IS428 2016-17 Term1 Assign1 Lim Zi Yu Jouta

From Visual Analytics for Business Intelligence
Jump to navigation Jump to search

Abstract

By looking at data for unit price and vacant stock of private property, I seek to gain insights to current policy influences on the housing market in 2015 to suggest improvements for young adults like myself in need of housing to be able to better afford it without overheating the market with excess demand.

Problem and Motivation

As a fourth year student approaching graduation, my batchmates and I will be soon facing problems that most of us as students would usually not have to be concerned with. Most of us living in Singapore has had the luxury of staying with our parents without worries about housing rent, especially given the central location of our University, which facilities easy travel from most areas of Singapore. But as we step out into the working world and rely less on our parents’ support, we will soon find ourselves facing financial problems that our parents used to handle in place of us. As independent adults, most of us would take into consideration of moving out and having our own home, especially for those in committed relationships who wish to move in together having their own home.

The Singapore housing market has seen many changes since the beginnings of the Singaporean economy. Housing prices used to be low enough for the average Singaporean to own their own permanent residence, instead of opting to only rent a temporary property to keep a roof above their heads. Compared to the past where people can own properties easily, we now find ourselves unable to do that easily. The increase in housing prices compared to the increase in people’s salary in Singapore has led to people being unable to own their own private property. Hence, we seek to investigate the housing market in 2015 to assess the current situation. I will focus on the data for private properties instead of public housing such as HDB as public housing with subsidies would not portray an accurate reflection of the housing market for all individuals given the restrictions on HDB purchases resulting in ineligibility for some people. Data used will be for landed property such as the detached houses, semi-detached houses, terrace houses, and also non-landed property such as the apartments and condominiums. Executive condominiums are part of HDB property hence would be excluded from this analysis. With this, I hope to be able to come up with some policy recommendations on how to improve the situation for young adults like myself.

Approaches

I have chosen to present the property price data in the format of their unit (area) prices as the area of each property type is inconsistent across the different property types, making it difficult to make fair and accurate comparisons across the board if the absolute transacted price was used to make comparisons. To make each type of property more differentiated, I have chosen to represent each box plot with a different color to make it easier to identify. The lower whisker of unit price for the different types of private properties show that generally landed property offer the lowest unit price compared to non-landed property, with the exception of one outlier for the apartment property type. Despite that, we also see that the upper whisker of unit price for the different types of private properties show that landed property also offer the highest unit price compared to non-landed property. This shows that landed property unit price has a larger spread compared to that of non-landed property. Following that, we also observe that the non-landed properties have significantly more statistical outliers on the higher end of the pricing compared to landed properties. These two points could be attributed to the fact that foreigners may only own non-landed private properties, and also premium non-landed property in the central business district and Sentosa cove that are often owned by wealthy foreigners. The median line on the box plot graph shows the median across all property types, whereby it can be observed that non-landed properties medians are closer to the global median whereas the landed properties have a comparatively lower median than the global median.

To look at the share of private property supply, I have chosen to look at the dataset for vacant stock sorted by the type of private property. I chose vacant stock as the variable of choice to express supply in the market as ready stock, instead of supply in the pipeline which would still be under construction and yet to be ready for people to move in. Given the short time period to analyze, I feel that looking at current vacant stock would give a more accurate reflection compared to pipeline supply which would involve a lot more speculation and would have reflect more strongly in long term decisions. The percentage change in vacant stock per type is place next to that of the change in vacant stock to allow easy viewing of the graph from top down and make comparisons. Over the year, vacant stock has increased for all property types, as the vacant stock in Q4 exceeds that of Q1 in all cases. Condominiums has had the largest absolute increase and percentage change in vacant stock compared to all other types of private property. This is likely attributed to how condominium blocks are built and launched together in large numbers, resulting in large number changes for its vacant stock. All three landed property types have had a U-shaped pattern in its vacant stock over the course of the year, whereby they see a decrease in number in either Q2 or Q3 before increasing again in Q3 or Q4. The same has also been observed for apartments. It could be possible that it is a cyclical pattern in the housing market for private properties hence observed by these 4 types excluding condominiums which would see such effects eroded by changes in its vacant stock as explained above.

Infographic

Dashboard 4.jpg

Despite the different types of property and the different supply level based on the vacant stock per type, we see that the distribution of the property price per unit area is largely similar. Hence, we could conclude that the transactional prices of each property would be largely influenced by the area specifications of each type of property. Despite large changes in vacant stock for condominiums, condominium prices don’t see a wider spread in distribution compared to the other types of properties unlike what is usually expected.

The reason why such effects were not that pronounced might be due to the cooling measures in place by the government which effectively dampened demand on the housing market. Measures such as the Total Debt Servicing Ratio (TDSR) effectively restricts your ability to make loans for housing payment. Given the already high price of housing per unit area and the resulting price per property in addition to TDSR, private properties start to become out of reach for many people.

This is particularly detrimental to young adults like myself who have to service a tuition fee loan while earning the basic graduate starting pay. Restrictions set as a result of cooling measures cause people in need of housing to become unable to purchase them instead of simply affecting speculative actions is doing more harm than good.

Firstly, I believe that the TDSR can be relaxed for first time private property buyers seeking to buy their first housing. Measures to cool the housing market should not deter people wishing to buy their first housing, but instead should only affect buyers buying their second property onwards. One way to ensure this can be done appropriately is to relax for first time property buyers, only if they have no owned any sort of property before. By doing so, we can price differentiate the market to different groups of buyers so as to not cause unnecessary harm to different groups.

Secondly, private properties should be sold with a more differentiated unit area pricing so as to target different segments of the market with different property types beyond just pricing it differently due to the build specifications of each category of property. Apartments and Terraces could be priced with a lower unit price so as to reduce its transaction prices and make them more affordable as they are considered the more basic type of their class as non-landed and landed private property.

Lastly, to dampen effects of speculation and heating up of the housing market due to foreign flows of purchases, tax rates on the non-landed property could be adjusted such that condominiums could have a lower tax rate to invite more foreign purchase given the steady high vacant stock of that group of property. On the other hand, Apartments could have a high tax rate so that demand from aboard for it would not be as high to keep the market price low for local buyers.

Overall, the 3 policies suggested should allow locals who are in need of housing to be able to afford some basic form of private property without being restricted by cooling measures meant to keep the speculative heat off.

Tools Utilized

Tableau