In this research note, we examine the opportunities and challenges associated with implementing a quant factor process in Japan. The information is presented in a question-and-answer format.
What are the advantages of implementing a quant factor process in Japan?
The primary reason we like Japan is it offers high “breadth”. This is crucial for any quant factor process given predictive power is low, and it’s important to utilise the law of large numbers to generate consistent alpha. Having a large number of stocks also makes it easier to balance the portfolio across the various risk dimensions we monitor.
Based on the breadth criterion, Japan is the second-best market globally after the United States.
For our investment process, breadth is a function of the number of liquid stocks with analyst coverage that are also shortable.
There are 779 stocks listed in Japan with 2 or more analyst providing coverage that have a median daily turnover over $US2m, and most of these stocks have a cost of borrow less than 0.5% per annum. There are more large and liquid companies listed in China, but shorting these stocks is problematic given the high borrow costs (Table 1).
Table 1: Number of stocks within the Asian markets with different borrow rates Source: OQ
Japan also offers extremely low transaction costs. Bid-ask spreads are low (Chart 1), liquidity is high (Chart 2), brokerage rates are low (we pay 1 basis point), and there are no stamp duty costs.
Chart 1: Bid-ask spread for Asian stocks >US$1m turnover Source: OQ, Bloomberg
Chart 2: Average turnover for Asian stocks with at least 2 analyst coverage Source: OQ, Bloomberg
Being able to transact frequently in a cost-effective way is extremely important for our style of investing. And it’s becoming more important as we increase our trading frequency to exploit short-term alpha signals. For more information, refer to: https://www.oqfundsmanagement.com/post/why-we-execute-a-large-number-of-daily-trades
Chart 3 shows a 6-month moving average of daily trades since the fund was launched.
Chart 3: 6 month rolling average of daily trades Source: OQ
What are the challenges?
Japanese equities have some unique characteristics that can potentially undermine the predictive power of our investment process.
First, it’s important to consider shareholder perks. We discussed this issue in the following market insight: https://www.oqfundsmanagement.com/post/asian-stock-market-idiosyncrasies
“In Japan, shareholder perks are highly sought after – more so than in any other stock market - and companies that offer attractive perks often look expensive based on standard valuation measures.
Based on finance theory, companies should be valued based on the present value of future dividend payments. In Japan, it’s also necessary to include the value of shareholder perks.
Consider Aeon (8267 JP), one of the largest retail companies in Japan which operates a well-known supermarket chain. It has a reasonably impressive revenue yield, comfortably exceeding 250%. However, it’s gross profit margin is miniscule, and it trades on an earnings yield just over 1%.
It looks grossly overvalued - until you consider the shareholder perks.
Japanese retail investors purchase shares in the company for the shareholder discount, ranging from 3% to 7% for purchases, depending on the number of shares owned. This explains why the gross profit margin is so small – and why traditional valuation models aren’t appropriate for valuing the company.”
In the same market insight, we also discussed the distortionary impact the BOJ has had on
Japanese stock returns. For many years, the BOJ aggressively purchased ETFs based on nonrepresentative market indices such as the Nikkei 225 and Nikkei 400. This inhibited efficient price discovery.
Next, the composition of Japanese company balance sheets presents unique challenges.
Numerous Japanese companies have assets on their balance sheet which are priced well below market values. This is one reason why Japanese equities provide such abundant opportunities for activist investors.
These assets often comprise property, as reported by Bloomberg (16 April 2024):
“In Japan, there’s a huge gap — 22 trillion yen ($143 billion) by one estimate — between how companies value their real estate assets on their books, versus what those same properties would fetch if sold in the current market.”
Undervalued shareholdings are also common in Japan. Consider Keisei Electric Railway’s (9009 JP) holding in Oriental Land (4661 JP), as reported in Nikkei Asia (30 October 2023):
“Keisei's numbers are stark. Oriental Land has a market capitalization of 8.3 trillion yen ($55.3 billion), meaning that Keisei's 22% stake should be worth 1.8 trillion yen. The anomaly is that Keisei's own total market capitalization stands at just 952 billion yen, or half the value of its stake in Oriental Land. The anomaly is obscured on Keisei's books where the stake is carried at a value equivalent to just 5% of its market worth.”
This means that the outputs from valuation models that use net assets data are often compromised and need to be interpreted with caution.
Another unique characteristic of many Japanese balance sheets is net cash positions (ie cash and holdings in liquid securities which exceed financial debt). Of the 1081 companies with a daily median turnover greater than $2m per day, over 40% have net cash positions (19 April 2024).
This has implications for valuing companies. It’s important to consider valuation factors normalised using enterprise value as well as market capitalisation. It’s also important to consider if companies are using, or likely to use, their cash holdings to increase the shareholder yield, either via a higher payout ratio or share buybacks.
Finally, there is more quant factor money in Japan than other Asian equity markets. This largely comes from global quant funds which are attracted to the high breadth offered by Japanese equities. These funds have their own proprietary models, but they often target similar mispricing opportunities, making it more difficult to generate alpha.
These issues have impacted factor performance in Japan – and it’s a good segue us to our next question.
How well have quant factors performed in Japan?
We typically measure the predictive power of quant factors using regression analysis. This can be done in myriad ways using both univariate and multivariate regressions. For simplicity, our comparative performance analysis uses univariate regressions that measure cross-sectional correlations between monthly factors score and performance rankings (monthly Rank ICs).
The following chart shows the average monthly Rank IC for our generic composite quant score over different time periods for the key Asian markets in our stock universe. Recent performance has been strong, but it has been more subdued over the last 10 years.
Chart 4: Asian markets ICs over different timeframes Source: OQ
How well has our fund performed in Japan?
The performance contribution and statistics from our Japan portfolio are shown below .
T
Table 2: Full Japan Portfolio Performance
Chart 5: Japan Portfolio Performance VAMI Source: OQ
Japan has been a particularly strong alpha source over the last 18 months.
Table 3: Japan Portfolio Performance from Oct 2022
What is the outlook for our investment process in Japan?
There are several reasons why we’re optimistic that our investment process will continue to generate strong alpha in Japan.
First and foremost, recent enhancements to our investment process targeting short-term mispricing opportunities are ideally suited to Japanese equities given the low transaction costs. In particular, we have recently added intraday rebalance screens to our investment process which identify attractive short-term mean reversion opportunities.
Second, the BOJ has ceased purchasing share ETFs. Given the BOJ purchased Nikkei 225 and Nikkei 400 ETFs, neither of which use a cap weighted methodology, this removes an impediment to normal price discovery.
Third, we believe we will continue to generate strong alpha via our discretionary overlay in Japan. We have a lot of experience addressing issues such as shareholder perks, balance sheet undervaluations, and large net cash holdings.
Fourth, recent improvements in corporate governance and increased attention on capital efficiency represent positive developments for our investment process. This has been driven by initiatives from the Tokyo Stock Exchange and the Government Pension Investment Fund (GPIF), as well as by increased influence from activist investors who are gaining more traction with Japanese management and boards. These developments could potentially address one of the challenges we face in Japan: the identification of undervalued stocks that remain undervalued for extended periods. We believe these changes will help shorten the time required to close the gap between market and fundamental values, making it more compatible with our investment horizon. Man Group agrees with our view, arguing in a recent report that “any policy that is generally targeting companies with low valuations should inherently benefit Japan Value stocks” (www.man.com/maninstitute/corporate-japan-finally-getting-house).
Finally, we believe the amount of quant factor money in Japan has declined compared to the past 20 years. Our factor-based approach to investing is out-of-favour with many multi-manager firms which are increasingly dominating the industry in large markets such as Japan. Many of these firms are fixated on simplistic measures of idiosyncratic risk which comingle alpha and risk factors. We discussed this issue in detail in the following research note: https://www.oqfundsmanagement.com/post/the-potential-pitfalls-of-focusing-on-idiosyncratic-risk. The blending of systematic quant factor models with human intuition and insights also differentiates us from other funds. The upshot is we can exploit a differentiated alpha source in Japan.