Semiconductor shares have been on a tear lately: the SPDR S&P Semiconductor ETF (XSD) is up 24.7% ytd, vs. 18.6% for the S&P 500 (SPY). Because I own and/or follow several names in the industry, I recently prepared an analysis of valuation for analog semiconductor companies. This article presents my findings.
Texas Instruments (TXN), Microchip Technology (MCHP) and Linear Technology (LLTC) were included in my Synthetic Dividend Growth Portfolio when it was implemented in June. LLTC was sold on valuation for a quick profit, the other two are up 12.6% and 11.4% respectively. Analog Devices (ADI) is the second largest in the industry, and Maxim Integrated Products (MXIM) rounds out the top five by market cap.
The larger players are dividend growth stalwarts. The industry serves a diverse customer base in multiple areas of the economy. It's a competitive business, but features a large R&D component, high margins, and an advantage for the incumbents on many accounts. It's cyclical, and the current optimism of industry participants bodes well for the future of the economy.
The analysis relies on a multiple linear regression of factors that I consider significant. It starts with forward earnings, then considers various aspects of capital deployment: it can be returned to shareholders as dividends or buybacks, deployed in the business via R&D or capex, or retained as excess current assets.
All amounts are per share: R&D is TTM; dividend is current annual; capex and buybacks are three year averages; excess current assets is the amount by which current assets exceed 2X current liabilities. The companies reviewed are TXN's peers in the analog semiconductor business, per S&P, with the addition of Microchip, which has a meaningful and rapidly growing position in the segment.
Putting the data through a multiple regression (available free on the internet), a formula is developed, as shown in the following table:
Some of this is counterintuitive. A multiple of 17.4 times forward earnings seems logical, if somewhat optimistic. From there, it appears that paying a dividend reduces share price, on a multiple of 5.2. On the other hand, buybacks increase it, at 7.8X. I can hear groans from dividend growth investors, a majority of whom devoutly prefer dividends over buybacks.
R&D is rightly interpreted as an expense, a cost of doing business. All of these companies spend serious amounts on it, and I had half expected that higher R&D would be a positive. Increased efficiency on R&D seems elusive: many industry participants say every application is unique and it takes ten years to develop a capable high performance analog engineer. Some sort of modular or outsourced approach might be helpful here.
Capex is regarded very favorably. The market seems to consider money spent on property, plant and equipment as a worthwhile investment, awarding it a multiple of 15.8. This in spite of the fact that most industry participants favor holding it to a bare bones minimum.
Excess current assets is peculiar. I would have thought a multiple of less than 1.0 would be appropriate, on the grounds that the shareholder can't tap this asset freely. However, the software sees 2.0X, perhaps reflecting the market's preference for strong balance sheets, or a belief that the funds will be deployed profitably.
The prices developed by the formula have a 98.5% correlation with the actual, and are particularly accurate for the largest five players, which constitute 83.7% of the market. Causation is debatable, but descriptive power is not.
Dividend Growth Credentials
TXN, ADI, LLTC and MCHP are included in David Fish's CCC spreadsheet. MXIM skipped the increase during the financial crisis, but has been regular otherwise.
All of these companies have assets above a current ratio of 2:1, as shown in the first table. Innovation in the automotive and industrial end markets holds forth the prospect of real growth, while smaller competitors can be acquired from time to time.
Morningstar provides coverage on the five companies shown, and awards all of them a moat: narrow for TXN and MCHP, wide for LLTC, ADI and MXIM.
The expectation of increasing dividends is reasonable: capital appreciation is less likely, given today's strong valuations.
The high multiple on forward earnings is consistent with an industry EPS CAGR of 8%, by my estimate. The preference for deploying cash into capex or buybacks reflects optimism, that the money will be well spent in order to garner an increased share of this 8% industry-wide growth potential.
My usual valuation methods involve the use of PE5 (projected 5 year average EPS), to which I apply a multiple consistent with my estimation of quality. Applying that line of thinking to the 5 largest companies in the industry, I develop the following:
These stocks are already priced for wonderful things to happen during the next year and a half to two years. I would resist the urge to chase them higher.
Observations on Individual Companies
From the most recent TXN earnings conference call:
And as we said, strong cash flow, particularly free cash flow, means that we can continue to provide strong cash returns to our shareholders. In the second quarter, TI paid $309 million in dividends and repurchased $721 million of our stock. Our capital management strategy is it to return all of our free cash flow to shareholders except for what we need to repay debt. In the second quarter, we reduced our debt level by $500 million. We also benefit from proceeds from employee stock option exercises that have allowed us to return more than 100% of our free cash flow to our shareholders while maintaining our targeted level of cash.
For the past 12 months, we have paid $971 million in dividends or 34% of our free cash flow. Similarly, we have repurchased $2.6 billion of our stock or 90% of our free cash flow. In total, we returned 123% of free cash flow in the past 12 months.
The company seems to be defining itself as a cash cow, for the benefit of shareholders. When the largest player takes this tack, the others can chip away at market share and grow nicely.
From the Microchip conference call:
Moving to our analog business. Our analog business grew 6.2% sequentially in the June quarter to also achieve a new record and continues to perform exceptionally well. Analog revenue was also up 119% versus the year-ago quarter. Analog revenue represented 22.3% of Microchip's overall revenue in the June quarter. And at a $413 million annual sales run rate, it has quietly become one of the larger analog franchises in the industry. We are continuing to develop and introduce a wide range of innovative and proprietary new products to fuel the future growth of our analog business.
Based in these results, and the valuation model developed for this article, I think MCHP's recent surge was to be expected.
From Linear Technologies conference call:
Looking forward, over the last 2 quarters, our business has steadily improved at a measured, consistent pace. Customers are still cautious and, on average, have a neutral stance on the global economy. There is enough innovation, particularly in the industrial and automotive end markets to drive this steady improvement. However, customers generally have been ordering to meet immediate demand and not unnecessarily growing inventory.
For Linear on the positive side, June was a good bookings quarter. Bookings definitely improved from the previous quarter and we had a positive book-to-bill ratio. Our bookings were stronger across all major end markets. In particular, industrial and automotive continue strong. Both of these end markets are in innovation cycles: energy efficiency and smart manufacturing in industrial, and higher electronic content in automotive. In many ways, these markets are moving into a new electronics age as electronics are replacing many previously mechanical functions.
There is a potential investment thesis here: innovation in automotive and industrial applications of analog semiconductors, driving another growth ring for this mature industry. Because the stocks are fully valued on that basis, the primary importance at this point may be as a bullish indicator for the economy as a whole.