The Chinese interference in the A-Share market has been intentionally broken by Chinese regulators for months through a series of direct and indirect interventions. Many of the interventions will just play out in the inefficient allocation of capital – the lockups on institutional shareholders, the rigorous use of debt to finance equity purchases, etc.
But for U.S. investors actively investing in China, there's a much more fundamental problem: the trading locks. China has simply – seemingly almost at random – closed various A-shares for trading. As a U.S. investor not actually on the ground in mainland China, this leads to significant information gaps. Nowhere is this more evident than in the Deutsche X-trackers Harvest CSI 500 China-A Shares Small Cap ETF (ASHS).
To start with, let me say that I love that this fund exists as an option for the adventurous international investor. It directly holds the 500 smallest cap stocks listed in mainland China, which turn out to be not-that-small: the weighted average market cap is over $3 billion, and most of the portfolio would firmly be considered mid-cap by U.S. standards. In other words, despite the name, these are not fly-by-night penny stocks.
However, to halt the decline in the local markets, China has locked a host of these firms – which are current holdings of ASHS - out of the market – in some cases since 2014. The last recorded trade in Ningxia Zhongyin Cashmere Co., Ltd. Class A (000982-CN) was August 22, 2014 – 11 months ago. More typically, a huge swath of A-shares were suspended in from trading back in May, leading to over a hundred charts from the portfolio that look like this, top ASHS holding Eternal Asia Supply Chain Management (002183-CN).
And yet these stocks remain in of both the Harvest CSI 500 Index and ASHS.
This completely breaks any ability for investors to really understand the "fair" price of their holdings in ASHS.
Breaking It Down
First, let's take a look at what's happened with ASHS here in the U.S. over the last month.
The first point I'd make is that ASHS premiums and volumes have held up surprisingly well here. Volumes, while not stellar, are averaging well over 100,000 shares a day, making it entirely tradable for most investors. And in general, the price you pay for the ETF in the market has actually born a strong resemblance to what's being reported as Net Asset Value. Most of the time you would expect – even in a normal market – that an international ETF actually wouldn't track its own NAV all that well, because of the difference in time zones.
ASHS, however, is one of a class of ETFs that use Fair Market Value practices in calculating NAVs. From its prospectus: "If a security's market price is not readily available or does not otherwise accurately reflect the fair value of the security, the security will be valued by another method that the Adviser believes will better reflect fair value in accordance with the Trust's valuation policies and procedures approved by the Board."
That means that even though some of these stocks haven't traded in almost a year, the fund calculates a "best guess" based on comparable stocks and indices that are trading. This is a very good thing, generally speaking (and how the entire bond market deals with pricing as well). It's a very, very good thing when it comes to a crazy, half-closed, less liquid market like China A-Share Small caps. So what happened in the middle of July, when the traded price of the ETF differed significantly from the fair-valued NAV reported by ASHS?
Two things. First, the premium coincided with the maximum number of securities in the portfolio being closed.
On July 9, less than half of the portfolio in ASHS measured by weight was actually trading. That made the NAV being reported particularly hard for anyone other than the funds' accounts to be sure of, which makes it particularly hard for Authorized Participants to step in and arbitrage out any differences.
The second point is that this was going on precisely when the shorts were having a field day. Using Markit's short data via FactSet, we see the spike in premium and collapse in underlying tradability coincided with the height of short selling.
The story here – maybe – is that short sellers saw the "irrational" high price as the trading price of ASHS bounced from $37 on the 8th to $53 on July 10 and pounced to take advantage. Of course, to put on that big of a short, you have to be able to borrow it, and the evidence would suggest that the APs were more than happy to step in, create new shares, and loan them to the shorts for a pretty penny (ASHS has the most expensive rating for borrowing costs by Markit).
And once the trade started paying off and the premium collapsed, the shorts left, and APs redeemed those shares back into Deutsche, shown here happening on July 13.
The Moral of the Story
I acknowledge that forensics like this are imprecise. Both short-reporting and flows-reporting can be subject to lags, which are often impossible to tease out. My suspicion here is that the flows numbers we're calculating are lagged by one day, which is actually the norm for flows reporting. That would put the creations happening more cleanly with the highest shorting and the redemptions happening right when the short interest collapsed.
But importantly, at a micro scale, these kinds of transactions happen all the time. ETFs trade to slight premiums and discounts; short-sellers and APs look for short term opportunities, and ETFs' trading prices snap back to fair value.
The reason these charts look so dramatic is because the underlying market is broken, so moves and gaps that normally happen in pennies in basis points happened here in dollars and percentages.
And the coolest part is that this is evidence of the ETF structure doing exactly, precisely, and beautifully what it's intended to do. Just as the underlying market went to pot, the ETF became a price discovery vehicle. Once that happened, a whole raft of market participants stepped in to actually contribute to that process of price discovery – the short sellers were part of it, the APs were part of it, and yes, the enthusiastic buyers who drove the price of ASHS from $37 to $53 were part of it too.
Once the underlying market became, once again, at least slightly more liquid, the ETF snapped right back in line to advertised fair value.