macro notes: china + an incomplete section on bubbles and overvaluation

China

Peace is never an option, which meant that as the big western lords patted themselves on their backs (see: Shehbaz Sharif's endorsement for Trump's Nobel), the Chinese came out with tightened export controls on rare earths. China controls an extraordinary amount of rare earth production (61%) and processing (a mighty 92%!). The new rules require overseas firms to obtain Chinese government approval before exporting products containing even trace amounts of certain rare earths that originated in China.

Trump responded with a threat of 100% tariffs, along with export controls on critical software, causing the S&P to close down 2.7% Monday. This was followed by him placating the world:

The Chinese relationship has always been quite interesting (and clearly, quite impactful to the global economy). In fact, this great book I've been reading (Trading FX by Dirk Willer [Citi Macro Strategist]) has an entire chapter dedicated to China, and the fact that China is the only "outside" force in the last few decades to have a meaningful impact on Fed policy. It's the only market to have an effect on the Fed's primarily domestic mandate.

Additionally, this podcast episode by Brad Setser had some excellent points, noted down below:

• China - export + manufacturing giant
• Imports have hardly moved but exports have increased a lot
• Chile had to close its only seal foundry
• Malaysia can't compete on the high end (Ex. Highly technological goods) and has to compete on the low end - textiles and others
• Domestic weakness in China
• China used to intervene through PBOC but now using banks (they do interesting things (aka corporate lending, cross-currency swaps)mas opposed to just buying treasuries)
• It's very hard to move to balanced trade because Chinese goods are still cheap and third party imports are possible
• (Higher interest rates = weaker FX for EM, but not reflected in CHina)
• China is buying huge amounts of forex
• 
• Pharma and gold are not included in the tariffs
• The only changes are in the places of final assembly - no change in surplus, deficit, it has put pressure on European autos
• Save for gold and pharma the volume of imports hasn't changes - but they are tariffed at a higher rate
• Seems like importers are absorbing the tariffs - not paid by foreigners neither passed to consumers completely - over time 60-70% will be passed to the consumer (if tariffs stay where they are, which they clearly haven't)|
• China can’t really run a $1.2 trillion surplus without trade with the US
• 2 mechanisms to trade:
• -          The standard “high tech goods” mechanism
• -          More trade to Europe allows EU to be more free to export?
• The only sanction free asset is gold (and should be stablecoins in the future)
• Money is in sovereign funds/quasi sovereign investment funds
• China is effective at coordinating state-sponsored economy control
• Where does China put its interest income?
• CNY is undervalued by about 2.5-3%
• The deals don’t make sense – because if you believe in America first why would you want the future profits of the US capital investment flow to the rest of the world
• If one wants FDI, establish an exchange rate rather than government sponsored deals
• We’ve come into a backdoor globalist government

The takeaway from all of the above is that the overall effect of a rush to set up alternate supply chains (something already in progress) is just going to speed up. That said, Trump's non-focus on stopping the depreciation of the Yuan leaves the real problem of encouraging corporates to actually invest meaningfully in the US as is.

Next: Javier Milei and the Argentinian swap

Favourite Tweets

Incomplete Title: Howard Marks

Howard came out with this interview on CNBC a few days back. I've highlighted some statements below, especially the last one:

"My response to date has been that the valuations are ... high but not crazy."

"Expensive and going down tomorrow are not synonymous."

"There's no such thing as a price too high."

"The S&P 500 is expensive. The forward PE ratio is 24 compared to the historical average, BUT, the companies are better too (ergo, the ratio is justified, albeit expensive)." [Note that this is better outlined in the below memo, section discussing 'bull case']

Additionally, he refers to his latest memo:

The Calculus of Value

In his latest memo, Howard Marks sets forth the essence of value and price, as well as the critical relationship between the two.  

Oaktree Capital

As always, a brilliant piece of writing. I've been thinking a lot about that memo, in conjunction with the interview above, and here are a few points:

1. The bull case (justifying asset prices) is this:
   
   Compared to the past, today’s S&P 500 is increasingly made up of companies that (a) grow faster, (b) are less cyclical, (c) require less incremental capital to grow, enabling them to generate more free cash flow, and (d) have much stronger competitive positions or “moats.” Thus, they deserve above average p/e ratios.
   ...
   I just have no idea which of those two concerns is more valid today. But investors should bear three things in mind:
   - the enormous likelihood that AI and related developments will change the world,
   - the possibility that it is “different” for some companies – those that truly embody the factors listed above and will demonstrate the “persistence” I described in On Bubble Watch, but also
   - the fact that in most “new, new things,” investors tend to treat far too many companies – and often the wrong ones – as likely to succeed.
   
   Let's break down this down by the points mentioned above.
   Do companies grow faster? Yes, a quick "AI" estimate puts the average unicorn time (time to reach $1B) for companies founded between 1990-2005 at ~10 years against an average of ~4.5 years for 2018+ ones. That said, I'd argue that this lower statistic is largely driven by the fact that
   Are they less cyclical? Presumably, with tech stocks and changing business models.
   Do they require less incremental capital to grow? Highly debatable, especially in the current extreme capex cycle with AI growth.
   Are "moats" stronger? Yes, and the "consolidation" story, with large companies gobbling up usecases is part of this.

Some interesting comments from HackerNews:

But at least my perception of the idea that this is a "bubble" presently is rooted in the businesses that are created using the technology. Tons of money spent to power AI agents to conduct tasks that would be 99% less expensive to conduct via a simple API call, or because the actual unstructured work is 2 or 3 levels higher in the value chain, and given enough time, there will be new vertically integrated companies that use AI to solve the problem at the root and eliminate the need for entire categories of companies at the level below.

In other words: the root of the bubble (to me) is not that the value will never be realized, but that many (if not most) of this crop of companies, given the amount of time the workflows and technology have had to take hold in organizations, will almost certainly not be able to survive long enough to be the ones to realize it.

This also seems to be why folks draw comparison to the dot com bubble, because it was quite similar. The tech was undoubtedly world changing. But the world needed time to adapt, and most of those companies no longer exist, even though many of the problems were solved a decade later by a new startup who achieved incredible scale.

I imagine it'll be closer to a slow deflation as these companies need to turn a profit.

Fundamentally, serving a model via API is profitable (re: Dario, OpenAI), and inference costs come down drastically over time.

The main expense comes twofold: 1. The cost of train a new model is extremely expensive. GPUs / yolo runs / data

2. Newer models tend to churn through more tokens and be more expensive to serve in the beginning before optimizations are made.

(not including payrolls)

OpenAI and Anthropic can become money printers once they downgrade the Free tiers, add ads or other attention monetizing methods, and rely on a usage model once people and businesses become more and more integrated with LLMs, which are undoubtedly useful.