NVIDIA, Facebook And The EMH

The Efficient Market Hypothesis is a mainstay of academic thinking about financial markets. It is rejected by many traders and money managers. Warren Buffett, for example, famously said that he would be on a corner, selling pencils from a tin cup if markets were efficient.

I do not expect to settle this decades-long debate in a single blog post. Instead, I will share how my own personal thinking changed along with my career – from college professor, to financial analyst, and to investment manager. I hope to stimulate and to provoke; we can all benefit from some wise comments. I will also suggest a few ideas we might consider to exploit inefficiency.

Provocative Examples

It is often useful to have a specific example in mind. Let’s start with Facebook (FB), a company familiar to all. Here is a chart of recent stock action:

Let’s consider four time periods.

1. Before the Cambridge Analytics announcement. The stock was comfortably trading at the 185 level. Was the market efficient and rational? Was all information reflected? Try doing a custom Google search ending just before the news was announced. I used “Facebook use of personal data” ending on 3/15/18. The Cambridge story was different in magnitude, but not a surprise to anyone carefully following the stock.
2. Right after the announcement, with the stock at 173. Was the market efficient and rational? That was the market-clearing price for a day.
3. What about the Zuckerburg silence effect, taking the stock to 165 for a day or two. Was the market efficient and rational?
4. What about the current levels, with the (highly predictable) news that every attorney general trying to make a name – at any level of domestic or foreign government – is suing the company and demanding testimony. Is the market now efficient?

My question, and it is a question, is whether the market was efficient at both the start of this process and at the end. This seems inconsistent.

Let’s try Nvidia (NVDA).

1. The stock was trading around 250. Did that reflect an efficient market?
2. News broke about the tragic death of a pedestrian in Arizona, struck by a self-driving car. The stock dropped to the 240 range. Was that a reflection of market efficiency? Please keep in mind that the self-driving tests generally show superiority over human drivers. Are we really surprised that millions of miles of testing resulted in a death? This seems like a predictable event.
3. The stock rallied back to 250. Was this a rational market response?
4. The stock dropped to 235 and rebounded, once again, to 250. Does this reflect an efficient market?
5. Today’s trading took the stock to the 220 range. The company announced “a suspension” of self-driving tests. Viewers of CNBC got the analysis from two sources. First, the Nvidia story sparked selling in all artificial intelligence stocks. Then Art Cashin explained that the selling started with tech stocks and spread to the S&P 500 and then the DJIA. Was any of this efficient?

My Take

I try to separate pre-news trading from what happened later. I call it “local market efficiency.” It means that the dramatically different original price in the stocks might make sense, varying according to daily news – until the big story hits. It is a bit like The Big Short. Even though the heroes were accurate in their assessment, payday was years away. Disgruntled investors did not understand. The thesis was taking forever to play out, and the logic was difficult to explain. (In my reviews I try to make points not seen elsewhere, and I thought I really nailed it with this one).

Today’s trading unreasonably crushed the AI and supporting chip stocks. This is the kind of knee-jerk reaction that can be exploited both by traders and by long-term investors building a position.

I am also considering social media stocks that do not share the Facebook model of selling customer data.

I especially like artificial intelligence stocks. There will be some setbacks, but this is a dominant market direction. Here are some good ideas:

• NVDA itself;
• Some general AI ideas (Zacks).
• 10 Best, large and small (U.S. News)
• AI without FANG.

Conclusion

I hope that my thoughts about efficient markets were provocative, and I invite comments. I hope that others see the inconsistency involved in claiming that these markets were efficient both before and after news – drawn out and not completely unexpected.

Those of us who believe in exploiting inefficiencies got some great opportunities today!