A Strange New Bubble In Chinese A-Shares (A Is For Arbitrage)
Chinese stock markets have broken one of the basic “Laws of Finance” (as reported this week by the Wall Street Journal).
Or – it might be that the markets are demonstrating, quite dramatically, that the “Law” in question – the No-Arbitrage Principle, described as “the fundamental principle underlying much of financial engineering” and “a gold standard in quantitative finance” – simply doesn’t hold up here.
In any case, the situation we will describe below is deeply puzzling, challenging both formal theory and common sense. There is evidence of a deep dislocation within the Chinese financial system, which may portend market violence at some point in the future.
Background on the Chinese Equities Markets
China has 3 stock exchanges – “onshore” in Shenzhen and Shanghai, and “offshore” in Hong Kong. Or 4, if you count the Chinese companies listed or cross-listed in New York. Or 4+ if you include the new STAR exchange for high-tech Chinese companies (part of the Shanghai exchange, and the focus of the WSJ article).
Shares of Chinese companies are quoted in at least three different currencies: Chinese renminbi (¥), US Dollars ($), and Hong Kong Dollars (HK$).
There are many types of shares, including:
- A-shares – shares of Chinese companies trading in Shenzhen and Shanghai, quoted in ¥.
- B-shares, shares of Chinese companies which trade in Shenzhen and Shanghai, and quoted in foreign currency (US$ or HK$)
- H-shares, shares of mainland companies that trade in Hong Kong, quoted in HK$
- N-shares, shares of Chinese firms listed in New York, denominated of course in US$
(To keep things simple I am leaving out several other share types and trading venues.)
The A-shares dominate in terms of market share. There are over 3500 companies listing A-shares, compared to less than 100 that trade B-shares. Several hundred companies trade H-shares in Hong Kong, and a similar number trade in New York.
Dual A/H Listings
About 100 Chinese companies are dual-listed, trading on both the mainland and the Hong Kong exchanges. Their shares trade as A-shares in Shanghai or Shenzhen, and as H-shares in Hong Kong. Here an intriguing discrepancy arises. A-share valuations and H-share valuations should be the same. And for a while they were. From 2010 until 2015, H-shares traded more or less in line with the A-shares. But since 2015, the A-shares have carried a very large premium on the mainland exchanges, compared to Hong Kong.
This has been called the A/H Anomaly, a profit opportunity similar to the Value Anomaly or the Small-Cap (Size) Anomaly – a persistent mispricing that shouldn’t exist, or persist, but does, and is therefore exploitable for profit, in principle. It has been widely studied, and there is now a quoted index that tracks the A/H premium.
But this A/H premium is more that just a mispricing. A-shares and H-shares, for dual-listed companies, reference the same underlying economic assets. They are equivalent claims on the same cash flows. They should be priced the same, more or less (perhaps allowing from some currency frictions between the ¥ and the HK$). But as the chart shows, A-shares are almost 40% more expensive than the equivalent H-shares today.
What Does the A/H Premium Mean? (Multiple Non-functional Explanations)
This raises a number of questions. For one thing, it exposes the major “regime shift” in Chinese equity markets which took place rather abruptly in 2014. Whether the current situation represents an over-valuation of the domestic shares, or a penalty for Hong Kong shares, we cannot say with certainty. Several explanations have been proposed. All are problematic.
The Impact of “Stock Connect”
One theory is that it was the introduction of more liberal investment regulations in late 2014, the “Stock Connect” program that linked Shanghai/Shenzhen and Hong Kong and allowed easier access for Chinese investors to trade in HK, and for foreign investors to participate in the Shanghai A-shares market. This was a major Chinese initiative to move towards more liberal and mature financial markets.
Here is how one study described this significance of this change:
- “Before 2014, some companies in China opted to list A-shares on Shanghai stock exchanges (SSE) while at the same time, listing them on the Hong Kong stock exchange (SEHK limited) as H-shares. However, these markets were segmented for A-shares, and only domestic Chinese investors could trade A-shares on the SSE. Meanwhile, the access of overseas and HK investors to A-shares was very limited. Although Qualified Foreign Institutional Investors (QFII) have been allowed to trade A-shares since 2002, and Qualified Domestic Institutional Investors (QDIIs) have been allowed to trade H-shares since 2006, applied restrictions like investor minimum capital requirements and daily or periodic trading quotas still remained very tight. To promote the connection between the two markets, the Chinese government announced a new policy that launched on November 17th, 2014. Through this new policy, Hong Kong and overseas investors were allowed to trade eligible SSE listed A-shares through the Shanghai-Hong Kong (SH-HK) connect.”
The timing of the “regime shift” corresponds almost exactly to the initiation of the Stock Connect program. But — the effect should have been the opposite. Linking the two markets should have synchronized prices on both markets – instead, it put them out of joint.
- “As investors from both China and Hong Kong were permitted to trade on both markets at the same time. It was thought this might lead both markets to a price convergence effect. The Hang Seng AH premium (HSAHP) index should have converged to 100 [i.e., to “no difference”] according to the standard present-value asset pricing theory, but instead it diverged and arrived at almost 140 by September 2015. Then, it fluctuated between 120 and 140.”
The General Regulatory Environment
Another factor is the overall quality of financial regulation in a particular market. The conventional view, supported by a huge amount of academic research, would say that markets with stronger regulation have higher valuations than less regulated markets. This is the principal explanation for the “cross-listing premium” – the higher valuations that Chinese firms receive when they list on U.S. changes – the subject of two previous columns. The phenomenon is referred to as “bonding” – a company listed on an exchange with stronger overall regulation benefits from the greater credibility of the general market environment, and attracts a higher valuation. Hong Kong exchanges observe higher standards than Shenzhen and Shanghai, with better accounting and more transparent financial information, and are preferred by knowledgeable (foreign) investors. Accordingly, H-shares should carry the premium – but obviously they do not.
Foreign ownership of A-shares is only about 3%, whereas over 30% of H-shares are foreign-owned. Again, based on past research, one would expect to see higher valuations in markets that attract more foreign investors, but the A/H premium does not conform.
A-shares trading is mainly “retail” — that is, carried out by individual Chinese investors; H-share ownership is mainly institutional; the academic literature has tended to support the idea that higher institutional ownership should lead to higher valuations — but apparently not in this case.
A recent article suggests several other factors which could cause the A/H premium, such as liquidity differences, investor styles (e.g., different risk preferences – “In mainland equity market, investors, as mentioned before, prefer short-term speculative behaviors and tend to hold a positive view on future returns”), “different demand elasticities” (whatever that means), exchange rate effects, differences in trading systems in Shanghai and HK (rather minor, for the most part), and sector weightings (the H-shares market is dominated by financial companies, for example). Yet, none of these factors appears to be decisive and many should have the opposite effect (i.e., they should support a premium for H-shares, not A-shares).
The bottom line is that “domestic investors pay a much higher price than foreign investors for an identical asset.” It is not clear why this is happening.
The No-Arbitrage Principle, Rudely Violated
Meanwhile, there is a larger issue to deal with: the failure of financial theory.
The terminology is a bit exotic (it’s always fun to say the word “arbitrage”) but the idea is simple: it should be impossible in a developed market for the same asset to sell at two different prices. If gold is selling for $1000 an ounce in New York and $1500 an ounce in Philadelphia, smart investors will buy gold in New York and sell it in Philly – easy money, risk-free – this is “arbitrage.” Except that – poof! – the profit opportunity quickly vanishes. Alert investors will see the price discrepancy and pounce on it. A surge in gold purchases in New York will drive up the New York price. Selling gold in Philadelphia will drive down the price there. The two prices converge, and the risk-free trade disappears. The market, in short, will quickly find the equilibrium price that balances supply and demand, and it will be one price, everywhere. Thus, there is no true risk-free arbitrage opportunity in an efficient market system. This is the No-Arbitrage principle, also called the Law of One Price.
It can be stated even more simply, such that it seems mere common sense:
- “Imagine two prices being demanded in one market for exactly the same good—who but a fool would pay the higher price? Therefore if both sellers are to be able to sell they must charge exactly the same price. Hence, the law of one price.”
The No-Arbitrage rule is central to the academic view of how markets work. Stephen Ross, one of the giants of orthodox Finance Theory, called it “the fundamental theorem of finance.” He said that “the basic intuition that underlies valuation is the absence of arbitrage.”
NA is the conceptual foundation of quantitative finance, especially. It is routinely recited in the textbooks, and appears in the introductory paragraphs of countless academic papers – e.g.,
The Reality: The “Yes-Arbitrage” Law
Now, in fact, the pure no-arbitrage situation is never actually found in real markets. “Yes-arbitrage” exists on every stock exchange, all the time. There are always at least two prices for a share of stock – the Bid and the Ask, the buyer’s offering price and the seller’s asking price. As I write this, the Bid price for a share of Apple’s common stock is $501.71, and the Ask is $501.92. (A relatively large difference – but then it is Saturday morning and the markets are closed. Most of the time the Apple spread is just one penny.) This 2-price framework is not merely a technicality; this Spring the average spread of the S&P 500 surged to more than 20 basis points. Market makers (and many quantitative hedge funds) make their living by buying at the Bid and selling at the Ask (so to speak – yes it is a bit more complicated than that, but a few basis points is plenty of room for algorithmic traders and fast market makers to operate in.) “Yes-arbitrage” at this micro-level supports a highly profitable trading strategy. Harvesting these arbitrage profits is a sophisticated and lucrative business, worth tens of billions of dollars a year. But because the Bid/Ask differences are very minor – less than .001% in Apple’s case – the NA rule may still seem valid as a general principle, with “YA” for market makers as a special case.
However, the more important conceptual application of NA in Finance theory is to (incorrectly) deny the existence, or at least the persistence, of mispricings in the market.
A mispricing refers to a situation where there are also, in effect, two prices: the market price, and the “true” price. The latter is based on the (unobservable) “intrinsic value” of the company, which may be different from the market value. The argument of orthodox Finance theorists – repeated often, and evidently with conviction – is that true mispricings cannot exist, because if they did the professional investors would arbitrage them away, buying (say) if the market price were below the “true price” — which would drive up the market price until it matched the true price, and the arbitrage opportunity disappeared. Thus, in this view, the market price is the correct price, the true price. According to orthodox theory, there is no open profit opportunity, no arbitrage. “You can’t beat the market.” This idea underlies probably half (at least) of the investment programs in the market today.
However, important violations of this principle – i.e., persistent mispricings – do exist, and that is what the other half of the investment world tries to exploit. The study of market anomalies it a vast subject.
But the A/H anomaly is special. It is a particularly egregious violation of NA, for this reason: whether “normal” or “traditional” mispricings (like the “value anomaly”) are truly violations of NA can always be argued, because the “true price” is not observable. The 2-price framework applied to these anomalies compares a known valuation (the market price) with an assumed valuation (the true price). We can always contest the assumption. But with the A/H premium, we are confronted with two actual, observable, executable prices for the same thing (or nearly the same thing). The existence of more than one price for the same asset is not a matter of assumption or interpretation. It is an Inconvenient Fact.
So why don’t investors simply buy in Hong Kong and sell in Shanghai? Yes, there are trading frictions, political risks, and institutional obstacles which might impede the free functioning of the arbitrage operation between these markets. But not enough to explain a long-lasting A-shares premium of 30-40% – especially given all the issues discussed above which ought to support a premium for H-shares, not A-shares. It is especially significant (in my view) that the A-share premium exploded after the creation of the Stock Connect program, which was designed precisely to “expand market access” and (in the words of the Chinese goverment) “to promote two-way opening-up and healthy development of the capital market on the mainland and Hong Kong.” In short, the whole point of Stock Connect was to synchronize these markets by allowing a freer flow between them. It seems to have had the opposite effect.
The End of No-Arbitrage
The No-Arbitrage principle is one of those zombie concepts in Finance that should have been laid to rest long ago. We know that markets don’t always eliminate mispricings. But the continuing existence of a “naked” violation of the One Price Law in China is ominous.
Still, even in its mystery, the A/H premium offers insight, and a warning. It exposes a severe misalignment of some of the world’s largest equity markets. Some of these prices are wrong, clearly, but which ones? If the A-shares are overvalued, it may be a symptom of speculative delirium, and a dangerous “bubble.” Or, it may be that the H-shares – heavily skewed towards firms in the threatened banking sector – are depressed and discounted due to the political risk surrounding the future of Hong Kong as a financial center. There are many moving parts, as Beijing proceeds with its takeover of HK, as foreign investors respond and hedge, and as the Chinese domestic markets adapt to extraordinary government stimuli in the framework of the Covid economy and trade-war nationalism. There are of course political constraints on the flow of capital.
In sum – whatever the theory says, from a practical standpoint the A/H anomaly represents a worrisome coagulation of political and financial risk. At some point the system will unblock –and trillions of dollars in equity value will suddenly become unstable.