2023-10-16 14:40:05 ET
Summary
- IOO has a heavy skew towards megacap IT stocks, raising concerns about AI optimism and consumer electronics demand.
- Despite being marketed as a global and diversified ETF, IOO's exposure to US stocks and higher expense ratios make it less efficient.
- The ETF's performance is driven by AI fervor and trends in electronics demand, but there are concerns about the sustainability of these factors.
- In general, there is also the issue of higher costs of capital which have an outsized impact on higher-multiple issues. We are staying away.
The iShares Global 100 ETF ( IOO ) has a skew towards the megacap IT issues that dominate US indices. We reiterate or concerns with AI optimism, and moreover, we worry about trends in consumer electronics demand. We continue to believe that the economic setup is quite dire and that pain is coming for the world economy as CPI remains stubbornly high and job markets tight. There isn't much margin in equity exposures currently, and the IOO that captures broad markets doesn't strike us as terribly attractive or advantaged at this moment in time.
IOO Breakdown
Exposure in IOO is supposed to be global and more diversified, but US exposure is still 75% , and the fact that foreign larger-cap stocks tend to be in IT means that the skew towards IT in IOO is even more intense than for the SPY .
Despite the limited additions of foreign exposures and orthogonal allocations, expense ratios are still substantially higher at 0.41% instead of 0.03%, which is not terribly efficient. The only benefit to IOO is that the PE is 20x instead of 22x owing to more foreign exposures which tend to be lower multiple.
On the matter of tech, there is meaningful skew towards some of the largest cap tech stocks due to the fact that IOO has a lot fewer holdings at around 100 instead of 500 for the iShares Core S&P 500 ( IVV ).
Top Holdings (iShares.com)
Apple ( AAPL ), Microsoft ( MSFT ) and the other megacap names feature heavily. Broadly speaking, performance has been driven by AI fervour and by trends in electronics demand.
Bottom Line
There are some concerns in electronics demand to start. Outside of GPUs which are being bolstered by AI-based demand, electronics demand is not particularly strong, and it is being seen upstream in chemicals and basic materials and also downstream in electronics value chains. For Apple, some of the initial data from contract manufacturers doesn't look so good . Insiders are also selling .
As for GPUs, and all the related AI hype, there is likely to be some disillusionment . A portion of the current demand, it is difficult to say how much, are premiums being paid by companies for the AI business option. Not every company is going to be able to use AI to great effect. There are structural issues too. AI requires massive upfront investment, and even after that some of the best models are inexplicably beginning to perform worse in terms of generating working text, including executable code and performance in logical reasoning, which is concerning. While models aren't trained on causal relationships, just statistical relationships between words in documents, the ability to perform more sensitive generations seems to be declining. The running expenses are also extremely high in terms of computation, and the ongoing efforts to de-bias text can have unintended consequences on performance, but are also a necessary expense to avoid unsavoury generations that unchecked could create PR problems for customers. On the business side, there seems to be reduced traffic to the ChatGPT website. While API demand might be growing nicely, some element of the novelty is wearing off at least for retail users.
There is also the fact that the backlash against AI will only grow. In the EU, many touted AI use-cases are essentially not legal, and will be punitively sanctioned by GDPR. Even some of the intended use cases, which may be technically permissible, may open the door to abuse that will make it also unsanctionable or will subject it to substantial regulation so as to make it less competitive with human capital. The only reason you're not hearing more about it is because regulators are still coming to grips with understanding the technology. But it will be a nuisance throughout the entire development of LLMs in particular in Europe . Union strength in Europe is also going to massively limit the market compared to the ebullient expectations that have been able to bid up prices of equities despite decade-long records in capital costs for companies. In general, the power of AI will galvanise incumbent interests to regulate it which will impact growth prospects.
In addition to these idiosyncratic concerns, there is the issue that a higher-than-market PE at around 20x exposes the ETF to a greater degree to upward revisions in capital costs. Higher expectations inherent to higher multiple stocks have more scope for downward revision when those long-dated cash flows get revised downwards. We have every reason to believe capital costs will continue to surprise upwards as labour markets remain tight and inflation high, including from exogenous factors related to commodities. There is still scope for upwards revisions in those long-term rates as well with deglobalisation and other long-term inflationary factors to be concerned about. There aren't especially more reasons left in the tank to see upside in equities. We would be cautious on IOO and other ETFs capturing broad, high multiple market indices.
For further details see:
IOO: Softening Electronics Demand, A Lot Riding On AI