A distinguished group of quantitative equity managers have been extracting extraordinary amounts of alpha from onshore China equity markets, which are the world’s second largest based on market capitalization, and the most liquid measured by daily turnover that averaged nearly $180 billion and sometimes surpassed $250 billion per day in 2021. Annual long book alpha has sometimes exceeded 15% or 20% and returns in equity market neutral strategies have been in the teens or better even after the 5-12% annualized cost of shorting indices that historically populated or dominated short books. Increasingly, managers are also now generating alpha from a growing universe of single stock shorts, as well as running active long only strategies and contemplating more scalable passive or semi-passive long only strategies alongside their pure alpha products.
In July 2021, we pivoted from large to mid and small caps, after the Chinese government announced regulation of technology and education companies.
Lewis Prescott, Partner and CEO of International, Mingshi Investment Management
Mingshi Investment Management, which was one of the first quant equity managers in China when it was founded in 2010, started shorting single stocks in March 2020, and after some learning experiences and a restructuring of the process, the single name shorts generated an absolute return on capital of around 7% in 2021, according to a source familiar with the matter. In alpha terms, this was many times higher – and even higher than on the longs – since short index positions would have cost a multiple of the single name short profits.
Mingshi declined to comment on performance figures for this story.
Mingshi’s long alpha generation in the high teens for 2021 was almost entirely idiosyncratic or stock specific alpha. MSCI Barra’s widely used 21 factor model (covering analyst sentiment; beta; book to price; dividend yield; earnings quality; earnings variability; earnings yield; growth; industry momentum; investment quality; leverage; liquidity; long-term reversal; mid-capitalisation; momentum; profitability; seasonality; size; short term reversal) could not explain the returns of Mingshi’s market neutral OPTIMA strategy. In aggregate, these traditional factors netted out near to zero, leaving the residual “Mingshi factor”, which includes tens of thousands of dimensions, accounting for nearly all the returns. Barra is used to providing clients with industry standard performance attribution, but the true drivers of returns are more nuanced, subtle and esoteric variables.
These signals are unearthed through rigorous research and the DNA of the firm is clearly academic. It was founded by two renowned academics from the US and China, respectively: Professor Robert Stambaugh, Miller Anderson & Sherrard Professor of Finance at The Wharton School, University of Pennsylvania, who supervised the PhD of Professor Yu Yuan, the other founder. Out of 100 staff members, most of the researchers and risk experts have PhDs, mainly from US and Chinese universities, such as Wharton, Yale and SUFE (Shanghai University of Finance and Economics) with which Mingshi has a strategic partnership that both drives forward quantitative finance and raises Mingshi’s profile amongst potential hires. Historically, behavioural finance was one of the most fashionable disciplines but now computer science and machine learning are very sought after. Mingshi is competing not only with giant US hedge funds but also with global technology firms that are developing autonomous driving.
“There is a huge talent pool of domestic PhDs in finance, physics and engineering,” says Mingshi Partner and CEO of International, Lewis Prescott, an Australian CFA charterholder who is highly unusual in not having a PhD in a firm dominated by Doctors of Finance and Science. He joined Mingshi in August 2019, after a career trading, managing and structuring North Asian equities and derivatives on the sell side and buy side for Nezu Asia Capital Management, Mizuho Securities, Citi and Deutsche Bank. Prescott, who sits on the investment committee, is now applying his hands-on trading savviness to onshore Chinese equities.
5
Mingshi’s average holding periods of 5 days, ranging between 1 and 10 days, make the strategy “medium frequency” and mean that transaction costs are an important consideration
Mingshi manages around $2.5 billion, focused exclusively on around 2,500 of the c. 3,500 China ‘A’ share stocks listed on the Shanghai and Shenzhen exchanges, (and does not trade OTC stocks in China nor listings of Chinese firms on other exchanges in Greater China, Europe or the US). Within the onshore universe, mid-caps have often been the sweet spot for alpha generation, though Mingshi are opportunistic in moving between large, mid and small caps. “In July 2021, we pivoted from large to mid and small caps, after the Chinese government announced regulation of technology and education companies,” says Prescott. This was picked up by models reading headlines and one of the firm’s programs made 20% in July 2021. Mingshi’s research is dedicated to deep dives into the Chinese equity market, which has many distinguishing features.
Retail participation differentiates Chinese equity markets and is substantial partly because equities and property are the main options for local investors. Though institutional investors have been growing their proportion of equity market ownership, this is often longer-term holdings, and some 70% of daily turnover in China is active retail investors, and this leads to different dynamics versus the passive and systematic flows that dominate US, UK, and Japanese market flows. “For instance, some research has arrived at a stylized fact that single stocks more often show reversal than momentum patterns. Retail investors are active traders, contributing to a higher velocity of trading: turnover is 200% per year versus around 100% in the US. Patterns of retail investor behaviour are one example of a signal that Mingshi monitors, where the data needs to be gathered from disparate sources,” says Prescott. Retail investor behaviour can also change over time: historically, retail investors were sometimes heavily leveraged, but limits on margin were introduced after the market crash in 2014-2015.
Limited sell side broker coverage of stocks is another source of inefficiencies and bottlenecks in price discovery. “The major banks overlook many Chinese stocks and some 70% of the Chinese equity market is covered by less than two analysts, which means that news headlines or retail investor opinions can be more important drivers,” says Prescott.
In common with many emerging markets, standards of accounting, corporate governance and transparency in China will tend to be lower than in the US or UK, where the financial infrastructure is hundreds of years old. “This adds risk and potential return for those investors who can interrogate the data and discriminate between the quality of disclosures from different firms. We have a whole team dedicated to reading government, regulatory and earnings announcements, which are not as clean and easy to navigate as in developed markets. The approach is quite disjointed,” says Prescott.
Mingshi’s average holding periods of 5 days, ranging between 1 and 10 days, make the strategy “medium frequency” and mean that transaction costs are an important consideration. “Overall dealing costs in Chinese equities – including market impact, dealing commissions, and a stamp duty of 0.10% – are higher than in the US, though some components of dealing costs such as bid/ask spreads are not necessarily higher for all clusters of stocks. Strategies that might work in markets with lower costs could struggle to cover the higher fixed costs in China,” says Prescott.
The situation is fluid and is getting better for those investors who know where to locate more competitive trading and pricing: “The quality of execution does vary between onshore brokers, though it is improving as onshore brokers, such as Morgan Stanley, Goldman Sachs, BAML and UBS, get more onshore licenses. We have developed our own internal algorithms to automate execution,” adds Prescott.
Single stock shorting is the best alpha opportunity in global equities, partly because it is difficult and expensive to find inventory, and less crowded.
Lewis Prescott, Partner and CEO of International, Mingshi Investment Management
Shorting is also, on average, much more expensive than in other markets. Mingshi uses two index futures for hedging out market beta risk. The mid-cap CSI 500 is the most popular as it is the closest match for mid-caps on the long book. The large cap CSI 300 is also used. “Shorting costs do fluctuate somewhat, but are typically around 10-12% for the CSI 500 and 5% for the CSI 300,” says Prescott. At single stock level, shorting costs can be in the mid-single digits annualized.
The research papers on the Mingshi website discuss proprietary factors in areas such as sentiment, quality, size, value, and ESG at a very high-level thought leadership style, which provides some broad-brush perspectives on Chinese equity drivers. For instance, Mingshi’s proprietary quality factor has outperformed in China while its proprietary size and value measures also explain a high proportion of anomalies. The long-term phenomenon of low beta stocks outperforming high beta stocks, dubbed the “low volatility anomaly”, can hold true in China, but only once a valuation filter is added. “This published research could provide a framework for passive long only offerings rather than pure alpha. The hard-core in-house research is much more granular,” explains Prescott.
Mingshi boasts a team of 65 staff dedicated to researching the unique features of Chinese onshore stocks, including price histories, volume, behavioural investor patterns and alternative data. Even sourcing data is not straightforward and requires elaborate processes: “Gathering and cleaning share price data history, which cannot be easily downloaded off the shelf, is not easy. A cumbersome process has been developed to scrape environmental filings from government websites. Meanwhile Natural Language Processing (NLP) is used to scrape news headlines. In some areas, such as geolocation, data is not as advanced as in the US,” says Prescott. Mingshi has invested over 100 million RMB in computer power, including a supercomputer in Nanjing.
Researchers specialize in these sorts of activities. Rather than specializing in industrial sectors, like a traditional fundamental discretionary asset manager, Mingshi’s teams focus functions, such as NLP of news; transaction cost analysis; exchange pricing or volume data across the full universe of 2,500 names, drilling into each one. “The research is tailored to each individual stock, with tens of thousands of unique, discrete bespoke signals and factors used internally,” says Prescott.
ESG data in China is still at an early stage and takes a lot of intensive effort to gather. Though there have been some notable ESG scandals such as contaminated baby milk, and rice tainted with metals, Mingshi’s published research does show “clean” firms outperforming “dirty” ones, and there is not as much historical data in aggregate. “There is no useful third party ESG data for China since agencies are copying and pasting data from western firms. Therefore, all our ESG data is proprietary and self-built from company reporting and some local government carbon disclosures. Companies’ own reporting can be biased if they are more inclined to report positive data. Supply chain analysis can also be used to identify the regional sources of inputs,” says Prescott. The volume of data and analytics is growing exponentially: “ESG is evolving very fast: the number of ESG research papers has exploded from 3 to 300 per year over the past 10 years,” says Prescott.
Mingshi has developed a proprietary ESG China ‘A’ share index, which is currently used for internal analytics but might one day become investible, possibly even via an ETF.
Minghsi’s process is substantially quantitative, though the investment committee does exercise some discretion over such regulatory and compliance matters, such as weighting and updating signals. “Qualitative judgements adjust the weightings of signals and proceed cautiously with newer signals. The updates cycle includes daily optimization with minor updates weekly and major updates monthly. Signals and factors are also adjusted with shifting market regimes,” says Prescott. China’s equity market regimes can sometimes move in the opposite direction of the US, as seen in 2021, when the Chinese mega cap tech bear market coincided with new all-time highs for the Nasdaq 100.
Mingshi’s single stock shorts have recently grown to 50% of the short book. “Single stock shorting is the best alpha opportunity in global equities, partly because it is difficult and expensive to find inventory, and less crowded,” argues Prescott. The short-able universe has reached 80-90% of the major indices and should continue to grow. “Inventory is expanding thanks to the October 2020 reforms that initially opened up offshore QFII inventory for synthetic shorting, and will eventually allow onshore shareholders to lend offshore,” says Prescott.
Mingshi is acutely aware of the risk that overcrowding in some trades could erode alpha and is monitoring a new factor that gauges quant investor crowding in certain stocks, factors or sectors. “Some of the more popular positions have seen a performance reversal since September 2021, possibly due to synchronized deleveraging and increased correlations between positions,” says Prescott.
Mingshi has an onshore PFM (private fund manager) entity in China, regulated by the Asset Management Association of China (which is a regulator not a trade association) under China Securities Regulatory Commission (CSRC) rules, and an offshore Hong Kong entity, overseen by the Hong Kong Securities and Futures Commission (SFC). Some US rules and laws are also observed.
Mingshi aims to offer a best-in-class institutional quality operational and compliance infrastructure designed to satisfy investors’ operational due diligence criteria. Its annual reviews also provide clients with a high degree of transparency.
Onshore Chinese equities have attracted more inflows over the years since various reforms paved the way for them to become constituents of indices such as MSCI, but have seen some net outflows in the early months of 2022, possibly due in part to political and geopolitical sensitivities. Some prospectuses of funds investing in China do continue to flag up disclaimers around repatriation risks and tax uncertainties. Prescott argues that “an offshore fund structure, with a Hong Kong management company and a Cayman fund, provides some degree of bulwark against regulatory black swans”.