Ditto with government publications. The recently-released Singaporeans in the Workforce, issued jointly by the Manpower Ministry and the Statistics Department, had the facsimile of an honest statistical appraisal, but with every fourth sentence sounding highly defensive, it ended up looking like a meretricious rag.
Reading it, I got the impression it was a response to the rising criticism over the growing income inequality in Singapore, one of the issues that had set the tone for the general election in May this year. Peppering the report were regular references to the hoped-for takeaways: (a) things are improving and (b) things aren’t as bad as elsewhere.
Suspicion was aroused the moment I tried to scroll to the end of the report. A half-decent report that claims to interpret statistical data should have statistical tables in its appendix, the same way a company’s annual report should include a thick second section detailing its financial statements and notes after having presented its fancy bar charts in its front section (the rule of thumb being that every claim made in the fore section should be supported by an exhaustive table in the rear half). I scrolled. There was no appendix other than a half-page listing the many help schemes the government has laid out (“Workfare bonus”, “Utility rebates”, “Baby bonus” etc).
Here I discuss some of my key impressions:
Improvement in individual incomes
Section 5.1 of the report had this statement:
5.1 Despite experiencing three recessions over the decade, the median monthly income from work [gross salaries] of Singapore citizens in full-time employment grew by 29% or 2.9% p.a. from $2,000 in 2001 to $2,588 in 2010. Even after taking inflation into account, the median income rose in real terms by 11% or 1.2% p.a. over the period. Virtually all of the real income growth occurred in the later half of the decade.
The third sentence reiterates something that is common knowledge by now. Prior to the general election of 2006, median individual incomes had stagnated. Worse, those in the bottom two or three deciles suffered declining incomes.
(You would also have noticed the highly defensive opener “Despite experiencing three recessions”. . . )
Ah, but Section 5.2 said:
5.2 Even for low-income Singaporeans at the 20th percentile of incomes, the nominal income rose by 17% or 1.7% p.a. from $1,200 in 2001 to $1,400 in 2010. Nonetheless, after accounting for inflation, real income growth was flat.
So, the “improvement” for low income citizens was that of going from declining real incomes (after adjusting for inflation) to flat. Overall, since the median for all full-time employed citizens increased in real terms, this suggests to you that inequality would have grown – but it’s not a conclusion that the report wants you to reach, as you would have noticed from its silence.
Note also, that the above numbers refer to full-time employed citizens. They exclude part-timers and the unemployed. Thus, the “income growth” this statement speaks of must be carefully understood to reflect the situation of those who are in full-time employment. A society can well have increasing income for those who are employed, but if more and more people are in part-time employment, or unemployed, the overall situation is less rosy than depicted by the initial figures.
Is that the case? With respect to employment, the proportion of citizens aged 25 – 64 in employment has grown over the last ten years. It was 73.7% in 2001, and 77.0% in 2010. Table 2 of the report gave these figures.
Section 3.4 said:
3.4 Singapore citizens had one of the highest employment rates internationally. Nearly eight in ten (77%) Singaporeans aged 25 to 64 were employed in 2010, surpassing economies such as Hong Kong, Taiwan, South Korea, Japan, the United States, Canada and the United Kingdom.
As for how many of these were in full-time or part-time work, Section 4.3 added:
4.3 Nine in ten employed citizens were in full-time jobs in 2010. While the majority of women were also in full-time employment, they were more likely to work part-time7 (14%) than men (6.3%), primarily due to the need to manage work and family responsibilities.
What we have is a snapshot, that currently 6.3% of men and 14% of women work part-time. I couldn’t find from the report any indication of a ten-year trend. Without this and corresponding data either about part-time incomes over a ten-year period, it gets difficult to assess the overall incomes picture.
At this point I would also caution readers about the crowing within the report, when it said “Singapore citizens had one of the highest employment rates internationally”. I will come back to this further down.
Improvement in household incomes
The Highlights (or executive summary) of the report said:
The growth in individual incomes, coupled with an improving employment rate, also boosted incomes of citizen-headed employed households. The median monthly household income from work per household member grew by 40% or 3.4% p.a. in nominal terms and 20% or 1.8% p.a. in real terms from 2000 to 2010. Households at the bottom 20th percentile had flatter but still positive growth of 34% or 3.0 % p.a. in nominal terms and 8.1% or 0.8% p.a. in real terms.
Read carefully and you will be alerted to a measure – household income in households with at least one employed person, headed by a Singapore citizen – different from what was discussed above (individual incomes). That this definition requires 15 words should make us sit up and notice how contingent the subsequent numbers are. They aren’t measures of households in general, but measures of certain kinds of households.
This claim is repeated in Sections 5.4 and 5.5, followed by a Table 5:
5.4 With the improvement in employment rate and higher median income for employed citizens, the median monthly household income from work per household member among citizen-headed employed households grew by 40%, or 3.4% p.a., from $1,083 per household member in 2000 to $1,520 per household member in 2010. After taking inflation into account, the income grew by 20% or 1.8% p.a.(Table 5).
5.5 Monthly income from work of citizen-headed employed households at the 20th percentile also grew from $560 per household member in 2000 to $750 per household member in 2010. In real terms, this was an increase of 8.1% or 0.8% p.a. (Table 5).
So, even the poorer households saw life improve? How to explain the discrepancy with the earlier mention that at the 20th percentile of individual incomes, growth after inflation over the ten-year period, was nil?
The answer probably lies in the constitution of the household. It would appear that there was a marginal increase in the average number of employed persons per household, because
- more members of a household found work compared to ten years ago (which would be consistent with a 3.3 percentage point increase in employment rate among citizens aged 25 – 64), or
- fewer working members of a household are now moving out to form their own households, or both.
The latter points to some additional thoughts.
I mentioned above that maybe we shouldn’t crow that at 77%, we had more Singapore citizens working than in some other countries, for it can easily beg the question why we need to have so many people working to produce a standard of living not necessarily higher than those other examples. It may reflect our poor productivity. In fact, I pointed out in an earlier article Kaya toast lowers Singapore’s productivity, part 2, that we had the longest working hours and one of the lowest productivity rates in a list of developed economies.
More people working, and working the longest hours. Is this something to be proud of?
Moreover, work isn’t everything. More of it means less time for families, creative leisure and charitable causes. Our society becomes less rounded as a result. A very simple question would be this: Would our birthrate be better if we had less people working? As any zookeeper knows, when animals are stressed out (whether due to crowding in cages, lack of stimulation in the environment, no respite from the constant glare of visitors), reproductive rates fall.
Sections 5.7 to 5.9 of the report trumpetted the government’s generosity:
5.7 The redistributive effect of government transfers and taxes was similarly seen in the Gini coefficient. In 2010, government transfers and taxes reduced the Gini coefficient among citizen-headed employed households from 0.465 to 0.446 (Chart 5).
5.8 Citizen-headed employed households have a lower Gini coefficient than resident-headed employed households. Among resident-headed employed households, the Gini coefficient was 0.454 in 2010 after government transfers and taxes, compared to 0.446 for citizen-headed employed households.
5.9 Singapore nevertheless has a high Gini coefficient. This has parallels with several other global cities, such as Hong Kong and leading US cities like New York, Washington, Chicago and Los Angeles, all of which have Gini coefficients above 0.5.
I don’t know about you, but what is most significant from the chart is not the 2010 numbers that the text speaks of, but the relentlessly upward trend over the decade, with or without government transfers. There’s no reversal in sight. Consider this: Even with government transfers, we were more unequal as a society in 2010 than we were in 2000 without government transfers.
The text got me a little hot under the collar for several other reasons. Paragraph 5.8 said that inequality was greater among resident-headed employed households compared to citizen-headed employed households. The former (resident-headed) category means citizens + Permanent Residents; the latter category is citizens only. Feeding the schadenfreude that look, others are even worse off than we are! is quite meaningless. Because the state picks and chooses who becomes a Permanent Resident and who does not, the Gini coefficient for them is not a “natural” condition. In any case, look at the data when I re-present them:
The Ministry of Manpower’s website says that Workfare (the primary mechanism for government transfers) is for citizens only, so the more appropriate comparison is between the first two numbers (0.465 vs 0.454) rather than the second and third (as in paragraph 5.8 of the report). This means the inclusion of Permanent Residents lowers the Gini coefficient; they provide an equalising effect. Another way to say it is: There is less inequality among Permanent Residents than among citizens. Which is contrary to the impression that paragraph 5.8 tried to make.
The comparison with other cities is also problematic. Firstly, why choose to compare with cities and not countries? If you find the cost of living too high in say, New York or Washington, you are free to move to another county or state. Poorer Singaporeans cannot do likewise since we have no hinterland. How comforting is the statement, then, that those in other cities are worse off?
You might say, well, the situation is similar for a Hongkonger. And yet, if you give it a little thought, you’d find the report’s assertion even more troubling. How so? It’s got to do with the narrow category of citizen-headed employed households. Are the figures from those other places also for citizen-headed employed households? Or all households?
The report did not cite any source for paragraph 5.9, so I could not check. But a websearch of Hong Kong statistics indicated that in the territory, their Census Department uses different bases. It speaks of ‘usual households’ and ‘mobile households’.
This webpage from the Census Department explains:
The Hong Kong Population is measured on the definition of Resident Population, which comprises Usual Residents and Mobile Residents. Among the total population at mid-2011, 6 905 400 (provisional) were Usual Residents and 202 700 (provisional) were Mobile Residents.
“Usual Residents” include two categories of people: (1) Hong Kong Permanent Residents who have stayed in Hong Kong for at least three months during the six months before or for at least three months during the six months after the reference time-point, regardless of whether they are in Hong Kong or not at the reference time-point; and (2) Hong Kong Non-permanent Residents who are in Hong Kong at the reference time-point.
“Mobile Residents” are Hong Kong Permanent Residents who have stayed in Hong Kong for at least one month but less than three months during the six months before or for at least one month but less than three months during the six months after the reference time-point, regardless of whether they are in Hong Kong or not at the reference time-point.
Further surfing revealed that in Hong Kong household income surveys are collected from households of Usual Residents only, though this term encompasses about 97% of the total population of Hong Kong. As this government factsheet says (the numbers are slightly different from above because they are from a different year):
Population Size: At mid-2009, the population of Hong Kong was 7.00 million, including 6.80 million Usual Residents and 0.21 million Mobile Residents.
This means that Hong Kong’s household income data (and Gini coefficient derived from there) is comprehensive.
In Singapore’s case, the data that we’ve been discussing mostly referred to citizens and their households, and as you can see from this table I obtained from our Statistics Department website, citizens make up only about 64% of total population.
Why does this matter? Because in Singapore we have some 900,000 very low-wage workers on Work Permits. If we included their “households”, I suspect it would further aggravate the inequality depicted in the Gini coefficient. Between Singapore and Hong Kong we may not be comparing like with like. Conceptually, I suspect it’s like this:
The Singapore data only represents those in the red bulb, whereas the Hong Kong data appears to represent all “Usual Residents” households in the territory. The Singapore data excludes households with no employed persons, and also excludes low-wage Work Permit holders. That it produces a Gini Coefficient slightly more flattering than Hong Kong’s should hardly be a surprise. When the report spoke of Singapore’s Gini at 0.465 (or 0.446 after government transfers and taxes) being “better” than Hong Kong’s above 0.5, is it meaningful?
* * * * *
It is nagging doubts like these that taint the report. Is it a statistical analysis or a piece of propaganda?