Friday, October 09, 2009

Mah Bow Tan needs to be slapped ...

... for making me waste time to blog yet another entry because of him :P.

That little good-for-nothing minister is hiding behind nice-looking statistics that tell us *absolutely nothing* about the state of affairs in HDB housing. Here is what he says (I'll even cite this from the PAP website rather than the news) - "Hard to get that first flat? Not so" : National Development Minister Mah Bow Tan sets records straight on claims by couples.

1) He quotes 3 different pieces of data from 3 different data points. This is what he tells us:

Data Point #1 (Oct 7 2009) - 80% first try applicants offered BTO flat. 96% second try applicants offered BTO flat.

Data Point #2 (2007) - 67% of first try applicants reject offer.

Data Point #3 (May 2008) - 52% of first try applicants reject offer.

What conclusions can you draw? Absolutely nothing, unless you make assumptions about the relevant pieces of data for each data point.

2) His conclusions from the skimpy data above require huge leaps of logic, typical of PAP claims that "if the percentage of the problem is small, then there is no problem". Let me do part of his homework for him (he, or someone else, will need to fill in the blanks):

For any given application period,
F = first try applicants.
S = second try applicants.
O = other applicants.
X = flats offered.
Y = flats taken.

Using just Oct 7 data:
X = 0.8F + 0.96S + pO where p = fraction of other applicants who are successful.

Assuming May 2008 data applies to Oct 7:
Y = 0.48x0.8F + aSx0.96F + aOxpO
where aS = acceptance fraction of 2nd try,
aO = acceptance fraction of others.

Flats remaining available for next cycle = X - Y (assuming the decision based on demand is to build).
Number of first tries with no flat at the end of this cycle = 0.2F + 0.52x0.8F = 0.616F.

To make his conclusion, one needs to know at least F and S. If the numbers are small, then maybe it really isn't a problem for the "innocent" rejectees. But if the numbers are large, then 0.2F could very well be a serious problem. Why? Because a fraction of 0.616F will go on to become the next round's S figure. Again, we have no clue what fraction that is.

How many people give up? What are the numbers for X, F, S, p, O, aS and aO for the application cycles over the last 5 years? What are the trends? Are the S and O figures growing, indicating flat supply is not keeping up with demand?
How many of the "bad" F (0.416F) applicants go on to reject their flats when they become the S figure on the next cycle?

Where the rejections are concerned, how many of F, S and O got their first-choice of flats? How many rejected the offer because they were offered their 3rd or worse choices they could pick? How many were offered choices they had not even opted for in the first place (as is typical with NUS undergrad applications)? I remember very unhappy friends who got shoved into NUS Comp Sci merely because it had showed up at the bottom of the list of choices for their study major. How many applicants were "picky"? How many were simply not offered any reasonable choice?

Come on, Mr Mah, I'm sure you and your staff can do a better job than that with statistical data! Show me the meat upon which we can have a far more meaningful discussion!

5 comments: said...

On he ground, the demand is definitely much much more than the supply of flats.

Statistics are only statistics and there are many different ways of representations of them.

Ponder Stibbons said...

Great analysis. Could not agree more with your post title...

Kaffein said...

Exactly my sentiments too. In fact you wonder why you pay so much to have so little information fed to you?

Just a glimpse of his quoted data brings absolutely no value to anyone. Perhaps only bootlickers and aye-sayers.

Paul said...
This comment has been removed by the author.
Chee Wai Lee said...

Hi guys, thanks for your comments!

singaporeshortstories -

I think the Straits Times finally released some numbers supporting what you say: something like 12,000+ applicants for 2,000+ flats for the Oct 7 2009 data. That actually prompts more questions than it answers, I don't really have the time to do the math now. I will, however, say that I disagree with your view that "Statistics are only statistics". Well-collected statistics are very important for meaningful discussions about the state of the world around us.
Ponder Stibbons -
Thanks! Don't go acting on the title though ... :P

Kaffein -
It is hard to say if Mah Bow Tan has even seen the complete dataset. For all I know, he could have just been fed with the nicest-looking data by the HDB. In any event, this is a (yet another) clear case where transparency and accountability in government is sorely lacking. As citizens, we'll need to push harder for it.