Signal du Jour - Short vol in SMH

Earnings in NVDA and potential vol crush.

The main event for the week should be tomorrow, post-market when NVDA releases its Q1 2024 earnings. The stock generates wild expectations and debates about the industry. Is AI a bubble? When will it burst? Why does NVDA make money when others don’t?

We don’t have the answers to these questions, but if you’ve been following us for a while, you know we have some opinions as ex-AI practitioners.

However, our approach to volatility trading involves having no opinions: we sell something because it is expensive right now, based on the data, and our personal views don’t interfere with the process.

Today, we consider a trade-in SMH, the ETF that exposes investors to semiconductors. Interestingly, the lack of consensus and the wide divergence in views and opinions on NVDA makes SMH volatility expensive.

Let’s take a look.

The context

"Software is Eating the World" was published by Andreessen ” almost 13 years ago. The VC firm argued that with infrastructure at a maturation point (think hardware like smartphones, more powerful laptops, and, most importantly, networks like the Internet), the software industry would transform the daily lives of billions of people over the next decades. They were right.

Yet, after decades of capitalizing on software providers, we are now back in a hardware era: without strong microchips, forget about truly bringing AI to the public. That could explain the frenzy around semiconductors in a paragraph: since the low of October 2022, the market has rallied almost 200%.

Quite the ride for the hardware stocks in SMH.

And when there is a frenzy, realized volatility rarely stays quiet. We haven’t seen peaks like those observed in 2022, at around 50% annualized volatility, yet realized volatility spikes every quarter like clockwork as semiconductor companies release their earnings.

Considering that NVDA represents 20% of the holdings in an already quite concentrated ETF, it’s needless to say that when stock volatility rises, it impacts SMH.

It is well-documented that implied volatility rises before market events, particularly before earnings (refer to the excellent "Positional Option Trading" by Euan Sinclair for more details). We’ve seen traders in our Discord channel implement this trade for a while with some success.

It is also well-documented that implied volatility goes down right after the event. Therefore, considering NVDA's influence on SMH, let’s look into the options data to detect the part of the expiration that may be best suited to benefit from that expected volatility crush.

The data and the trade methodology.

First, let’s examine the volatility term structure in SMH using options data and recreate a VIX-like index at 9 days, 30 days, 3 months, 6 months, and a year.

The backwardation is pronounced as one would expect before NVDA earnings.

The backwardation between 9 and 30 days is quite pronounced, characteristic of anticipated market events. Once the earnings are behind us, the front-month volatility will decrease, and the term structure will be flatter, like the one observed in March 2024.