We use "big data" to short_Data Analyst Exam

Not to be a headline-grabber, but it's true, but it's not us shorting, it's us helping our clients short, and that was back when I was a full-time attorney at a law firm.

If you know anything about antitrust law, you know that if a concentration (e.g., a merger or acquisition) occurs between two or more companies, and the companies involved in the concentration reach a certain size (e.g., turnover reaches a certain standard), the concentration must be declared (e.g., to the Ministry of Commerce in China), and the concentration can be carried out only after the declaration is approved. The purpose of an antitrust notification is to prevent a concentration from disrupting competition in a relevant market to the detriment of consumers. For example, if Coca-Cola and PepsiCo were to merge into a single enterprise, the concentration would most likely result in the distortion and disruption of the competitive order in the cola market - competition in the cola market would be extinguished by the demise of the two major competitors, and the price of cola would most likely soar, to the detriment of consumers.

If Coca-Cola and Pepsi were to merge in a centralized way, that centralization would have to be declared to the government authorities of the relevant markets, and in the meantime there would be a lot of brokers, hedge funds, or other people deciding whether or not to go long or short the two companies' stocks. If the concentration is more likely to be approved, then the fundamentals of going long the stocks of both companies are high - while the concentration may hurt consumers, it is good for both companies, and the stock prices of both companies will go up, so there is a big win for going long. Conversely, if there is a higher likelihood that this concentration will be rejected, then the fundamentals of shorting the stocks of these two companies are large - because once the concentration filing is rejected, then the stocks of the companies participating in the concentration will fall, and thus shorting wins. Of course my example with these two cola giants is probably too typical to be relevant, as their concentration rejection is almost a foregone conclusion. So let's use a real-world example. But this example still has to do with Coca-Cola.

On Sept. 3, 2008, Coca-Cola announced plans to buy China Huiyuan Juice Group Ltd (01886.HK) for cash. Coca-Cola proposed a takeover bid of HK$12.20 per share and the equivalent of the convertible bonds and options issued. Prior to the announcement, Coca-Cola had obtained irrevocable undertakings to accept the offer signed by three shareholders of Huiyuan, who *** own 66% of Huiyuan's shares. The consideration paid by Coca-Cola for the acceptance of this proposed transaction is approximately US$2.4 billion. The deal, if completed, would have been Coca-Cola's largest acquisition in China up to that point, and Huiyuan Juice would have been de-listed.

Shares of both Huiyuan and Coca-Cola rose sharply after the announcement. But the problem was that Coca-Cola's purchase of Huiyuan was a concentration that should have been declared under China's antitrust law, and whether the concentration could be approved by the Ministry of Commerce was the X-factor in the deal, which a hedge fund approached us to analyze, and we collected data to do so according to our own formula and methodology for this kind of business (we won't get into the details of what kind of data and what kind of analytics we used). ). In any event, our final analysis was that the Antitrust Bureau of the Ministry of Commerce would not approve the concentration, and fortunately our analysis was correct. Accordingly, our clients who followed our advice and went short made money.

When we did the above case study seven years ago, there was no such thing as "big data" or "small data". In retrospect, what we did (and still do) is nothing more than data analytics. Of course, the data involved may not be that big in terms of total volume, but it's big enough for a specific project. Of course, it may be debatable whether these data can necessarily be viewed as what we now call "big data," and we will discuss this in a separate article later, which is why I put "big data" in quotes in the title of this article. In any case, considering that the Commerce Department has so far failed to file only 2 out of more than 1,000 antitrust filings, we are proud of our ability to accurately predict such a small probability of an event, which should be attributed to the accuracy of the data we collected and the accuracy of our analysis.

If we can regard the above successful short-selling as an effective analysis using "big data", then "big data" analysis seems to have the following characteristics, and we are trying to summarize the so-called characteristics here to To achieve the purpose of a brick to attract jade:

- Big Data analytics should first of all be a commodity. Regardless of the method of data collection and analysis, the final product should be purchased by someone with money. Big data or big data analytics products that have no commercial value are worthless, in other words, they can't be done.

-Big data analytics product development should be targeted to customers. Different customers have different needs for big data analytics products. Take the legal industry as an example, the above big data, big data and big data analytics products have a direct need for basically do foreign business law firms and international companies, so the above big data and big data analytics products are basically the working language of English.

- The vitality of big data analytics lies in its accuracy. Taking our above case as an example, Coca-Cola's acquisition of Huiyuan was rejected, and Huiyuan's stock price plummeted 42% throughout the day on the day immediately following the opening of the market. And before that the news of Coca-Cola's sky-high acquisition of Huiyuan had spurred Huiyuan's stock price to soar nearly 200 times. After Coca-Cola announced the acquisition of Huiyuan Juice, its share price on the New York Stock Exchange had a strong rise, but in the subsequent six months the share price fell by 20%, which is not unrelated to the failure of its acquisition of Huiyuan. It is conceivable that if our analysis had been inaccurate at the time, the client would have had to lose money. Of course, the success of our case cannot be said to have a certain degree of chance, so is there a certain degree of fault tolerance in big data analysis? I believe there is. If big data does not make mistakes, then its equivalent to God, but the error rate of big data is too high, then there is no commercial value, and even the entertainment value is gone.

At the end of the article, I asked a question: Is it malicious to use data (whether big or small) to analyze the conclusions of shorting? Maybe this question is a bit "irrelevant".

The above is what I have shared with you about our use of "big data" to do shorting related content, more information can be concerned about the Global Ivy to share more dry goods