Property markets in Hong Kong have seen major boom and bust cycles. This phenomenon is well known in economics. It is the product of fluctuations in demand and long lags in the supply of housing. The pattern was first observed in hog markets and came to be known as the Hog Cycle.
The basic mechanism is that a short-term increase in demand increases the price of pork. Pig farmers then decide to keep more pigs for breeding, which decreases the immediate supply so prices rise further. After a year, the increased production of pigs begins to be felt in the market and the price for pork begins to fall. In response pig farmers keep fewer pigs for breeding, which increases the immediate supply of pork and the price falls even further. After a year, the decreased production of pigs begins to be felt in the market. Another round of price and output cycles begins.
In the property market, the Hog Cycle phenomenon is worsened by much longer supply side lags in the production of housing. These lags are exacerbated by the lengthy time required to develop land due to the high and rising regulatory cost of making land available. Building a large land bank would address only one part of the problem, land availability; it would not tell us how much housing should be supplied.
Property Market Has Hog Cycle
Demand shocks in the property market also have a cascading effect that magnifies the initial effect of a shock. When demand increases, property prices rise initially. Existing property owners experience an increase of wealth as the equity value of their property appreciates. The wealth effect and the opportunity to refinance the original mortgage loan allow the homeowner to purchase a second property or a larger unit. The multiplier effect further fuels property price increases. The effects are similarly exaggerated when prices collapse. The feedback loop from demand to price, to wealth, and back to demand magnifies price changes.
The Hog Cycle phenomenon is found in many commodity markets. Futures markets are set up to allow suppliers and demanders to hedge against risk from the fluctuations and smooth out some of the ups and downs.
Hong Kong once had a short-lived and some very limited trading of the property price index, but the activity was so thin that it was mostly one-sided interests who were either longing or shorting the index.
A more sophisticated instrument to hedge against property development risk was the pre-sale of housing units before their completion, first introduced by Mr. Fok Ying-Tung. A secondary pre-sale market developed in trading these instruments known as confirmor transactions. I have written in more detail about these instruments in two earlier essays in the HKEJ. Although they were only useful for hedging against risk when the market was in an upswing, during such periods they proved to be valuable in conveying information on market expectations.
However, since they were only privately traded, the information conveyed to the public was necessarily imperfect and the timing of their release by the parties involved was probably intended to influence public sentiment. The information was biased upwards in a booming market, a problem that could have been easily corrected if the government had required price information on all such transactions to be publicly disclosed within, say, a day.
The government, however, viewed the trading of pre-sale units as speculative activity that was responsible for fuelling price increases in a tight property market. Since the 1990s the government has on numerous occasions imposed punitive measures to dampen transfer transactions of pre-sale housing units. Today the volume of such transactions is minimal. The secondary pre-sale market has lost its role in conveying information on market expectations and the public has to make transactions without the benefit of this information.
2002 Review of the Institutional Framework for Public Housing
My colleagues at the University of Hong Kong have researched the consequences of curbing speculation in pre-sale housing markets. They found that the volatility of property prices increased signi?cantly after the secondary market from pre-sale housing units was curtailed by regulatory control measures. Anti-speculation measures therefore had increased the information cost in the market. More transactions now take place at “wrong” prices and price volatility in the market has increased.
How, then, can property price volatility be addressed? In the depths of the Asian Financial Crisis the government reexamined its role in housing. The policy document Review of the Institutional Framework for Public Housing (2002) set out revised goals for a long-term housing strategy.
1. To ensure that adequate housing is available to all households;
2. To promote and help satisfy the demand for home ownership;
3. To provide 50,000 subsidized housing units each year, including Public Rental Housing, Homeownership Scheme (HOS), and Home Purchase Loan Scheme (HPLS) units;
4. To reduce waiting time for Public Rental Housing to 3 years;
5. To suspend the Homeownership Scheme (HOS), but maintain the building of 2000 HOS units each year;
6. To secure the most effective use of the resources of the private sector in housing production.
7. To publish periodically the long term demand forecasts, not as targets, but as indications of planning requirements only;
8. To adopt a private sector led strategy in the development of private housing units; and
9. To suspend scheduled land sales via public auctioning and replace it with an Application List System (ALS).
The Review broke some new ground on private housing, but it also renewed government’s commitment to the public provision of subsidized housing on a scale similar to that of previous administrations under Youde, Akers-Jones, Wilson, Patten and Tung.
In retrospect almost none of the objectives relating to subsidized housing was achieved, with the possible exception of reducing the waiting time for Public Rental Housing. Only about 600 HOS units were produced and the number of new Public Rental Housing Units each year also fell far behind target. In 2011-12 fewer than 13,000 Public Rental Housing units housing units were built; the corresponding number for private housing units was just over 13,000.
Flaws in the Application List System
However, a significant change was seen in the rethinking of the long-term strategy for private housing units. For the first time the government recognized that a command approach towards private housing was futile. Long-term demand forecasts were needed, which would prepare an adequate supply of land for use by property developers, while not being seen as production targets for private developers
The ALS effectively achieved this by shifting the initiative for private development from government to the market and abandoning the old system of scheduled land sales via public auctioning. The guiding principle was to let the private sector lead private housing development. Government’s role was to set the reservation price for the sale to be effective and to prepare land for inclusion in the Application List according to long-term demand forecasts.
The ALS uses both tenders and auctions in a single system, the difference being that tenders use sealed bids and auctions use open outcries. It works like this: the government updates and announces the sites available for sale in the next financial year. A developer interested in a site on the application list can make a sealed offer to the government. If the trigger offer is accepted then an auction for the site is arranged in about10 weeks. The triggering developer is required to pay a deposit immediately.
The accepted offer (or trigger price) is the starting bid of the auction. If the final bid at auction exceeds or equals the trigger price, the deposit is returned to the triggering developer (without interest). If the final bid falls short, the triggering developer’s deposit is forfeited. The government can also withdraw the site if the auction fails to reach the updated reservation price (based on assessed open market value).
The use of ALS over scheduled auctions has a distinct advantage for the government when market sentiment is very weak because it eliminates the need to impose a moratorium on land sales in order to stabilize the market. Let me explain. When the market sentiment is very weak, developers will make very poor offers or no offer at all. If the government accepts or rejects a low offer, then the weak sentiment gets confirmed. The government is in the unenviable position of acting to destabilize the market.
By comparison, the ALS lets the market trigger land sales. This system, though, has also had its drawbacks. In most years the ALS has not worked to increase land supply because developers have been generally reluctant to make initial offers even in a rising market. For example, no residential sites were successfully triggered for almost a year from October 13, 2004 to September 26, 2005.
This situation arose because the government initially set the trigger price at equal to or higher than the reservation price, which was a very high price for triggering developers who had to pay upfront. They were assuming all the risks but receiving neither compensation nor advantage. Developers urged the government to lower the trigger price.
The government responded by modifying the ALS in 2005 to reduce developers’ commitment. The trigger price was set as low as 80% of the reservation price, the deposit was standardized at 10% of the trigger price with a cap of $50 million, and the time between a successful trigger and the auction was shortened from 10 to 7 weeks. But these measures only had a short-term effect. Three sites were triggered and the auctions held in September 2005, but no other sites were triggered in 2005-2006.
Developers saw two solutions: further lower the trigger price or reinstate scheduled auctions. But my colleague Professor Stephen Chingoutlined a far better solution.
Professor Ching argued against lowering the trigger price to avoid selling land too cheaply. He feared if a low trigger price was set then result in the highest bid in the subsequent auction may failing to reach the reservation price. Government would then have to withdraw the site and precipitate negative market sentiment.
The government was leaning towards reinstating scheduled auctions (at least for some sites), presumably because it also believed that the ALS was not working as intended. But this concerned Professor Ching for two simple but related reasons. Scheduled auctions would set back the goal of letting the market drive the land sale process and once again put the government back in the driving seat of putting land up for auction. The government would once again be at risk of having to withdraw a site if the bid price ended up lower than the reservation price, a situation it had wanted to avoid from the very start.
Professor Ching pointed out that one of the problems with ALS was it suffered a classic problem of the “winner’s curse”. The central idea is that the true value of a site is uncertain because every developer has his own assessment and none can be certain that it is correct.
In an open auction the bids of all developers are known to all. Developers can estimate their bids using information not only from their own assessments but also those of other bidders.
If the offers are sealed then no developer knows what the others are bidding.
In a sealed offer each developer does not know what the other developers are offering.
Under the ALS you get a mixture of the two systems to the detriment of the trigger developer. Developers make offers based only on their own assessments of the value of the site and the developer with the highest valued assessment will trigger the auction. He is then obliged to make this offer or stand to lose his deposit. But while all the other developers know he has made the highest offer, he does not have information about their assessments of the site. This is the “winner’s curse”.
Developers are of course aware of this problem even if they have not heard the term “winner’s curse”. To avoid falling into this trap they submit offers that are lower than their own assessments. This explains why sealed offers submitted by developers are very low and often fail to trigger an auction. The conservative offers work against the seller’s interest. The “winner’s curse” backfires and becomes the “seller’s curse”.
Professor Ching observed that the triggering developers’ concern over the reasonableness of their bids could be partly alleviated if the developer knew that at least one other developer was willing to make a similar offer – a situation that occurs in auctions. He therefore proposed to modify the triggering mechanism for an auction to require two offers instead of one. An auction would only be triggered when two independent offers each exceeding 80% of the reservation price had been made.
Proposal For Amending the ALS
If the government received an acceptable offer, both the developer and the public would be informed. The identities of the developers and the amounts offered would not be disclosed. Consider Professor Ching’s reasoning in the following two situations:
Would a second developer be more willing to make an offer high enough to trigger an auction knowing that an earlier offer had been accepted? Very likely because he now could surmise that two somewhat similar offers were on the table. Each of them would be committed to making a bid in the auction equal to the accepted offers they submitted, or stand to lose their deposit. The probability of over offering would be approximately halved because there would be another developer who was very likely making an acceptable offer of similar magnitude. This would ease the “winner’s curse” and induce the second developer to bid more aggressively.
Would an initial developer be more willing to make an offer high enough to be acceptable even when no such offer had yet been made? Any developer making a first offer would know that an auction would not be triggered even if he made a high offer unless the government accepted a second offer. The two developers would have similar commitments and the relevant considerations would be the same as in the previous situation. The “winner’s curse” would again be eased.
Professor Ching recommended that the first developer to make an acceptable offer should be required to pay a deposit only when a second acceptable offer was received. He also recommended that the deposit be halved between the two developers. This would further ease the “winner’s curse” and encourage more aggressive bidding. He also suggested that the government should pay interest on the deposits.
Another recommendation was to allow the first developer the option of triggering an auction alone without waiting for a second acceptable offer. In this case the first developer would have to pay the full deposit. If he chose to wait for a second acceptable offer, then he would only have to pay half the deposit.
It is important to acknowledge that the low number of Application List sites that were successfully triggered in previous years may have been the result of design flaws in the system. These flaws can be remedied. But it would be incorrect to conclude that the market driven system has failed or that there has been collusion among developers. This would be far too sweeping a conclusion. Reverting back to scheduled auctions would be premature and carry other risks if market sentiment turned negative either because economic prospects had dimmed or policy uncertainty had increased.
Over the past few weeks we have considered the history, problems and challenges in both subsidized and private housing. Now the question is, what is to be done about Hong Kong’s long-term housing strategy? Should there be homes for Hong Kong residents only?
(to be continued)
Stephen Ching, “Unlocking the Winner’s Curse in the Application List System”, HKCER Letters, Vol. 85 January-March 2006.
S. K. Wong, C. Y. Yiu, M. K. S. Tse and K. W. Chau, “Do the Forward Sales of Real Estate Stabilize Spot Prices?” Journal of Real Estate and Financial Economics, (2006) 32, 289-304.
Y.C.R. Wong, “Why there is Speculation in the Market for Pre-Sale Housing Units” (新樓預售為何會有炒賣), HKEJ, 16 and 23 November 2011)