Monday, 13 January 2020

Navigating the Options Cycle on Tesla

Tesla is one of the most volatile large cap stocks in the market. It's a battleground stock with the war between longs and shorts waged in each buy or sell order, and each resulting fluctuation of its stock price. In about a year the price has gone from $377 in December 2018, halving to $177 in June 2019, subsequently tripling to close at $525 most recently on the 13th of January 2020.

The stock war is waged not just in the direct purchase or sale of the stock, or in the media, but in the options market too, where insufficient attention is paid by the ordinary investor, and where much descriptive value may be found. The options market can exhibit reflexivity, a circular relationship between cause and effect. The strike price of price of an option can act as a magnet pulling the underlying stock price, creating a virtuous, or vicious, cycle depending on where you have placed your own bets. The options activity in Tesla is unique in several respects, and understanding this will aide understanding of the volatility in the underlying stock price.

According to the Options Clearing Corporation the open interest in put options on Tesla stock amounted to 160 million units on the 13th of January. This is more than the number of shares in Tesla's float, an unusually large number of open options that suggests a large degree of influence over the underlying stock. There were 80 million call options on this date for a 2:1 Put-Call ratio. This is in contrast to an approximate 1:1 Put-Call ratio across the market and suggests a high degree of bearish sentiment toward Tesla stock.
The Put-Call ratio is a common, albeit crude, measure of investor sentiment. In the case of Tesla it is is significantly misleading, as less superficial analysis will reveal. Not all options are created equally and not all options should be weighted equally.

Two of the key factors in the relationship between a call or put option and the underlying stock price are the strike price of the option, and the expiration date. This is where two important facts will emerge.

1) Most put option volume on Tesla stock is currently deep out of the money, and most call option volume is in the money, as the following volume chart segmented by option strike price reveals:
Based on the 13th of Jan closing price of about $525 per share of Tesla stock, more than 99% of put options are out of the money, and 65% of call options are in the money.

2) Nearly half of call and put option volume expires on the 17th of Jan, as the following volume chart segmented by option expiry date reveals:
This means that there is only a narrow window of time remaining for movements in the underlying stock price to effect the intrinsic value of the outstanding options.

Combining the above facts under the assumption that counterparties have hedged their exposure, we can determine what has been the impact to the underlying stock demand of these contracts, and how this might change in response to changes in the underlying stock price. Options pricing theory provides just such a mechanism - Delta Hedging. For our purposes the most important thing to know about Delta Hedging is that the more likely an options contract is to be exercised, the greater the number of underlying shares the counterparty to that contract needs to transact in to hedge their exposure.
The chart above contrasts the gross number of options reported with the number of underlying share transactions that would currently be required to offset counterparty exposure to that option contract. The 80 million call options equate to 50 million share purchases. The 160 million put options equate to 5 million short sales.

The same options pricing theory that provided us with Delta-Hedging, the Black-Scholes model, also provides a mechanism to calculate the value of the call and put options, and the calculated values of these options track closely to recent trade prices of those options. This implies this is not simply a theoretical exercise but has some basis in reality with regards to the any underlying share transactions to offset the exposure on these options, and the observed activity in Tesla stock.

It's important to recognise that the number of shares required to offset an options position is not static and changes, most importantly, with the price of the underlying stock. The following chart shows the Delta Hedged volumes segmented by different underlying stock prices:
The chart shows that between the underlying stock price of $500 and $525 the number of share purchases required to offset the call position increases from 45 million to 50 million as more call options are deeper in the money and more likely to be exercised. Conversely, the number of short sales required to to offset the put position reduces from 8 million to 5 million as more put options are deeper out of the money and less likely to be exercised. This equates to a net increase in demand for Tesla stock of 8 million shares. Such a price move therefore induces as much demand for Tesla stock as the third largest institutional investor in Tesla, the Saudi Public Investment Fund, has accumulated over time.

The actual stock price increase of $47 on the 13th of Jan is enough to stimulate demand for nearly 17 million shares of Tesla stock, more than the single largest institutional investor in Tesla, Baillie Gifford, has accumulated, and all in a single day. This reflects a virtuous (or vicious) cycle where a stock price increase induces further demand for Tesla stock which all else equal will push the price up further as further counterparty share purchases are required to offset their options exposure. The reverse is also true, as a reduction in the share price would have the exact opposite effect. This self-reinforcing cycle explains the volatility and periods of momentum we have seen in Tesla stock to a significant degree, but there are caveats. There will be differences in timing and extent to which counterparties hedge their exposure. And other factors influence the stock price too, principally market participants deciding whether to continue to participate, or to place their chips elsewhere.

- Vincent

[Disclosure: long time holder of Tesla stock]

Sunday, 24 November 2019

The Man Who Solved the Market

By now many of you will be familiar with the investment record of Renaissance Technologies. A book on the subject came out earlier this month.
For the uninitiated their Medallion Fund has compounded at 66% p.a. for multiple decades, 39% p.a. after the most delicious fees in the industry. $1 invested in 1988 would be worth $19,500 in 2018. That is if you were fortunate enough to remain invested, as many early LP's were not. The fund size has apparently been capped at around $10 billion, in order to preserve high percentage returns.

We can never completely rule out that this is a blind-folded monkey throwing darts. But if that is the case then it is one hell of a monkey. Or it's an 800 pound gorilla.

This performance is even more striking because the fund apparently employs few, if any, finance domain experts. Mathematician Isadore Singer described Renaissance as having "the best physics and mathematics department in the world". This capability is combined with a data-driven, quantitative approach across multiple asset classes, and a reported 2 day average holding period.

It's a powerful story of just what is possible with data, and I love it because of that. But unless you also have a team of world class scientists, their approach may not be one that you want to try to emulate exactly. There may or may not be room for a few more players deploying similar strategies to the Medallion fund, and at similar scale. But their very existence suggests that for many others with a quantitative focus, their efforts may best be expended by doing what the Medallion Fund does not.

- Vincent

Tuesday, 27 August 2019


Another day, another great investment site. This one is Pininvest which curates investment themes, amongst other things. It's a great place to generate ideas and insight into domains which are topical and which you find interesting.

- Vincent

[Disclosure: uses Sharadar data]

Tuesday, 15 January 2019

Latest Updates

Here at Sharadar HQ we are pretty obsessed with expanding our datasets and making them increasingly valuable to our customers. A downside of the singular focus on expanding our datasets is that we are not particularly good at taking the time to tell our customers about these improvements.

This post is an effort to remedy this situation. We'll list recent updates to the datasets, and we'll keep adding to this same post for future updates. We have many future updates in the pipeline, so you can keep checking back here. At some point we'll also backfill the full history of updates, because we're sticklers for completeness.

This post can be read in conjunction with our progress post which charts the high-level long term trend in our expansion efforts.

Latest updates:

15 January 2019 - Added current & historical S&P500 constituents as a free update for our fundamentals customers (here).

15 January 2019 - Added corporate actions as a free update for our fundamentals and equity prices customers (here, here, here and here). Focused initially on stock splits and expanding in the future to cover a broader array of corporate actions.

5 January 2019 - Passed the 20,000 ticker milestone for coverage in our equity and fund prices offering (here).

26 December 2018 - Added Exchange Traded Debt (ETD) as a free update for our fund price customers (here and here).

2 November 2018 - Added daily resolution of price-based metrics as a free update for our fundamentals customers (here and here).

14 October 2018 - Added preferred stock and stock warrants to coverage as a free update for our equity prices customers (here, and here).

7 June 2018 - Passed the 15,000 ticker milestone for coverage in our equity and fund prices offering (here).

7 June 2018 - Launched our EOD fund price offering (here, and bundled with equity prices here). Focused initially on ETFs, CEFs and ETNs and subsequently expanded to include Exchange Traded Debt (ETD).

24 April 2018 - Added indicator (field) descriptions that are retrievable via API as a free update for our fundamentals, insiders, institutional holdings, equity prices and the bundle customers.

24 April 2018 - Added rich ticker/company meta data that is retrievable via API as a free update for fundamentals, equity prices and the bundle

- Vincent

Wednesday, 28 November 2018


Another excellent investor tool joins the ranks - Koyfin

I've learnt a lot from using it, and recommend that you check it out too.

- Vincent

[Disclosure: uses Sharadar data]

Wednesday, 24 October 2018


Udacity has launched a new nanodegree program focused on artificial intelligence for trading.
It uses Sharadar data for part of the curriculum and so of course I think the course is great. But in this case it is actually true. It has been a pleasure to deal with Udacity throughout. It was also kinda cool since Udacity played a big part in me personally upskilling from an advanced spreadsheet jockey to computer programmer all those years ago. I still recommend their free CS 101 course as the best way to get started.

Update: sadly it appears Udacity no longer offers the same introductory course I took years back, the prior link will now redirect you to a replacement course which is quite a bit different. The course I took and recommended was titled "Building a Search Engine", taught by the excellent David Evans.

- Vincent

Tuesday, 2 October 2018

The Street Is Probably Not As Short On Tesla As You Think

It's common knowledge that Tesla is among the most shortest stocks around. Jockeying for the top spot with the likes of Amazon and Apple, companies with 20x the market cap. Each price fluctuation spawns countless headlines and as many forgettable articles about the "billions" that have been made or lost in short span by those betting against the company.

There are a few problems with this.

The underlying Nasdaq published short interest data evidently looks at the "total" short position and ignores any long position that the short seller might simultaneously hold in Tesla stock. i.e. these are gross short positions rather than net. And then we don't get to see who are the actual holders of the short positions. Instead we are subjected to the media musings of a cast of self-identified short sellers whose portfolio sizes range from tiny to small, and who as such cannot possibly account for a material portion of the money that we are told is short Tesla.

Fortunately, due to the varying mechanics of establishing a short position there is another data source that offers us more visibility - institutional holdings data which is disclosed to the SEC via form 13F and which allows us to look at individual investor holdings of stock, call options and put options. Peaking behind the curtain at such data we find that the top put holder also happens to be the top call holder. The same is true of the number 2 put holder. This is a surprising result with potential implications for everything that is written about short selling of Tesla.

But first let's quickly motivate why the SEC institutional holdings data can be used as a complement or substitute to the Nasdaq short interest data. There are two main mechanisms for going short a stock: (1) by borrowing and short selling directly (represented by Nasdaq data); or, (2) by buying put options (represented by SEC data). For more detailed explanations of these mechanisms see here. Many of the Tesla short sellers identified on television or in the pages of Seeking Alpha are the holders of put options.

Importantly, buying put options and direct short selling are not unrelated since the counterparty to a purchase of put options may offset their exposure by themselves establishing a direct short position. Therefore, to the extent that a short position is established via put options, and the counterparty offsets their exposure through a direct short position we should expect their to be overlap between the Nasdaq short interest data and the SEC put data.
The question is therefore, in the case of Tesla how much overlap is there between the Nasdaq short interest data and the SEC put data?

It's not a question that can be answered definitively since the Nasdaq data are completely opaque. However, viewing the long term trends in each source does allow us to make an educated guess.

To the extent the above trends are aligned the data sources are likely to overlap each other. Clearly there are periods of difference, which may themselves be instructive, but the sources are of the same magnitude and trend with a tendency to converge and intersect. The sources are more alike than they are different and this is unlikely to be accidental.

If we accept that the SEC put data are similar enough to be used as a proxy for the Nasdaq short interest data we can derive the "net" short position, as above, after adjusting for long positions held by put holders. This net short position is drastically reduced and, on numerous occasions, negative.

This suggests that not only are short positions and their reported wins and losses greatly exaggerated in the countless articles on this subject, but they may even be directionally incorrect at times!

Last, there remains the curious question of why the largest short position (put holder) would also be the largest call holder?
The firm in question, Susquehanna International Group ("SIG"), is a quant focused shop with $291 billion of reportable holdings at the end of June. Not quite the same scale as Tesla institutional longs such as Fidelity and T Rowe Price, but a lot closer than any of our media musing short sellers. $4.5 billion of this is short Tesla via put options, $2.7 billion is long via call options, and nominal amounts of Tesla stock and debt are held too. SIG's net position has varied over the years and has been net long on occasion. Interestingly, SIG is also the largest put and call holder in Apple, Amazon, Google, Netflix and Facebook (and even QQQ) - suggesting that the overall short positions of these stocks are also likely to be greatly exaggerated.

We don't get to see the durations and strike prices of SIG's options. Nor do we get to see the details of any call or put options that SIG has written. But at an aggregate level from what we do see it appears that SIG has most recently been betting on price declines *or* volatility. Bets which would probably have rewarded in recent days:
Some quick housekeeping before I go. (1) I am personally long Tesla stock, and have been accumulating shares since prior to the Model S release; and, (2) the SEC data used in this article can be found here.

- Vincent