How do property prices in Singapore vary spatially and temporally?

pangyyyyy
6 min readApr 17, 2021

--

The Housing & Development Board (HDB) was set up on 1 February 1960 to solve Singapore’s housing crisis. Prior to HDB establishments, many people were living in unhygienic slums and crowded squatter settlements. Today, over one million HDB flats have been completed across the island, providing affordable housing options for generations of Singaporeans throughout the decades.

But just how much have HDB home prices changed spatially and temporally throughout the years?

Data source:

This project uses publicly available records of successfully-transacted HDB resale flat prices from Data.gov.sg. Data from Jan 1990 to Dec 2020 was downloaded and it contains over 850k rows of transactions.

Table 1. Data fields for resale flat prices data. Geocoding based on the postal code was done to obtain the latitude and longitude of the HDB transacted.

How did property prices change temporally?

Fig. 1 shows a line chart, which represents the average price of HDB resale flat each year. This plot is enhanced with the shaded area (also known as bands), which shows the minimum and maximum limits of resale price transacted each year.

Overall, prices of HDB resale flats increased between two periods (1) 1990 to 1996 and (2) 2006 to 2013. Notably, prices begin to plateau or drop after 1996 and after 2013.

1990–1996

In the early 1990s, the government viewed Singaporean homes as an asset and potential source of security for their old age. Implemented policies included the deregulation of the HDB resale market including housing loans for HDB resale flats which facilitated mobility, physical upgrading of HDB flats and neighborhoods as well as the introduction of demand side subsidies in the form of CPF housing grants. These policies partly contributed to the rapid escalation of housing prices in the early half of the 1990s. (Phang and Kim, 2013).

1996–2006

On May 15, 1996, the government introduced a suite of anti-speculation measures to curb real estate speculation. These included capital gains tax on the sale of any property within three years of purchases and reducing the house loans to 80% of property value. In early 1997, regulations were further tightened- where HDB flat buyers were limited to two loans from the HDB (where there was previously no limit) and the time before flat lessees are eligible to purchase a second new flat from the HDB was extended from five to ten years.

The initial intention of these measures was to cool the property market, but it entered a slump with the onset of the Asian economic crisis in 1997. Housing prices continued to decline until about 2004.

In the post global financial crisis period, rapid population increase, low interest rate, and high global liquidity of Central Banks led to the accelerated price increase of Singapore property. In addition to the numerous demand curbing measures, the government has been ramping up the supply of HDB flats since 2011.

Fig 2. Resale price of flat types from 1990 to 2020

Fig 2. shows that HDB resale price stabilized after 2013. These findings are consistent with a press release by HDB on 23 August 2013, titled “Measures to Further Stabilise the HDB Resale Market” (Housing Development Board, 2013). According to the article, the government has imposed cooling measures such as tighter loan-to-value limits to curb the impact of increasing demand on HDB resale price. It is likely that this slew of government measures have led to the effect as shown above.

Together, these figures show that prevailing market conditions as well existing government property laws are intricately tied to property prices on the resale market.

How did property prices vary spatially?

Fig 3. Price per square metre for 4-room flats (the most popular type of flats)

From Fig 3., we can see that HDBs in the central area, Bukit Merah and Bishan were amongst some of the highest priced flats. However, interpreting this plot might be challenging for audiences who might not have the geographical knowledge of the different town locations in Singapore, these town labels might be confusing. There are too many towns to visualise (>20), and the color could not be easily differentiated between them.

Fig 4a and b. Visualisation of housing transactions using Kepler Gl. In KeplerGl, users are able to pan around to view the data from different angles and height.

The same data was visualised with KeplerGl, an open-sourced geospatial analysis tool for large-scale datasets by Uber. It allows users to aggregate and visualise spatial data in different formats, and the current data (Fig 4) was aggregated using hex bins. The price per square metre was encoded visually by the height and the color of the bars.

Fig 4 seems to suggest that areas closer to CBD generally fetch higher resale while tower further from the CBD would fetch lower resale price. Since the CBD area is considered by many as a strategic city centre where most of the financial and business activities take place, HDB flats with a shorter distance to this area are hence more desirable and in demand.

How did property prices vary spatially and temporally?

While KeplerGl allows us to better visualise the spatial distribution of housing prices, it is not as intuitive to visualise the temporal changes.

Fig 5. Time slices in ranges of 5 years were used to capture the temporal variation. Do note that this visualize show only the quantiles.

Some observations can be made about how the spatial distribution of resale prices change over the years.

In 1990–1995, housing prices (normalized by area) appears to be randomly distributed, expect for a cluster of high resale prices in the East. After 1995, the regions with the most priciest housing have shifted towards the city center, as we can see a significant reduction in yellow and orange areas in the East and a growing cluster of yellow areas in the South.

We can also notice the development of new HDB estates in the outskirts of Singapore in the North and West from 1990 to 2020. A direct comparison between time slices 1990–1995 and 2010–2015 reveals that there are new establishment in Sembawang, certain areas in Jurong West, Punggol and Seng Kang which appeared after 2000s. These new towns are located far from the city centers, and this would likely have fewer facilities as compared to mature estate, and this might explain why their resale prices generally fall in the lowest quarter.

In the recent years, the housing prices (normalised) seems to have shifted in the north-east direction whereby certain towns like Punggol saw an increase in resale price. This might be due to increased amenities, promotion of waterfront living and new developments that might have made it more desirable to live in.

Conclusion

While each property on the resale market has its own distinctive traits, there is a mix of underlying factors that determine the spatial patterning of prices on the resale market. In particular, the importance of accessibility to important amenities (i.e. malls, MRTs) could be one key factor in influencing flat price that is worth exploring.

This project is part of the NUS Module CS5346: Information Visualisation.

References

Housing Development Board. (2013). Measures to Further Stabilise the HDB Resale Market. Retrieved from https://www.hdb.gov.sg/cs/Satellite?c=HDBArticle&cid=1383799805536&d=Touch&pagename=InfoWEB%2FHDBArticle%2FPressReleaseLayout. Accessed on 10/4/2021.

Phang, Sock Yong and KIM, Kyunghwan. Singapore’s Housing Policies: 1960–2013. (2013). Frontiers in Development Policy: Innovative Development Case Studies [Workshop, Seoul, November 21–22]. 123–153. Research Collection School Of Economics. Available at https://ink.library.smu.edu.sg/soe_research/1544

--

--

No responses yet