We explain in detail how the Tap mechanism will draw index shares from the Government's holdings when the market moves upwards, dampening the upside on the market. At the current rate, we expect HK$24bn of tap next year. We also look at the redemption and subscription mechanism and its likely effect on the trading prices of the TraHK. Finally, we demonstrate how the HSI has been boosted relative to the rest of the market by the Government's intervention.

TraHK on Tap
1 November 1999

Contents
No ordinary unit trust
The Tap
Case 1: the shareholder
Case 2: the futures player
Other types of arbitrage
Uncertainty over NAV
What's the trading range?
Will the tap affect the market?
Tap size
Outlook for HSI
Intervention boosted HSI

With the closing of the retail offer of the Tracker Fund of Hong Kong (TraHK) only 3 days away, we explain the obscure "Tap" mechanism and analyse the cost of the arbitrage opportunities that it may create and its likely dampening effect on the market level. We'll also look at the complicated subscription and redemption "in kind" mechanism, and this will lead us to an estimate of the trading range of the TraHK relative to its net asset value. At the end of this article, we'll demonstrate how intervention pushed the Hang Seng Index up, and look at the market downside.

No ordinary unit trust

As the Government has mentions in the subtitles of its sweet commercials, the TraHK is not like a normal unit trust. In a normal unit trust, there is no real need for a listing on a stock exchange because no trading takes place in the units except with the fund manager. The manager handles subscriptions (when an investor puts new money into the fund in exchange for units) and redemptions (when an investor sells units back to the fund in exchange for cash).

In the case of TraHK, redemptions can only be done "in kind". What this means is that in return for the units, you get a basket of the 33 shares which make up the Hang Seng Index, in about the same proportions as in the index. Because of the problem of odd lots (see previous articles), redemptions will only be accepted in multiples of 1 million units (or about HK$13m). Even then, an investor may not get whole board lots of constituent stocks, depending on how precisely the basket weightings are calculated by the manager.

Subscriptions can normally only be done "in kind" (by presenting the manager with the right numbers of shares in the index basket) and again only for multiples of 1 million units. However, there is also a cash subscription capability known as the "Tap".

The Tap

The Tap is basically a mechanism to allow the Government to sell further units in TraHK when the market is rising. The way it works is that an investor may apply (again in multiples of 1 million units, known as a "Creation Unit") to pay cash for new units in TraHK, and the manager will use the cash to buy an index basket of shares from the Government. The price of the basket will be the IBEAS (Index Basket Exercise and Alignment Settlement) price. This prices the index based on the average price of each member stock taken at 5-minute intervals during the trading day (excluding the 10:00 and 14:30 opens and the 16:00 close). That is, there are 47 readings, 30 in the morning session and 17 in the afternoon session.

In practice, the tap will be used by investment banks and large institutions to "arbitrage" the market. In a fast rising market, or one which rises suddenly late in the day, the IBEAS will be lower than the price of the index at the end of the day. If this difference is large enough to cover the costs involved, then it provides a potential arbitrage opportunity using the Tap.

To determine how wide the gap has to be to activate the Tap, we'll look at two cases. The first represents a holder of index shares, the second a futures player.

Case 1: the shareholder

If you are a large investment fund then you may already hold the stocks which make up the Hang Seng Index (or you can top up easily with the smaller ones you don't own). In that case, when an arbitrage gap opens, you can sell the shares in the market and subscribe for TraHK units under the Tap. The cost of selling the shares will be 0.125% stamp duty, 0.011% transaction levy and, unless you are a broker, 0.25% commission (if you get rebates or "soft dollars", then deduct that and don't tell anyone). The Government will also charge you 0.125% stamp duty on the shares purchased under the Tap to create your new TraHK units.

You may also find that your own selling cannot be achieved at nominal market prices, because to act fast you may "hit the bid" side of the bid-offer spread, whereas in a rising market a lot of the prices will be dealt at the "offer". The minimum spread is determined by the increments in which bids and offers are permitted on the Stock Exchange. The current "spread table" is:

Stock price from Spread Spread as % of price
$0.01 $0.001 10%-0.40%
$0.25 $0.005 2.00%-1.00%
$0.50 $0.010 2.00%-1.00%
$2 $0.025 1.00%-0.50%
$5 $0.050 1.00%-0.17%
$30 $0.10 0.33%-0.20%
$50 $0.25 0.50%-0.25%
$100 $0.50 0.50%-0.25%
$200 $1.00 0.50%-0.10%
$1000 $2.50 0.25%-

In practice, the top band is not used. HSBC was above $200 before its stock split in July. The top 10 companies in the HSI, which account for 80% of its value, have share prices between about $15 and $100, so you can expect average minimum spreads of about 0.35%. Therefore, if you have to hit the bid, and half of the stocks have last traded at the offer, you can assume a discount of about 0.18% on the last index value. This is a hidden cost of dealing in Hong Kong, and small brokers have resisted narrowing the spread table for obvious reasons. However, with the increasing presence of ECNs (Electronic Communication Networks) we hope that the newly merged HKEC will narrow the spread table at an early opportunity to compete with them for pricing quality.

Additional uncertainties are introduced by the fact that trading computers cannot be directly linked to the Exchange, and all orders must be input to terminals by hand. So it takes time to hit 33 stocks in one go. Time equals uncertainty. That won't change until AMS3 is introduced later next year. 

Once you've sold your shares and got your TraHK units, then to get back to where you started you would have to redeem the units for shares (in kind) or sell TraHK in the market and buy back the shares you sold. Redemption is cheaper. Most fund managers won't want to pay State Street 0.1% per annum when they could be managing the shares themselves, and in some cases their trustees won't permit them to hold more than a certain percentage of their investments in other funds (such as TraHK), since that leads to the institution's clients being "double charged" for fund management. So they can't just hold the TraHK forever.

Redemption of TraHK results in a transaction fee of $30,000 per application. For the minimum size of 1m units (about $13m), that works out at 0.231%, declining for larger transactions. However, bear in mind that larger transactions will have greater "market impact" - selling the stock may cause prices to go down. So it may be necessary to accumulate TraHK through several arbitrage transactions and then redeem in a larger batch. In our model we'll assume redemptions in batches of 5m units, reducing the cost to 0.046%.

There is also a cancellation fee of $1 per board lot of TraHK (500 units), or about 0.015%. All in all, the arbitrage costs for a broker and non-broker are estimated below: 

Tap arbitrage by a shareholder Broker Non-broker
Sale of index shares - spread (estimate) 0.180% 0.180%
Sale of index shares - stamp 0.125% 0.125%
Sale of index shares - brokerage   0.250%
Sale of index shares - transaction levy 0.011% 0.011%
Subscription of TraHK - stamp 0.125% 0.125%
Redemption of TraHK - transaction fee (5m units) 0.046% 0.046%
Redemption of TraHK - cancellation fee 0.015% 0.015%
Total costs 0.502% 0.752%

Given the uncertainties on the spread, and allowing for a worthwhile profit, we are unlikely to see brokers arbitrage until the IBEAS is at least 0.70% less than the Hang Seng Index, and fund managers won't come in until a 0.95% gap exists (in each case offering a 0.20% profit). It is also possible that Participating Dealers may charge a commission for processing subscription applications, adding to the costs.

The whole point of arbitrage is to remove market anomalies, so we expect the investment banks (with their own brokerages) to normally move in before they become wide enough for non-broker institutions to take advantage of them.

If the index has been steadily moving upward during the day and gained at least 1.4%, or has been flat followed by a late spurt of more than 0.8%, then this would create such conditions.

In both case 1 and 2, we have not allowed for the possibility that an arbitrageur might take a position and then wait for the mis-pricing to reverse itself (e.g. until futures go to a discount to fair value or TraHK goes to a premium to its net asset value) before closing the position, since there can be no certainty over when or if this might happen.

Case 2:  the futures player

In this case, the would-be arbitrageur would sell futures contracts and buy TraHK through the Tap. Such an opportunity might be created not just by a move in the index, but by an over-pricing of the futures contract which can emerge from time to time. Apart from the price of the underlying shares, the "fair value" of a futures contract is determined by interest rates, the dividends due on the shares, and the time remaining until expiry.

Hang Seng Index futures are still dealt with by open-outcry, which has inefficiencies of its own. The move to screen-based trading won't come until early next year (or whenever they get it working). As each TraHK unit represents 1/1000th of an index, the minimum creation of 1m units (a Creation Unit) is equivalent to $1,000 per index point, or 20 futures contracts (which are each worth $50 per point). There would be some market impact from selling several Creation Units.

In order to "close out" his gains, our trader will have to buy back the futures contracts and either sell the TraHK or redeem it for underlying shares and then sell the shares. He may find that TraHK is trading at a discount to its NAV in which case he'll have to redeem the units rather than sell them. Either way, he pays stamp duty when he sells the units or shares in the market. The minimum brokerage on futures contracts is $100 each side, plus transaction levy of $11.50. In addition, the minimum spread is 5 index points. If the last trade was at the offer, then hitting the bid would thus cost 0.038% (if the bid were large enough in volume). Apart from this, the costs of buying TraHK, redeeming it, and selling the index shares are the same as in Case 1.

Typical costs would then be:

Tap arbitrage by a futures player Broker Non-broker
Spread on sale & purchase of futures 0.076% 0.076%
Round trip commission on futures 0.031%
Round trip transaction levy on futures 0.004% 0.004%
Other costs as in Case 1 0.502% 0.752%
Total costs 0.582% 0.863%

Again, we have ignored any commission charged by Participating Dealers on subscriptions or redemptions. We've also ignored interest loss on the margin deposit for the futures contracts.

Other types of arbitrage

The TraHK itself is open to arbitrage if its pricing drifts too far away from fair value, even when the Tap for that quarter has been exhausted. For example, if TraHK is trading at a discount relative to its net asset value (based on latest market prices) then it should be possible to buy TraHK units in the market and present them for redemption, and sell the underlying shares in the market. However, it takes 2 trading days to redeem the units (once you've got them), so to avoid market risk, you'd have to sell index shares you already own, or borrow them (more costs) until you can redeem TraHK, or sell futures. If you already own index stocks, then the estimated cost is as follows:

Stock/TraHK arbitrage (no Tap) Broker Non-broker
Stamp on purchase of TraHK 0.125% 0.125%
Transaction levy on purchase of TraHK 0.011% 0.011%
Commission on purchase of TraHK 0.250%
Sale of index shares - spread (estimate) 0.180% 0.180%
Sale of index shares - stamp 0.125% 0.125%
Sale of index shares - brokerage 0.250%
Sale of index shares - Transaction levy 0.011% 0.011%
Transaction fee on redemption (5m units) 0.046% 0.046%
Cancellation fee on units 0.015% 0.015%
Total costs 0.513% 1.013%

Conversely, if an investor holds (or borrows) TraHK and it is trading at a premium to net asset value, then  he can sell it in the market and at the same time buy the underlying index basket of shares, then present the shares to TraHK to create new TraHK units if desired.

Uncertainty over NAV

Whenever an investor want to buy or sell TraHK, one challenge will be to figure out what the current net asset value (NAV) is. This is because it will not be as simple as just dividing the HSI by 1,000. The TraHK will be accumulating dividends and operating costs (paying out the dividend net of such costs every 6 months) and in addition, it will accumulate a "tracking error" due to changes in the index weightings.

The NAV will only be calculated and published once per day through the Stock Exchange (based on the previous night's closing prices) and it probably won't distinguish between the index shares and other net assets in the fund. So for the private investor, as the day goes on and the market moves, it will become difficult to figure out the exact NAV. The simplest estimate will be to multiply the previous night's NAV by the current HSI and divide by the previous night's HSI.

Fortunately, it is unlikely that other net assets will ever be more than 3% of the fund, unless one of the index stocks pays out a very large dividend.

What's the trading range?

Based on our analysis of the arbitrage costs, we would expect the TraHK to spend most of its time in a range of -1% to +1% relative to its current NAV. However, if the Tap is open and the market has risen steeply that day,  expect the TraHK to trade at a larger discount in the afternoon, because the big guys will not pay more for units in the market than they can pay at the Tap.

When the Tap is closed, the range will hold because investment banks with direct access to the stock and futures exchanges will arbitrage to remove the difference. It will take a few weeks for these banks to get a grip on the mechanics, so we can expect a slightly wider range until things settle down.

As for liquidity, don't expect miracles. If people are long-term holders then you will at first be looking at a relatively small amount of stock (perhaps $2-3bn) trading in the market. That's nothing like Hutchison or HSBC. But as an individual small holder of TraHK, you will probably be able to get 98-99% of NAV any time you want.

Will the Tap affect the market?

We can expect the Tap mechanism to be activated whenever the market rises steadily by about 1.4% or more during the course of the day (leaving the IBEAS at about a 0.7% discount to the closing prices) or if the market is flat followed by a late surge. Tap subscribers have until 16:15 each day to apply for units at that day's IBEAS, so in practice they can wait until late in the afternoon session when the IBEAS is almost known. Applications are made through CCASS terminals which every broker has, although they have to agree to be a "Participating Dealer" in the scheme.

Of course, the very act of arbitrage selling will have a negative effect on the price of the shares. In effect, as long as the Tap is open, the net demand for index stocks will be met by the Government supply of TraHK units under the Tap. This shifts the supply-demand balance and will provide downward pressure to stock prices in the absence of other factors.

The Tap Size

At least 10 business days before the beginning of each calendar quarter, the Government will announce how much stock it will make available for the Tap. The 10th business day before the first quarter of 2000 is 16-Dec-99. In the case of the partial quarter from 12-Nov-99 until the end of the year, they have said it "is not expected to exceed" HK$3bn, an amount which will be announced by 9-Nov-99. After that, your guess is as good as ours. As the first tap only runs for half of the quarter, we could expect a Tap size of at least $6bn for the first quarter of 2000. Presumably the Government wishes to sell as much as it can through the Tap to reduce its equity holdings towards the target level, without denting the market too much. It owns $198bn of stock (as of 20-Oct-99, when the index closed at 12,499). If it keeps tapping at the current rate, that would see another $24bn of stock returned to the market during 2000, in addition to the planned privatisation of the MTRC.

Outlook for HSI

The size of the Government holdings represents a considerable overhang of stock on the market. Whenever the market rises, the Government will be tempted to open the Tap wider for the next quarter, so the market upside will be limited by this knowledge. That gives us a reason to be bearish on blue chips over the near term. In addition, we regard stocks on Wall Street as seriously over-valued, and you should know that if Wall Street crashes, Hong Kong will crash too. Professionals call that "correlation".

Now you might be thinking that the HK Government, having done it once, would always intervene in the event of a major market fall, as they did last Summer. Think again. The Government last year bought 15% of the free float (the shares not held by controlling shareholders) and would put itself in a far more precarious position if it were to do the same thing again. Owning 30% of the free float would eliminate any residual claim to be a promoter of free markets. So expect any Government support in the event of a crash to be very limited.

Intervention boosted HSI

It is worth comparing what has happened to the Hang Seng Index (where the Government intervened) with the wider Hong Kong market (where no purchases were made). We can do this by looking at the Hang Seng Mid-Cap Index (HSMCI), which represents the stocks approximately ranked 101st to 150th by size.

The Government intervention began on 14-Aug-98. If we look at the period from 2-Jan-98 (the base date of the HSMCI) to a week before intervention (7-Aug-98), the two indices moved almost in line, the HSI falling 34.3% and the HSMCI falling 34.0%. We ignore dividends in both cases.

However, in the period from 7-Aug-98 to today, we can see that the HSI has increased by 89.8% while the HSMCI has increased by only 41.8%. That's a huge out-performance for the HSI of about 34%. If the HSI were to fall back into line with the MCI then it would fall by about 25% to just below the 10,000 mark.

2-Jan-98 7-Aug-98 1-Nov-99 From 2-Jan-98
to 7-Aug-98
From 7-Aug-98
to 1-Nov-99
HSI 10,680.57 7,018.41 13,322.11 -34.3% 89.8%
HSMCI 1,000.00 659.64 935.65 -34.0% 41.8%

Bear those figures in mind. You might think that the difference is explained by mid-cap companies having a tougher time then their larger counterparts, but that's an awfully big difference which did not seem to be present during the first 7 months of 1998, when the economy was probably in greater distress than it is now.

We don't buy that story - we think that, apart from the bull-run on global stock HSBC, the main reason for the out-performance is the simple scarcity of supply of stock which is about to be remedied. Intervention pushed it up, disintervention will push it down.

© Webb-site.com, 1999


Organisations in this story

Topics in this story


Sign up for our free newsletter

Recommend Webb-site to a friend

Copyright & disclaimer, Privacy policy

Back to top