Is the era of quantitative hedge fund outperforming Shanghai index AI alternative fund manager coming?

category:Finance
 Is the era of quantitative hedge fund outperforming Shanghai index AI alternative fund manager coming?


A-share quantitative hedge fund outperformed Shanghai index by more than 10%

Influenced by the plummeting sentiment in the European and American markets, the A-share market has also experienced a significant correction since the end of the festival. As of March 23, the Shanghai index has risen by - 10.63% since the Spring Festival, while the Shenzhen Composite Index and gem index have risen by - 9.27% and 5.22% respectively.

Back to the public fund market, wind statistics show that 28 quantitative hedge funds (A and C shares are calculated separately) with performance statistics in the whole market have an average increase of - 0.09% since the Spring Festival (as of March 23), and an average return of 1.49% in the year.

Its not hard to see that 28 public quantitative hedge funds have shouldered the A-share correction - their average performance has outperformed the Shanghai index by more than 10%.

Specifically, when there are severe fluctuations in many global markets (the performance since February 3 is counted here), 15 of them have borne the fluctuation, with an increase of 0-3.62%; while the performance of 13 public quantitative hedge funds has been withdrawn, but the maximum withdrawal of 5.04% is also lower than that of A-share main index.

The reporter of daily economic news noted that several quantitative hedge funds with more performance withdrawals mainly track the US three-year treasury bonds, while several funds with relatively stable performance mainly track the one-year term bank deposits announced by the peoples Bank of China.

According to the 2019 annual report, the largest heavy position stocks of 19 funds are Ping An of China; secondly, Guizhou Moutai, China Merchants Bank and Industrial Bank are also the preferred heavy position stocks of these funds.

Background review:

December 27, 2019 is the last Friday of 2019, and China Securities Regulatory Commission has approved quantitative theme funds of several fund companies. The approved fund companies are: Debang, Huaxia, Fuguo, Haifutong, Jingshun Great Wall

Since the approval of Huatai Bairui quantitative hedge fund on January 21, 2016, the approval of public offering quantitative hedge fund has been suspended for many years.

Quantitative trading contributes 25% of market volume

In fact, in the A-share market, the most impressive thing for investors is that in June 2015, the bull market on leverage suddenly plummeted, and thousands of shares fell for a while. During the stock market crash, a Russian used the quantitative high-frequency system to pick up money for nothing in the A-share market. Although the relevant personnel were subject to regulatory penalties afterwards, at that time, it was sad that the domestic investment institutions had no power to fight back!

Chinas capital market is opening wider and wider. Foreign capital is coming in, but its not only capital, but also advanced financial instruments and trading methods.

Chinas quantitative investment started after the introduction of stock index futures in 2010. In 2015, when the stock market was in turmoil, many active funds adopted Liquidation Strategies, which had a great impact on the market. However, in the period of stock market turbulence, the performance of quantitative investment is very stable, which has been affirmed by the market, and the steady return of quantitative hedging has attracted the attention of investors, and a large number of funds have poured into this field.

In 2018, it was difficult for new fund issuance to raise funds, but quantitative fund issuance bucked the trend. According to the data, the number of newly established quantitative funds reached 66 in 2018, exceeding the number of established funds in 2017.

According to China Merchants Securities Research Report, by the end of 2019, there were 311 public offering quantitative products, including 191 active quantitative products, 102 index enhanced products and 18 hedging products. By the end of 2019, the total scale of public quantitative products reached 160.507 billion yuan, including 68.237 billion yuan of active quantitative products, 84.898 billion yuan of index enhanced products and 7.372 billion yuan of hedging products.

By the end of 2019, the number of public quantitative fund products

Total scale of public quantitative fund products by the end of 2019

According to the data of private placement platoon.com, by the end of February 2020, there are 6 quantitative strategies of 10 billion level private placement in more than 30 securities, namely, technetium in 2011, Jiukun investment in 2012, Ruitian investment in 2013, Lingjun investment and Mingyi investment in 2014 and magic square quantification in 2015.

According to incomplete statistics, as of December 2019, the overall return of 1369 quantitative strategy funds with performance records since 2019 is 15.89%, second only to the yield of stock strategy 22.29%. Among them, there are 1186 positive return funds, accounting for 86.63%; 11 products have doubled their income, with the highest income of 373.61%. In addition, there are 396 fund returns between 20% and 100%. The performance of ten billion level quantitative private placement is also good. In terms of the overall yield, in addition to the yield of individual institutions in 2019 under 15%, the other five ten billion level quantitative private placements all have a yield of more than 20%, of which the yield of magic square quantification and Minghe investment is more than 30%.

After two years of development, 25% of daily trading volume in the Chinese market is made by quantitative trading. Zhou Lefeng, assistant to the president of Xiangcai securities and general manager of the brokerage branch, shared this data in his speech on Ai quantification.

In the semi annual report published on August 23, 2019, CITIC Securities disclosed its self investment information, which has a key description of quantitative investment: in the first half of 2019, the companys alternative investment business actively responded to market changes, based on macro analysis and judgment, with quantitative trading as the core, flexibly used various financial instruments and derivatives for risk management, and developed multi market diversification Our investment strategy effectively disperses the investment risk and enriches the source of income. The latest technology of artificial intelligence / machine learning is widely used in strategy development, and the results are achieved. At present, the businesses or strategies that have been carried out include: stock index arbitrage, stock long short, macro hedging, block trading, statistical arbitrage, fundamental quantification, convertible bond arbitrage, convertible bond strategy, commodity strategy, option strategy, etc.

CITIC Securities has always been a myth in stock index futures trading, and investors are very concerned about the layout of its multiple short positions. And the mystery is finally solved. The trading partner of the investor is actually a computer!

CITIC Securities obviously has high hopes for quantitative trading. In the semi annual report, there is another sentence:

In terms of alternative investment business, the company will increase investment in artificial intelligence / machine learning, further research and develop new strategies, build a more efficient trading system, grasp various investment opportunities in the market, and steadily improve the investment return rate.

Chinas top securities companies attach great importance to quantitative trading and artificial intelligence, which is evidence of the rapid development of quantitative trading in the domestic market, and also subverts many peoples understanding of investment. The combination of artificial intelligence and quantitative trading is the next development direction of quantitative investment - AI quantification.

Mathematicians + it elites cross border attack

On December 23, 2019 in Sichuan, China, model China u00b7 2019 top ten new economic leaders in Sichuan award site, 10 presidents and CEOs wearing red scarves stepped onto the podium, more than half of them didnt wear suits, and there were many people wearing jeans and sneakers to receive awards. Maybe this is the model that the new economic leaders deliberately stand out -- not stick to one style, different and subversive Unification...

One of the winners, Liang Ju, CEO of a technology company in Chengdu, wrote five words in the wechat circle of friends at 19:29 that day: growth in winter.

Just three days ago, the reporter of daily economic news just heard a promotion meeting, heard a word Ai quantification for the first time, and also got a message - most of the people who are engaged in AI in-depth learning and image recognition now come from a group of people who are engaged in Bing Search engine in Microsoft Asia Research Institute, and Liang Ju is one of them.

AI quantification is a quantitative investment transaction using artificial intelligence, which is an upgraded version of quantitative transaction. AI automatically searches for rules without defining rules.

In 2019, we will achieve balance of payments, with revenue increasing by 300% compared with 2018, Liang Ju said in an exclusive interview with reporters from the daily economic news According to the year-on-year data of revenue growth, the AI quantitative trading market is growing rapidly in 2019.

In the context of quantitative hedge fund approval, the cross-border model of mathematician + IT elite is coming. This is a group of people who use cloud server to trade. In their eyes, the investment problem is a mathematical engineering problem. Looking for reliable factors, correlating the cause (x) with the result (y), establishing mathematical model to predict the price, and quantifying the trading work can be vividly called the password to crack the fluctuation of capital market.

Can Ai + quantification surpass human beings in investment?

How to let the computer learn to choose a good watermelon from a bunch of watermelons?

Another afternoon in the sun, daily economic news reporter wanted to understand how AI chooses stocks. Mr. Li, a senior strategy engineer of science and technology, used watermelon as an example.

Engineer Li, who speaks very fast and has a low voice, has a typical it masculine temperament. He is obviously different from the people in the financial circle who have been contacted by reporters. This is a man of science and engineering, a programmer, and a person who works in the company on Saturday

However, the daily economic news reporters question is more simple and direct: give me some stocks that will rise in the next few days. When the program runs, a few minutes later, the reporter saw the names, codes and trading directions of several stocks.

Just these? Dont need to continue to analyze direct brainless buying? Is it reliable? The reporter asked in silence, embarrassed to speak out of politeness, only nodded calmly.

Its the first time Ive seen auto screened stocks. I dont need to intervene and think at all. Its amazing! The reporter is now an investment Xiaobai.

Reporters saw a long series of data, the cumulative yield of this strategy is 157.30%, starting on August 28, 2018. In the past week, the yield was - 1.65%, 2.36%, 19.61%, 121.57%, 63.40%, 15.23% and 1.69% respectively

This is the first time that journalists have come into contact with the real quantitative strategy. For a long time, journalists mistakenly believe that quantitative trading is another way of saying procedural trading, but this concept was corrected by Liang Ju at the beginning of the interview, this can only be regarded as an expert system.

Quantitative investment includes: quantitative stock selection, quantitative timing Whats the difference with value investment, technology analysis and index investment? You can see it at a glance from the diagram below.

Is it complicated to quantify with AI? Do you remember the selection of watermelon just now? This is just one of the algorithms. Li said.

There are only seven steps to create a quantitative model:

Step 1: set the data range of training set and test set

Step 2: set goals

Step 3: find factors

Step 5: model training + stock forecast

Step 6: back testing

Step 7: view and analyze the results

Quantitative investment history: the worlds first AI fund and God

Computers know cats, but can they recognize a valuable stock? In the United States, a man named chidananda khatua asked such a question.

A few years ago, while attending a lecture on hedge funds at business school, chidananda katua was suddenly inspired. If we can combine the messy information and accurate financial data in the annual report and news articles, it will produce powerful results.

In 2017, aipowered equity ETF (hereinafter referred to as aieq), the worlds first fund known as the use of artificial intelligence stock selection, appeared in the United States. Chidananda katua founded a company called equbot, which is the fund manager of aieq.

Equbot, based in San Francisco, has few employees and 17 programmers and statisticians in Bangalore, India. The system inputs 1.3 million texts a day, including news, blog posts, social media posts and SEC documents, Forbes said. IBM Watson system digests and absorbs language, selects knowledge points, and inputs a knowledge map of 1 million nodes. Each node to be linked may represent a company (15000 in total), a keyword (e.g. FDA), an economic factor (e.g. oil price). There are countless possible arrows to connect them. The computer uses the neural network which simulates the connection of brain neurons to make trial and error, so as to give weight to the important arrow. Then, the system continues to explore which fluctuations in the input data will affect the stock price one week, one month or one year later. In a busy day, the calculation of equbot is 500 trillion times.

The first years performance of aieq is remarkable. On October 18, 2018, the yield of aieq was 11.81%, outperforming the S & P 500 and Russell 200 indexes. In 2018, according to Bloomberg data analysis, aieq beat more than 87% of actively managed fund managers.

However, on Christmas Eve 2018, the price plummeted into a large rollover site. Aieq bought netsapp and newrelic, which may be that aieq read news a lot, and responded to peoples enthusiasm for cloud computing with buying behavior. The two stocks fell sharply, which greatly affected the performance of aieq.

But katua, the founder, says it doesnt matter. Neural networks learn from mistakes. In 2019, AI seems to have found a feeling, and aieq is surging. However, comparing it with the S & P index, reporters from the daily economic news can see that S & P is clearly going to be stronger. According to Forbes, at present, the investment managed by equbot is only $120 million. It is too early to draw a conclusion whether it will succeed or not.

According to public information, in March 2019, Goldman Sachs launched five new ETFs, relying entirely on machine trading and AI algorithm to get rid of human initiative control. Art and investment are the most intelligent areas of human beings. Once AI makes a breakthrough in the field of investment algorithms, it may mean that the job of fund managers may be ruined.

What stocks do quantitative investors buy? When can I buy it? Its all a mathematical engineering problem. The representative of quantitative investment is James Simmons, who is believed to be the only one who can beat Buffett.

James Simons

James Simmons is a mathematician and hedge fund manager. Before his investment, Simmons worked for the military before he was 29. His job was to decipher passwords. Because of his big mouth, he was fired later, and then there was another great investor in the world.

As a mathematician, his research results are named Chens Simmons theorem, which is a result of joint research with Chen Shengshen, a famous mathematician in China.

As a fund manager, from 1989 to 2009, the average annual return of the medal fund managed by James Simmons was as high as 35%, more than 20 percentage points higher than that of the S & P 500 index in the same period, and more than 10 percentage points higher than that of Soros, the financial giant and Buffett, the God of stocks. Simmons has never been absent from alpha financial magazines global fund manager rankings for 15 years, and has been at the top of the list, making $23.5 billion during that period.

In 1978, Simmons set up a fund called limroy.

According to the legend, when the limroy fund was just established, he was also a subjective trader, studying the fundamentals. However, he soon found out that this method didnt work, and then he used the most proficient Mathematical Engineering to solve the investment problem. This time, he was very successful.

In March 1988, the grand medal fund was officially established to replace the limroy fund. Simmons completely stopped fundamental analysis and became a thorough quantitative investor relying on the model. As of December 1999, the total net return was 2478.6%, ranking the first in the same period, more than twice that of Soros quantum fund, while the S & P index rose only 9.6% in the same period.

Reporters note: who are your counterparties?

A routine interview inadvertently opened the door of quantitative investment, and found that there are such a group of people in the capital market, who study the factors affecting the market, design the trading strategy, place an order through the cloud server, quietly. Today, the volume generated by the execution of procedures accounts for 80% in the US and a quarter in the A shares.

Cross border, disruptive innovation is taking place in the capital market. An alternative investment corps of mathematicians, it elites and programmers is coming across the border. Barbarian at the gate with the new AI weapon is trying to crack the password of market fluctuations. This is a dimension reduction blow. For fund managers, the alarm has been sounded. Before quantification, investment was an art, an inspiration, a perception, an understanding of human nature, or a qualitative category. After the emergence of quantification, investment is technology, by factor, by strategy, by artificial intelligence, and began to evolve into mathematical engineering problems. 3777 A-shares, all analyzed once, how long does it take? Can we establish a copper price forecasting model by analyzing the data of LME copper inventory, smelting and processing cost, operating rate and spot price for 30 years? This is the short board of human intelligence, which is the strength of artificial intelligence. Quantitative investment + AI is subverting the rules of the game. How to survive? Source: editor in charge of daily economic news: Yang bin_nf4368

Cross border, disruptive innovation is taking place in the capital market. An alternative investment corps of mathematicians, it elites and programmers is coming across the border. Barbarian at the gate with the new AI weapon is trying to crack the password of market fluctuations. This is a dimension reduction blow. For fund managers, the alarm has been sounded.

Before quantification, investment was an art, an inspiration, a perception, an understanding of human nature, or a qualitative category.

After the emergence of quantification, investment is technology, by factor, by strategy, by artificial intelligence, and began to evolve into mathematical engineering problems.

Can we establish a copper price forecasting model by analyzing the data of LME copper inventory, smelting and processing cost, operating rate and spot price for 30 years?

This is the short board of human intelligence, which is the strength of artificial intelligence. Quantitative investment + AI is subverting the rules of the game. How to survive?