Ask anyone to name the top investor of all time. Chances are you’d hear names like Benjamin Graham, Peter Lynch, and of course the Oracle of Omaha – Warren Buffett. Buffett is known for espousing the principles of value investing.
In the 1988 Berkshire Hathaway shareholder letter, he mentioned that ‘our favourite holding period is forever’. Behind this hyperbolic expression is Berkshire’s attitude towards being able to hold a stock over extended durations rather than speculating over truncated periods of weeks, months, and even years. Curiously, Berkshire appeared to have breached this code by buying and dumping a stock within a short span of three months.
In its 13F filings to the US Securities and Exchange Commission (SEC), Berkshire bought over US$4.1 billion worth of shares in Taiwan Semiconductor Manufacturing Company Limited (NYSE: TSM) in the third quarter of 2022. Shortly after, the investment firm sold over 86% of its TSMC shares. Neither Berkshire Hathaway nor TSMC provided comments on the share sale. Today, we examine whether this trade reversal was a questionable decision, as well as explore plausible reasons.
Possible reasons behind Berkshire’s sale of TSMC shares
The goal of any investor is arguably to make a profit. TSMC’s share price went up by 9.3% during the period which Berkshire held its stock. This suggests that the bet on the chipmaker paid off, except the NYSE Composite Index rose by 12.7% during the same period. Given the relative underperformance, it makes one ponder if disposing majority of its stake in TSMC is indeed a better call than simply ‘holding forever’.
We first assess the overview of the semiconductor industry. The chipmaker has been and continues to hold the crown as the market leader in the semiconductor foundry space, with a near 60% market share that is approximately five times that of the next challenger, Samsung Foundry. The industry faced an inventory correction in the second half of 2022. This will persist in 2023 as demand in both consumer and enterprise end market segments decline. Recovery will ensue in the second half of 2023.
Correspondingly, the chipmaker’s revenue follows a similar trend direction as global demand. Short-term declines can hardly be reasons for ditching a stock as the length of the semiconductor business cycle averages more than two years from peak to trough. Similarly, capital expenditure plans also take several years to bear fruits. For example, a fabrication plant takes an average of two to three years to complete construction. Berkshire’s disposal of shares in barely three months implies that their initial conviction and subsequent U-turn has little to do with the prospects of the semiconductor industry.
We now look at the fundamentals of the business over the past five years. Between 2018 to 2022, TSMC has a consistent trend of growing its revenue, gross profit, operating income and net income. Its margins also rose across the years, signalling increasing profitability. To sustain this growth, capital expenditures climbed upwards commensurately. Despite higher dividends over the same five-year period, the payout ratio nearly halved in the most recent year due to soaring net income.
According to Berkshire’s 13F filing, the firm sold off TSMC between September and December 2022. During this period, TSMC released its third quarter results. Thus, besides comparing the foundry’s yearly financial performance, we also look at its 3Q2022 scorecard. Analysis on the bases of year-on-year (3Q2021 vs 3Q2022) and quarter-on-quarter (2Q2022 vs 3Q2022) revealed similar findings: the company’s metrics improved, margins went up, and the dividend payout ratio fell. Moreover, the company managed to outdo its profitability guidance in 3Q2022 while keeping net revenues within ballpark estimates. With TSMC’s track record of healthy fundamentals, it makes it highly possible that Berkshire’s sell thesis comes from elsewhere.
A plausible reason for the sale could be due to tense relations between China and Taiwan as the former looks for unification. China has also not ruled out the use of military force to achieve this end. This sparked concerns among investors as it may cause customers of TSMC to diversify away from Asian foundries, adversely affecting the performance of the company. This is a nebulous argument for Berkshire’s sale for two reasons.
Firstly, TSMC and Samsung Foundry are perhaps the only two entities which can produce 3-nanometer (3nm) chips in the foreseeable future. Alongside the current 5nm chips, these chips are critical for mobile and high-performance computing (HPC). In turn, their applications are wide ranging, some of which prove critical to countries such as defence and natural resources. One can thus expect significant customer stickiness, a sentiment echoed by TSMC’s CEO who stated that technology leadership remains as the most important factor for foundry customers. Secondly, geopolitical tension already existed before Berkshire’s purchase of TSMC shares in September 2022. Arguably, there have been no material escalation that signals a fundamental threat to the company in the last three months of 2022.
If it is neither the business fundamentals nor its risks catalysing Berkshire’s short-term sale, what could possibly prompt a deviation from its ‘hold forever’ strategy? One conjecture lies in Berkshire’s succession plan. In 2018, Buffett spoke about how he progressively cedes control over business decisions to executives – Greg Abel, Ajit Jain, Ted Weschler and Todd Combs. Berkshire’s investment holdings have also dropped hints that his lieutenants are increasingly responsible for stock picking. A few years before this in 2016, Berkshire bought several airline companies. This contravenes his previous distaste for airline stocks which he verbalised in multiple interviews.
Another speculation could be Berkshire adapting its blueprint to today’s reality. The long-only strategy has shown signs of fading popularity. Average holding periods for stocks are on the decline from the peak of five years back in the 1970s to an average of 10 months as of 2022. This may be due to a host of reasons such as emergence of high-frequency trading and falling lifespans of large corporates. To quote Charlie Munger from the latest Berkshire Hathaway shareholder letter, ‘When the world changes, you must change.’ Some of Berkshire’s shorter duration holdings may simply be an adaptation to the way the market works today.
A third hypothesis could be Berkshire shifting its TSMC exposure to the Taiwan Stock Exchange (TWSE), where TSMC is listed under the ticker TWSE: 2330. The underlying cause is the slight pricing discrepancy between the stock listed on NYSE (known as an American Depository Receipt or ADR) and the stock listed on TWSE. Theoretically, both should have equivalent valuations and thus the law of one price should hold. Yet research suggests this is not the case, giving rise to arbitrage opportunities. To support this notion, the ADR listed on NYSE has higher valuation metrics including higher trailing and forward P/E, as well as higher price/earnings-to-growth (PEG) ratios than the stock listed on the Taiwanese bourse. For a value investor like Berkshire, it makes sense to buy the comparatively undervalued TSMC stock on TWSE instead of the ADR.
The fifth perspective
Unfortunately, there is no way to ascertain Berkshire’s motivations for selling a seemingly robust market leader in a growing space. Those who blindly follow the decisions of famous investors like Berkshire do so at their own peril. There are substantial reasons against ‘mirror trading’ discussed in another article. Instead, devising and staying true to your own investment philosophy which is guided by parameters such as one’s risk tolerance level is preferred.
I would put it down to the considerable uncertainty from US aggression in SE Asia against China.. The US will not allow TSMC to pass technology to China.
A peaceful reunification will result in the US sanctioning any suppliers of TSMC, effectively putting the company out of business.
A forced reunification will result in the destruction of the factory. Biden has already hinted it.
Maintaining the status quo is not acceptable long term, as the US wants to move advanced chipmaking to the US and out of the sphere of Chinese influence. Taiwan declaring independence would send China ballistic.
There are absolutely no signs that relations between the US and China will improve in the foreseeable future.
I currently do not have any optimism that TSMC in Taiwan has a future, except in non-cutting edge chip manufacturing.
Cheers,
Bob.
Thank you for the comment, Bob.
Fully agree that a reunification (forced or otherwise) will almost certainly invite sanctions from the US. TSMC has taken steps to diversify its production through building facilities outside of Taiwan, in both Japan and the US. In fact, cutting edge chips such as the 3nm will be produced in the planned Arizona plant. These diversification moves should reduce the geopolitical risks that TSMC faces, which have been present even before Berkshire bought the stock. Perhaps Berkshire’s assessed that tensions have materially escalated which makes the long term situation unfavorable to the chipmaker.
Best,
Ke Xuan