Regular readers will forgive me for recycling a bit of my own copy. There are only so many ways to introduce the same set of data — only so many times I can paraphrase myself.

With that in mind, let me reiterate that when it comes to global growth canaries, you could do worse than Asian exports and factory output. In 2023, Taiwan is my preferred canary. As I’m fond of putting it, the island is a barometer for two kinds of hard landings: Economic and geopolitical.

So, I like to update industrial production figures and export data out of Taiwan each month. Early this week, figures from the island’s economic affairs ministry showed IP fell 16.63% in June. It was the 13th straight monthly decline.

A subindex of manufacturing output declined 17%. It too has fallen for more than a dozen consecutive months.

The problem: Slower external demand, which is to say global demand in this case.

The same issues plagued exports in June, when shipments fell a 10th month.

The near 25% decline was the second-largest of the current slump, the worst since the financial crisis.

Note that the six-month decline (so, the first half drop) in industrial production was likewise the deepest since the GFC. The IP index plunged 17.6% during H1 compared to the first half of 2022.

I’ll take this opportunity to highlight the commentary from Taiwan Semi’s call last week that raised a few eyebrows, as well as a couple of soundbites from analysts.

From Taiwan Semi’s call:

The short-term fancy about the A.I. demand definitely cannot be extrapolated for the long-term. Neither can we predict the near future, meaning next year, whether the sudden demand will continue or flatten out.

Three months ago, we [were] probably more optimistic, but [not] now. For example, China’s economic recovery is actually weaker than what we thought. And so the end market demand actually did not grow as we expected. Put it all together: Even [though] we have very good A.I. processor demand, it’s still not enough to offset [the] macro.

Higher inflation and interest rates impact end-demand in all market segments, in every region in the world. Under such conditions, our customers are more cautious in their inventory control in the second half of this year.

Mike O’Rourke, JonesTrading:

TSMC is the most important Tech manufacturing company in the world. Five companies represent more than 50% of TSMC revenues, and that is an exclusive list that includes two of the Magnificent Seven. According to Bloomberg, Apple is 23% of its revenues, Qualcomm is 9%, AMD is 8%, Nvidia is 6% and Broadcom is 5%. Thus, when TSMC management highlights slowing tech spending, it should not be ignored. TSMC management was clear that it is seeing slowing relative to their expectations. The slowing is global, with China being a key part, and the novelty strength of A.I. is not nearly enough to offset the slowing.

Mike Wilson, Morgan Stanley:

Artificial Intelligence has provided a very constructive theme for investors to adopt. In addition to the clear winners like NVDA and MSFT, many investors are getting excited about the longer-term productivity benefits from such technologies. While we share that excitement, we also believe it is premature to extrapolate such benefits across the entire economy and company earnings for this year, particularly given we are still facing growing risk of cyclical headwinds, as data are suggesting the business cycle is increasingly at risk of a slowdown. This past week, Taiwan Semiconductor helped to confirm this sentiment as it relates to the semiconductor industry. More specifically, the company confirmed that A.I. chip demand is booming, but it’s not enough to offset the cyclical slowing they are experiencing across the broader semiconductor landscape. These comments and meaningful reduction in guidance were a surprise to many [but] this dynamic fits with the title and main theme from our mid-year strategy outlook—“Cyclical Bear within a Secular Bull.”