Summary
- With all end-markets enjoying strong momentum, ChipMOS y/y comps should accelerate, potentially leading to a positive multiple re-rating.
- Our analysis suggests upside to current consensus margin estimates.
- Shares trade at substantial discount to Taiwan-listed holdings.
- 1Q earnings on May 14th should provide a near-term catalyst.
- If shares don't appreciate near term, significant excess capital supports buybacks that should drive shares to fair value.
(Editors' Note: ChipMOS reports earnings on Wednesday, May 14, before the market opens).
There are times as an investor when one feels they are living in the filmGroundhog Day. Such has been our experience with ChipMOS (IMOS), which we have written about multiple times over the last several years. Seemingly, on an annual basis, shares pull-back, creating a compelling entry point. The 14% sell-off from recent highs, coupled with a rise in shares of ChipMOS Taiwan (8150.TW) has created such an opportunity. At current levels, we believe shares offer a compelling risk/reward profile and also offer near-term catalysts, starting with earnings on Wednesday. ChipMOS shares are cheap on a relative and absolute basis (double-digit free cash flow yield), is enjoying accelerating top-line growth and margin expansion, and is levered to several of the best secular growth segments in technology. In addition, based on our analysis of the recently filed 20-F, we are optimistic that an important headwind will become a tailwind, driving upside to current Street gross margin estimates. With excess cash on its balance sheet, and significant, sustainable, free cash flow generation, we believe that if shares fail to perform near-term, they will ultimately rise due to the implementation of healthy dividends and/or buybacks. We expect shares to trade at $30+ by year-end - approximately 6x EV/EBITDA, and 50% upside from current prices.
Investment Summary
Shares of ChipMOS are up 4-fold over the last 2+ years despite modest growth and several negative earnings revisions over that time frame - which illustrates how cheap IMOS was in December 2011. Shares remain inexpensive and meaningfully undervalued, with a double-digit free cash flow yield. With growth accelerating and earnings beats/raises likely, there is substantial room for multiple expansion. In addition, excess capital that will be used for buybacks/dividends should drive shares higher and provide downside protection. Our confidence is bolstered by the continued ownership of ChipMOS' largest shareholder, the Baupost Group, which filed its most recent 13-F on Friday after the close, indicating they continue to own 11% of the company. We believe shares will trade 50% higher by year-end.
Thesis
1. ChipMOS is exceptionally well positioned, with its entire DRAM, mixed-signal, NAND and NOR customer base all enjoying strong secular drivers, which should result in several quarters of double-digit y/y growth, potentially attracting a new set of growth investors. For investors looking for exposure to mobile, 4K2K TV, SSDs, wearable consumer, and the Internet-of-Things, ChipMOS at 4x EV/EBITDA is a very cheap way to play these trends, and if the company meets/raises estimates, as we expect, there is significant room for multiple expansion.
2. Current 2014 margin estimates are likely to prove too low. Despite ChipMOS shares increasing 65.9% in 2013 and up 7.7% year-to-date, we believe there were fundamental challenges that muted 2013 results. Specifically, the recently released 20-F reveals the dramatic decline in test revenue had a disproportionate impact on margins. Current Street models don't adequately reflect the impact of a recovery in test, which we believe will result in 2014 gross margins 100-150bps above current views.
3. ChipMOS shares trade at a 17.4% discount to Taiwan-listed shares, when they should trade at par or above. We believe that recent US tech/small-cap weakness has created a growing gap between ChipMOS US and ChipMOS Taiwan that is likely to close or reverse when a single shares structure is created. Notably, IMOS holders own 60.2% of the Taiwan entity and therefore would only support a deal that is favorable. In addition, given the tax benefits and cost savings of a consolidation, a premium is warranted.
4. Deployment of excess capital on buybacks/dividends should boost shares over the next several quarters. We expect positive fundamentals (beginning with Wednesday's earnings) to drive shares higher, however, if shares fail to rise we believe the deployment of excess capital will provide a positive catalyst to propel shares higher.
Background
We will keep our background on ChipMOS brief given we have written several articles on ChipMOS over the last few years, and SA contributor Jaret Wilsonhas written recent articles (and here).
ChipMOS is a leading outsourced semiconductor assembly and test (OSAT) company focused on the memory (DRAM and Flash) and LCD driver-IC segments. The company was on the brink of bankruptcy in 2009 following the bankruptcy of its largest customer Spansion (CODE) and meltdown of its second-largest customer ProMOS. Since then, the ChipMOS has been extraordinarily disciplined with capital spending, resulting huge cash flow generation and a massive de-levering of its balance sheet.
(click to enlarge)
ChipMOS was traditionally an OSAT provider to the very challenging commodity DRAM industry, but has diversified to support niche/specialty DRAM, flash and LCD driver-ICs.
At the same time, new applications have emerged positioning ChipMOS very well for growth (discussed below). While ChipMOS' business is what we would consider low-tech tech, there are significant barriers to entry including customer relationships and substantial capital requirements. We believe few, if any, new credible players are likely to enter the industry.
In addition, just as key customers such as Micron (MU) are benefiting from being in an industry that has become an oligopoly, we believe ChipMOS stands to sustainably benefit from the consolidation experienced in its core segments.
(click to enlarge)
Details
1. ChipMOS is exceptionally well positioned, with its entire DRAM, mixed-signal, NAND and NOR customer base all enjoying strong secular drivers, which should result in several quarters of double-digit y/y growth, potentially attracting a new set of growth investors.
Shares of ChipMOS increased 124.8% in 2012 and an additional 65.9% in 2013. To achieve such a performance, one would guess they must have knocked the cover off the ball with huge upside to guidance. The truth is far from the case. ChipMOS frequently fell short of its own projections and analyst estimates. For example, in early 2012 the company guided to 10% top-line growth for the year and ultimately grew just 5.5%. In 2013 ChipMOS grew revenues less than 1% and saw gross margins expand 500bps, a healthy increase, but well below analyst expectations at the start of the year.
We'd suggest the stock has worked because it was so unbelievably cheap back in 2011. Our thesis in 2011 was that as depreciation rolled off from excess cap-ex years earlier, investors would take note of ChipMOS' earnings improvement, balance sheet de-levering, and cash flow generation. This has all occurred. We also expected healthy growth - which has not occurred. Despite this, shares are still cheap.
If shares of ChipMOS can perform as well as they have with tepid growth, we're very confident that there is substantial upside if the company can produce real growth and positive earnings revisions.
Over the past few years as one segment of ChipMOS' operations were strong, another was weak. In 2012 LCD drivers grew significantly, but DRAM was weak. In 2013, DRAM and LCD was healthy, but NOR flash was terrible. Finally, today, all segments are strong.
ChipMOS' top customers are Novatek (20% of sales), Micron (15%) and Winbond (8%). Novatek recently guided to 16-21% sequential growth for 2Q driven by panel demand and 4K2K TV uptake (4K2K uses 3x the driver-ICs of previous TVs). Micron is enjoying well-reported strength and stability. Winbond, which has been weak over the last few years, manufactures specialty DRAM and NOR flash, and recently revised its cap-ex to 40k WPM from current 33k.
The fact is, between mobile devices and 4K2K penetration, LCD is well positioned for 2014, with LCD driver companies currently seeking out additional foundry capacity. Similarly, specialty DRAM's strong prospects have been well-documented with Digitimes recently noting expectations of 46% non-PC bit growth for 2014. Finally, after several years of deterioration, NOR flash should benefit from supply rationalization and new growth drivers including automotive, industrial, wearable consumer and the Internet-of-things. NOR test customers such as Spansion are expecting a resumption of growth in 2014. As we explain below, a recovery in test revenue should drive outsized margin performance in 2014.
In addition to these segments, it has been reported that ChipMOS has won Toshiba NAND test business, which we believe would have a meaningful positive impact to margins starting late 2Q. Also, ChipMOS is seeing growth in new segments such as fingerprint sensors through Validity, now owned by Synaptics (SYNA). In sum, there is no segment we can point to that is weak for ChipMOS, and, in fact, several areas appear poised for secular growth.
As a result of these strong trends, ChipMOS enjoyed above seasonal 1Q revenue of $165.4 million, up 2.0% sequentially (please note, our revenue is based on the USD-NTD exchange rate as of April 30). The quarter was also up 12.9% y/y - the best y/y performance since 3Q2012.
We believe ChipMOS will report double digit sequential and y/y growth in 2Q, and double digit y/y growth in 3Q.
(click to enlarge)
ChipMOS reports 1Q results on Wednesday May 14th. While we expect the story to play out over the next few quarters, we believe earnings will provide a positive catalyst. Given the ongoing fluctuation in USD-NTD (New Taiwan Dollar) exchange rates there is always headline risk, but we believe that when analysts parse through the data, the results will be well received. Specifically, both Craig-Hallum and Cowen forecast $167.4 million for 1Q, above the number already reported, but in NTD the revenues are in line at up 2% (both have up 2.2%). More important is that the "consensus" for Craig-Hallum/Cowen is up 6.5% q/q and 7.5% y/y for 2Q, which should prove far too low. It's worth noting that most Taiwan analysts have IMOS growing 5-6% q/q and 6-7% y/y for 2Q. Our confidence about 2Q was bolstered by this past Friday's release of monthly sales data in which it was revealed that April sales were up 13.4% y/y - which we believe would have elicited a far more positive reaction for a more widely followed stock. The monthly sales figure was released after the close of the Taiwan market on Friday, but before the US market open. Notably, IMOS shares were up modestly on Friday, while ChipMOS Taiwan shares hit a new 52-week high this morning on strong volume.
None of the above is intended to be overkill about the quarter. This is not a short-term trade. The point is that ChipMOS has missed a lot of numbers over the last few years. Growth has been tepid, bordering on anemic. We believe expectations are very low for ChipMOS to deliver upside and in the event they do, and are able to produce double-digit growth, a new group of investors are likely to get involved in the stock.
2. Current 2014 margin estimates are likely to prove too low.
Last year ChipMOS' gross margins improved from 12.8% to 17.8% on just 0.7% growth in revenues. While this might appear impressive, the gross margin expansion was completely a function of a decline in depreciation expense. The recently filed 20-F reveals a dramatic decline in core profitability.
In the below analysis we add back depreciation and amortization, which reveals that on a cash basis, gross profit declined 300bps from 2012 to 2013.
Evaluating the deterioration in "core" cash gross margins does not look bullish, and would support the contention that ChipMOS might be facing some pricing pressure. However, a deeper view into the recent 20-F makes it clear that the entirety of the decline in gross margins (and EBITDA) is due to a decline in test revenue.
The above indicates that test margins increased 550bps y/y, but cash gross margins declined by over 1000bps. Looking more closely, test cash gross profit dollars declined $32 million, the same as the decline in test revenue. For anyone familiar with the OSAT industry, this makes a lot of sense. A tester is a high fixed cost item ($1-$5 million) which is essentially leased for several hundred dollars per hour. The incremental cost of running a tester for 20 hours per day instead of 16 is miniscule, and can result in incremental margins close to 100%. Clearly, that has impacted ChipMOS. If test revenue in 2013 had been flat with 2012, then, by our estimates, corporate gross margins would have exceeded 21%, and EPS would have been at least $0.50 higher. (Test gross margins would have been 30%+ versus the 15.6% reported)
We expect test revenue to grow in 2014 (as does the sell-side) however we believe that sell-side models don't adequately reflect the leverage inherent in test.
Assuming test revenue grows 10% in 2014 (which would result in test revenues still 10% below 2012 levels), due to a recovery in NOR, modest growth in specialty DRAM and LCD, and several million in incremental Toshiba revenue, we forecast corporate gross margins of 23%. Of course, more significant growth in test would have a material positive impact on margins.
Our model is roughly 100-150bps above consensus which is in the high 21% to 22% range. More importantly, our analysis gives us further confidence that estimates will be revised higher, which has not been the case for the last several years.
3. ChipMOS shares trade at a 17.4% discount to Taiwan-listed shares, when they should trade at par or above.
ChipMOS recently up-listed its Taiwan operating business to the Taiex from the GreTai (essentially moving a US listing from the Pink Sheets to the NYSE). The listing of the operating company has been part of a plan to simplify the corporate structure (from a Bermuda domiciled entity) and reduce costs. Following the listing on Taiex, IMOS owns 60.2% of ChipMOS Taiwan, or 522.1 million shares. We believe that Bermuda currently has approximately $75 million of cash (it had $85 million, spent approximately $30 million on Thailin share repurchases, and will shortly receive $20 million in dividends from Taiwan following a dividend of 1.4NTD per share announced on March 10th).
Based on Monday's close in Taiwan, IMOS should be valued at $25.09 per share, meaning IMOS trades at a 17.4% discount.
(click to enlarge)
We believe that IMOS should trade at equal or above the fair value. Based on our understanding, given IMOS shareholders 60.2% ownership in ChipMOS Taiwan, a move to consolidate IMOS into an ADR/ADS cannot occur without IMOS shareholder approval. ChipMOS Taiwan shareholders stand to benefit from a consolidated entity due to overhead savings and a lower tax rate (we believe tax rate will move from high 20s to low 20s). We expect ChipMOS Taiwan to offer a combination of 8150.TW stock and cash in exchange for IMOS shares, and believe part of the reason ChipMOS has accumulated so much cash, instead of pre-paying debt, has been to be prepared for a sizable cash outlay. We expect the discount to close as investors become more aware of ChipMOS management's plans to consolidate the entities.
4. Deployment of excess capital on buybacks/dividends should boost shares over the next several quarters.
With $75 million sitting in Bermuda, IMOS should pursue a buyback immediately. If they used $60 million, they could repurchase almost 3 million shares at current prices, reducing shares outstanding by 10%. However, we believe it's unlikely that there would be many sellers at these levels.
If a buyback at current prices could be accomplished, it would be preferable since it would be accretive. Specifically, with approximately 30 million shares outstanding, each share of IMOS would exchange for 17.4 shares of ChipMOS Taiwan - or as we stated previously, is worth $25.18. If 3 million shares were repurchased, each share of IMOS would be worth 19.3 shares of ChipMOS Taiwan or approximately $25.75. In addition, in the case of a dividend, we'd also have to pay taxes on the dividend.
While ChipMOS management has demonstrated itself to be slow in deciding and then acting on buybacks/dividends, the fact is that within the next several months a decision will come. Certainly, a sizable buyback or dividend will be warmly received by the company's shareholder base. Moreover, given the exceptional free cash flow generation expected for the foreseeable future - likely approaching $80 million in 2014 - significant ongoing dividends are likely, or buybacks should shares fail to appreciate.
Risks
While we believe ChipMOS is as well positioned as the company has ever been, the company faces risks.
Should the dynamics and apparent stability of the DRAM industry suddenly change, that would have a meaningful negative impact on results.
There have frequently been periods of apparent strength that were actually double ordering, followed by slower periods. We don't believe that's currently the case, however visibility in the OSAT industry is often very limited.
Management may continue to be slow to issue dividends or buybacks.
We believe that ChipMOS Taiwan will not be able to move forward with a consolidation into a single share structure without the approval of IMOS holders, however it is possible that management acts to the detriment of IMOS holders, although we note that ChipMOS CEO SJ Cheng owns over 500,000 shares of IMOS, and therefore has a vested interest in maximizing value of IMOS.
Conclusion
We believe ChipMOS is extremely well positioned and shares are inexpensive. We expect a positive tone for 2014 to be set on Wednesday morning's earnings conference call.
At 6x EV/EBITDA, ChipMOS Taiwan would trade at 50NTD, implying a value for IMOS shares of $31.39, or 50% upside from current prices. As a further sanity check regarding valuation, it's notable that Taiwan companies face high taxes on retained earnings, and therefore tend to issue generous dividends. On March 10th ChipMOS announced a $1.40 NTD dividend for 2014 - a 43% payout. We believe this relatively small payout reflects the continued build-up of cash for a consolidation of shares. We'd expect a payout of 70%+ in 2015. Using the 70% metric and EPS of 3NTD (a dividend of 2.1NTD), IMOS would receive $45 million, or $1.50 per share. Assuming a 5% dividend yield, shares of IMOS would trade at $30.
The bottom line for us is that ChipMOS remains too cheap. This industry has seen private equity players and consolidation. Siliconware (SPIL) is a significant owner of ChipMOS and by our estimates could acquire ChipMOS Taiwan at 50NTD per share and enjoy significant accretion to earnings. Further, given the new found stability of the memory sector, we believe that following the simplification of the ChipMOS corporate structure, if shares remain at current levels, private equity will take a serious look.
Additional disclosure: We conduct thorough research on our ideas, but our views are our own. Please do your own research.
No comments:
Post a Comment