How to Navigate the Semiconductor Cycle

This article was prompted by a Morningstar GrowthInvestor subscriber e-mail that asked my thoughts on Growth Portfolio holding  Maxim Integrated Products  and how best to play the semiconductor cycle. As many of you know, the semiconductor industry is cyclical, with alternating periods of growth and contraction. While many value investors have traditionally avoided technology, I think semiconductor industry is fertile ground from which to harvest profits. The stocks often overshoot intrinsic value on both the downside and upside based on short-term sentiments, and mispricings tend to be corrected fairly quickly when the cycle turns. If you are interested in exploiting these discrepancies between price and value, here is my playbook to profit from the semiconductor cycle.

Don't Buy at the Top of the Cycle
This seems like an obvious truth, like "buy low, sell high," but investors often get themselves in trouble when fits of fear or greed overcome their common sense. The chart below is a five-year history of the Philadelphia Semiconductor Index, commonly known as the SOXX. If you had purchased the SOXX at the peak of the cycle in early 2002, you would still be underwater. Now, I am a patient investor, but five years is a long time to go without a return. Even buying good semiconductor companies like  Linear Technology  and Maxim would have netted you losses of roughly 35% and 45%, respectively, over the past five years.

All this of course, begs the question: How do we know when the cycle is peaking? Here are two strategies that can help flash the warning lights.

Recency Bias
You should take no comfort in standing with or against the crowd, but with semiconductor stocks, it pays be a contrarian. Investors tend to place more emphasis on recent events, regardless of whether those events are representative of the future. This is called "recency bias" in the lingo of behavioral finance. The peak in semiconductor stock prices is likely near when companies have reported several quarters of excellent financial results. As victims of recency bias, investors have extrapolated current benign conditions far into the future and priced the stocks as if a cyclical downturn will never again occur. When the rest of the world loves semiconductor stocks, you should stay far, far away.

Reverse Engineering
At Morningstar, we value stocks by building a discounted cash-flow model with our expectations for a company's future growth and profitability. By inverting this process, we can ask an important question: What assumptions about the future are embedded in the current stock price? If you can only arrive at the current stock price by entering improbably optimistic forecasts in the valuation model, it is best to stay away.

Buy at the Bottom of the Cycle
Yes, I know I am again stating the obvious and there is no way to know prices have bottomed except with the benefit of hindsight. This rule is more accurately stated as "buy when the stocks are cheap and unloved." So what are the signposts that semiconductor stocks are ripe for purchase? We again can take advantage of other investors' recency bias. If semiconductor companies have reported poor financial results for many quarters and stock prices have declined significantly, it is time to get interested.

Negative sentiment, however, should not induce you to buy unless prices are well below intrinsic value. Morningstar's semiconductor team provides fair value estimates on more than 75 semiconductor stocks, and these estimates can help you assess the attractiveness of a particular stock. In addition, if you have a valuation model available, you can reverse-engineer the stock price. If you have to use implausibly pessimistic forecasts to arrive at the current stock price, then you should become even more interested.

I realize many of you do not have a valuation model at your disposal, so let me suggest another tactic. The market may act strangely in the short run, but over the long term, it is very good at getting prices right. If you compile a stock's price history relative to an appropriate valuation metric, you may see a pattern. For example, below is a summary of  Applied Materials' (AMAT) price/book ratio at the end of 136 different months versus the stock's return during the subsequent 12 months.

Applied Materials
	Price/Book Below 3	Price/Book Between 3 and 6	Price/Book Above 6
Observations	45	49	42
Mean 1-year return	105%	55%	17%
Median 1-year return	95%	51%	-13%
Number of positive returns	43	41	15
Number of negative returns	3	8	27
Success rate	96%	84%	36%
Morningstar data

As you can see, buying Applied Materials at a price/book below three has been a fairly good bet over time. On the other hand, buying the stock at more than 6 times book has been a losing proposition. The only reason the mean return is 17% for this category is because the great technology bubble of the late 1990s is included in the sample. Whether this relationship holds in the future depends on the relevance of your valuation metric. I think price/book makes sense for Applied Materials because the company has an exceptional balance sheet. For other companies, a different metric might be more appropriate.

Stock Prices vs. Business Fundamentals
I want to highlight an essential point here. There are many smart people who spend a lot of time trying to determine the exact timing of the peaks and troughs of the underlying semiconductor business cycle. You should ignore all of them. They are concerned about timing, whereas we are concerned about pricing. All you have to do to be a successful semiconductor investor is to recognize the extremes of pricing and sentiment, and then act accordingly.

The cyclical upturns and downturns will happen when they happen. Your job is to consider purchases when investors price in a never-ending downturn and consider sales when investors price in an eternal upturn.