Market On Close (MOC) Mechanisms, Part 3: What Do Traders Need to Know About APAC MOC Facilities?

Winnie Khattar:
There is investor interest in Asia Pacific, which is improving and will continue to trend in that direction, particularly where reforms and market reforms translate into real execution confidence.
Peter Haynes:
Welcome to TD Cowan's podcast series, Bid Out: A Market Structure Perspective from North of 49. My name is Peter Haynes, and today for episode 82, we're going to dedicate our attention exclusively to the mechanics of market-on-close mechanisms in the Asia-Pacific region. This is the third of our three-part series on changes to closing auctions globally, albeit the first two episodes covering the Americas and then Europe were done last year. It took some time to find the right guests, but I can assure you it was well worth the wait. Joining me today to help educate us on APAC MOC are Winnie Khattar, Head of Market Development for BlackRock in the Asia-Pacific region based in Hong Kong; and Khashayar Sirti, Head of APAC trading for UBS Asset Management and based in Singapore. Winnie and Khash, thanks for joining the show today and making time for the discussion.
Winnie Khattar:
Amazing. Thanks for having us.
Khashayar Sirti:
Thank you very much, Peter.
Peter Haynes:
Okay, so it was the goal of the first two discussions to play like trading tourists and pretend like our audience knows a little bit about each of the markets that we talk about today and visits from time to time with order flow. But none of us really know exactly all the local conventions or the landmines that need to be avoided, especially around the end of day benchmarking, so Khash, I want to start with a general question for you. Given that the APAC region has a whole bunch of heterogeneous markets, I'm curious how you manage the trade-off between relying on local broker relationships where there may be a little bit more concern about information leakage as opposed to using global relationship broker relationships where there may not be as much local market knowledge of conventions about where the bodies lie, but more of a trust factor that information won't be leaked.
Khashayar Sirti:
Yeah, Peter, that's a very relevant question given that we have 13 to 14 markets in this region and very few brokers who operate in each and every market. There are some of them that we know of that operate in every market, and in cases where the trades are not very complicated and we would like to use only one broker for an entire basket, we would use those brokers.
However, in cases where trades are complicated, you either go to a local broker, which we rarely do for our passive business, or you can also alternatively go to brokers who we face on the trade on a contractual basis, but they execute through a local broker's pipes. In these cases, we tend to benefit from two ways. The leakage is minimized to a certain extent because the end broker certainly doesn't know who the client is. It's on a no-name basis, and therefore we manage to extract the best of both worlds where at times we can access the local liquidity without giving away too much. So it depends on the type of the trade, and it depends on the situation. We use different brokers to achieve different outcomes.
Peter Haynes:
Khash, are there any of those marketplaces where the global broker you're dealing with who is then dealing with a local broker is required to give you up as the client? Is there any of those markets where they can see through to you?
Khashayar Sirti:
There are ID markets where they end up seeing the ID. The executing broker will end up seeing the ID, not necessarily on the sales trader level, but on the back office level when you are confirming trades, in ID markets it's inevitable that your ID is disclosed to the person you're settling with.
Peter Haynes:
Okay, so I think India was one of those markets, and that's the market that I want to shift to next, and I just have a question for Winnie. So given the nature of the region being heterogeneous in terms of the marketplaces, end-of-day auctions, as you folks well know, generally come in one of three shapes. They'll either be done on a parallel basis to the continuous market, the way things work in North America. They can be standalone blind, i.e. taking place after continuous, but blind and limited in the amount of transparency. That would be some of the smaller markets in Europe. And then of course, standalone transparent, which is the larger European markets. And now I believe all of the APAC MOCs are created in this fashion, as Indonesia I think was the last one to convert to transparent. There is one market though that doesn't yet have a market-on-close mechanism in the region, and that is India. Winnie, can you explain to our listeners on the discussions that are happening with the local exchange and regulators about a new MOC facility which I'm told is coming in August?
Winnie Khattar:
India is definitely a market which is near and dear to my heart, or I'd say both our hearts. Before we dive into India, maybe I'll take two minutes to give a bit of background on why closing auctions are so important. And I must say in the last decades, it's been fascinating to see how much markets have evolved, but also along with the market, the market design products and participants. As a result, closing auctions have been a critical part of this market design evolution, be it equities, ETFs, and lately there are discussions on closing prints for bond markets as well.
So why do closing auctions actually matter? They matter not just because they set the close price, which is a very important benchmark price, be it equities or derivatives, but they also concentrate day's biggest liquidity into a single transparent benchmark option, which enables all investors, be it active investors, index investors, or retail to be able to trade in size and minimize their impact when there is liquidity maximization in an auction period.
And over the years, the design of closing auctions has evolved, so has the trust in the matching progress and price discovery. Globally, when we look at the overall volume that trades in closing auctions, it has already doubled since 2070. And it's not just indexing what I mentioned earlier, it is both active and index investors that are trading and leveraging these closing auctions due to liquidity and price discovery achieved through auction mechanism. As we see this evolution, India obviously is one of the largest EM markets, and you rightly pointed out, one of the last markets to enable closing auction. It's been great to work on India. Khash and I have both spent a lot of time over the last two years guiding the closing auction design in India.
So if we look at global closing auctions, they roughly account for about 15% of daily volume, whereas when we look at India, about 20% of daily volume actually occurs in that last 30 minutes of trading. And how close price is currently set in India is weave over the last 30 minutes of trading, which means there is higher volatility in that close price formation, and that creates challenges for both index and active investors who are participating to be able to benchmark that price accurately. Also, the volume is spread over 30 minutes versus a smaller period of time, so the element of liquidity maximization is absence of a full former auction mechanism.
So over the last two years, what we've done, we've had a number of very active discussions with the SEBI, with the exchanges, industry bodies like SFMA have been very closely involved, Q15 as well, and Indian investment community, local asset managers as well in both educating them on the role as well as what is an efficient design of a closing auction. And I would highly commend everyone for their efforts, especially SEBI, for taking the time, their thoughtful analysis in putting together the proposal and designing a closing auction that is going to be implemented in August this year.
Peter Haynes:
So Winnie, let's look at the specific design of the MOC model that's coming in India. What will it look like and how does it compare to what is in place currently?
Winnie Khattar:
The new model that is being proposed for going live in August this year, it's a 15-minute auction period that starts at 3:20 PM local time. There is five minutes for order entry period. Board market orders and limit orders are allowed during this period. There is a no cancellation period as well in the closing auction design, which lasts for another five minutes, followed by a randomized close in the last two minutes of the trading session. India's design also has five minutes of auction matching, which is after this 10 minutes auction, i.e. five minutes order entry, five-ish minutes of no cancellation plus last two minutes of randomized close. There is an order matching period which lasts from 3:30 PM to 3:35 PM, after which an official close price will be established.
Like a good auction design, there is a plus/minus 3% price limit ban as well, which limits how much prior orders can move during the closing auction period itself. We really like the idea of having both non-cancellation period as well as randomized close, which reduces risk of any kind of gaming in the auction. Any orders, any unexecuted orders from the continuous trading session will be automatically carried forward into the close auction session, which is another very good feature of the closing auction. And finally, I think what is most important in any auction, any good auction design is transparency of information so there will be market data that will be disseminated to market participants on indicative equilibrium price, the cumulative buy and sell, order quantity, the imbalance quantity that really helps investors decide how to continue pushing new orders and modify their orders into the closing auction and really supports that price discovery process.
Peter Haynes:
You mentioned in your answer randomized close. Do all the markets that we're going to talk about today on the APAC discussion, do they all have randomized closes, or is this becoming more popular?
Winnie Khattar:
Randomized close actually is becoming more popular. So India's auction design will have a randomized close. I know the Australian auction design also has a randomized close in the auction period. Japan is planning to introduce a randomized close into their closing auction logic. Japan actually launched their closing auction only in 2024, so them actually introducing randomized close is big change that will take place and comes after many months of advocacy across the industry on introducing an anti-gaming feature. There are however markets that do not employ randomized close in Asia. For example, markets like China doesn't have randomized close. Korea doesn't have a randomized close either.
Peter Haynes:
When you're having discussions with regulators in the countries that do not currently employ that, are they receptive to thinking about the importance of the anti-gaming feature of a randomized close?
Winnie Khattar:
Yeah, definitely. I think the policymakers exchanges are open to improving the design of closing auction overall and randomized closed anti-gaming feature is definitely a positive. There are other features as we discussed through this. We will discuss through the design of various different markets. I'm happy to point out there are areas where improvements can be made to the design of closing auction that may improve the flexibility on how liquidity and how much liquidity matches on the closing auctions. If you look at it from a principles approach, I think an important principle, the first most important principle of any closing auction design is attracting liquidity and that providing that price point that matches and maximizes liquidity.
Second important feature is price discovery. Then an important feature is making sure that when there is sufficient liquidity matching, there is a design that supports price discovery, allows participants to send and price their orders efficiently. Then how do we protect from them being gamed?
So comes the third design feature of anti-gaming logic. And then the finally, also thinking about how any unexecuted orders are actually being executed or they're being like post-trade sessions that allows unexecuted orders to be trade in a post-closed session. So varying degrees of engagements on different auction designs depending on where they are and where they need to be evolved in order to improve the overall design of auction across these four principles.
Peter Haynes:
So Winnie, I want to stay with you for a sec here and move on to China, which is obviously a very important market in terms of the global landscape. From time to time, we read about regulatory concerns about closing price volatility in China and in some cases have read about certain participants being told they can't trade at the close. That hasn't happened lately, but it was maybe a couple of years ago. Can you tell our listeners, are there any limitations on participant closing activity today? And can you explain how the order entry, and non-cancel period works for the auction model in China?
Winnie Khattar:
China is a unique market in history. In the past, there have been instances of window guidances in China market overall. What we do see is very high level of surveillance from the exchanges in the closing auction itself. The close in China is designed to dampen volatility and reduce the ability to gain the closing print by canceling orders just before the price is established. So during the order entry window, which is the first two minutes, participants can submit, amend, and cancel orders. And then the non-cancel window, which is the next one minute, new orders can still be accepted under the rules, but cancellations are restricted so that there is indicative margin stabilization to establish a reliable closing price. This one-minute period tends to be longer in other auction designs, and is really to attract and can attract new liquidity as price becomes more stable after the first two-minute order entity period.
In practice, that means you need to be comfortable or investors need to be comfortable in sending their risk earlier on, and use indicative price imbalance signals to appropriately size rather than relying on last second or last minute cancel or replace tactics, which is why there is that non-cancellation period in that one minute. So we like this is a common design feature of good closing auction mechanisms to have a non-cancel period.
Peter Haynes:
Just to be clear, is the auction a two-minute order entry and then a one-minute no cancel, that's the entire length of it?
Winnie Khattar:
Order entry is two minutes, and there's one minute non-cancel period. Yes, that's correct. It's a three-minute auction.
Peter Haynes:
And there's a rule that says you can't put a limit order in more than plus or minus 2%, but given that every stock has a different level of volatility, and obviously there's a lot of volatile names in China, does that ever affect your ability to manage your on-close risk?
Winnie Khattar:
I think that definitely limits the overall price move. Again, what I started with, the goal of China's closing auction has been designed to dampen volatility in the auction, so the price bans are relatively tight compared to other auctions in the market. As a result, what you do see is closing auction only accounts for 1% to 2% of days volume on normal days and two to 3% of volume on rebalance days. Participants have to limit participation rate to avoid unintentionally impacting the price due to this plus/minus 2% versus last traded price ban. And as a result, what you do end up seeing is market participants end up trading earlier in the day, especially on event days to manage their risk of non-completion.
Peter Haynes:
I've never traded in that market, don't know anything specific, follow it from a distance. I was unaware of the limited amount of volume that's happening in the close. Given there's so much activity in the Chinese market that's retail in nature, do you ever have a worry that there won't be enough offsetting liquidity when you have institutional size volume due to some form of rebalance or name being added to the index?
Winnie Khattar:
Given the volume of retail in China market, it is pretty significant. I think roughly 50% to 60% of China market is retail, so that hasn't been the case, but I think that's something definitely to keep in mind as overall institution volume and trading liquidity continues to expand. Overall liquidity in China market is actually very, very large.
Peter Haynes:
At some point, we're going to be revisiting the percentage of A shares in the index. I'm sure it's not a front topic right now, but I'm sure in the next few years it will be again, and that'll mean even more institutional activity in the local market.
Khash, let me ask you a question about China just as a follow-up here. When you're trading in the Chinese market, again, one of these many heterogeneous markets, do you have one rule of thumb that you might provide as a general rule of thumb to foreigners that may be trading in that market that they should always be thinking about?
Khashayar Sirti:
In terms of MOC, yes. Like Winnie mentioned, we also deployed the execution strategy of completing the order, or not completing the order necessarily, but certainly starting the order before the close. This is to ensure that we have 100% completion on the order. Just like you mentioned, China's a highly retailed market. Plus, there are rules, window guidance around how much you should participate in the close. You can be more than a certain percentage of volume if you're not impacting, but if you are impacting, then you have to scale the participation down, which means that you could be left with a portion incomplete. There are small occasions, certain occasions where we don't have completion on our orders, but over time we refined our execution strategy at UBS AM to ensure that we achieve completion. And the numbers that we need shared, 1% to 2%, on regular days at 2% to 3% on rebound days.
Sometimes, some rebound days are slightly higher. We go with that guidance and we ensure that we achieve completion. The other important thing to note in China is with India, you cannot cross. So even if you are aware or even if you have the other side of the floor yourself, you can't cross, so you have to make sure that your order goes into the market and achieves 100% completion.
Peter Haynes:
So in other words, a broker has to expose one side of the order for a certain period of time and then match that order after giving a certain amount of time for others to participate. Is that how a "cross" would take place?
Khashayar Sirti:
You can't cross. So one side of your trade could get done and the other side potentially may not get done if the market moves too much. Yeah.
Peter Haynes:
Yeah. No, I get it. Okay, that's good to know. So Khash, let's move on to Taiwan. It's a good trivia question to name the only stock outside of the United States that's actually top 10 in market cap rankings around the world. And as you well know, it is a stock that is in Taiwan, Taiwan Semiconductor, TSMC. And what is tricky, and we just talked about some of the limitations that you're dealing with in China, what is tricky as I understand it is that you have so much exposure in Taiwan to what is otherwise a volatile sector and that's semis, semiconductors, and yet there is a rule in the local market that no stock can move more than plus or minus 10% per day, including in the MOC. How often does this limit impact your ability to manage risk during the auction? And can you also explain why Taiwan has the shortest regular trading hours in the region at four and a half?
Khashayar Sirti:
Do you want to add another trivia question? Let me ask you, what's the lot size of TSMC, if you wanted to buy one lot of TSMC? And that's the important thing that you have not discussed yet, but I'll come to that later. It's $75,000. So assuming you wanted to buy one lot, which is the one minimum size you can't buy, it has to be $75,000. There are times when you have flows in accounts where you don't end up buying a whole lot of TSMC, and trading those at the close in fact is even harder.
But outside of that, yes, Taiwan has a 10% limit on the day in stocks. In TSMC, that's rarely hit. So when you're trading TSMC, we've almost always managed to complete, but there are other names which are more exciting and more volatile. And especially around rebalance days, we are usually left with a handful, I would say between five and 10 names at times which are not completed because of the stocks closing it up or a low limit against us, which is if you're selling, it's closing down, we can't sell, and we have to carry the order forward to the next day.
So it does happen. It is one of the other downsides of trading in these markets, whereas China and Korea have window guide, India has no auction, et cetera. In fact, India also has tight limit prices in certain cases. There is this element of Taiwan which is hard, especially the odd lot market where the odd lot closing price is nowhere close to the actual underlying stocks closing price. There are many times when we have odd lots being executed really far from the market. So that's the other problem that I've faced recently, especially since Taiwan has moved to the new odd lot execution.
With respect to hours, to be honest, Peter, I don't think extending hours increases market volume. You can keep markets open for 24 hours. Yeah, maybe there'll be some people trading, but I think the four and a half hours of Taiwan has never been a problem for me. In fact, in one way, I like the fact that Asian markets have different hours and closing times because it staggers our MOC, especially on big days starting from Australia all the way to India. We have markets closing at different times allowing us to pay attention to each and every market as it comes up for close.
So it's obviously coincidence that Taiwan closes at a unique time when no other market closes, but that gives me 30 minutes to focus on Taiwan solely. I don't know why it's that short, and there used to be a time when [inaudible 00:20:44] was shorter, they've extended it, but their market still trades $100 million and they can extend it even further and it's not going to change much. So I don't think the hours drives the market volumes and Korea is a perfect example. They extended the markets by 30 minutes, but the volumes have gone up four times in the last two years.
Peter Haynes:
I don't want you to get the impression, Khash, that I'm sitting here wanting longer trading hours. I am 100% in the same camp that you are, that going to 24-hour trading is not going to be helpful for the institutional community. But I will say that as we do move towards a 24-hour trading environment, which we're going to see in North America here at the end of the year, I'm curious if there is much off market hours activity in TSMC given how important it is as a benchmark stock around the world. And how important is the activity that happens in the DR, the depository receipt, on the New York Exchange for price discovery in TSMC? Does it matter at all?
Khashayar Sirti:
TSM has a big liquid DR in the US, so effectively there is a small period of time when there's no trading happening in TSMC, and after that you have the US market. So effectively, we are covered for 15, 16 hours of the day. And it is quite important, absolutely. Every morning, every broker note mentions how much the USDRs in every stock that is available, how they traded, et cetera, so it's a nice guidance. However, I've seen over the last, I would say, post-COVID years, the Asian markets are not necessarily taking cues from the US markets. The reliance on the DR to indicate as to whether stock could trade has been lower and lower time. And I believe that's primarily because of the increase in retail participation and the whole buy-the-dip mentality that we've seen in the last seven, eight years has driven a change in the way the markets work.
Because institutions will look at the US say, "Okay, it's down there, it's going to open down there. I'm going to wait off on a bit." And people think of it that way. Whereas the retail moves in ahead, it's totally different. And the correlation over time has dropped in terms of how the stock would trade versus how the DR closed.
Winnie Khattar:
I would actually agree. The overnight market, and 24 hours trading, I think it creates... I mean, there's a lot of demand driven by on the back of retail wanting to participating in ATS's example, Blue Ocean front and center and others as well. What we do want to be cognizant of, it definitely gives a price point for people to react when there are big event news. But on a day-to-day basis, it is primarily retail-driven activity that drives the volume that we see in overnight trading. And to Khash's point, it is very important when exchanges look at trading hours to see what is the local liquidity here and what extending trading hours would extend local liquidity. Because the retail, for example, in Taiwan can participate and does participate in the four and a half hours of trading. So extending trading hours for Taiwan wouldn't necessarily increase the volume that's trading in Taiwan because the demand from the investors for overnight trading is looking at overnight US exposures and US stocks that they want access to.
Taiwan is also unique, and I was going to add to Khash's point on when you mentioned volatility and when the stocks went limit up earlier this year during the Chinese New Year on the back of tech-driven rally. Another unique thing about Taiwan is Taiwan has short-sale quotas, which is about 30% of stocks ADB, which means that short-selling quantity is limited and then that reduces amount of quanta liquidity, counter liquidity that can be available during a limit of event. Because once that short-sale quota and capacity is exhausted, brokers and market makers cannot keep supplying as much sell-side liquidity for shorting. So very much to do what you were saying while these controls are designed to protect markets and investors, in such situations they can have the unintended consequence of limiting investor participation and price discovery.
Peter Haynes:
Yeah, for sure. We saw that in Great Financial Crisis when there were short sale limitations put in place in financials around the world, and we're going to talk about the rules in Korea in a minute here.
Just one final question, Khash, before we leave Taiwan. If you're ever working an order on TSMC, would you ever consider extending the working of that order to the DR line, or would you always stay local and stay in the local trading hours for any of your demand, supply needs in that name?
Khashayar Sirti:
Oh, we would stay in the local hours for more than one reason. First of all, for passive investing, we just track the benchmark, the stock that's in the benchmark. And if it's 2330, then that's the one that we use. Even though they may be fungible, even though they would track down to each other's price, we still use the local line for that. For the other reason being if you're long 2330, you'll have to borrow TSMC to shorten the deal and then converge that. So you just can't start selling in Asia and then continue selling the TSM line without owning it. Then you have to change the order type, et cetera. And that's not something that we do. However, there are some PMs and some funds who probably sit in the US and own TSM, and then they'll just keep trading TSM because they don't want to open an ID market either.
And then there are some cases where we just go and buy TSM if you need to in the night if there's some news, but rarely do we have an order that goes from Asia hours to the US hours and back to Asia because first of all, it's not even the same instrument, and that's not how we execute stuff. But we own both names in certain portfolios, and some portfolios we only own 2330 TT and some portfolios we only own TSM US.
Peter Haynes:
It will be interesting to see if a move towards 24-hour trading, the marketplaces can figure out how to address the issue you're referring to.
Khashayar Sirti:
You'll have to chain the CDR, right? Because if the CDR's different, then it's different instruments. So eventually, yes, if you can have a connect instead of a DR, which is you're trading the same security.
Peter Haynes:
I do wonder about the future of DRs in a 24-hour trading environment. Hypothetically, why can't we get to a world where a global broker is connected to the Taiwan market, and the Taiwan market is operating 24 hours a day allowing you to trade TSMC on their one marketplace on a more continuous basis through North American hours without there being any liquidity slippage? I think about that. I understand it seems like a leap of faith, but with tokenization, and the other things that are happening, we'll see whether we get to that world.
Khashayar Sirti:
It's absolutely possible. There are some futures like the Japan equity futures with trade in Chicago, which trade in Singapore, which trade in Japan, and you can trade it on them. The other thing I would like to add is there are times when we execute DRs in Asian hours, which is where we go to the broker and ask them to trade the underlying and deliver the DR to us. And that is possible. So to what you're saying, it's possible, but it's done rarely. However, in a 24-hour world, I'm sure the market makers are quite enterprising, and I'm sure they'll find ways to help us make more informed liquidity decisions and execute around the clock.
Peter Haynes:
Oh yeah, for sure. Winnie, let's move over to Hong Kong, your home market. It's an interesting market structure history. I recall years and years ago, the local government was supporting the market buying, I think futures actually, Hong Kong futures or Hang Seng futures, and then eventually got long the market to the point where they wanted to unwind that exposure, and the way they did that was to offer local investors exposure to Hong Kong at a discount via the creation of an ETF. And then you had a market-on-close mechanism at one point, and then the local market shut it down afterwards, a big move on one of the stocks in the index. And then you brought it back again in 2016. How does the auction work today and how did it used to work and why did we have all of this disruption?
Winnie Khattar:
Happy to chime in. And Hong Kong, I think I must admit, it's been a great success story after they launched their current closing auction mechanism, which I though was 2016, maybe in 2017. Should double check. So previously, Hong Kong's close price was established using median price of five prices at 15 seconds interval in the last one minute of trading. The mechanism, as you would suspect, was vulnerable to abrupt price dislocation, especially near the close.
Fast-forward, there's always market events that result in realization that there's need for a more robust and a more resilient market design. And Hong Kong did a great job in bringing the industry together, running consultations, running working groups to come up with a design, the current design of closing auction, which is a single volume maximizing closing auction with much stronger guardrails. It runs for about seven minutes. It includes a non-cancel period late in the session where cancellations are restrictively, again, similar to the design we discussed in India and is common in other markets. You have to be sure about your order before the end of the day because there is a non-cancel period and you can't cancel your order. And then this is followed by a randomized close in the last two minutes of trading.
The price ban that applies in Hong Kong closing auction is plus/minus 5%, which is to reduce the risk of extreme price movement. So net-net, what we see is it's a more transparent, more control auction mechanism compared to the previous design, and it definitely improves the benchmark integrity while limiting the tail risk that existed in the previous design. And we can see that on how the closing auction volume has been showing up as well. It's been steadily on the rise and roughly accounts for eight to 10% of daily volume in the close today. What's interesting in Hong Kong is the same auction logic also applies to ETFs as well.
And if I were to look closely, what we see is beyond equities, there's a wider range of instruments that are trading in the close. There is more liquidity forming in product outside of securities markets. It's great that the same auction logic applies to ETFs as well. However, unlike equities, ETFs do not see the same kind of volumes in the closing auction. Then you can sometimes see price moves of plus/minus 5% in the closing auction, which can be significant for an ETF. If we were to think about how we can improve this design in auction, I would say probably leveraging the design in other global markets, there is room to refine and setting these price bands based on the liquidity of instruments. It doesn't need to be same plus or minus 5% for both single stocks and ETFs during the closing auction. They can be different price band depending on how much, what are the appropriate guardrails for based on the type of instruments.
Peter Haynes:
As an ETF issuer, there's probably nothing worse than having the ETF price closed dramatically away from NAV. And we think a lot about how to fix that, and I wonder, one of the options was to come up with a way of closing the ETF maybe one minute before you close the stocks. I know it sounds ridiculous, but just some way of allowing the dealers to hedge whatever residual risk is left. There has to be some way to figure this out on the close or some sort of NAV functionality. When you're talking about your 15-second median price, all I could think about is the way NASDAQ used to be before they had a MOC facility, which was they took the print that occurred 15 seconds after four o'clock, that was the closing price. So we've come a long way in both Hong Kong and in the US since those old rules existed.
Winnie Khattar:
Yeah, 100%. And I think to your point on ETFs and closing ETFs one minute before the rest of the market closes, I think what's being done well, we can double check the data, but Europe has a very good framework where there is different price bands that apply to ETFs in European markets versus single stocks. And that works really well because there is again, different liquidity, different tiers of price bands that should apply based on liquidity of the stock, which I feel is a really good model. But at the same time, you do want your single stocks and ETFs to actually close at the same time so you don't end up with a risk that there could be a different price formation or dispersion in price formation at the close of the ETF versus the actual fund now by closing the two things at different times.
Peter Haynes:
Winnie, moving on to another market, as I was doing my prep work for this pod, this single market kept coming up as an outlier, and that was Korea. First of all, the regulators have been messing around with short-selling bans, and of course it's well known that there are some limitations to trading the local currency. That said, the country says it wants to do it, and it needs to do to get promoted to developed for MSCI's benchmarking, but it seems every time they take a step forward in development, there seems to be an accompanying step backwards. For instance, two or three years ago, the local regulator started to provide guidance on what is allowed during the market-on-close auction, and they suggested that no one client could be more than 25% of the auction if that client was using limit prices that were more than 1% from last sale. Can you tell us how does this limitation impact order entry in practice? And can you speak to the specific design of the Korean market-on-close auction?
Winnie Khattar:
Yeah, sure. Why don't I stick to the design of the auction itself first and then we can come to how the limitations around volume and price actually have an impact? So I think overall the Korean's market-on-close design is actually fairly stranded and pretty neat and clear, but relatively tighter guardrails than most other developed markets. The continuous trading session ends at 3:20 PM. The closing auction order collection period starts from 3:20 to 3:30 PM. It's a 10-minute period during this period. Orders are accumulated but not mashed continuously. It's a periodic call auction mechanism. All eligible orders are then mashed at a single equilibrium price, which is your official closing price, disseminated and disclosed at 3:30 PM. Both market and limit orders are allowed in the closing auction itself. What I mentioned earlier, there is no anti-gaming feature in Korea's auction design, there is no random close or non-cancellation period.
Now, where Korea becomes a bit of an outlier, as you mentioned, is not the mechanism itself, but the constraints around it, particularly around pricing, the limit pricing and the participation rate. These were introduced by the regulator to actually prevent price distortion and concentration risk in the close. Now, good idea in theory, but in practice, due to the 25% participation rate and 1% price move limit, there is less flexibility for investors to express their sizable interest going into the auction. And as a result, what we see happening, large flows, risk is being scaled back or forced to trade earlier in the day in the lit market before the closing auction starts. The risk obviously is elevated during rebalance days or other market event days when volume going into the close can actually be significantly higher.
So to conclude, yes, these constraints actually reduce predictability and liquidity at the close, especially for global investors who rely on closing price as benchmark for tracking because the more volume that you're trading earlier in the day, the less you're tracking close exactly to the close price. It doesn't actually break the system per se, but it does limit the efficiency of executing in size versus markets versus developed markets like Hong Kong or Japan or European markets where price formation is more unconstrained.
Peter Haynes:
That's a very interesting point, and it leads me to ask you the important question. In your opinion, with these constraints that exist and the things we know about currency limitations, et cetera, in your opinion, FTSE has, Korea has developed. Do you think Korea is ready for primetime and MSCI should in fact move it to developed or are these constraints still too much?
Winnie Khattar:
Look, I would say Korea as a market is economically developed, but operationally, it is still emerging. There's the positive side of Korea that we all know and we've seen a lot more of it this year. There's definitely a lot of liquidity and a lot of scale in the market. And you're clearly seeing a very strong equity performance as well, very much driven by global AI and tech leaders when you see names like SK Hynix, Samsung, and so forth. The market has made a lot of positive changes as well. Short selling, lifting and removing short selling bans was a very progressive decision in Korea, and very well appreciated. Obviously, there are things that can be improved on short sales as well.
But overall, I think what MSCI's hesitation has been is very consistent and I would agree to that, it's about accessibility and predictability and not just the size. So three key things that I think keep coming up in Korea when we think about it from an accessibility and predictability standpoint. One, currency restrictions. Korean won is still not fully convertible offshore. In fact, the liquidity of Korean won tends to drop, especially outside of the continuous trading hours in Korean market. Regulatory consistency, particularly around short selling are very repeated bans, reverses have created lack of trust from international investors. There's still a lack of clarity around punitive treatments around short sellings and accidental short selling when there are real admin errors is a point that needs clarity across the street.
And then for the broader market access frictions like broker funding, we're seeing more recently a lot of constraints on how much capacity brokers and street can offers to invest in that market. FX liquidity, account structures, closing option design, all of these collectively inform overall market accessibility. Korea, I believe is accidentally fixing some of these, expanding FX trading hours, improving disclosures and a number of other things that have been on the radar for MOU that they have discussed more widely with the industry as well recently. But our recommendation to Korean authorities would be to set up small targeted working groups to actually seek feedback and practical solutions, and then make appropriate changes, and oversee not just from announcement of the change, but all the way to the implementation as they look towards going into a DM market. I think index providers like MSCI tend to wait until reforms are fully tested in practice and not just announced, so definitely look forward to that.
Peter Haynes:
Yes, there's nothing wrong with taking a balanced and slow approach to these things. It's very important that those issues that you've addressed are dealt with. And as you mentioned, a couple of the high-flyers around AI, SK Hynix getting a lot of attention globally given at one time here recently was a top 10 name globally in terms of market cap, and that's obviously putting the Korean market on the map. As I understand it, there's a group of emerging markets that are known as the ASEAN nations, if I'm pronouncing that correctly, and that includes Thailand, Malaysia, Philippines, Indonesia, Singapore, and Vietnam. And each one of those countries has functioning market enclosed facilities. I'm curious, is there anything about these markets as a group that investors should know?
Winnie Khattar:
ASEAN markets broadly have functioning closing auctions and MOC facilities. And you started the discussion earlier with the example of Indonesia and I think so taking that example, the direction of travel for ASEAN markets has been in the right direction, and moving towards the global standards. Indonesia moving from a blind auction to a transparent order as soon as 2021 with real time indicative price and volume indications is exactly the kind of evolution you want to see as these markets mature.
I'd say what's important to understand for ASEAN region is that these are still very retail-driven ecosystems and that shows up when you look at the market microstructure. You have wider spreads in general, you have higher volatility, more episodic volatility, especially in markets like Thailand, Indonesia, and the Philippines. You also see it in the way liquidity shows up at the close. So globally, closing auctions on average can be about 15% off the daily volume, but in ASEAN markets, they're typically lower on normal days, and then they would spike up significantly on index or rebalance events. For example, Indonesia can go from low single digit participation on a normal day, up to 40% to 45% of daily volume during MSCI events. Malaysia can similarly push up to 40% to 50%. So there is liquidity, but it tends to be more event-driven rather than consistently deep.
What we have been seeing at the same time in ASEAN markets is turnover across the ASEAN region has been picking up and very much driven by what little as well as more increased retail participation, which means absolute liquidity during the day as well as at the close has increased even if percentages may not have changed as much. And as we continue to see these volumes pick up and focus shift from operation to integrity of the auction, making sure that there are right safeguards in place, things like volatility collars, volatility extension, having the right queue priority for market versus limit orders, and obviously ability to manage toxic liquidity and avoid toxic liquidity is going to be things that are key that should be on exchanges and industries' minds, especially for ASEAN markets that are relatively thinner.
Peter Haynes:
I'll tell you what, there's a future for market structure knowledgeable people like both of you in the Asian region, given the complexities that we're discussing in each of these markets. And what I can tell just in this conversation, it seems to me is there's just a genuine excitement in terms of the evolution of these markets, and there's so much growth in them. One of those markets in ASEAN region is Vietnam, and Vietnam is scheduled to be upgraded by FTSE from frontier to emerging starting in September of 2026. My understanding is that this is ahead of the upgrade the local exchange just did on its trading engine, it now uses the same technology as Korea. With all of the retail participation that you discussed earlier, I'm curious if you're comfortable that Vietnam is in fact ready for primetime and will be able to handle the increased end-of-day message traffic that will accompany its upgraded index inclusion.
Winnie Khattar:
It's really fascinating to see the market's evolution in APAC region. We discussed Korea earlier and Vietnam because it comes with a very strong momentum from the policymakers' exchanges to really engage the industry on what are the changes that we need to move and really driven market reform. I would say how Vietnam is is definitely, again, with very strong momentum, but it is still work in progress or proving phase operationally. There has been very significant progress that has been made and FTSE upgrade definitely reflects that in their decision-making. I'd say the non-brief funding requirements is working, settlement workflows are improving, and turnover has definitely been rising with index inflows expected to come through the course of this year.
Now, as this market activity increases, Vietnam already we see is punching over its weight in the region. Its share in ASEAN liquidity has climbed up to roughly around 20%, which puts it closer to some of the bigger ASEAN markets like Thailand and Singapore. And closing auction in Vietnam as well is accounting for roughly 15% of daily trading volume as a close, which is a good base for index inclusion. I know you mentioned KRX trading system, which was a long journey. And I know Khash, you were part of a lot of discussion on getting the system live and up and running for Vietnam. And there's been a lot of focus on getting the basics right. The next phase is really going to be about implementation and what gets built on top, having the broker connectivity, more flexible, dictating models, and really the ability to absorb index-driven flows consistently. And we see that stress dynamic in other markets on index closing auctions. Volumes can jump three to five times of average normal volume.
So I think that will be the test for Vietnam as flows start to ramp up through the course of this year. And that will be the consideration that index providers, excuse me, like MSCI would also take into account on how the actual implementation is progressing and taking shape. So to your question, whether Vietnam is ready directionally, yes, 100%, but upgrade itself will be the test. So market now needs to prove if it can handle sustained institutional scale liquidity without frictions.
Peter Haynes:
One of the other ASEAN nations that the index providers have been paying attention to is Indonesia with suggestions that they're going to get downgraded and each of the two providers going back and forth on that. Can you tell our listeners a little bit about what's been happening in Indonesia that's causing so much consternation?
Winnie Khattar:
So MSCI back in February this year raised three concerns about Indonesia's equity market, which was high ownership concentration, low effective free float, and weak market surveillance and enforcement. These issues triggered MSCI's decision to freeze any index updates for Indonesia market, and potentially suggesting a downgrade from emerging to frontier market. What has happened pretty quickly since obviously at the risk of market outflows we've seen Indonesia make a measure of moves to reform, first create transparency and improvements by lowering their disclosure threshold from 5% to 1% on individual companies.
Second, to reduce the shareholder concentration risk they've started publishing this data, which is publicly available to investors as well as index providers as well. What Indonesia also did is it increased its minimum free fraud limit from 7.5% to 15%. Once again, similar to what we discussed on some of the other markets, all of these market structure refunds are steps in the right direction.
What would be incredibly important from here is for Indonesia and the market to actually leverage their market surveillance and enforcement as and when they see specific events or cases of price impact coming through because of shareholders, especially large head shareholders making decision on part of their companies and impacting stock price, which we've seen in the past in some of the index rebalance event and taking action, taking enforcement actions against those companies. So that will be an important step to track how the progress is being made in Indonesia going forward.
Peter Haynes:
It's always interesting in these marketplaces want to keep complete control of everything in terms of foreign ownership limits and then at the same time they seek foreign investment, and there has to be a trade-off there at some point. They have to gain some comfort to let go a little bit.
Khashayar Sirti:
I'd just like to add actually two points there. Vietnam is also one of those markets where there are foreign limits imposed by companies, and not necessarily by the regulator, which makes it even harder for foreign investors to buy the stocks, and makes the index formation very difficult. So that's one of the things that we've been explaining to the regulator, that we need to have wider foreign limits if you want foreign participation because there used to be a time when there was a stock called as Media, which was in China and was enlisted in Hong Kong and it used to hit their 28% limit quite often so every now and then you would have the inclusion, and then you had to sell the stock down, and it used to be in and out, in and out, and you couldn't buy the stock until it came back to 25% and that just kept on happening, and I fear Vietnam may run into that problem.
One thing I would like to add about Vietnam since you asked, will they be able to handle the amount of volumes coming into the close? It's one of those few markets where you can't merge orders today. You have to trade on ID by ID. So even in other ID markets, you can actually merge your orders and send one line. Vietnam is one of those markets where you can't. Once again, we've been explaining to the regulator that it'll do them a lot of good if they could move away from this format of execution and allow foreign institutions who could have 25 funds buying the same fund to send one line to the exchange instead of sending 25 lines because you can imagine the volume that'll go through otherwise.
Kudos to them, they are quite receptive. As a matter of fact, I think they're coming to Singapore in first week of June, and they'd like to meet with us again to understand better how they can go about executing that. So there are these things, but like Winnie said, Vietnam regulators, in my opinion, one of the most receptive regulators in the region, and their intent and their execution both show up.
Peter Haynes:
Well, they have to listen to the institutional community, and I'm glad to hear they're listening to you and hopefully they can get those changes made for the industry. It seems a bit ridiculous to have to put line by line instead of the bulked order.
Khash, we're going to move over to Australia now. The ASX employs a 10-minute auction after 4:00 PM end of regular session ends, and the auction has a random close that occurs within 59 seconds after 4:10. Again, all these markets seem to have random close. Over in New Zealand, they have a market-on-close system as well that's similar, but it's 15 minutes in length, and for most of that period, it's kind of a sleeper. Are you comfortable with how these two auctions work, and are there any refinements that are in the works for either Australia or New Zealand?
Khashayar Sirti:
Okay. So New Zealand is, yes, one of the smaller markets. Even though it's a DM market, we participate infrequently there when we have flowed some of our ETFs. Australia's a much bigger market, trades a lot at the close. It's also one of those unique markets, I think along with Singapore, which doesn't have price limits in the auction. Therefore, the volatility in price discovery is quite high. You could see stocks starting off as far as 5%, 7% away from the reference price, and then it pulls back as more volume comes through. So it's one of those markets where we've seen a lot of volatility within the price formation. Eventually, the stock doesn't close that far, but there are moments during that period where you see big moves. It's also one of those markets where there's no non-cancel. So not to say that there is people out there who are misusing it, but it's ripe for misusing if there's somebody who wants to do that.
The other thing is it doesn't allow market-on-close orders either. So it's one of those unique markets where there could be reform. They did have some refunds, but they were marginal in chain, especially around the opening auction more than the closing auction. Winnie, she's been watching that market as well for a long time, but we would like to see the volatility in the price formation of Australia come down if possible.
Winnie Khattar:
No, Khash, I think you've covered it really well. I think Australia is one of the markets that has a very large volume in closing auction. Even on regular days, you see 25% to 30% of volume trading on the close. So I think exactly the points you mentioned, the volatility that we see during the price formation period. And we were talking about price bans earlier. So having 5% price bans in Hong Kong versus 3% that is considered in India, 10% in Taiwan. Australia doesn't have any price ban in the closing auction. So as a result, plus it doesn't allow market orders. So as a result, you se investors, market participants putting orders at very, very wide limits.
Eventually, obviously it compresses closer to the last traded price and there comes to be an equilibrium price, which is not very far and not very volatile, but it would definitely really help to incentivize participants to participate early on, help narrow the price range early on if there was a market orders because that will definitely converge organically liquidity at the prevailing market price and potentially maybe the price bans as well to reduce the amount of overall volatility or range in which participants can set their limits in the auction design.
Peter Haynes:
So Khash, just as a follow-on, and this is relevant for us folks here in Canada, is CBOE has put its assets that it has in Canada and Australia, the two marketplaces, up for sale and they did so late last year. And recently, it was announced that Canada's primary exchange operator, the TMX, has actually won the auction for those assets subject to regulatory approval. I'm curious, how important is the CBOE, which was formerly Chi-X in Australia, to the local market? And is the existence of a second marketplace considered positive or negative for local market structure?
Khashayar Sirti:
We were very excited when CBOE came in and they had a very nice product to offer and we were quite looking forward to it. However, unfortunately they could not gain market share and therefore they decided not to pursue any further I believe also in Japan, but I'm not sure about that.
Peter Haynes:
Correct.
Khashayar Sirti:
What happened, this is we have other markets and a few places such as Japan where we have a lot of PTS, multiple venues, and that has really helped the microstructure. As you can also see in Korea, you had a new exchange come in, NXT came out with a newer product, new offering. They themselves didn't expect to have such a high performance of 30%. They hit that within six months. Everybody was shocked. Now it's getting KRX to change. So competition is always great, and it brings about change. Australia though, is it a harder market to crack in and would be interesting to see how TMX after it wins is our regulatory approvals will be able to change the microstructure in Australia, so we are looking at it very keenly. And I guess the proof of the pudding will be in the ET in this case.
Peter Haynes:
I know that I've been following from a distance. The ASIC has been very, very vocal in its frustration towards the primary exchange operator in Australia, criticizing them. And I do think there's other marketplaces in the world where they could be argued the same criticism that they've put too much money in the hands of their shareholders focusing on shareholders over stakeholders, and there's been technology outages and the problems with the settlement system. Overall, do you get a sense or what you've read in Australia that the TMX is a name that the local market is happy to see competing with the primary exchange? Has that been well received so far?
Khashayar Sirti:
Yes, I haven't seen any negative press about that, to be honest. And as you mentioned, the regulator there wants change, wants competition, wants Australia to move into the right direction, and that's a great thing and we totally support that. And as I mentioned earlier, competition is great. There has been no negative press about DMX, so that's a positive as well. We would again have to wait and watch as to what they come in with, what changes they bring in, and how we can participate across both exchanges seamlessly. That's very important for us as well in terms of settlements, et cetera. So we are quite excited about the change, it's one of those DM markets where we are seeing change in terms of exchanges. Hong Kong Exchange is also a listed entity. Singapore Exchange is a listed entity. So there are other exchanges which are listed and commercial, but they managed to do the right thing as well.
Peter Haynes:
Yeah. Actually, you mentioned the number about Korea and the competition there. I wasn't aware that the second marketplace had chewed away that much market share off of the primary. That's interesting.
Khashayar Sirti:
Phenomenal. And now KRX is doing something which is not great because you're opening at six o'clock to seven o'clock and then they're stopping for one hour and stuff like that. But yeah, it's forcing KRX to wake up, and it's been phenomenal.
Peter Haynes:
Yeah, that's great to hear. Winnie, my last market I want to cover here is Japan. We can't forget about it. I think people, when I started my career in the late '80s, early '90s, Japan was bigger than the United States. I just don't think that's even comprehendible today. But up until the fall of 2024, the Tokyo Stock Exchange auction was the one APAC market that was out of this convention, which used a closing auction that ran concurrent to continuous. Since then, it has now created a standalone post-close, and they also introduced a new trading engine as part of that. Can you tell us about the new model in Japan compared to the old model? And are there any tweaks that you would like to see with this new model to further improve the auction process in Tokyo?
Winnie Khattar:
You're right. The closing auction move to a standalone auction in Japan probably the most recent. Became effective in 2024, and the move from standalone auction after the regular session brings Japan more in line with global practices. Once again, it concentrates closing liquidity into a dedicated call, makes auction participation more intuitive and reduces operational scramble that can happen when auction mechanics are intervened with the continuous trading. Market has already seen steady growth in the adoption of this new design as well. Closing auctions in Japan roughly trade around 15% on normal days, and 30% to 40% on event days. Under the old models, participants sometimes saw limited effective matching and tendency for market to default back to last continuous price. Sometimes we did not have a close price in the way that last three minutes of closing price formation was designed in the old model, which is definitely not ideal in rebalanced days.
So in the new model, especially with the new engine, it improves the actual price formation. It actually gives confidence that there will be a closing price on all securities and ability to actually pair off opposing close in the closing auction in the current design in Japan is significantly better than before. Compared to other auction, it's a five-minute auction period that runs between 3:25 to 3:30 PM. Orders can be entered, amended, and canceled, and at 3:30 PM, they identify the single closing match price. What is not there in a closing auction currently today is some form of anti-gaming logic, which always is a very good feature of any closing auction design. And following discussions with the Japan Exchange, they are definitely open to creating an anti-gaming logic and potentially introducing a non-cancel period or a randomized close. I think more likely a non-cancel period as they look to evolve their auction design, which I think would be definitely beneficial in Japan.
Peter Haynes:
These market-on-close mechanisms are living organisms and should constantly be tweaked in order to make them better, not just for the sake of trying to make them better, but where you can actually add things that are truly beneficial, as you mentioned, anti-gaming techniques. So final question for both of you, you've given me so much time here and our listeners so much information. Both of you are working for prominent global asset managers that are active in the APAC region. I'm curious whether or not you would say the investor interest that you're seeing from your customers is on a positive trajectory and what would you say is needed for that momentum to continue to be positive in the region's flows? Khash, I'll start with you.
Khashayar Sirti:
Yes, absolutely. As you can see, and as you mentioned, TSMC, Korea, they are all the rage right now. Before that, it was India and China, so the good thing about this part of the world is you have some form of rotations in the emerging markets. And after, I would say, a 15-year period, the emerging market is finally having its moment in the sun, not only in Asia, but also in Brazil, et cetera. So it's a great thing to see right now that there is a focus back in this part of the world. It's driven by technology at the moment. However, Asia is a huge growing economic block, and I don't foresee that stopping for the next few decades. So I do believe we will keep increasing our market share in terms of global market cap, et cetera. What do we need for it to continue? I guess as long as the risk stays on, money will keep flowing. What we do need from our side is stability in regulation, stability in policy, and just welcoming flows in a sensible format. That's my take.
Peter Haynes:
What would you add, Winnie?
Winnie Khattar:
I agree, Khash, you've covered it really well. There is investor interest in Asia Pacific, which is improving and will continue to trend in that direction, particularly where reforms and market reforms translate into real execution confidence. In APAC specifically, this means easier access to markets that have been operationally challenging to access historically, more transparent auctions, fewer ad hoc regulatory interventions or guidances, all of which together enables smoother foreign investor participation.
There is without doubt, strong momentum in the region to marketinize market structure as that transition happens from one index classification to next, all the way from Japan and to Korea and Vietnam as well, there are visible tangible improvements being made to market rules in terms of improving access. What will continue to be important, and we touched on this earlier is consistent regulatory messaging, which is definitely key to building that investor trust. I am definitely very delighted and feel very fortunate to be a partner in helping to shape these rules in these regional market, which eventually enable BlackRock's vision of being able to deliver more investments to more investors.
Peter Haynes:
Well, the number of times that you've said back and forth that each of you have been working with regulators as part of an advocacy effort to make sure that market structure evolves and gets better. Thank you on behalf of all of us in the markets for your advocacy work in the region, which as you say is growing. I mean, Vietnam alone has 100 million people. There's significant percentage of the world's population located in the regions you cover, and it's only a matter of time before they continue to grow, and I'm glad you're helping with that process. Thank you very much on behalf of TD Securities. To both of you for joining today, really appreciate the information, and look forward to follow-on discussions.
Khashayar Sirti:
Thank you, Peter. It was a pleasure.
Winnie Khattar:
Thank you, Peter. It was great to chat with you.