Home > Economics > Just some interesting posts netizens had countered PAP’s claims

Just some interesting posts netizens had countered PAP’s claims

Talk about the power of the internet and social media. As I am personally interested in the housing policy (my undergraduate econometrics project actually shows that government policies have the greatest impact on housing prices), I am quite intrigued by this post on facebook that rebutted Minister Mah Bow Tan’s claims that Hougang resale flats are not as valuable in comparison to the ones in Aljunied. To be fair, I can’t judge on the article without thorough analysis of the data set. After all, statistics modeling is prone to abuse. Simply changing the denominator of the base year can have a great impact on GDP growth figures. Still, it’s an interesting read for statistics geeks out there. For those who are not familiar with statistical techniques, please bear with me on this post:

The link: Click here.

*****************************************************************************

No evidence that flats in Opposition-held wards are worth less

by Giam Xingli on Thursday, May 5, 2011 at 11:10pm

We refer to the report “Aljunied flats ‘fetched higher resale prices’ than those in Hougang: Mah Bow Tan” (TODAY, May 4), where the Minister for National Development gave figures to show that 5-room resale flat prices in the Hougang Town part of Aljunied GRC were 16% higher than those of Hougang SMC. These figures are misleading about the values of PAP versus Opposition-held wards for two reasons.

The first flaw in the argument is that Mr Mah looked selectively only at flats completed between 1980 to 1985, i.e. flats 25 to 30 years old. Using the same “public info” from the HDB website that Mr Mah referred to, we can see that flats 15 to 20 years old show the opposite picture of what Mr Mah attempted to paint: prices for these flats in Hougang SMC were nearly 4% higher in value than the same type of flats in Aljunied!

Choosing specific groups of flats is a poor way of controlling for “flats with similar attributes”. We use commonplace statistical techniques to control for age of flat, size of flat, type of flat, and number of stories, which are all factors that co-vary strongly with the value of a flat. We find that there is no significant difference: a resale flat in Aljunied-Hougang can be sold on average between $22.6k more to $3.6k less than Hougang SMC.

The second fatal flaw in Mr Mah’s attempt is that he only compared Aljunied GRC with Hougang SMC. Using our same methodology, a resale flat in the Hougang Town portion of Ang Mo Kio GRC is worth on average $26.5k less than a flat from Hougang SMC. Similarly, a resale flat in the Hougang Town part of Pasir Ris-Punggol is worth $18.8k less than a flat from Hougang SMC, which supports the conclusions of Mr Low Thia Khiang in his rally speech! Does this then say anything about values of flats in the Prime Minister’s and Deputy Prime Minister’s wards compared to Workers’ Party’s sole ward?

Our point is that Mr Mah’s figures and comparison portrays a biased picture of values of HDB flats in Opposition versus non-Opposition wards. In this age of higher education and free information flow, the PAP needs to be careful about the data that they pull from their hats. Note that our whole dataset uses 356 flats transacted between 1 May 2010 and 30 April 2011, which is even larger than Mr Mah’s “fairly large number of transactions”.

Please see attached figures and detailed methodology.

 

Methodology

 

We modelled flat resale prices in Hougang Town (as defined on the HDB database website; n=356 from May 2010 to Apr 2011) as a function of 1)size of flat, 2)level, 3)age, 4)type of flat (i.e., A, I, S, various Maisonette etc), and lastly 5)constituency (i.e., Hougang SMC, Pasir Ris-Punggol GRC, Aljunied GRC). Our generalised linear model (with Gaussian errors and identity link) explained 51% of the variation in the data. The variation not explained may be due to other factors such as interior design, quality of furnishing, flooring, as individual differences in seller and buyer behavior.

Next we performed a multiple comparisons test (Hothorn et al. 2008) which tested the hypotheses if resale flat prices in Hougang SMC are different from the prices in 1) Aljunied GRC, 2) Ang Mo Kio GRC, and 3) Pasir Ris-Punggol GRC, after controlling for size of flat, level of flat, age of flat, and type of flat. We report 95% confidence intervals for differences in flat value after controlling for the above-mentioned variables.

All statistical analyses were conducting using R (R Development Core Team 2011). R is an open source statistical environment which is used in academic as well as industrial research and modelling.

References

Hothorn T., Bretz F.,Westfall P. 2008. Simultaneous Inference in General Parametric Models. Biometrical Journal, 50, 346–363.

R Development Core Team. 2011. R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing.

Authors: Chong Kwek Yan & Giam Xingli

Feel free to repost. Statistical code and dataset (publicly available anyway) are available upon request.

Figure 1
Figure 2
Advertisements
Categories: Economics
  1. November 4, 2011 at 5:16 pm

    Above all, a well written post!

    • December 4, 2011 at 2:53 pm

      Thank you! But time is getting shorter at my side due to work. I would so love to write a bit more. If only God permits me more time. haha.

  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: