The residential collective sale market jolted back to life with the sale of Shunfu Ville, Raintree Gardens and Harbour View Gardens last year with a total value exceeding $1 billion. The prospect for collective sale remains encouraging with the sale of One Tree Hill Gardens for $65 million to a unit of Lum Chang Group this month.

Amid a tight supply of private housing sites on state tenders, developers are setting their sights on the collective sale market for alternative land sources. Nearly 30 private housing sites were launched for sale in 2011, excluding executive condo sites. In comparison, less than 10 private housing sites were launched for sale in 2016.

Homeowners stand to reap attractive sale proceeds from collective sales. This is because for developments that are 10 years old or older, at least 80% of owners must agree in writing to put up the property for sale in the open market. The potential sale proceeds serves as a carrot for homeowners to sign their consent.

On the other hand, a collective sale deal is likely to fall through if the asking price is too high. Developers must reap a reasonable profit margin from the collective sale with a view of redeveloping the property into a new project for sale in the future.

Besides asking price, there are other factors that would determine the success of a collective sale. To identify these factors, The Edge Property constructed a logit model, also known as the logistic regression*. 

Based on the model, there are three variables that increase the success of collective sale. They were statistically significant at 5% level: 

  1. Age (older developments have better success rate);
  2. Number of units in the development (smaller developments have better succes rate); and
  3. Premium of new sale and sub-sale price over the collective sale/ asking price, within the same year and district.

Interestingly, tenure (freehold versus leasehold) and location (whether the property is located in Central or Non-Central Region) were statistically insignificant. 

 

Former HUDC estate Rio Casa (left) and Eunosville (right), which were put up for collective sale, have a success probability of 79.7% and 62.9% respectively based on the logit model

Photos: Knight Frank, OrangeTee

 

Methodology

The dependent variable "y" is the success of the collective sale, where "1" represents success and "0" represents failure.

The independent variables "x" comprise:

  • Tenure where “1” denotes freehold and 999-year-leasehold, and “0” denotes leasehold. As developers have to fork out millions of dollars to top up the property’s balance lease to a fresh 99 years, a freehold development may have better odds in collective sales.
  • Age. Homeowners in ageing developments could be more motivated to sell due to higher expenses to maintain the property and to repair wear and tear. Landlords may also face difficulty to rent out their units due to competition from newer projects.
  • Units. This refers to the total number of units in the development which is a proxy for the number of homeowners in a development. For a collective sale to materialise, a majority of homeowners must reach a common agreement in terms of asking price and apportionment of the sale proceeds. It would be easier for a small group of homeowners to come to an agreement.
  • Segment where “1” denotes Central Region and “0” denotes Non-Central Region. Due to scarcity of vacant sites in the Central Region, developments located in this segment are more likely to be sought after by developers.
  • Enhancement**. This refers to potential plot ratio enhancement. When the plot ratio in the prevailing master plan is higher than the development’s existing plot ratio, the land can be redeveloped into higher GFA although development charge is payable. 
  • Premium. This refers to the premium of new sale and sub-sale price over the collective sale price in the same year and district. This is a proxy of the potential upside and profit margin that the developer stands to reap if they were to redevelop the site into a new project for sale in the future. 

 

Factors that matter – Model 1

The model comprises 212 successful collective sales and 38 unsuccessful deals from 2005 to 2016. Unsuccessful deals refer to developments that were put up for collective sale but remain unsold to date.

These deals were selected as complete information for the independent variables can be obtained from reliable sources including URA and the media. They do not represent the actual number of successful and unsuccessful deals in that time period.

The logit model finds that “Age”, “Units” and “Premium” to be statistically significant variables at 5% level (see Table 1).

Table 1: Logistics regression, Pseudo R2 = 0.35

Success

P>|z|

Tenure

0.18

Age

0.00

Units

0.00

Segment

0.09

Enhancement

0.77

Premium

0.02

  _cons

0.34


Factors that matter – Model 2

Model 2 defines success as obtaining the required consent from owners for the site to be put up for collective sale, even if the deal fell through. This is because obtaining the required consent from owners is the first hurdle to cross towards a successful collective sale.

There were 250 success cases in model 2. They comprise the 212 successful and 38 unsuccessful deals in model 1. Failure in model 2 comprise 50 randomly selected developments island-wide that were never put up for collective sale.  

Model 2 finds “Age”, “Units”and "Enhancement" to be statistically significant at 5% level (see Table 2).

Table 2: Logistics regression, Pseudo R2 = 0.55

Success

P>|z|

Tenure

0.95

Age

0.00

Units

0.00

Segment

0.70

Enhancement

0.00

Premium

0.29

  _cons

0.06

Conclusion

The logit model found the following factors - age, number of units, plot value enhancement and the premium of new sale/ resale prices over the collective sale price – to increase the success of a collective sale. However, potential homebuyers looking to reap big bucks should note that collective sale is not a sure bet and a long process if it does materialise.

Tan Hong Boon, regional director for capital markets at JLL, says a collective sale process can take between 18 months to four years. “Discontentment over asking price is a common source of resistance. There are owners who are against the sale because finding a replacement property of similar size and location could be a challenge. Some owners may be liable to pay seller’s stamp duty if the collective sale takes place within three or four years from the date they bought the property. Foreign nationals would also have to fork out additional buyer’s stamp duty if they were to buy a replacement property,” Tan explains. However, when the asking price is too high, the collective sale would be unfeasible for potential developers.

Beyond a development’s inherent traits, there are also external factors that would impact the success of a collective sale including market cycle and availability of quality sites on the Government Land Sales programme. 

Potential homebuyers are advised to research the development’s zoning, land area, potential plot ratio enhancement, and if there were past collective sale attempts and why they failed. “Finally, the sale has to be carried out in good faith, yield the best price, with proper apportionment of the sale proceeds, so as not to create any ground for objection. A good collective sale committee, lawyer and consultant play an important role in the success of a collective sale,” Tan says. 

Which developments have the most en bloc potential based on this regression model? Watch this space!

 
* A logit model is usually adopted when the dependent “y” variable takes the value of “1” or “0” which represent binary outcomes such as success or failure. The probability density function of logit distribution is similar to the normal distribution, except that the latter has heavier tails, denoting higher kurtosis. 
**As information on existing plot ratio and GFA for old developments are not readily available, total strata area was adopted in place of existing GFA as a proxy for the model.