Do Valuation Shorts Work?
Reasons to short a stock fall broadly into two “buckets”: fraud and valuation. Fraud shorts are stocks where the (short) investor believes the company is misleading the public about its business. This can range from mild and small-scale, such as aggressive but GAAP-defensible accounting, all the way up to massive and brazen fraud, such as fictitious revenue and profits. Let’s Gowex is a recent and spectacular example of the latter. Valuation shorts, on the other hand, are stocks that the investor thinks are simply too expensive based on the underlying value of the business. For example, I read a pitch about a year and a half ago that pointed out that Tesla (TSLA) was trading at over four times what Geely (a Chinese automaker) paid to acquire Volvo, and there was no way Tesla could be worth four Volvos . . . I remember thinking the pitch sounded interesting. It was certainly well-researched. I’m still not sure Tesla’s current market cap is justified (apparently, neither is the CEO, Elon Musk), but I’m so glad I passed on this short 18 months ago: TSLA is up more than 500% since the pitch.
Valuation shorts have a bad reputation on Wall Street. You may be right in the long run, but you may not be able to hold the position long enough to get there. As David Einhorn puts it, “We have repeatedly noted that it is dangerous to short stocks that have disconnected from traditional valuation methods. After all, twice a silly price is not twice as silly; it’s still just silly.”
Valuation shorts are a dicey proposition on intellectual grounds, too. John Hempton, who is the chief investment officer at Bronte Capital and publishes one of my favorite investment blogs, puts it this way: “In a valuation short we are working on the same information as everyone else has. This makes me uncomfortable. There is an arrogance in suggesting we can analyze the information better than anyone else. We find it harder to answer the question of what we see when others don’t and hence harder to justify the position at all.”
I was curious whether valuation shorts work as a whole, and have recently had occasion to test this question using a new research service called Activist Shorts Research. They have compiled data on more than 400 campaigns by noted short-sellers from 2002 to the present. The returns look like this:
The mean price change (not including dividends), indicated by the blue line, was −14.2% over an entire “campaign,” which can be arbitrarily long. Additionally, 65% of campaigns were “successful” in the sense that the price of the target stock dropped since the campaign was announced. In 4% of campaigns, the price dropped 99% or more. These figures sound quite good, but it is important to note the sample is biased because the service does not cover all short-sellers, only the “best” and “most public” ones. These two groups likely overlap but not completely.
Despite the biases of the overall sample, the question I was most interested in was whether valuation shorts work better than fraud shorts. For each campaign in the dataset, we have a “Primary Allegation” which is the reason the short-seller used to publicly justify the short call. The reasons provided are many and varied, but I have grouped them into the two buckets we are interested in. The results are stark:
| Primary Allegation | Fraud | Valuation | Total |
| Mean Return | −30% | 3% | −14% |
| Campaigns | 229 | 219 | 448 |
Short campaigns that allege a stock is overvalued are wrong as a group: the target stock rises 3% over the life of the campaign, on average. Shorts that allege fraud are much more effective: the target stock drops 30% on average. We can see this result in some more detail by comparing return distributions:
Fraud short campaigns are much more likely to identify stocks whose prices ultimately drop to zero. This skews the entire return distribution to the left, as compared with that of valuation shorts.
In conclusion, it seems the poor reputation valuation shorts have is at least partially deserved. Even when we look at some of the best-known and most-memorable short campaigns of the past decade, valuation shorts significantly under-perform fraud shorts.
If you liked this post, consider subscribing to the Enterprising Investor.
Please note that the content of this site should not be construed as investment advice, nor do the opinions expressed necessarily reflect the views of CFA Institute.
Photo credit: ©iStockphoto.com/Jeff_Hu
Alon,
Thanks for this great write-up! My old boss (who was previously a tech fund manager in the late 90’s) always used to talk about this, but it’s great to finally see some hard numbers around it. He used to talk about how hysteria in asset classes could last way longer than many think and that alone makes it difficult to short on valuation concerns.
Best,
Mike
There is another thing in my opinion. Who would like to invest in the company that cheats? Fraud is some kind of cheating even if it is compliant with some laws and regulations. So you take this personally and You have this negative feeling for it and when most investors do than there is a quick run away from the company’s stock.
With companies that underperform financially is a different story. The company might come up with something that will rocket high they price or some VC or PE investor might come and see some potential in it ,or with so much money chasing after investments due to QE it will be easy for them to restructure their finances or company.
Alon,
Thanks for the post. I found this line of interest “Valuation shorts are a dicey proposition on intellectual grounds”. I disagree–I think this formulation is more accurate “valuation shorts are a dicey proposition as a money-making strategy”.
Intellectually, the bet may be defensible, but, a winning bet isn’t based on merit as much as sensing what others will do. Overpaying for a house in a rising market can pay off while one is in a valuation bubble. The money-making bet is, when will others share the same sense of where things are.
What is considered “normal” plays a huge role in market perceptions, of course; I hope to foster a new norm of thinking about valuations–both intellectually and as a money-making strategy.
I’m working on a book to promote the idea of a performance-based capital structure for companies that raise venture capital via a public offering. I call it the “Fairshare Model” because it balances and aligns the interests of investors and employees. A slide deck with the high concept is under the Resources tab at http://www.fairsharemodel.com
I’ll share a draft of my first chapter with anyone with interest. My email is [email protected]
Karl
Alon,
Great article! I particularly like the Einhorn comment, “twice a silly price is not twice as silly, it’s just silly.” The market get it wrong for far longer than you can remain solvent.
I am not sure if the sample size is large enough to dissect this any further but I’ve been interested in two related items.
1. If you overlay an element of technical analysis to identify when the market is starting question the valuation, do the results change? (i.e., if you only short once there is a cross below the 50 day MA, etc.)
2. If you go long a corresponding position in either an industry index or market index and net the gain/loss, what do the results look like?
Regards,
Neal
“In a valuation short we are working on the same information as everyone else has. This makes me uncomfortable. ”
Can’t we say the same thing about a valuation long? By extension this is saying that one should not expect to beat the market unless he has better information than participants in general.
nycnitte,
I presume a “valuation long” is buying the stock and holding it. That’s an assertive expression of optimism.
A short is an assertive act of pessimism. It appeals to a sense of fear, not hope. In politics, such acts–in the form of negative advertising–work very well. As a species, we are wired to respond to fear quickly; we show more variation in how (and how quickly) we respond to hope. If fear works in politics, why doesn’t work in markets (i.e., valuation shorts)?
I wonder if its that many participants are valuation unaware. Also, accepting that fear has costs to those holding a stock–they may not sell–so, a valuation short may not work out.
Karl
If one considers shorting as a stand-alone strategy, then the expectation of a high positive return over the long-term should probably be fairly muted.
If however one considers short positions as an additional source of capital to be invested on the long side (e.g. 130/30 portfolio), then the potential return drag from these shorts should be compared to the cost of other capital, bearing in mind that the short portfolio does a better job of asset/liability matching (assuming an all equity portfolio) than other sources of leverage.
Interesting. I partly agree. Valuation by itself is never a good reason to go short. But an excessive valuation is a tailwind that helps the shorts. Valuation plus a declining business or a business showing early signs of trouble is indeed a good starting point for short ideas. Also, I’ll point out that there are more types of shorts than frauds and expensive stocks. Secular decliners, flawed business models, and growth stories starting to break also factor into the picture. In the latter case, an excessive valuation is like downside leverage when the growth story starts to break. Otherwise an interesting article. Thanks for posting.
Excellent article!