Does TSMC’s Share Price Signal the Burst of a U.S. Stock Market Bubble?

I enjoy reading the occasional essays written by Owen Lamont, a portfolio manager at Acadian Asset Management (“Acadian”), a global, system-based asset manager headquartered in Boston. In his latest piece, “TSMC: Totally Stupid Market Chaos”, he argues that the ADR of TSMC listed in the U.S. is trading about 20% more expensive than the common stock listed in Taiwan, and claims this is evidence that the U.S. equity market has entered bubble territory. I wanted to share his line of reasoning with our members.

To start with the conclusion, I do not agree with his view that the U.S. stock market has entered bubble territory (more precisely, my reaction is closer to “So what if it is a bubble?”), but I do think it is an interesting perspective.

As of October 2024, the price gap between TSMC’s common stock (listed on the Taiwan Stock Exchange) and its ADR listed on the New York Stock Exchange is said to be around 20%. In other words, you are paying 20% more in the U.S. for a share that represents the exact same claim on the same company. Considering that TSMC’s market capitalization easily exceeds 1,000 trillion won, this implies an arbitrage opportunity of some 200 trillion won. Contrary to Eugene Fama’s Efficient Market Hypothesis, which holds that markets become more efficient over time, this gap has widened from around 0% to about 20% over the past two years.

In plain terms, because it is institutionally impossible to buy TSMC shares in the U.S. market and then deliver them to the Taiwan market for sale, mechanical arbitrage cannot take place. As a result, the law of supply and demand does not push prices toward a single unified level. The same mechanism explains the so‑called “kimchi premium” in Bitcoin, which arises due to foreign exchange regulations in Korea.

He suggests that this price gap can be explained by four possible scenarios:

In reality, shorting TSMC’s U.S.-listed shares is possible and not particularly expensive. This suggests that fund managers refrain from shorting the ADR mainly because they are worried about scenarios (3) and (4).

In conclusion, unless there are legal restrictions or extremely high transaction costs on buying the Taiwan-listed common stock, the author effectively views investors who buy the U.S.-listed TSMC shares as “stupid,” and seems to be saying that the price gap has grown this large precisely because there are now “so many” such investors.

LOOP is an economic principle that in an efficient market, identical goods should trade at a single price. He cites a couple of academic papers showing that LOOP has frequently been violated when markets were in a bubble state.

The chart shows the premium (price gap) of TSMC’s U.S.-listed shares over the Taiwan common stock, as compiled by the author. It reminds us that the premium hit an all‑time high in the 70% range during the dot‑com bubble around 2000 before collapsing, and that it also peaked right before the 2008 financial crisis and again in late 2021.

Of course, Owen’s argument could ultimately turn out to be right. Until just a few days ago, AI‑related names seemed to rise without end, and the sharp sell‑off in tech stocks—despite strong earnings from the big tech companies yesterday and today—feels almost like an eye‑opener after reading his piece.

But let’s not forget that the macro economy is shaped by countless variables we cannot fully enumerate. AI may now be able to gather and analyze even unstructured data, but it is still humans who must look at the output and interpret it. And humans are famously consistent at one thing: failing to predict the future.

The following is an excerpt from the conclusion of “2024 Major Market Outlook.”