Revisiting the Questions for Written Answers: 12 February 2007

In an earlier post, I set out my Questions for Written Answers filed for 12 February 2007. This post explains why I filed them, and my thoughts about the results.

The first WPQ was about the median and mean monthly household per capita income figures from 2001 to 2005 (the Minister kindly provided the figures for 2001 to 2006), for all resident households in Singapore and also for each quintile (20% divisions) of resident households broken down by income.

There were 3 tables in the answer. Unfortunately, I simply can't get the tables to display properly on this blog. So I'll simply state the relevant figures. If anyone wants to have the tables themselves, drop me an e-mail and I will send the complete answers to you.

I've always been curious about these figures, because up to 12 February 2007, the Department of Statistics had only ever published mean figures for household incomes, without a per capita breakdown. In the lead-up to the Budget, there was some talk about help for the "sandwiched class", and one of the criteria mentioned by commentators (including myself) for assessing eligibility was household per capita income. I for one saw this as a useful metric, because it indirectly measured the number of dependents (and hence the costs that a given income must bear) as well, which a flat household income figure does not.

But there was just a lot of guesswork involved here, because of the lack of official figures. So I filed the question, to get them out once and for all.

The numbers published raised a couple of interesting facts. Firstly, the gap between the mean household per capita income for the 61st-80th and 81st-100th quintiles was substantial -- the figure for the 61st-80th quintile was 42.92% in 2001, steadily dipping to 40.28% by 2006. Secondly, while the mean and median figures for the bottom 4 quintiles were relatively close, the median for the top quintile in 2006 was $3,940, compared to a mean of $5,090. In percentage terms, the median was 79.81% of the mean in 2001, dropping to 77.41% by 2006.

Taken together, these two facts suggest that there is a relatively broad and flat middle class in Singapore going all the way up to around the 80th percentile (or maybe even higher). But after that, the income seems to spike significantly after the 90th percentile. In fact, within the top quintile, there seems to be a widening income gap as well.

And that is backed by the DOS occasional paper released on the same day. In table 5 at page 6 of the occasional paper, the DOS breaks down per capita household income by decile (10% divisions). In 2006, the 81st-90th decile had a mean per capita household income of $3,120, compared to $6,880 for the 91st-100th decile.

Let me repeat this in a different way: the mean for the 81st-90th decile was 45.35% (less than half!) of the mean for the 91st-100th decile. The corresponding percentage for 1997 was 49.24%. So not only is there a widening income gap between the rich and the poor, the rich are also getting increasingly richer than the well-off and the upper-middle class.

That's one headline that the newspapers certainly didn't see fit to publish.

And since I'm on the DOS paper, another headline that you didn't see in the media was that not only are the rich getting richer faster than everybody else, they also seem to be experiencing slower price increases than the rest of the population.

If you look at Tables A1 and A2 at pages 10 and 11 of the DOS paper, you will see that the real income figures for the top quintile for 2006 were higher than the nominal figures, suggesting that their purchasing power had actually increased since 2000 (since the real figures were based on 2000 dollars). This should be contrasted with everyone else in 2006, whose nominal figures were higher than the real figures. This difference arose because the DOS does compile different CPIs for the lowest 20%, middle 60% and top 20% of households, and these different CPI figures were used to calculate the real income figures in these tables.

The DOS' Monthly Digest of Statistics for January 2007 is consistent with this conclusion, at least with respect to 2005 and 2006. Using 2004 as the baseline year, by June 2006, prices for the top 20% had risen only 0.2% since 2004, versus 2.5% for the bottom 20% and 1.3% for the middle 60%. That would explain why the complaints about the rising costs of living have appeared to come predominantly from the lower-income.

The second WPQ was about the mean and median monthly amounts of CPF funds used for mortgage payments. I filed this question because I was concerned about people using the increased CPF contributions they were going to get for mortgage payments, on the theory that people may maximise the use of CPF contributions for mortgage payments so as to increase the amount of cash available. If the Government then reduces the CPF contribution rate in a downturn a few years down the road, they would then be caught out.

The answer was that, as at the end of December 2006, 690,067 CPF members were making mortgage payments using CPF contributions, with the mean and median monthly payments being $662 and $500 respectively.

In retrospect, I probably constructed the question erroneously, because it did not tie in the amounts contributed by persons to CPF. But extrapolating from the DOS occasional paper on household income trends, a household at the 40th percentile would have two income earners making a total of around $3,800 per month (based on a mean per capita household income of $1,000 and a household size of 3.8). (I selected the 40th percentile on the rough assumption that the top quintile would be living in private property, hence the 40th percentile represents the approximate median of the HDB-buying population.)

Their CPF contributions to the Ordinary Account (which is what can be used for housing) would be around $836 if the income earner(s) are below 35, and $760 if they are below 35-45 (and lower for older workers). But the mean and median figures of $662 and $500 are for individual CPF members, whereas the figures of $836 and $760 are for two income earners combined.

This does suggest that people are in fact maxing out the use of their CPF contributions to pay for mortgages. That is worrisome, because the Government has shown a historical willingness to use CPF contribution rates as a counter-cyclical economical tool. So in a downturn, precisely at the time when people (and really, the economy) needs their cash the most, many could well be forced to use their cash to service their mortgages because of a shortfall in CPF.