Leaves on the track
At more buyers or sellers we have a general rule for life that if everyone is doing something, we're not doing it. The premise is that this way we will never end up stuck in queues, and have many more opportunities to achieve something interesting (which could be good or bad, but should help us to avoid ending up like the man referenced in the title).
Some examples? Let's see:
Everyone else uses a free VPN app on their phone. The company that developed this app then sells to hedge funds the usage data for all apps on the phone. Our phone doesn't even have a mail app installed. We frequently get junk texts addressed to Marylin, a lady who we believe died two years ago. Reggie would be proud.
Everyone else speeds on straight roads and slows down to take corners. We don't speed, except possibly in corners, where all the fun is. This delays us on highway journeys, but does mean that we don't get snarled in traffic. (Honest, just try driving at 60-65 on the highway and see what a different experience it is.)
Everyone else (at least in our part of the world) thinks that Autumn is the best season. They presumably enjoy getting stuck in traffic queues going to see leaves. Our favourite season is Spring, when we don't need to tidy leaves, and we get helpers to cut the grass:
We can see why you might want a mail app on your phone, and to get to distant places faster, but are you seriously telling us you don't like being visited by baby bunnies? And have you seen baby foxes?
Other than some gratuitous cuteness, what is the point here? Well simply that if everyone is long a particular stock, at some point they might well find themselves stuck is a big queue trying to get out of it. How can we determine if this is the case, and thereby do the opposite? Let's investigate.
I didn't get where I am today without recognizing another slight wobble when I see one
It would be great and super if you could read this heading in the voice of CJ. Apologies if this is a bit of a rabbit hole. We're sure you didn't get where you are today by going down rabbit holes.
One of the nuances of sell side research is that analysts are disincentivized to argue that stocks should fall. (Nuance, bias; po-tay-toh, pot-a-too). A cynical view of this might be that if they are correct, there will be no bonus pool with which to reward them. However, as you may know from reading analyst certifications, pay isn't determined by the views expressed in research reports. You may laugh for once, but the reality may indeed be that it is in the analyst's career interests to avoid telling all their sales colleagues who fill out year-end reviews that their product stinks.
To avoid scaring people, when an analyst thinks a stock is likely to fall, they will often forecast 'volatility'.
Why is volatility synonymous with stocks falling?
We think it is a manifestation of risk-averse behaviour.
Specifically, when adding to risk, it is typical to try to avoid moving the market too much, thereby destroying any alpha. However when getting out of risk, particularly on a stop, a more aggressive order might be preferred to ensure it is 100% filled. If there is a big queue of people trying to put on the same aggressive order, that will serve to magnify the size of the price moves. Therefore, we can infer the general state of positioning by measuring the size of price moves in each direction.
Let's look at an example from last week in DKNG.
A short term mean reversion model might look at this series of prices and think that the market is probably short and so the current range should break to the topside. However, in that range, most of the large returns are all red bars. This means that the stock is still trading as if there are long positions that are being sold aggressively.
Sure enough, on Friday, the range broke below with a bang. After this capitulation, DKNG looks much more attractive to us in the short run, for reasons we'll discuss next.
How to avoid being 11 minutes late
Using large price moves as a reverse indicator of positioning is perhaps counterintuitive to you. If you were long but then sold aggressively, you are no longer long, so the measure seems like it is always a little late. This is a fair point, and it means that using this indicator systematically is challenging. Fortunately, options markets are here to opine on what future volatility will look like.
As a hypothetical example, suppose that you own shares in a blancmange manufacturer and are concerned about a slight wobble. Earnings are next week, and if they come out soggy, this wobble may become a collapse. You may wish to purchase a short dated put option, just to get you past this event.
If you do so, you have effectively transferred a portion of your blancmange temporarily to a market-maker. This market-maker will likely hedge the risk, causing a small sell-off in the stock, and will raise their asking price for puts (to avoid getting more of them).
The market maker sees all the customer flow in the blancmange stock and so has a good idea about positioning. If the market is particularly long, they may choose to raise the price of the puts further in anticipation of additional demand. Therefore we may infer that an increase in the cost of puts that is not explained by the stock price selling off is a signal about positioning from the market maker. We can combine this with our measure from past price moves to get a more complete picture.
You may argue that if you buy a put, you are no longer a forced seller, so this could be a bullish signal. This is certainly a possibility, although we try to account for this by looking at the unexplained changes in the prices of options. The way it should play out however is that it is only bullish after the event risk has passed, or we have seen a dip. The reason for this is that the largest buying this put foreshadows is when you monetize any gains on it by selling the put after the event or after a dip.
Shockingly, I wasn't able to find any blancmange manufacturers with earnings this week to use as an example, so let's look at an equally old fashioned company in GE.
GE reports on Tuesday (26th). As with DKNG, the larger price moves in the recent range have been down, so it appears that the market is long. On Thursday there was a slight wobble in the form of an outside reversal day. On Friday it retraced higher, but the activity in options was telling, with puts being bought for the 29th across the whole spectrum of strikes. This points to initial weakness this week.
However, we expect a dip on earnings to be bought in GE. In part this is due to the the dispersion of strikes, which suggests that many people may be hedged. Moreover, while the 1 week options became more expensive, options at a 1 month expiry cheapened. Abstractly, we interpret the longer dated volatility as a measure of risk premium, and the fact that it did not increase suggests that the view of fundamentals for GE has not deteriorated.
A similar analysis of option activity and positioning is supportive of BAX,CAT,EBAY,GM,LLY and PINS ahead of earnings this week, but is less keen on MSFT,NOC and NOV.
Now what about DKNG? Well they report on November 5th, and someone appears to be expecting fireworks, because on the dip on Friday, they purchased 5,000 55 calls for that date. We think a concentrated bet like this is more likely to based on good information than a similar volume spread across multiple strikes.
Market view: Don't trust the easy chairs
Last week we thought that the market would be volatile on Monday, but after initially trading lower, this volatility was all to the topside, which set the tone for the week.
The sentiment picture from Reddit again points to a volatile start to the week, having dipped substantially on Friday.
This week our positioning models are not as supportive of the market. The cost of puts on SPY rose