Liquidity vs Fundamentals

What’s driving share prices?

To state the obvious, share prices are driven by liquidity flows. Put simply, when there are more buyers than sellers, share prices rise - and when there are more sellers than buyers, they fall.

Over the long term, liquidity flows reflect company fundamentals. Good companies attract more net inflows than bad companies.

The problem is the long term can be very long. We subscribe to the view outlined by Cliff Asness in his paper The Less-Efficient Market Hypothesis (https://www.aqr.com/Insights/Perspectives/The-Less-Efficient-Market-Hypothesis) that markets have become less efficient. Here’s an excerpt:

I believe markets have gotten less efficient over the 34 years since the data in my dissertation ended. I believe it’s likely happened for multiple reasons but technology, gamified 24/7 trading on your phone, and social media in particular are the biggest culprits.

While there are likely structural forces making markets less efficient, there are also temporary fluctuations driven by investor sentiment. Extremes of optimism and fear - manifested as exuberant buying or panic selling - tend to produce the largest deviations from intrinsic value.

Currently, investors are bullish, particularly regarding AI, and we’re witnessing some interesting liquidity driven price moves.

What are some examples of liquidity driven price distortions?

Softbank (9984 JP)

SoftBank is notionally classified as a Communication Services company, but in reality, it operates more like a closed‑end investment fund with numerous holdings, predominantly in the technology sector. For example, it has a large stake in Arm Holdings (ARM US) and interests in other AI‑related companies such as OpenAI.

Like other closed‑end funds, its share price can deviate significantly from its net asset value (NAV), and the discount/premium to NAV is the key valuation metric. Earlier this year, it traded at a discount of around 60% to NAV. Currently, the discount is in single digits and, if AI euphoria continues unabated, it could even trade at a premium to NAV in the not‑too‑distant future. SoftBank Group founder Masayoshi Son has now reclaimed the crown as Japan’s richest person after a four-year gap.

The fund previously held a long position in SoftBank, but we now maintain a relatively large net short exposure following the rapid narrowing of its NAV discount. If AI sentiment deteriorates, the share price could decline sharply. Not only will the value of its investment decline but the discount to NAV is likely to increase sharply.

TSMC US vs 2330 TT

In our September newsletter, we highlighted the significant premium at which TSMC’s ADR was trading relative to its home listing in Taiwan. At the end of September, the premium exceeded 30%, and it currently stands at 25.3%.

Chart 1 shows the extent to which the premium has changed over time.

Chart 1: TSMC US Listing vs TSMC Taiwan Listing Premium

The premium reflects differences in liquidity pools. U.S. investors - particularly retail investors - are all in on the AI theme. However, there isn’t enough institutional money in Asian equity markets to match this enthusiasm. While this continues, the ADR premium may remain at elevated levels or increase further.

Tesla (TSLA US)

Tesla is a company I write and think about too often, given that it's not an Asian company and I manage an Asian fund. However, I can’t help myself, as I find it a fascinating example of the irrational exuberance particularly prevalent among U.S. retail investors.

One factor that differentiates Tesla from the other Magnificent 7 stocks is its high level of retail ownership. Below is ChatGPT’s analysis of retail ownership across this group of companies.

Given Tesla’s strong retail investor base and Musk’s cult-like following, it bears many of the hallmarks of a meme stock. What amazes me most is that, unlike other well-known meme stocks such as GameStop and AMC Entertainment, Tesla is an extremely liquid mega-cap company. More recently, we’ve seen wild swings in Beyond Meat (BYND US) based on retail investor flows, but this is a relatively small company. The fact that retail investors can meaningfully influence the share price of a $1.5 trillion company with an average daily turnover of around $25 billion underscores the sheer scale of retail participation in the U.S. market.

Last week, Tesla reported earnings that came in below expectations, despite a buying frenzy ahead of the expiration of the U.S. EV tax credit. It didn’t matter. The next day, the share price rose - presumably because retail investors were more focused on Tesla’s AI initiatives, such as autonomous driving and robotics.

In my view, it’s extremely unlikely the company will live up to the hype and the company’s share price in 5 years’ time will be much lower than it is today. What happens to the share price between now and then, however, is highly uncertain. It may more than double, resulting in a market cap exceeding $3 trillion, before correcting sharply.

Aeon (8267 JP)

Aeon Co has the highest proportion of retail shareholders among Japan’s large-cap companies - roughly one-third of its shares are held by individuals. Most large‑cap Japanese companies have retail ownership well below 20%, as their shareholder bases are dominated by institutional investors, financial institutions, and cross‑holdings.

As I’ve previously noted, many retail shareholders bought their shares because of Aeon’s attractive shareholder perks. They’re entitled to discounts of up to 7% on purchases, with the exact rate depending on the number of shares held. Clearly, they didn’t buy the stock because it was cheap - it’s by far the most expensive supermarket company in the world.

I find it painful to write about this stock, as it has detracted significantly from the fund’s performance. It’s our largest net short position (we hold larger gross short positions, but those are paired with dual listings on other exchanges). The share price has outperformed markedly in recent months; just this month alone, it has risen by 31%, detracting approximately 0.25% from the fund’s return.

I’ve always known that shareholder perks are a big deal in Japan - it’s one of the key Asian market idiosyncrasies I documented in the following research note: https://www.oqfundsmanagement.com/post/asian-stock-market-idiosyncrasies. Still, I couldn’t resist the temptation to short a low-growth company burdened with significant debt that remains inordinately expensive by every valuation metric we monitor, except price-to-sales. It was a mistake, and I plan to reduce this short position when the timing is appropriate (hopefully when the share price is lower).

AH Spreads

AH spreads vary widely over time. There was an expectation that dual A‑ and H‑share listings would trade close to parity following the launch of Stock Connect in November 2014. Indeed, when Stock Connect was officially confirmed, the spread between A‑ and H‑shares narrowed sharply, only to widen again as investors realised that separate liquidity pools and the absence of fungibility would result in persistent pricing discrepancies.

The aggregate A‑H spread is largely driven by investor sentiment. When overseas investors turned massively bearish on Chinese equities in 2024 - due to irrational concerns over investability - the spread spiked to an all‑time high. More recently, it has narrowed significantly and now sits well below its long‑term average. Chart 2 shows the Hang Seng China A‑H Premium Index since the implementation of Stock Connect in 2014.

Chart 2: Hang Seng China A‑H Premium Index since 2014

The fund maintained a net short position in A‑shares and a net long position in H‑shares when the A‑H premium was near its peak, but we have since reversed this positioning. This illustrates how we seek to exploit the impact that liquidity flows and investor sentiment have on relative share prices.

What’s the best investment strategy in this market environment?

When you're betting on long-term pricing relationships normalizing and there’s no clear catalyst, breadth helps a lot. We have over 1,300 stocks in the portfolio and are targeting a vast array of mispricing opportunities.

If the objective is to generate strong risk adjusted returns, breadth is also important in terms of trading frequency. Consider a hypothetical and stylised example where Stock A looks undervalued relative to Stock B. Rather than waiting for entry and exit criteria before trading, it makes more sense to trade constantly based on relatively valuation shifts.

The above example makes sense when targeting stock pairs (such as A‑H pairs). However, most of our positions are based on factor scores, where each stock is evaluated relative to others with similar risk characteristics. Nevertheless, the same concept applies - it is appropriate to regularly fine‑tune position sizes.

Consider another, more realistic, stylized example. You could build a long portfolio consisting of stocks that appear relatively cheap, have strong sentiment scores, and exhibit positive momentum, while constructing a short portfolio of stocks with the opposite characteristics. You could then rebalance this portfolio based on a set rebalance schedule (eg weekly) with trade lists determined before the start of trading.

This approach is flawed. It will likely generate alpha but there will be drawdowns. Quant factors don’t always generate alpha and uncorrelated alpha sources aren’t always uncorrelated. Indeed, we’re currently experience such an environment based on a story I’ve just read on Bloomberg (Fast-Money Quants Stumble as Momentum Bust Roils Strategies).

A more effective approach is to maintain core positions based on quantitative factor scores and adjust them regularly in response to short-term signals. These signals include real-time, intraday updates of technical momentum measures and short-term sentiment factors - such as earnings revisions - that have strong but rapidly decaying predictive power.

We incorporate numerous rebalance screens into our investment process, which are triggered intraday and allow us to continuously fine-tune position sizes. These screens target both stat-arb and pair alpha generation opportunities.

The following chart shows the average daily trading frequency.

Chart 3: 6-month Moving Average of Number of Daily Trades

Trading frequency has increased markedly as we have gained a greater appreciation of the alpha that can generated by maintaining high trading breadth.

There is one important caveat: it’s essential to trade based on information rather than noise. This is particularly relevant after significant “news events” such as an earnings surprise or a profit alert. The market reacts immediately to such events, but analysts take some time to update their estimates. During this interim period, we lack robust factor scores to guide trading decisions.

Minimising this lag is crucial. To address it, we calculate a Flash Consensus based on the most recently updated analyst estimates. This process is not as straightforward as simply recalculating the consensus using only the updated estimates. Instead, we use all available analyst estimates and employ an algorithm to adjust those that haven’t yet been revised. By using this approach, we reduce the period during which we lack reliable estimates for our quantitative factors to approximately two to three trading days following news events.

In simple terms, rather than targeting mispricing opportunities solely based on value, sentiment, momentum, and other quantitative factor scores, we also pursue stat-arb and pair alpha opportunities through our rebalance process, which incorporates technical and short-term sentiment indicators. We believe this approach enables us to generate more consistent alpha than a traditional multi-factor quantitative investment process would. This is particularly important in the current market environment where we’re witnessing some extreme liquidity driven price distortions that are adversely impacting the predictive power of quant factors.

Based on your discretionary overlay, where do you see the best short term mispricing opportunities?

There are numerous attractive long-term investment opportunities. For example, at some point in the future - it’s hard to say when - SoftBank will likely trade at a material discount to its NAV, and TSMC’s ADR price will be more in line with its listing in Taiwan.

Markets are likely to normalize once the AI euphoria fades. This process could take a couple of years - or even longer - which is problematic given the fund’s investors, myself included, have limited tolerance for drawdowns.

It’s important to also target opportunities that are likely to be alpha-accretive in the short term and that don’t rely on prevailing AI sentiment. We believe there are A/H-share mispricings across our universe of Chinese banks that are unlikely to persist and therefore represent an attractive source of alpha.

We analyse A-H mispricings based on cross-sectional and time series analysis.

A-H Commercial Banks

China Merchants Bank’s H‑share (3968 HK) currently trades at nearly a 9% premium to its A‑share (600036 CH). This is the second‑highest H‑share premium after CATL (3750 HK), which, as we discussed in our August newsletter, is an unusual outlier.

The current H-share premium is also high based on our time series analysis. It is more than 2 standard deviations above the 3-month average and more than 1.5 standard deviations above the 12-month average.

At the other end of the spectrum, consider Agricultural Bank of China. Its H-share (1288 HK) is trading at more than a 30% discount to its A-share (601288 CH). The H-share discount is more than 0.5 standard deviations above the 3-month average and more than 1.5 standard deviations above the 12-month average.

For both these banks, the A‑shares and H‑shares are highly liquid and offer the same dividend and voting rights. We don’t believe there is any sound justification for the current mispricing. We can understand why broad‑based discrepancies can exist between A‑share and H‑share valuations, given the differing liquidity pools and the susceptibility of H‑shares to negative Chinese sentiment. We can also understand why some H‑shares trade at substantial discounts due to relatively low liquidity. However, in this case, we’re dealing with liquid stocks in the same sector, and there is no reason why this pricing anomaly should persist.

Several other Chinese commercial banks also display unusual A‑/H‑share pricing anomalies. In particular, the A-share listings for Bank of China (601988 CH) and ICBC (601398 CH) look relatively expensive compared to their H-share listings (3988 HK and 1398 HK). The A-shares for both these banks also have relatively low borrow rates. We have significantly increased our gross exposure to these names while keeping our net exposure close to 0%. It’s one of the most attractive investment opportunities I’ve seen for a long time.

A-H Investment Banks

For several years, the A‑shares of investment banks have traded at a substantial discount to their H‑share listings. They are high‑beta companies, and overseas investors have generally been cautious about gaining exposure to them.

Recently, however, we’ve observed an interesting development. CITIC Securities’ H‑share (6030 HK) is currently trading at less than a 3% discount to its A‑share (600030 CH). This represents a difference of more than two standard deviations from both the three‑month and twelve‑month averages. Across all the time periods we monitor, the z‑score is more than 1.5 standard deviations from the mean.

Conversely, China International Capital’s H-share (3908 HK) trades at more than a 45% discount to its A-share listing (601995 CH). Both the A-share and H-share are extremely liquid, and the A-share has an exceptionally low borrow rate of 10 basis points per annum.

We use quantitative tools to cast a wide net and identify the most promising alpha generation opportunities in Asia. These opportunities are then evaluated through the lens of our investment experience and expertise. When our quantitative analysis aligns with our discretionary view—and sufficient liquidity is present—we believe it’s appropriate to allocate significant capital to fully exploit the opportunity set. This approach underpins our current net long and net short positioning in Chinese banks.