Mechanical, Technicals, And An Election
Submitted by Peter Tchir, Academy Securities
Another hectic week! This market is not getting any easier to manage.
I want to refocus attention on a subject that continues to come up in conversations and that we tried to address in last week’s report: Stocks & Credit: Cheap but Getting Cheaper? Energy: Not So Much? That subject is why energy prices are still high, if, according to reporting, as much oil is getting out of the Middle East as it was before the war started. We addressed this in detail last weekend, so please refer to last weekend’s report for more details, but here it is, in a nutshell, why these reports are misleading:
- The reports don’t include Iranian oil, which, due to the blockade, has slowed to a trickle. That is probably between 1.5 million and 3 million barrels per day less.
- They don’t account for increased shipping time and shipping costs, helping elevate prices and slow delivery to refineries across the globe.
- It is difficult to measure how much is making it through the Strait, with transponders off. What seems to happen is periodically a “convoy” makes it through, boosting flows, which are then dramatically reduced due to attacks and the time it takes for another group of ships to attempt it.
- Finally, and more importantly, diesel, LNG, and fertilizer are not being exported at even half of the pre-war levels (by even optimistic estimates).
That question came up so often, it seemed important to address it again.
Before jumping into this week’s report, let’s quickly talk crypto. With allegations that a North Korean-sponsored entity hacked a crypto company for over $6 billion worth of crypto, it seems like a good time to re-engage on the concept of Could the U.S. License Privateers to Combat Crypto Theft?
Technicals
Let’s just start with an updated version of last weekend’s SOX (semiconductor index) chart.
Markets initially bounced off of the 200-day moving average, but on Friday, they retested that level and failed. It is often the 3rd retest that is the problem, so a bounce again is quite possible, but breaking the 200-DMA rarely is a good thing.
The Nasdaq 100 is solidly through its 100-DMA and is about 1.6% from testing its own 200-DMA. On Friday, the S&P 500 just broke its 100-DMA for the first time since November of 2025. The 200-DMA is 3.5% away on the S&P 500, which should be support.
We are resorting to technicals more than usual because:
- The technicals are aligning with our views on risk and pricing quite well.
- In a market that is whipsawing investors of all types on an almost daily basis, there is a tendency for risk managers to rely on technicals more than they would otherwise. That increased focus on technicals tends to make them work better than they otherwise do.
From what I see anecdotally and through some simple charts I like to use, the market is not exhibiting strength, and the risk is to further downside.
One positive is credit markets didn’t suffer as badly as stocks on Friday (spreads did NOT finish at their wides) and none of my indicators for “disorderly” selling pressure have triggered. I am watching those triggers as that is a risk that is still in play.
Mechanical
There was solid “buying the dip” in SOXL (a 3x leveraged ETF on the SOX index). Those inflows create demand for the stocks in that index. TQQQ (a 3x leveraged ETF on the NDX index) saw some buy-the-dip trading as well, but not on the scale of SOXL.
We have argued that the mechanical nature of leveraged ETF balancing is disproportionately large in the semiconductor space. Not only is there SOXL, but there are also a large number of other ETFs over $1 billion with leveraged exposure to the index or single names.
Let’s put the SOXL rebalancing in context. These are approximations but illustrate the point effectively. For our example, we will assume the ETF size is $20 billion on Thursday (about right for SOXL).
Let’s say the index it references is at 100 (for SOX the official index was 11,518 on Thursday, but it is easier to work with 100; I am a metric sort of person anyway).
This table shows how the fund should be positioned Thursday after the close for Friday, so that it would have met its required 3x return for holders on Friday. But, most importantly, it shows how the fund would have to be positioned after the close for the start of trading on Monday.
To provide 300% of the change on Friday, the ETF effectively borrows $40 billion and buys $60 billion of shares of the underlying index (so 600 million shares). The ETFs tend to use swaps and other ways to get their exposure, but this is “close enough” for the mechanical part.
Let’s say on Friday the index dropped 7% (it dropped 7.28%, but 7% makes the math easier and this isn’t meant to be a difficult mathematical exercise, even though it is starting to look like one).
The index is 93. The fund assets are 600 million shares now worth $55.8 billion. You repay the loan of $40 billion, and you have fund assets of $15.8 billion, which is exactly what the fund should be worth. Losing 21% on $20 billion is a $4.2 billion loss, leaving the fund at $15.8 billion.
Ignoring inflows and outflows (which we will do for the moment), the fund has to mechanically prepare to deliver 3x the daily change on Monday.
The ETF size is down to $15.8 billion on Friday (ignoring the inflow/outflow component). The index it references is now at 93.
The ETF borrows $31.6 billion (to get the leverage) and has $47.4 billion to buy the index ($15.8 billion *3). At an index value of 93, it purchases just under 510 million shares.
This ETF, which owned 600 million shares at the start of the day, is forced to sell over 90 million shares at the close, to be properly rebalanced for Monday.
On a 7% move in the underlying index, this hypothetical ETF had to sell 15% of its holdings!
90.3 million shares sold at $93 is $8.4 billion. Is $8.4 billion of mechanical selling in a market this size meaningful? I would argue that it is, because the total size of mechanical rebalancing is bigger than just SOXL. Also, this is selling on a day that there must be overall selling of the names in the sector, so it just adds insult to injury (or whatever the opposite of insult to injury is on good days).
This rebalancing is purely mechanical. There is no subjective behavior available to the ETF. This is what it must do if there were no inflows or outflows. It is one of the reasons that if an index referenced by a leveraged ETF is volatile (large daily moves) but goes nowhere in price over time, the leveraged ETF will underperform an unleveraged position (buying higher and selling lower takes a toll); this is why every one of the leveraged ETFs has a “warning” about being intended to be traded, not owned.
In any case, on Friday in the real world, SOXL had inflows. So, the selling pressure would have been lower (if there were net sellers rather than buyers, the selling would be even greater). The net creates and redeems of the ETF occur throughout the day, so that buying or selling isn’t necessarily concentrated at the end of the day, unlike the rebalancing.
Every single day, on the way up, and on the way down, this happens. Mechanical rebalancing by leveraged ETFs amplifies moves in both directions. The amplification is more noticeable on big days, but it is there every single day.
Every leveraged ETF in the tech space, whether linked to a single name or an index, has to follow these mechanical procedures. This segment of the market, with a proliferation of leveraged ETFs, helps drive prices higher in good times and lower in weaker times.
In a separate but related thought, I often think of tech as growth. Mag 7 is growth. SOX is growth. Yet VTV, the Vanguard Morningstar Value ETF, has about $200 billion in market cap. Its largest holding is Micron at 4.56% of the fund. One thing I’ve only been starting to take notice of is how what I would have considered growth has worked its way into value.
There is absolutely nothing wrong with that. Though it does potentially leave investors who thought they had an allocation to growth and an allocation to value with little overlap, possibly having more overlap than they thought. Many of these growth companies are also meeting the requirements to be owned by “value” investors. I suspect that managers with a lot of leeway will be overweight these names because you needed them to perform and attract AUM.
Probably not a big deal, but I suspect it makes risk management more difficult as what you may have thought were two distinct factors, aren’t as distinct as you possibly thought.
That is an issue we have seen with AI in general, where AI exposure is in so many factors (growth, momentum, etc.) that managers are trying to figure out better ways to examine their true exposure to AI.
These mechanics only accentuate moves and are not in and of themselves responsible for moves, but I think we are in a moderately more dangerous place than we would be without them at the moment.
This sector is incredibly important to markets and the economy! Yet it is a concentrated sector that makes up significant parts of various indices and has a relatively high proportion of leveraged assets attached to it (nothing like some of what we see in the KOSPI, but I’m not sure that is a market we aspire to).
Election
By Wednesday morning we should know how people actually voted.
Will the validity of polls take yet another beating? Possibly.
Will prediction markets turn out to be useful? Increasingly I question the value of prediction markets in an election. Do investors “bet” with their heart vs their mind? Are there biases based on who trades in those markets? Heck, if I were running a PAC, I might put some money to work in the prediction market instead of buying more ads.
- I think it is safe to say people are sick of all the ads, so what is the marginal value?
- If you bet right, you have the money for the next go.
- My gut is that you buy your candidate, making them look more likely to win, hoping that energizes that side and demoralizes the other. Though maybe that would backfire? Maybe you bet against your candidate?
Using the prediction markets to try to influence outcomes is probably not happening, but I wouldn’t rule it out altogether. BTW, prediction markets were about as wrong as possible on the elections in Brazil, which is maybe why I tried to steer away from that particular geopolitical subject last Monday on CNBC. Since we mentioned this in the interview, this is as good a place as any to mention that European stocks outperformed the U.S. this week (despite the headlines that have been largely negative).
I am much more concerned about the elections for the markets than I think I’m hearing from others. Though maybe some fear of the election also contributed to recent weakness?
I am less concerned about who wins and loses what (though that will influence markets), but I am much more concerned by how those who win or lose behave.
That is a risk that certainly a decade ago didn’t really cross my mind. Get some concession speeches. Get some victory speeches. Rush to get some things done if there is a transition. Wait until the new year if your position strengthens. All within what we have come to expect.
That should still happen and probably will, but I do have a small concern that we could be in for more volatility from the elections than people are pricing in.
Bottom Line
We didn’t spend much time on yields; in part, we covered it in The Fed and Compute Spend on Wednesday. I think, for now, Treasuries continue to do better when there is concern about the “compute” spend (less competition for bonds and some risk of a slower economy, and vice versa). I think we can safely say the Fed is hoping to not have to hike and will fight the data as much as they can, to avoid hiking.
Credit issues remain directed and trading remains orderly. That is key for equities, but I am watching to see any signs that credit market structure integrity is breaking down. I’m nervous about markets due to technicals and the mechanical nature of much of the trading in and around some of the most important sectors.
I don’t think Tuesday night’s elections are going to decrease volatility. Typically, “knowing” the outcome helps, but I’m not sure it will. I do suspect that by Wednesday morning we will have more questions about polls and prediction markets than we did before! I am scheduled to guest host on Bloomberg TV on Wednesday morning; it could be an interesting hour.
Have a great weekend and hopefully we can all navigate these markets well!