Taiwan Semiconductor Manufacturing Company (TSMC) is the world’s leading semiconductor manufacturer. TSMC produces most of the chips that we use in our smartphones, computers, and tablets. They also make chips that are used in high performance computing to power the cloud that Amazon, Netflix, and Google use. Multiple clients outsource the manufacture of chips to TSMC.
We have covered TSMC for over 15 years. We have met the management team many times and visited their factory in Taiwan multiple times. I spent hours with award-winning analysts from various global investment banks to understand the business with the hope that they could help me predict how the stock price would behave. We understood all of the technicalities of the business. We understood that TSMC was a good company with solid fundamentals and way ahead of its competition—TSMC had a rock-solid balance sheet, rising margins, high capex productivity, market share, free cashflows, dividends, etc. We knew how far behind the various competitors were. Management was excellent at execution. There is a huge barrier of entry for any new competitors who want to get into this industry.
I thought I knew everything I needed to know about TSMC
Fast forward to March 2020 during the COVID-19 outbreak…
The main concern was the US China trade war, and Huawei had to stop outsourcing to TSMC. Along with the market, the stock dropped from NT$340 to NT$270.
Was I willing to put this into the portfolio? Probably not. Why? I was anchored to my past price / valuation of the company and could not break out of it. This is anchoring bias. This happens when we are anchored to a certain price we bought / sold the stock at. I had a preconceived idea of what it was worth.
When Intel gave a profit warning, the street realized that AMD, which outsourced the manufacturing to TSMC, would be the main beneficiary. Subsequently, the stock rose for various reason: better demand for chips during lockdown, stay-at-home effect, higher capex guidance, Apple deciding to stop using Intel chips, and Intel had a new CEO considering outsourcing to TSMC.
In fact, in 2008, AMD exited its chip manufacturing business, focused on chip design, and outsourced the manufacturing to the likes of TSMC.
It was old news. AMD’s steady climb on the technology curve for the mid-range desktop CPU had been well-known in the industry for at least the past three years.
This narrowing of the gap between AMD and Intel started in 2017 with faster, cheaper, lower power consumption in the mid-tier CPU chips used in everyday laptops and computers. Here is a chart of desktop CPU market share of Intel versus AMD. On the server and datacenter front, Intel still dominates. If you had been covering the semiconductor industry, you would have known all of these things but not been able to predict when the stock price would move up significantly, which it did in the second half of 2020.
Data from Passmark
Hanging out with the best analyst doesn’t mean you will get the call of the stock correct. Some analysts may get their earnings per share prediction right but miss the big picture on why the stock went up or down.
With all of the technical knowledge and experience, one would expect us to be able to predict the stock price movement with better accuracy. In fact, one of the catalysts for the stock move was Intel’s profit warning and its implication for TSMC. This news had been in the market for two to three years, but the market didn’t recognize it. I could not have predicted the timing and extent of the market’s reaction. During the COVID-19 lockdown, who would expect that the stay-at-home movement would drive up so much demand for electronics?
Humans are biased creatures; we are anchored to a price of a stock.
These biases could manifest in other ways like availability bias. This describes the current environment where savers allocate 95% of their cash into fixed deposits, despite record low interest rates.
With these inherent biases, we asked if we could create portfolios that would address this. At 8Vantedge, we use simple software algorithms to help us.
Analysts will assess the 3Ms of the business: does it have a barrier of entry (Moat), what is the competitive landscape and addressable market size (Market), and does it have a competent management (Management)? Then we let the algorithm construct the portfolio.
The algorithm helps us :
- Hang on to the winners and cut away losers with no emotional attachment
- Have no memory of your past target prices or other anchors
- Does not look at others for affirmation
Of course, with an algorithm to construct portfolios, we set certain limits on the portfolio to ensure adequate diversification and risk controls. In the last eight months, we’ve managed to maintain a medium to high weight to TSMC stock. This has helped the portfolio. TSMC’s Taiwan stock price has risen from NT$330 to NT$650 by February 2021.
At the end of the day, we imagine that we know the stock better by our experience and knowledge, but I’d rather not fool myself!
