Intel a Standout Buy as Deutsche Curbs Chip Enthusiasm
By Tiernan Ray
Deutsche Bank’s Ross Seymore today cut his view of the semiconductor industry to Neutral from Overweight, writing that the 35% rise in the Philadelphia Semiconductor Index (SOX) means that chip stock valuations have “already baked in a strong recovery in 2014.”
The SOX is 3% below its five-year average versus the broader market, closing a 10% gap a year ago, observes Seymore.
Seymore’s one positive stance is on Intel (INTC), which he rates a Buy, with a $28 price target, stating that its valuation remains depressed on an historical basis.
After growth of 5% in total semiconductor revenue worldwide this year, 2014 may see growth of 8%, he thinks, helped by macroeconomic improvement boosting overall demand. What’s more, “relatively low capital intensity should limit supply growth with stable-to-improving ASPs,” he believes.
“By end-market/product segment we expect all categories to see growth in 2014, with improving conditions in Computing/Comms, stable growth in Autos and slowing in Industrial and Memory (after a strong 2013).”
However, “we see upside beyond our current estimate being unlikely as it already assumes the second best growth year in nearly a decade.”
Modoff notes that semiconductor inventory remains elevated, though some of that may burn away this quarter and next as production is throttled back by chip makers:
At the end of 3Q13 inventory on semiconductor companies’ balance sheets grew +4% q/q. DIOs increased three days q/q on a weighted basis but declined marginally on an un-weighted basis, still remaining above historical averages. On a weighted basis, 3Q DIOs at 87 days increased slightly (+1% q/q) and remained well above the historical average (81 days). Overall, DIOs were ~2% higher than the historical peak average (85 days). On an un-weighted basis, DIOs fell -1% q/q to 100 days in 3Q13 and remained well above the historical average of 93 days and near the historical peak of 102 days. We believe the primary reasons for continued high levels of DIOs is a combination of normal 3Q seasonality (initial preparation for 4Q demand) and eventual weaker-than- expected demand in 3Q/4Q caused by the still uncertain macro environment. Looking forward, semi companies have cut utilization in 4Q, and therefore a slight normalization in DIOs in 4Q13/1Q14. Customers are cautiously maintaining lower-level inventories as evidenced by channel inventory remaining flat q/q in 3Q, significantly lower than average seasonality of +4% q/q. As customer sales grow seasonally faster than their semiconductor suppliers in 4Q, slight channel inventory burn is expected but at a slow rate as end-demand remains tepid and visibility stays low. On an EV/Sales basis, most stocks under coverage have experienced multiple expansion, especially notable in the Analog group as evidenced by TXN, ADI, and LLTC. The EV/Sales largely ignores actual revenue/EPS growth and focuses on perceived benefits/risks of end market exposure as many of the companies with the best revenue growth over the past few years have experienced significant multiple compression and a number of companies that have not grown revenues have seen multiple expansion.
Valuations of chips stocks have expanded “significantly,” he writes:
Semiconductor valuations expanded significantly in 2013 on expectations of better fundamentals and a slowly improving macro backdrop. Today the group trades at an EV/TTM sales multiple of 2.7x, up +15% y/y from 2.3x a year ago. The current multiple is roughly 4% above the 5-year average (2008-2013). On a company-by-company basis, valuation multiples in the semiconductor industry diverge widely across both P/E and EV/Sales. The P/E is largely unrelated to operating performance and often reflects operating and/or financial leverage. This focus on end market exposure also is appearing to trump the historically high correlations between EV/Sales and gross margin (GM). We note many of the firms trading at a premium to their historical EV/Sales average have not experienced a significant improvement in GM. TXN, which has enjoyed significant multiple expansion and currently trades at a 50% premium to its 5yr average EV/S, has experienced a modest 100bps in GM expansion. Meanwhile, NXPI, which has grown GM +690 bps over the same period, only trades at a 12% premium to its 5yr average. Similarly, NXPI grew OM +740 bps while TXN’s OM dropped -340 bps over this period.
Regarding Intel, Seymore writes:
Intel trades near its historical low points on both P/S and P/E multiples. We believe further upside potential remains as investors see accelerating growth in 2H13. Our price target is based on an EV/S ratio of 2.2x our 2014 revenue estimate, a ~10% discount to the five-year average. On a P/E basis, our price target is based on ~12x our 2014E EPS. We note the straight P/E is well below the large-cap average within our coverage group (16x) and a discount to the broader market multiple (S&P500 at 14x CY14E).
