The competition in commodity semiconductors is so extreme that for some products package costs can be similar to the cost of the semiconductor itself and saving a small amount of plastic or using slightly thinner leads influences profit margins. The product life of many commodity products extends to decades. The market for proprietary semiconductors is somewhat more complicated than for commodity devices. Intel is the prototypical example of a company that makes mostly proprietary devices, though Qualcomm, Xilinx, and others exist. Some companies such as Texas Instruments are a sort of hybrid, offering both commodity and proprietary products. It would be a mistake to assume that proprietary vendors have no competition, because substitutes are typically available. A smartphone vendor can select a high end ARM-based microprocessor from Qualcomm, make its own, or buy from any number of licensees selling similar devices, depending on the market segment and price range it is targeting. This has the effect of limiting the price of a proprietary device and the associated margins. As with any M&A transaction the opportunity arises to take associated restructuring charges, write-downs, and all manner of "one-time" items which can make "non-GAAP" earnings look better than before. Similarly, management may decide to cut costs by reducing R&D and other expenses to improve near-term performance at the expense of long term results. Company managers typically highlight "synergies" and "complimentary businesses" when selling their latest M&A transactions. Nevertheless, it is rare that the combination of two semiconductor companies actually amounts to something greater than what the two were apart. Instead, what tends to result is a mix of products and activities with varying degrees of margins and growth potential. Like any overly diversified portfolio, the combined companies are more likely to grow at the same rate as the industry than to become high-tech powerhouses. In summary there is no reason to believe that organic revenue growth will arise as a consequence of any particular semiconductor M&A transaction and it is far more likely that revenue growth and margins will trend towards the mean for the industry, setting aside the impact of "non-GAAP" adjustments. Why Is There A Buyer's Panic? As we have shown, in most cases industry consolidation will not provide much in the way of operational leverage to the consolidator's results. Similarly there is little reason to believe that companies which remain independent will be affected positively or negatively from the trend.2 This raises the question of why these transactions are occurring at such a frenetic pace. Most likely the answer has more to do with capital market trends than objective business decisions. Investors have elected to reward high tech companies for financial engineering on an equal footing with organic growth (i.e. innovation), and the capital is very cheap nowadays (Chart 3). As we have addressed previously, increasingly imaginative "non-GAAP" financial presentation means that overpaying for an acquired product line is better for the bottom line than developing it in house, so managers are focusing more on financial engineering than actual engineering. Cheap capital and less-than-rational capital markets mean that companies become acquirers or targets. As companies get larger, the targets need to be large enough to "move the needle" with respect to financial impact. This goes all the way down the food chain as mid-cap companies buy small-cap companies and large cap-companies buy mid-cap companies. There are a finite number of target firms for any given company and this creates a sort of "buyer's panic" which stimulates the buyers to move quickly before the target is acquired by a rival (Chart 4). As acquirers get bigger they become the targets of larger acquirers, as they are now large enough to provide the illusion of growth. Chart 3Capital For Financial Engineering Is Cheap bca.tech_wr_2016_11_15_c3 bca.tech_wr_2016_11_15_c3 Chart 4'Buyer's Panic' In Semiconductor Industry "Buyer's Panic" In Semiconductor Industry "Buyer's Panic" In Semiconductor Industry Non-S&P 500 Semiconductor Companies Are Also Targets Unless the cost of capital rises significantly or investors suddenly get concerned about actual operating results rather than "non-GAAP" performance, consolidation will continue until there is a relatively modest number of large semiconductor companies. As we explained above, this does not mean these companies will have superior margins or revenue growth. Indeed we believe the end of the consolidation period will have negative impact for semiconductor industry valuations because: Opportunities for financial engineering of revenue growth and managing "non-GAAP" earnings will be limited; Balance sheets will typically be highly leveraged; and Valuation premiums associated with M&A activity will disappear. Until the consolidation phase runs its course investors should be able to profit by assembling a portfolio of smaller names since these are more likely to be acquired. This is a major reason we have most of the smaller members of the S&P 500 Semiconductor sub-index rated Overweight. Table 1Summary Of Potential Semiconductor Targets Feeding Frenzy: Semiconductor M&A Feeding Frenzy: Semiconductor M&A We have identified 14 additional small semiconductor firms that are not included in the S&P 500 as likely targets (Table 1). This list is not exhaustive but represents companies which are both likely to be acquired and large and liquid enough to be investible. We selected the most attractive companies based on the Price-Earnings-to-Growth (PEG) ratio, which attempts to adjust valuation for growth prospects. These companies, which have a PEG ratio close to or below 1, are bolded in Table 1 above. We are adding these 10 companies to our Overweight list, and will track these recommendations as an equally-weighted index (Chart 5). Chart 5Small Semiconductor Companies Should ##br##Outperform Due To M&A Small Semiconductor Companies Should Outperform Due To M&A Small Semiconductor Companies Should Outperform Due To M&A However, due to the "buyer's panic" described above, companies that appear expensive or exhibit deteriorating financial performance are also potential acquisition targets. As such, it is important to note that our recommendations in this sub-sector are not driven by company fundamentals. Alternatively, investors might consider playing consolidation through the iShares PHLX Semiconductor ETF (SOXX). The structure of this ETF limits the weight of each constituent to approximately 8%, effectively overweighing smaller firms. Brian Piccioni, Vice President Technology Sector Strategy brianp@bcaresearch.com Paul Kantorovich, Research Analyst paulk@bcaresearch.com 1 Please see Technology Sector Strategy Weekly Report, "Tech Company Red Flags Part 2: Intangible Assets And Restructuring Charges," dated July 12, 2016, available at tech.bcaresearch.com 2 That would not be the case if, for example, Hynix and Samsung, two major DRAM manufacturers, were to merge which would be problematic for the #3 player Micron. However, we doubt regulators would permit such a merger.

The frenzy in semiconductor stocks has reached an overshoot phase. Relative performance had been buoyed by a flurry of M&A activity, but that has since waned without a similar response in share prices. If fundamentals return as the main driver, then a setback is probable. Our concern is that the inventory overhang continues to linger. Our proxy for global semiconductor inventories continues to grow, albeit at a slowing rate. Still, the semi sales-to-inventory ratio is deep in negative territory, which is typically deflationary and a harbinger of semiconductor profit contraction (bottom panel). That is confirmed by ongoing sluggishness in Taiwanese and Korean export and export price growth (second panel). The implication is that the surge in semi stocks is vulnerable to an abrupt reversal and an underweight position is warranted. The ticker symbols for the stocks in this index are: BLBG: S5SECO - INTC, QCOM, TXN, AVGO, NVDA, ADI, MU, SWKS, LLTC, MCHP, XLNX, QRVO, FSLR. bca.uses_in_2016_10_20_002_c1 bca.uses_in_2016_10_20_002_c1

Housing activity should accelerate in the back half of the year given the drop in Treasury yields. Buy home improvement retailers and add to long homebuilding positions.

Asian exports (volumes and prices) have been contracting, as global trade has hit a wall. While this broad deflationary backdrop has taken a toll on Asian DRAM prices (bottom panel), global semiconductor supply/demand imbalances best explain the industry's dwindling pricing power. Not only are global semi sales shrinking, but also BCA's global semi inventory proxy is surging. Taken together, our global semi sales-to-inventories (S/I) ratio is contracting at an accelerating pace, signaling that an inventory liquidation phase is looming. Historically, the S/I ratio has been an excellent leading indicator of semi earnings and the current message is to expect a significant drop in profits (middle panel). Bottom line: Steer clear from the broad tech sector, continue to underweight the tech hardware, storage & peripherals sub-index and we reiterate our high-conviction underweight status for the S&P semis index. The ticker symbols for the stocks in the S&P semis index are: BLBG: S5SECO - INTC, QCOM, TXN, AVGO, NVDA, ADI, SWKS, XLNX, MU, LLTC, MCHP, QRVO, FSLR. The ticker symbols for the stocks in the S&P technology hardware, storage & peripherals index are: BLBG: S5THSP - AAPL, EMC, HPE, HPQ, SNDK, WDC, STX, NTAP. bca.uses_in_2016_04_28_002_c1 bca.uses_in_2016_04_28_002_c1

Bearish sentiment is a red herring, as most other measures of investor positioning point to a strong undercurrent of bullishness. That is contrarily worrying.

An Insight yesterday showed that the overall technology sector was likely to record its worst quarterly earnings performance in four years. The highest beta components of the sector are most at risk. For instance, the semiconductor industry is losing its main source of support, namely an M&A premium. Last year's mini-M&A frenzy is petering out, which will put the onus on profits to support relative performance. However, global chip sales continue to deteriorate, and leading indicators such as Chinese electronics imports and Emerging Market currencies continue to warn of tepid chip demand. With chip producer inventories still growing at a historically rapid clip, there will be downward pressure on average chip selling prices. TSMC's profit warning earlier this week likely provides a good read for the overall industry, and we reiterate our high-conviction underweight rating. The ticker symbols for the stocks in this index are: BLBG: S5SECO - INTC, QCOM, TXN, AVGO, NVDA, ADI, SWKS, XLNX, MU, LLTC, MCHP, QRVO, FSLR. bca.uses_in_2016_04_15_001_c1 bca.uses_in_2016_04_15_001_c1

The self-driving car, or Autonomous Vehicle (AV), will have a profound impact on a variety of industries. However, expectations for the timeframe of commercial AV availability are too optimistic. The greatest near-term impact is likely to be from advanced safety technologies developed on the path to full autonomy. In today's <i>Special Report</i>, we discuss our expectations for the timeframe of AV development, and the effect of advanced safety technologies on the Insurance, Health Care, Semiconductors, and Automotive industries.

The self-driving car, or Autonomous Vehicle (AV), will have a profound impact on a variety of industries. However, expectations for the timeframe of commercial AV availability are too optimistic. The greatest near-term impact is likely to be from advanced safety technologies developed on the path to full autonomy. In today's <i>Special Report</i>, we discuss our expectations for the timeframe of AV development, and the effect of advanced safety technologies on the Insurance, Health Care, Semiconductors, and Automotive industries.

The previous Insight showed that semiconductor top-line growth remains under siege. Worse, there appears to have been little effort to realign cost structures to slower sales. The latter will become even more critical in the coming quarters, because pricing pressures are set to intensify. Our global semi inventory proxy is accelerating. Slowing demand has not been met with a sufficient output reduction to rebalance the market. Utilization rates are hitting new lows, and Taiwan export prices have plunged. These trends are a significant pricing power threat, and will compound profit margin pressures. We continue to recommend a high conviction underweight. The ticker symbols for the stocks in this index are: INTC, TXN, AVGO, MU, ADI, SWKS, LLTC, XLNX, NVDA, MCHP, QRVO, FSLR. (Part II) Semiconductors Are Losing Their Charge (Part II) Semiconductors Are Losing Their Charge

Semiconductor stocks finished last year on a strong note, supported by a surge in M&A and hopes that low oil prices would spur an increase in consumer spending, particularly on electronics. While the latter has improved, the M&A backdrop is becoming more hostile as the cost and access to capital become more restrictive. We put this group on our high-conviction underweight list to reflect our concern that once M&A euphoria faded, a renewed focus on fundamental profit drivers would trigger a de-rating. Apart from better spending on electronics, the data continues to support our bearish call. Global semi sales are shrinking, with key producing countries like Korea and Taiwan suffering from a steep export contraction. That implies heightened liquidation pressures, which will undermine profit margins. Worse, semiconductor companies have been slow to downsize despite threats to top-line growth, adding to profit margin pressures, please see the next Insight. The ticker symbols for the stocks in this index are: INTC, TXN, AVGO, MU, ADI, SWKS, LLTC, XLNX, NVDA, MCHP, QRVO, FSLR. (Part I) Semiconductors Are Losing Their Charge (Part I) Semiconductors Are Losing Their Charge