Growing divergence

Given new all-time highs in Japanese equities, corporate governance reforms, resurgent takeover activity, and the general emphasis on ESG, it seems perhaps surprising that governance alpha has not at least decayed over time, if not been arbitraged away. In fact, Khan observes the gap between good and bad corporate governance growing wider: “We find that the proactive companies are steadily improving, and the obstructive companies are barely changing. Reactive companies are somewhere in the middle. The gap between good and bad governance is widening. This is true in pretty much every sector”.

Majoring on short alpha

It is relatively unusual for such strategy to generate more short than long alpha since arithmetically longs have unlimited upside while shorts cannot drop more than 100%. Khan has however made 16.7% short and 7.1% long alpha defined as return on capital relative to the TOPIX 500 index. In terms of absolute performance contribution approximately 4.1% of the 5.9% annualized net alpha came from the shorts and 1.8% from longs.

A good nose for fraud has helped the short book: “We have been short a number of companies that have had frauds. These companies had multiple red flags. We do not name individual names,” says Khan, who published a “scandal watchlist” in his former role as Head of Japan Research at Jefferies.

Though Khan is confident in his shorting prowess, he is circumspect about the risks of single stock shorts and they are sized appropriately: “We keep the short risk from any individual name limited. Having two shorts half the size of each long makes portfolio construction easy. It also makes it easy to replace names if we hit a stop loss”.

Other shorting risks are also under control. The strategy has not had a short recalled nor paid  above 2% to borrow any name. “We do not take positions in illiquid, hard to borrow or small cap names,” points out Khan.

Net beta has fluctuated between plus and minus 0.1, and Khan does not view the portfolio as being vulnerable to a “dash for trash” rally in speculative stocks.

Factor exposures

Analysis shows two thirds of risk is stock specific. Style and factor risk is usually below 25%, and has been somewhat long of value, earnings yield, dividend yield, liquidity, profitability and momentum, and short of leverage factors. The long profitability and short leverage bias also creates a degree of quality factor exposure.

Governance (i.e. being long of good governance) remains the key style and factor exposure, which explains a good part of the long term positive returns – and the worst drawdown. “Our quantitative team’s analysis shows that in 1H 2022 the “governance” factor return was minus 11%. During that six-month period the drawdown was minus 8%,” Khan recalls.

Khan explains how governance also has some indirect degree of interest rate sensitivity. “With good governance companies you believe in their long-term future, hence a lot of the value is in the terminal value. If interest rates go up, you have to discount the terminal value more to come up with the present value. With bad governance companies, you do not believe in the long-term future, but rather you are betting on the short-term outperformance for whatever reason for e.g. they currently have a hit product, the sub-sector is seeing great demand, the FX rate is particularly beneficial, etc. The impact of interest rates going up is not a factor in your decision to invest.”

The growth style factor also has some interest rate sensitivity, but this should not be conflated with governance, and Khan is very precise about where the risks lie: “The risk systems as well as the ex-post-performance show that there is almost zero growth-value bias in our portfolio”.

Objective and subjective assessments of governance

Careful judgment is needed to arrive at sensible conclusions about Japanese corporate governance. The objective criteria that Khan applies – shareholder alignment, outside influence, diversity, board entrenchment and skills of independent directors – are only the start of the process. “Our well-defined rating criteria could be applied by a good data scraping program combined with AI. However, the numerical scores are only an indication and point us to areas where we need to dig more deeply. Judgement calls are very important and include the history of each company’s governance changes as well as a lot of non-quantifiable factors, including meetings with the company and opinions of sell side analysts, etc” explains Khan, who previously managed absolute return portfolios at Eiken Capital.

Dynamism and fluidity in long and short book categories

Moreover, the strategy is selective and adaptive within the corporate governance categories. Khan has broadly categorised the universe into approximately 20% of companies that are “proactive” regarding corporate governance, 30% that are “obstructive”, and 50% in the middle ground “reactive” group.

Khan is not only appraising current governance but is also forecasting change: “We include both companies with good governance and improving governance in our long holdings. We see both as a source of alpha”.

The long book split between good and improving governance varies: “During Covid, when the stock market was trading at much lower levels and activists were not a major factor, most of our longs were in the 20% Proactive group and most shorts were in the 30% Obstructive group. Since 2023, when activism has become a bigger factor and especially after the 2023 TSE reforms as well as the 2023 METI new takeover guidelines, we have focused more of our long book on companies in the middle Reactive group that we see signs of changing and improving”.

The long book has better corporate governance than the short book, but given the fundamental filters, the balance between corporate governance categories does vary opportunistically and according to valuations. Some of the most proactive and even some reactive companies have become too expensive to go long of – and some reactive ones might even be short candidates. “The shorts mostly are still from the 30% Obstructive group, though on some occasions we have shorted Reactive companies that have run up far too much and become very expensive. This has particularly been the case when many Obstructive companies in the sector were low PBR [price to book value ratio] companies subject to the TSE low PBR campaign,” explains Khan.

Wagering on and against activist success

Khan is neither publicly nor privately activist, but he pays close attention to activists in the Japanese equity market: “We do take long positions in companies that we believe would be attractive targets for activists”.

Some activists might be able to prompt or accelerate improvements in corporate governance, which could be a positive driver for longs – or a risk factor for shorts. Khan takes a case-by-case view on the probability of activist success, and he could sell a long, initiate a long, cover a short, initiate a short – or indeed do nothing – after activists take a stake. On several occasions Khan has accurately anticipated that activists would fail and put on shorts after they announced holdings.

Sector neutrality

The strategy is virtually sector neutral according to the broadest industrial sector definitions but can have some subsector positioning. “All three types of companies – proactive, reactive and obstructive – exist in each sector. We keep each sector net exposure within +/- 2%,” says Khan. An important nuance is that sectors here are Khan’s customised definitions, which derive from the MSCI definitions, subdividing some MSCI sectors and amalgamating others: “We split the MSCI Consumer Discretionary sector into two sectors: Automotive and Consumer. We split the MSCI Industrials sector into three sectors: Machinery, Transport and Electronics. We split the MSCI Materials sector into two sectors: Chemicals and Materials. We combine MSCI Energy sector into the Materials sector (Energy is a small sector in Japan). We combine the electronics companies in the TMT sector into the Electronics sector. We combine the Telecoms companies in the TMT sector into the Utilities sector,” he explains.

Subsector contributions

Overall, some sectors and sub-sectors have made a marginal performance contribution. “We keep net exposure to each sector within +/-2% thus there is a little bit of exposure left. But there may be sub-sector risk. For example, in the automotive sector, you have auto OEM, tire and auto parts. Within the consumer sector you have leisure, restaurants, speciality retailing, department stores, etc. We are not sub-sector neutral, so a tiny bit of such exposure would show up in the risk system,” explains Khan.

Enhancing stop losses and increasing turnover

The stop loss policy is humble enough to accept that the governance and fundamental analysis may sometimes not be appreciated by the market and also recognises that rising tides can lift all boats. “We apply the stop loss when both the absolute and relative performance is minus 15%, so a short simply going up in line with the market would not be stopped out,” explains Khan.

These stop losses have increased portfolio turnover. “Since we adopted the tighter hard stop loss rule in September 2022, turnover has increased from 50% a year on long and short books, to 100% on the long side and slightly higher on the short side,” notes Khan. This may however overstate the changes in names since some can be revisited after stop-outs: “Removing names stopped out and re-entered after 2-3 months when momentum reverses, the average holding period is probably closer to 1.5 years,” Khan explains.

ESG ratings and Japanese exceptionalism

This strategy is mainly focused on governance spreads as an alpha driver, but its long book also scores above its short book on broader ESG metrics. “There is significant correlation between G and E and S. For example, on the MSCI ESG rating system, we find that our long holdings have average ESG scores between AA and A, while our short holdings have an average score between B and BB. The average E and S scores also show similar differentiation, but not quite as large a gap. We believe that G is the foundation of ESG. If you cannot believe in the governance of a company, how can you trust in their E and S policies,” argues Khan.

As with industrial sector definitions, there are some exceptions to the correlations between public ESG scores and Khan’s proprietary ratings where his local expertise again comes into play: “There may be 60-70% correlation between our overall rating and public ESG ratings such as Sustainalytics, MSCI, etc. However, in a significant number of cases we find very large differences between our ratings. I believe this is because they use frameworks developed in the US or Europe that do not take into account Japan specific issues,” explains Khan.

Minimal macro sensitivity: Yen, interest rates and yield curve

As net exposure has averaged 0.3%, there is essentially zero currency exposure in the base currency USD share class. “The Yen share class suffers as the hedge cost from USD into JPY is higher than the yield on the T-Bills,” points out Khan.

The Bank of Japan ending yield curve control and ceasing ETF purchases is not an important factor, and nor has the return to some inflation and slightly higher interest rates made a difference. “Our DCF (discounted cash flow) valuation model uses 20-year JGB yields to compute the discount rate. We did this to explicitly avoid the impact of the BOJ’s yield curve control policy, which focused on yields up to 10 years. And we always assumed 2% inflation in our models for periods past financial year 5,” explains Khan.

Regardless of the macro backdrop, Khan is confident about extracting alpha from the governance spread for years to come.