In recent years, the scale of domestic quantitative private equity has grown explosively, and a number of outstanding quantitative private equity managers with outstanding performance and stable excess returns have emerged, and Century Frontier is one of them.
The two founders of Century Frontier, Wu Di and Chen Jiaxin, met while studying at Chinese University of Hong Kong, and in 2013, the two began to develop and trade cross-border arbitrage strategies for commodities in Hong Kong.
In 2015, A-shares just experienced a bull market, the stock index market was active, is a fertile ground for quantitative high-frequency strategies, Chen Jiaxin and Wu Di were keenly aware of the opportunities contained in the A-market, so they founded Century Frontier and began to do high-frequency trading of stock indexes.
In 2016, with the restriction of the stock index, Century Frontier launched a high-frequency strategy of commodities. In 2017, Century Frontier migrated its high-frequency strategy and experience to the company and began to engage in programmatic T0 trading.
In 2019, Century Frontier shifted from self-operated to asset management, launched a neutral and index-increase strategy, and in the following two years, it ranked among the tens of billions of private equity with its excellent and stable performance. In 2021, the scale of Century Frontier management will exceed 10 billion, and in 2023, the scale will exceed 20 billion.
Century Frontier has a deep accumulation in the field of high-frequency trading, can achieve full-band coverage, outstanding trading execution advantages, and is one of the few companies in the industry that has reached the forefront of the market in long-term alpha and high-frequency execution.
The team currently has a total of 77 people, and unlike the private equity that adopts the multi-PM model, Century Frontier adopts a close collaboration model, information sharing and infrastructure sharing, a high level of mutual trust between team members, and high stability, even if there is no relevant equity incentive, the core personnel still maintain a zero turnover rate.
How exactly does Century Frontier maintain its competitive edge?How to mine the effective factor?What is your experience in risk control?What do you think about next year's market?Recently, Wall Street News conducted an interview with Century Frontier.
Wall Street News: Can you tell us a little bit about the company and the team?
Century Frontier:Founded in August 2015, Hainan Century Frontier Private Equity Management focuses on quantitative investment research and asset management services in the domestic market. At present, the company has a team of 77 people, with Wu Di and Chen Jiaxin as the founders, and nearly 70% of the investment research and development personnel, mainly from Tsinghua University, Peking University, Fudan University, University of Science and Technology of China, CUHK and other universities majoring in physics, mathematics, quantitative finance, computer science and other universities, and the core strategic members have more than 10 years of experience in the quantitative industry. The team focused on proprietary trading (** high-frequency and alpha strategies) before 2019, and transformed into asset management at the end of 2019, with a current management scale of more than 20 billion.
Founder of Century Frontier and CEO of Core Investment Manager: Wu Di.
Wall Street News: The turnover rate of Century Frontier's core employees is zero, how is this done?How to ensure a reasonable distribution of benefits within the team?
Century Frontier:The team is not in PM mode, adopts a close collaboration model, information sharing and infrastructure sharing, a high level of mutual trust between team members, high stability, and even if there is no relevant equity incentive, the core personnel still maintain a zero turnover rate. We not only pay attention to salary competitiveness, but also pay attention to the fairness of assessment, and maximize the objective assessment and distribution, so that colleagues can feel fairness and flattening the company culture, and emphasize equality, respect and unity among employees, so that outstanding talents feel respected and have a sense of belonging here.
Founder of Century Frontier and CIO of Core Investment Manager: Chen Jiaxin.
Wall Street News: What is the product layout and investment logic of Century Frontier?Excess returns are mainly attributable to **
Century Frontier:There are mainly ** strategies (300 500 1000 index increase, market neutral, 300 500 1000 neutral, market-wide stock selection strategy), CTA strategy, mixed strategy, etc.
The cutting-edge strategy covers all frequency bands, of which the high frequency contributes 10-15%, and the medium and low frequencies each contribute about 40-45%.
Our quantification is to use quantitative methods to find opportunities for ineffective pricing in the market, over-allocation is undervalued, under-allocation or over-valued, and obtain excess returns. Specifically, by obtaining and analyzing relevant data, such as stock prices, transaction-by-transaction transactions, fundamental data, news and announcements, industry analysts' opinions, etc., to find out the law of ineffective pricing in the market, establish quantitative models, and realize the strategy of automatic operation, so as to capture the opportunities of ineffective pricing in the market.
These ineffective pricing opportunities may stem from the psychological factors of the participants, such as greed and fearIt may be due to the continued imbalance between supply and demand caused by the trading demand of large funds;It may be due to the process of updating perceptions and pricing caused by information diffusion. The reasons behind these laws are structural and don't change easily. Finding these laws supported by structural factors can ensure the continued effectiveness of research results and generate continuous excess returns.
Wall Street News: In 2021, Century Frontier completed the expansion and upgrade from high-frequency to mid-low frequency, how was this iteration completed?
Century Frontier:The medium and low frequency should be the main track for quantitative growth in A-shares in the future. This is because the mid-low frequency has the characteristics of large capacity, strong diversity, and relatively wide income. In particular, with the increase in alternative data, quantifying the overall scale**, the strength, diversity, and stability of signals in the mid- and low-band bands may more fundamentally affect large-scale asset management strategies. In 2021, we increased the proportion of mid-frequency strategies (intraday), which can switch faster when styles are quickly switched** in order to better respond to the rapidly changing market environment, which was completed at the end of January 2022. Follow-up to continue to iterate the low-frequency strategy: in terms of volume and price, fundamentals, alternative factors and other aspects, we will dig out some factors with considerable excess capacity, expand the underlying factor pool, and continue to adhere to the concept of long-term, diversified, and large-capacity.
Wall Street News: What kind of market environment is the products of Century Frontier more suitable for?Have there been any significant drawdowns in past investments?What is the reason?How to fix it?
Century Frontier:Cutting-edge products are more suitable for environments with active trading volume and active markets. The retracement in September-October 2021 was mainly due to the volatility of the index itself. There are two main phases of the drawdown of the excess part. The first phase is from the last week of September 2021 to the week after the National Day. During this period, the market characteristics are mainly manifested in the fact that after the statement of the coal stock, the valuation logic of cyclical stocks has changed, and cyclical stocks have entered a rapid unilateral retracement, and the market style has shown obvious changes. This change in valuation logic triggered by policy changes has adversely affected quantitative strategies that rely on historical data to find patterns.
The second phase of drawdown mainly occurred at the end of October and the first week of November 2021. During this time period, the volatility cycle of alpha accelerated, and there were some quantitative redemptions and liquidations in the market. These actions lead to a stampede on the alpha side, and the excess drawdown mainly depends on the correlation between the strategy and the closed product. After the end of the closing flow, the market gradually returned to its normal state, and the alpha returned to its usual level.
After this drawdown, we will be placing greater emphasis on extreme market risk at the portfolio level. Specifically, we will apply stricter constraints to each style and continue to delve into the risk** model to improve the performance of our risk management mechanisms. In addition, we have upgraded our trading systems and strategies to increase the proportion of mid-frequency strategies. Compared with the low-frequency strategy, the mid-frequency strategy can adjust more quickly when the style is quickly switched**, thus reducing the drawdown of the strategy.
Wall Street News: There are few and fine factors at the forefront of the century, what is your standard for judging the quality of factors?How do I find usable data by removing noise from alternative data?Will there be any concerns about the failure of factors caused by intensified competition in the industry?How do I determine the weight of a factor?
Century Frontier:In the course of our research, we always strive to find high-quality factors and characteristics with solid logical support. We avoid the data mining mode that roughly generates massive factors, but through a rigorous inspection process, we ensure that all real factors reflect the supporting laws of market structural factors. Our goal is not to pursue the quantity of factors, but to pursue the purity and effectiveness of factors.
In order to screen out high-quality factors, we use a series of strict evaluation criteria such as Sharpe ratio, drawdown, and correlation for backtesting and storage. These factors must be of significant validity and have a low correlation with risk factors. Once a factor has passed the rigorous selection criteria and is successfully stocked, we will try to keep its parameters unchanged. When a new and better factor appears, the original factor will be replaced. Machine learning model parameters will be dynamically adjusted on a monthly basis based on the results of the last 180 days of backtesting. We update the real factors on average once a quarter, and the number of factors eliminated or updated will be closely related to market changes, factor performance and research progress.
In this way, we are able to deal with factor failures and ensure that our models always perform at their best.
Although the proportion of alternative data in the strategy is relatively low, the information it brings is significantly different from the traditional volume, price, fundamental, and analyst data, and has good complementarity. Due to the nature of alternative data, special storage and calculation methods are required, which helps researchers to carry out strategy development, research, and update iterations more effectively, so as to improve the overall strategy level and provide customers with a better holding experience. In a mature market environment, alternative data strategies may perform better.
The research team conducts factor mining work on a daily basis based on the logic of volume and price information and fundamentals. Once a new factor is generated, we backtest it through a series of rigorous evaluation criteria, and the qualified factors are included in the factor pool. Subsequently, the portfolio team will combine these factors, mainly using multi-** and multi-period information to combine, while strictly controlling **, market capitalization and industry risk exposure.
Factor weights are assigned according to historical performance, and the combination of factors is updated daily through a combination of linear and non-linear, so the proportion of each factor changes dynamically.
Wall Street News: High frequency has requirements for capacity, and some companies only apply high-frequency strategies to proprietary disks, while Century Frontier is to raise funds and operate together with proprietary disks, what are the considerations?
Century Frontier:There is a certain misunderstanding of high-frequency strategies in the market. Due to the impact costs such as commissions, high-frequency strategies have no room to profit if the market trading opportunities are not prominent enough. That is, the market has fewer opportunities for high-frequency strategies to capture, so its capacity is smaller.
Century Frontier's high-frequency strategies are mainly applied at the trading level. To put it simply, the model uses low-to-medium frequency factors to screen out high-quality targets, if it is set within an hour**, then within this hour, the high-frequency strategy can use the trading information to make a profit, without considering the impact cost, after all, the cost has been decided to pay. As a result, high-frequency strategies can capture more opportunities and expand their capacity.
Wall Street News: Compared with other quantitative institutions, the biggest advantage of Century Frontier is in the world
Century Frontier:The level of awareness in the market is constantly improving. Maintaining a comparative advantage in the market requires continuous R&D investment. In the long run, the core competitiveness of quantitative managers must be the ability of continuous research and development.
1.The strategy covers all bands, and the advantages of trade execution are outstanding. Century Frontier is one of the few companies that has reached the first echelon of the market in terms of long-term alpha and high-frequency execution. The more important role of high frequency is to reduce or even offset the transaction impact cost in the case of excess reduction in the future, and obtain trading income through high-frequency trading while obtaining position income through stock selection, so as to improve portfolio stability and enhance excess. The value of the mid-range will be mainly reflected in the difficult market environment of quantification, and the ability to switch faster when the style is quickly switched**. The low-frequency factor and the medium-frequency factor complement each other and have a low correlation, which will better enhance the long-term stock selection ability and the ability to cope with the short-term style switch.
2.At the level of risk control, Century Frontier has always attached importance to the stability and reliability of earnings. We do not do proactive style exposure to ensure the long-term effectiveness of the strategy. The degree of freedom of the model fluctuates within a reasonable range, the overall sharpness is high, and the alpha is stable.
3.From the perspective of the development model of the company model, some other teams in China adopt the multi-PM model, which will make rapid progress in the early stage. However, in a highly competitive market, involution is inevitable, and PM A and PM B make similar models squeeze each other. Century Frontier adopts a close collaboration model, the overall collaboration is transparent, the internal communication is smooth, which is more conducive to the team atmosphere (the American Renaissance has always adopted this model), more information sharing and infrastructure sharing in the collaborative mode, a high level of mutual trust between the team, and high team stability, from 15 years to the present, no core investment and research personnel have left, and the long-term will be able to do more in-depth.
Wall Street News: In terms of risk management, how does Century Frontier control risks?Can you give an example?
Century Frontier:Adopt a scientific and unified risk quantification method, establish a complete risk control system and process, and realize scientific risk management. The self-built trading system has a risk control terminal to monitor positions in real time, generate risk reports after hours and submit them to the risk control committee for attribution analysis, and decompose the exposure of each portfolio to different risk factors. The company establishes and promotes the integration and implementation of risk control systems and risk control measures and methods in the company's various positions and business processes, and comprehensively supervises its implementation to supervise risk control personnel.
Wall Street News: This year is a big year for quants, why can quantitative strategies achieve excellent results?Will the subjective strategy fail to outperform the quantitative strategy in the future?
Century Frontier:This year's **, in terms of average daily trading volume, is better than in 2022, the market sentiment is relatively warm, the quantitative performance is better than in the past, and the trading activity is also accelerating driven by the ** recovery. The turnover level of major indices has been the best this year, and the negative impact on the cost and quality of quantitative strategy trading execution has continued to decline. The overall dispersion and rotation of the market style will also be relatively easier to obtain by quantifying the excess. Compared with last year, the overall environment is relatively favorable to the quantitative strategy to obtain excess returns, and the quantitative strategy manager has performed better in terms of flexibility and volatility control.
Subjective and quantitative benefits** are different in time and space, each has its own advantages and disadvantages, quantitative is a strict risk budget in advance, the benefits are very clear, and subjective risk control is mainly a compliance constraint. The performance of the two in the future depends on the market heat and style, the investment methods and returns of the two are different, quantitative is naturally the pursuit of shareholding breadth to obtain excess, so the more bids, the more obvious the quantitative advantage, and the subjective more attention to the vertical growth value, if you look at a single period of time or a certain year, compared with some excellent subjective managers, quantitative in terms of income may not necessarily be better than subjective. But it is true that on the whole, because of the advantage of width, the quantitative volatility control is more ideal. Subjective strategy can do a deeper understanding of the fundamentals and the company's long-term strategic development, and can see farther in the long-term investment direction. Therefore, it is impossible to judge in advance whether the future performance of the two will be higher or lower.
Wall Street News: There's a saying that no strategy works for a long time, so how do you see the problem of excess decay of quantitative strategies?
Century Frontier:The distribution of quantitative overweights tends to be uneven, and it can be low from month to month, and high from month to month, but it will accumulate throughout the year. We have to accept that the market environment is not good in a certain quarter, and it is difficult to do excess returns. But if the period is slightly extended, it can be seen that this excess fluctuation is quite normal. Overall, the market is still a long-term positive process, but the process is certainly not smooth sailing, and it must be accompanied by a variety of things that may be within expectations, but more likely to exceed expectations. Quantitative trading is not a magic wand, there is no winning strategy or tactics, it is just a tool to assist in trading, quantitative trading cannot be too mythical, or it has to return to the core of trading and strategy.
Investments are long-term, short-term fluctuations are normal, and investment returns take time to be maximized. When the environment is unfavorable, the strategy may underperform the index periodically. But filtering out these short-term fluctuations, the long-term overload of each strategy is cumulative, and long-term investments can often achieve more stable and sustainable growth by leveraging time value and continuous compounding.
Wall Street News: Quantitative trading has become an important force in the market, and people once aimed the spearhead of "helping the rise and killing the fall" at quantification, what do you think about this problem?Will there be a ceiling in the quantitative industry in the domestic market?
Century Frontier:Quantification is always full, no matter when the market is skyrocketing or **, quantification always provides liquidity to the market and plays a role in stability. Quantification is absorbing and suppressing market fluctuations, so that the market target returns to a more reasonable **, so there is a misunderstanding of the statement that quantification helps the rise and fall.
The quantitative industry currently accounts for about 20% of the transaction in the A** field, which is still a certain distance from the more than 50% of the transaction proportion in the U.S. ** market, and from the perspective of income mining, there are many things that can continue to be cultivated and mined, such as alternative data mining, the application of new nonlinear models and large models, etc. It is believed that in the future development process, the quantitative industry will continue to maintain its development momentum, and at the same time repay investors, it will play its unique value in China's bottom position construction, volatility stabilization, and discovery.
Wall Street News: How do you think about the changes brought by the AI model represented by ChatGPT to quantitative investment?
Century Frontier:ChatGPT has led a new wave of development in the field of artificial intelligence. Its strong capabilities are mainly due to two aspects: one is to have huge data resources, and the other is to have excellent control over data. Although ChatGPT's underlying model has been around for several years, its application in the field of artificial intelligence has led to significant innovations. Compared with traditional search engines, ChatGPT uses the inductive reasoning ability of deep learning to integrate a large amount of data and information, thus providing more logical answers.
The direct application of ChatGPT in the field of quantitative investment research is still in the exploration stage, but its unique advantages in processing text data have been recognized. We have been applying deep learning technology in this field for many years. Since 2020, we have established an AI team and recruited a number of machine learning experts from well-known Internet research institutions. At the same time, we have also invested tens of millions of dollars in the construction of machine learning computer rooms, and these experts use the neural networks built with logical features provided by our researchers to further find out the deeper laws of market structure. Over the past two years, these technologies have contributed significantly to our excess returns.
Wall Street Insider: How do you see the market next year?
Century Frontier:The current market's ** valuation is at a relatively low level, at a historical low. We are not pessimistic about the long-term performance of the market in the future, from the perspective of the cost performance of stocks and bonds, the current market is also more valuable, and in the second half of this year, the state has introduced a series of policies to encourage and guide the market to recover, I believe that it will gradually play an effect over time. On the one hand, it is conducive to obtaining excess returns, on the other hand, there is a greater probability of ushering in the market's first, when the market recovers and the trading volume is active, the follow-up performance of quantitative excess is still worth looking forward to.
Wall Street News: How does participating in the Galaxy Trading Contest empower and help your company?Compared with other institutions, what are the characteristics of Galaxy's private equity services?
Century Frontier:Participating in the Galaxy ** competition can not only enhance the company's popularity in the industry, but also inject new vitality into the company's existing investment and research database reserves with the help of various high-quality research services provided by the Galaxy**. More importantly, the cooperation with Galaxy ** can allow the company to provide investors with better asset management services in the long run, get wider attention, and thus attract the attention of more investors.