predictors
Contrasting previous studies for the U.S. market, liquidity emerges as the most important predictor, leading us to examine the impact of transaction costs closely.
This finding is in contrast to previous studies in the U.S. market (Gu et al. 2020b), where classical trend indicators are the main drivers of predictability. However, we nd notable differences across models. (这一发现与之前在美国市场的研究（Gu等人，2020b）形成对比，在美国市场，经典趋势指标是可预测性的主要驱动因素。然而，我们发现不同模型之间存在显著差异。 这里 Gu et al .2020b 中采用的也是机器学习模型，说明美国市场中，机器学习模型普遍支持经典趋势指标)
In particular, in addition to liquidity, neural nets tend to favor momentum and volatility factors over fundamentals. (特别是，除了流动性，神经网络倾向于支持动量和波动因素，而不是基本面因素。这说明中国市场中不同的机器学习模型支持的predictors 类型是不一样的 )
对于风险度量组成的 predictors，包括特殊回报波动率（INDIVOL）、总回报波动率和市场贝塔（beta）。 我们的研究结果与 Gu等人（2020b） 在美国市场的研究结果形成了鲜明的对比和差异。 他们发现，传统的价格趋势指标是最具影响力的预测指标，对中国股市来说， 除了最近的最大回报（maxret）外，其他指标的重要性都较低。 这一观察结果与之前应用线性因子模型预测中国股市的文献有很好的共鸣（参见，例如，Li等人（2010年）和Cakici等人（2017年））。 然而，基本面因素的突出作用令我们惊讶，因为根据 顾等人（2020b） 的说法，这些因素对美国市场的重要性微乎其微。 更具体地说，当我们从 Gu等人（2020b） 的 figure 5中提取前三（十）个因素时， 他们在中国市场的平均排名为41（34）。因此，两个市场在预测因素的重要性上存在很大分歧。
Cite from: Machine-Learning in the Chinese Stock Market.
retail investors
The retail investors’ dominating presence positively affects short-term predictability, particularly for small stocks . Cite from: Machine-Learning in the Chinese Stock Market. ===========================
Chinese stock market is dominated by retail investors. The speculative and short-term trading motives of many retail investors may lead to increased turnover.
- According to the 2019 yearbook of the Shanghai Stock Exchange, there are 214.5 million investors in china, 213.8 million are individual investors, and 0.7 million are institutional investors
- the value of shares traded stood at 224% of market capitalization in 2019, compared to 108% for the U.S. market.
=================================================
creates heightened volatility that may disconnect share prices from
the underlying economic conditions.=================================================
Especially the large proportion of small investors with speculative short-term behavior, make this market a highly attractive target for applying modern machine learning techniques
=================================================
!!! retail investor 和 machine learning 的关系: 我们将短期可预测性（尤其是小型股）归因于散户投资者在中国股市中的突出作用。 正如我们稍后将看到的，在较短的视野内，神经网络更加重视小盘股的波动性和动量相关变量， 这可能反映了散户投资者的短期投机行为，以及他们的知名度也就是喜欢优先交易小型股。 (We attribute the short-term predictability, particularly for small stocks, to retail investors’ prominent role in the Chinese stock market. As we will see later, at shorter horizons, neural nets put more weight on volatility and momentum-related variables for small stocks, which may reflect the short-term speculative behavior of retail investors, together with their well-known preference for trading small stocks)
Cite from: Machine-Learning in the Chinese Stock Market.
large stocks and state-owned enterprises
Another feature that distinguishes the Chinese from the U.S. market is the high predictability of large stocks and state-owned enterprises over longer horizons. The out-of-sample performance remains economically significant after transaction costs.
It seems that predicting SOEs’ returns requires a highly flexible method that can account for nonlinear effect. This additional complexity may be required since SOEs are controlled by the state, having two primary objectives: to generate profit and to carry out state policies. (似乎预测国有企业的回报需要一种高度灵活的方法，可以考虑非线性效应。 由于国有企业由国家控制，有两个主要目标：创造利润和执行国家政策，因此可能需要额外的复杂性 ) However, our results contrast earlier studies that argue that predicting stock returns for Chinese SOEs is not easy due to their nancial opacity and low informativeness of share prices(然而，我们的结果与早期的研究形成了对比，早期的研究认为中国国有企业的股票回报并不容易，因为它们的“财务不透明”和信息量低股价 )
Cite from: Machine-Learning in the Chinese Stock Market.
SOEs
Political connection may boost the SOEs’ performance through various channels.
(1) SOEs’ highly concentrated state ownership;
(2) SOEs’ highly concentrated financial opacity (金融不透明) ;
(3) SOEs’ low informative share prices
(4) SOEs’ lack of corporate governance mechanisms could potentially exacerbate the crash risk for these firms (缺乏公司治理机制可能会加剧这些金融机构的崩溃风险)
Cite from: Machine-Learning in the Chinese Stock Market.
Transaction costs:
For the Chinese market, the cost of A-share transaction mainly consists of three components:
(1) commission 佣金: In the Chinese stock market, commission fee for institutional investors was around 5 bps in 2012 but decreased quickly after then. In recent years, commission fee is usually 2-3 bps for retail investors and even lower for institutional investors.
(2) stamp tax 印花税: The stamp tax is set to 10 bps since 2008 and is collected unilaterally for sellers.
(3) slippage 滑点： slippage requires a more careful investigation as it is often diffcult to execute all transactions at the pre-specified price without affecting market price due to the liquidity issue （滑点需要更仔细的调查，因为通常很难在不因流动性问题影响市场价格的情况下以预先指定的价格执行所有交易） In some rare cases, such as 2015 Chinese stock market turbulence, the scale of slippage can be quite large as the stock market goes up or down flercely right after stock market opening. However, in such cases, the signs of buying and selling slippage are likely the same, which could partly reduce the actual slippage that investors face (在一些罕见的情况下，比如2015年的中国股市动荡，随着股市开盘后的剧烈涨跌，滑点的规模可能相当大。 然而，在这种情况下，买卖滑点的迹象可能是相同的，这可能会部分减少投资者面临的实际滑动。)
Cite from: Machine-Learning in the Chinese Stock Market.
Daily price limits
China’s market imposes daily price limits of 10% on regular stocks listed in Main Board and Second Board (20% on stocks listed in Second Board since Aug 2020), 5% on special treatment (ST) stocks, and 20% on stocks listed in Sci-Tech Innovation Board. For Chinese market, Chen, Gao, He, Jiang and Xiong (2019) find that price limits incentivize large investors to pursue a destructive strategy of pushing up stock prices to the upper price limit and then selling on the next day. Hence, they argue that this unintended effect renders daily price limits as counterproductive.
(发现限价激励大型投资者采取破坏性策略，将股票价格推高至上限，然后在第二天卖出。 因此，他们认为，这种无意的影响使每日限价适得其反。)Cite from: Machine-Learning in the Chinese Stock Market.
small stocks:
There are three main reasons for excluding the small stocks.
(1) small stocks are well-known for their high price volatility in the Chinese stock market, making it difficult for investors to find appropriate buying points
(2) the bottom 30% stocks often suffer the so-called shell-value problem caused by the IPO constraints in China as documented in Liu et al. (2019).
(3) large stocks in general have higher levels of liquidity and lower price volatility, and thus are less affected by the daily price limits of 10% in China. (大型股票的流动性水平一般更高， 并且有着低的价格波动率，因此受到 10% 涨跌幅的影响比较小。)
Cite from: Machine-Learning in the Chinese Stock Market.
market size
As of October 2020, the total value of China’s stock market has climbed to a record high of more than $10 trillion, as the country’s accelerating economic recovery from the pandemic has surpassed the previous high reached during an equity bubble five years ago, making it the second-largest in the world, after the U.S. at nearly $39 trillion.
Cite from: Machine-Learning in the Chinese Stock Market.
politic
(1)For example, the process of IPOs and seasonal stock offerings is highly political, and companies cannot predict when the market value will be high.
(2)On the other hand, listed companies, especially state-owned enterprises (SOEs), are prevented from shares buy-back when share prices fall below fundamental values. These automatic market correction mechanisms are therefore affected by government-oriented restrictions
(3)Given that China has been experiencing a highly dynamic development through a series of structural breaks, implementing various financial reforms, and expanding its capital markets’ openness, we conjecture that highly fiexible methods are required to account for the Chinese market’s specificity.
Cite from: Machine-Learning in the Chinese Stock Market.
IOP
China has been adopting an approval-based IPO system ever since its stock market opened, and it is wellknown that the China Securities Regulatory Commission often suspends or reduces the volume of IPOs when the market is down, making it reasonable for ntis to play an important role in predicting monthly returns. (中国自股票市场开放以来一直采用基于批准的IPO制度，众所周知，中国证监会经常在市场下跌时暂停或减少IPO数量， 这使得NTI在预测月度回报方面发挥重要作用是合理的)。
long-short strategies
While most of the literature on factor investing in U.S. and European markets rely on long-short strategies, such a strategy is less realistic for Chinese market. Hence, we also analyze long-only portfolios which are more relevant from a practitioner’s viewpoint.
Given the short-selling constraints in China, we wonder how much value-added can be enjoyed in long-only mandates. Many of the previous literature results relate to the performance of portfolios that include long and short positions. While such practices allow us to evaluate a signal’s predictive power, not all stocks are available for shorting at all times, and the costs of shorting can be substantial. (以前的许多文献结果都与包括多头和空头头寸的投资组合的绩效有关。虽然这种做法允许我们评估信号的预测能力， 但并非所有股票都可以随时做空，做空的成本可能相当高。 )
Cite from: Machine-Learning in the Chinese Stock Market.
factor investing
As of today, there is no large database of factor returns available for the Chinese market. Therefore, we contribute to the research on empirical asset pricing in China by building a unique and comprehensive set of factors
Cite from: Machine-Learning in the Chinese Stock Market.
machine learning
A rapidly growing number of studies examine the cross-section and the time-series of stock returns with machine learning tools, predominantly focusing on the U.S. market.
Their results suggest that machine learning improves the description of expected return and, when applied to portfolio construction, performance improvements arise most prominently among the more sophisticated models and are due in large part to the allowance of non-linear predictor interactions that are missed by simpler methods
Although machine learning may substantially improve the pricing of assets, Arnott et al. (2019), Gu et al. (2020b), and Israel et al. (2020), among others, also point out that these methods cannot identify deep fundamental economic principles.
Cite from: Machine-Learning in the Chinese Stock Market.
R^2
Comparing the out-of-sample R^2 with studies in the U.S. market, the Chinese market reveals substantially more predictability.
R^2 has some limitations for model selection (R^2 并不能作为衡量所有机器学习模型的一个普适的指标)
We also find that predictability of SOEs (国有企业) in terms of out-of-sample predictive R^2 is weaker than for non-SOEs at a monthly prediction horizon, which confirms the SOE’s reputation of being non-transparent (国企的声誉并不透明).
It is noteworthy that the OLS model performs much better in China’s stock market than in the U.S. stock market. The R^2 for the latter has been reported to be negative (-3.46%) in Gu et al. (2020b).
Such significant gaps in R^2 further motivates us to consider the fundamental difference between these two markets, which we conjecture, can be attributed to two critical aspects. First, the Chinese stock market is characterized by a large fraction of retail investors and their preference for small-cap stocks. Second, the Chinese stock market is influenced by the prevalence of state-owned enterprises, which are less transparent than private firms.
Cite from: Machine-Learning in the Chinese Stock Market.
events
2015 crash
COVID-19 pandemic
political risk related to a trade war between the U.S. and china.
Three key features in chinese market
Chinese stock market is dominated by retail investors.
a key characteristic of China’s financial system from an institutional perspective is that it is centrally controlled, bank-dominated, and uniquely relationship-driven.
The chinese market has a limited history of short sales.
Cite from: Machine-Learning in the Chinese Stock Market.