reporter2
16-01-20, 10:55
What’s the link between developers playing golf with each other and their bids for government land auctions? An NUS study has the answer
Researchers found that bosses at developer firms who played golf together tended to make lower winning bids for residential development land put up for auction by the Government than those who did not play golf.
By JANICE LIM
Published 15 JANUARY, 2020
UPDATED 16 JANUARY, 2020
SINGAPORE — By studying the public records of all golf players in Singapore and the bids which developers put in for government land auctions, among other data sets, a group of university researchers believe that they have found “evidence of insider trading in the land market”.
The findings of their research paper were released by the National University of Singapore (NUS) Business School on Wednesday (Jan 15).
The study found that the top managers of Singapore’s real estate developers played golf with each other more frequently after the announcement of a Government Land Sales programme, and their winning bids for the tenders were 14.4 per cent lower than the winning bids put in by those who did not play golf.
The researchers estimated that the shortfall in government revenue from the lower land prices was about S$147 million on average for every year between November 2010 and May 2014.
They noted that information on land bidding could be exchanged among these managers during golf, which was then used to make more informed bids to buy land parcels from the Government. However, the study was unable to determine if the developers were behaving as a cartel.
“This paper is among the first to find evidence of insider trading in the land market and to show that insider trading causes negative spillovers to other properties in the neighbourhood,” the research paper stated.
Professor Sumit Agarwal, the lead researcher of the study, told TODAY that he has always been interested in understanding how social networks form since his days as a banker, when he observed that deals were often made on the golf course.
In a previous study on the “glass ceiling” faced by women in the corporate world, he found that women who played golf were more likely to climb the corporate ladder. Prof Sumit is the Low Tuck Kwong Professor at the School of Business, and he is a professor in the departments of economics, finance and real estate at NUS.
He added in a press release: “While information sharing is not prohibited by law, governments whose fiscal revenue relies on a large proportion of land sale revenues must be prudent of developers’ behaviour on the golf course to ensure a competitive land auction market.”
The other authors of the study are Professor Sing Tien Foo, Assistant Professor Qin Yu and Ms Zhang Xiaoyu from the Department of Real Estate at NUS.
In response to TODAY’s queries, the Competition and Consumer Commission of Singapore (CCCS) said: “As part of its ongoing market surveillance efforts to ensure compliance by businesses, CCCS will monitor market developments closely and will open a formal investigation when it has reasonable grounds to suspect the Competition Act has been infringed.”
HOW THE STUDY WAS CONDUCTED
The researchers looked at four data sets:
Golf records of all golf players in Singapore from 2010 to 2014
Government land bidding records for residential land auctions from 1990 to 2016
Transaction records of all private residential properties from 1995 to 2018
Top managers of all registered developer companies in Singapore
After a matching process to ensure that the four data sets would correspond to each other, the final sample for their study was:
774 top managers from real estate firms who played golf
103 land bids from 895 bidders
Seven Government Land Sales announcements
The sample period was from November 2010 to May 2014.
Among the 895 bidders, the researchers put the bidders' golfing behaviour into four categories:
Type 1 bidders: Those who did not golf and therefore had no information exchanges through golfing. This group was used as the reference group in the study.
Type 2 bidders: Those who played golf alone or with others who were not from real estate firms.
Type 3 bidders: Those who played golf with managers from rival companies that did not bid in an upcoming land sale programme.
Type 4 bidders: Those who played golf with managers of other real estate firms, who later competed in the same land auctions
“If information exchanges take place when they are together on the same golf course they are likely to be better informed about each other’s bidding strategies… The fact that they both choose to bid indicates that they agree with each other’s valuation. Thus, their valuations are more likely close to the true value (of the land),” the paper stated.
Real estate firms were not fixed into these categories, and might be a Type 1 bidder in one particular land auction but behaved as a Type 4 bidder in another, for example.
THE FIVE MAIN FINDINGS
Top managers of real estate companies played golf with each other more often after the Government announced the land sites that it was releasing to the private sector.
Those who played golf more often — Type 4 bidders — submitted winning bids that were lower than those who did not play golf — Type 1 bidders.
Private homes built by Type 4 bidders were sold at a price that was lower than those built by Type 1 bidders.
Private homes surrounding land parcels acquired by Type 4 bidders were sold at cheaper prices than those located further away.
The lower winning bids led to revenue losses for the Government.
Frequency of playing golf
The study found that the number of golf games developers played with each other rose 14 per cent in the first week after a Government Land Sales announcement, compared with the number of games played in the week before, and by 24 per cent in the second week after the announcement.
The researchers also found that those who bidded more actively were more closely connected to other active bidders.
“Given that active bidders’ profits are highly dependent on land bidding outcomes, they are more likely to share true information with other bidders because sharing false information may lead to punishments by other bidders.
“The intensity with which development firms’ top managers play golf together increases only after the land bidding has been announced,” the researchers said.
Lower winning bids
Managers of property developers who played golf more often with their rivals after the land sites were announced were found to be placing winning bids that were 14.4 per cent lower than those who did not play golf.
The winners’ curse — a tendency for the winning bid in an auction to exceed the intrinsic value or true worth of an item — for Type 4 bidders was also found to be 5.8 per cent lower than Type 1 bidders. The curse was often measured by the gap between the winning bid and the highest losing bid.
This means that Type 1 bidders who won at an auction tended to overbid, while Type 4 bidders who won would likely have had more information to make a more accurate valuation and reduce the effect of the winners’ curse.
This suggests that “bidders exploit the information acquired from other bidders through golfing”, the researchers said.
Lower home prices
Because of their lower winning bids, Type 4 bidders could sell their new development at a cheaper price.
The researchers found that new projects by Type 4 bidders sold for 7.6 per cent less than those built by Type 1 bidders.
While developers shared some land cost savings with homebuyers, the researchers said that they shared only part of these savings and still reaped 6.8 per cent more in profits on these projects when compared with other developers.
The study also found that Type 4 bidders were able to sell their projects faster. In the first week after a project was launched, they could sell 10 per cent more units than projects by other developers. Three months after its launch, they could sell 20 per cent more.
Effect on competing projects
The low winning bids of Type 4 bidders affected the projects surrounding the site as well.
In the first 30 days after the announcement of a land sales result, properties within 500m of land parcels acquired by Type 4 bidders were sold 9.9 per cent cheaper than those further away, the researchers noted.
However, the negative effect lasted only for three months, which could possibly be a result of overreactions by homeowners and developers, they said.
Loss in government revenue
The researchers found that the lower winning bids by Type 4 bidders resulted in a loss of 0.2 per cent of overall government revenue and 1 per cent of land sales revenue, on average.
This translates to about S$147 million a year on average from November 2010 to May 2014.
“Losses in land revenue, which could otherwise be invested in infrastructure improvements, translate into more welfare losses for the whole city,” the researchers said.
IS THERE A CARTEL OR NOT?
The researchers also tried to find out whether the developers who played golf with each other more frequently could be part of a cartel.
The assumption is that a sophisticated and effective cartel could mimic the behaviours of competitive bidders, yet ensured that its members would have a higher chance of winning subsequent land auctions after playing golf together.
They found that two bidders would be more likely to bid for the same land parcel after a golf game, but it did not increase the chances of either one winning the bid.
The researchers have two explanations: Either the cartel was ineffective or there was none at all and that both developers were competing head-on.
IS IT ILLEGAL?
Without referring directly to the study, CCCS said in its response to TODAY on Wednesday that the exchange of non-public commercially sensitive information among competitors can violate the Competition Act.
“Such exchange will affect commercial decisions which should be made by individual market players independently and reduce competitive pressure among them,” the commission said.
In January last year, CCCS fined four hotels S$1.5 million for exchanging commercially sensitive information.
Anti-competitive agreements can also include collusive actions among bidders to discuss bid prices or coordinate bid submissions, or both. The objective of such actions is to restrict competition, the commission said.
Researchers found that bosses at developer firms who played golf together tended to make lower winning bids for residential development land put up for auction by the Government than those who did not play golf.
By JANICE LIM
Published 15 JANUARY, 2020
UPDATED 16 JANUARY, 2020
SINGAPORE — By studying the public records of all golf players in Singapore and the bids which developers put in for government land auctions, among other data sets, a group of university researchers believe that they have found “evidence of insider trading in the land market”.
The findings of their research paper were released by the National University of Singapore (NUS) Business School on Wednesday (Jan 15).
The study found that the top managers of Singapore’s real estate developers played golf with each other more frequently after the announcement of a Government Land Sales programme, and their winning bids for the tenders were 14.4 per cent lower than the winning bids put in by those who did not play golf.
The researchers estimated that the shortfall in government revenue from the lower land prices was about S$147 million on average for every year between November 2010 and May 2014.
They noted that information on land bidding could be exchanged among these managers during golf, which was then used to make more informed bids to buy land parcels from the Government. However, the study was unable to determine if the developers were behaving as a cartel.
“This paper is among the first to find evidence of insider trading in the land market and to show that insider trading causes negative spillovers to other properties in the neighbourhood,” the research paper stated.
Professor Sumit Agarwal, the lead researcher of the study, told TODAY that he has always been interested in understanding how social networks form since his days as a banker, when he observed that deals were often made on the golf course.
In a previous study on the “glass ceiling” faced by women in the corporate world, he found that women who played golf were more likely to climb the corporate ladder. Prof Sumit is the Low Tuck Kwong Professor at the School of Business, and he is a professor in the departments of economics, finance and real estate at NUS.
He added in a press release: “While information sharing is not prohibited by law, governments whose fiscal revenue relies on a large proportion of land sale revenues must be prudent of developers’ behaviour on the golf course to ensure a competitive land auction market.”
The other authors of the study are Professor Sing Tien Foo, Assistant Professor Qin Yu and Ms Zhang Xiaoyu from the Department of Real Estate at NUS.
In response to TODAY’s queries, the Competition and Consumer Commission of Singapore (CCCS) said: “As part of its ongoing market surveillance efforts to ensure compliance by businesses, CCCS will monitor market developments closely and will open a formal investigation when it has reasonable grounds to suspect the Competition Act has been infringed.”
HOW THE STUDY WAS CONDUCTED
The researchers looked at four data sets:
Golf records of all golf players in Singapore from 2010 to 2014
Government land bidding records for residential land auctions from 1990 to 2016
Transaction records of all private residential properties from 1995 to 2018
Top managers of all registered developer companies in Singapore
After a matching process to ensure that the four data sets would correspond to each other, the final sample for their study was:
774 top managers from real estate firms who played golf
103 land bids from 895 bidders
Seven Government Land Sales announcements
The sample period was from November 2010 to May 2014.
Among the 895 bidders, the researchers put the bidders' golfing behaviour into four categories:
Type 1 bidders: Those who did not golf and therefore had no information exchanges through golfing. This group was used as the reference group in the study.
Type 2 bidders: Those who played golf alone or with others who were not from real estate firms.
Type 3 bidders: Those who played golf with managers from rival companies that did not bid in an upcoming land sale programme.
Type 4 bidders: Those who played golf with managers of other real estate firms, who later competed in the same land auctions
“If information exchanges take place when they are together on the same golf course they are likely to be better informed about each other’s bidding strategies… The fact that they both choose to bid indicates that they agree with each other’s valuation. Thus, their valuations are more likely close to the true value (of the land),” the paper stated.
Real estate firms were not fixed into these categories, and might be a Type 1 bidder in one particular land auction but behaved as a Type 4 bidder in another, for example.
THE FIVE MAIN FINDINGS
Top managers of real estate companies played golf with each other more often after the Government announced the land sites that it was releasing to the private sector.
Those who played golf more often — Type 4 bidders — submitted winning bids that were lower than those who did not play golf — Type 1 bidders.
Private homes built by Type 4 bidders were sold at a price that was lower than those built by Type 1 bidders.
Private homes surrounding land parcels acquired by Type 4 bidders were sold at cheaper prices than those located further away.
The lower winning bids led to revenue losses for the Government.
Frequency of playing golf
The study found that the number of golf games developers played with each other rose 14 per cent in the first week after a Government Land Sales announcement, compared with the number of games played in the week before, and by 24 per cent in the second week after the announcement.
The researchers also found that those who bidded more actively were more closely connected to other active bidders.
“Given that active bidders’ profits are highly dependent on land bidding outcomes, they are more likely to share true information with other bidders because sharing false information may lead to punishments by other bidders.
“The intensity with which development firms’ top managers play golf together increases only after the land bidding has been announced,” the researchers said.
Lower winning bids
Managers of property developers who played golf more often with their rivals after the land sites were announced were found to be placing winning bids that were 14.4 per cent lower than those who did not play golf.
The winners’ curse — a tendency for the winning bid in an auction to exceed the intrinsic value or true worth of an item — for Type 4 bidders was also found to be 5.8 per cent lower than Type 1 bidders. The curse was often measured by the gap between the winning bid and the highest losing bid.
This means that Type 1 bidders who won at an auction tended to overbid, while Type 4 bidders who won would likely have had more information to make a more accurate valuation and reduce the effect of the winners’ curse.
This suggests that “bidders exploit the information acquired from other bidders through golfing”, the researchers said.
Lower home prices
Because of their lower winning bids, Type 4 bidders could sell their new development at a cheaper price.
The researchers found that new projects by Type 4 bidders sold for 7.6 per cent less than those built by Type 1 bidders.
While developers shared some land cost savings with homebuyers, the researchers said that they shared only part of these savings and still reaped 6.8 per cent more in profits on these projects when compared with other developers.
The study also found that Type 4 bidders were able to sell their projects faster. In the first week after a project was launched, they could sell 10 per cent more units than projects by other developers. Three months after its launch, they could sell 20 per cent more.
Effect on competing projects
The low winning bids of Type 4 bidders affected the projects surrounding the site as well.
In the first 30 days after the announcement of a land sales result, properties within 500m of land parcels acquired by Type 4 bidders were sold 9.9 per cent cheaper than those further away, the researchers noted.
However, the negative effect lasted only for three months, which could possibly be a result of overreactions by homeowners and developers, they said.
Loss in government revenue
The researchers found that the lower winning bids by Type 4 bidders resulted in a loss of 0.2 per cent of overall government revenue and 1 per cent of land sales revenue, on average.
This translates to about S$147 million a year on average from November 2010 to May 2014.
“Losses in land revenue, which could otherwise be invested in infrastructure improvements, translate into more welfare losses for the whole city,” the researchers said.
IS THERE A CARTEL OR NOT?
The researchers also tried to find out whether the developers who played golf with each other more frequently could be part of a cartel.
The assumption is that a sophisticated and effective cartel could mimic the behaviours of competitive bidders, yet ensured that its members would have a higher chance of winning subsequent land auctions after playing golf together.
They found that two bidders would be more likely to bid for the same land parcel after a golf game, but it did not increase the chances of either one winning the bid.
The researchers have two explanations: Either the cartel was ineffective or there was none at all and that both developers were competing head-on.
IS IT ILLEGAL?
Without referring directly to the study, CCCS said in its response to TODAY on Wednesday that the exchange of non-public commercially sensitive information among competitors can violate the Competition Act.
“Such exchange will affect commercial decisions which should be made by individual market players independently and reduce competitive pressure among them,” the commission said.
In January last year, CCCS fined four hotels S$1.5 million for exchanging commercially sensitive information.
Anti-competitive agreements can also include collusive actions among bidders to discuss bid prices or coordinate bid submissions, or both. The objective of such actions is to restrict competition, the commission said.