A friend asked me – with all the hoohah on “fake news” – got “fake statistics” or not?
Well, from my experience analysing statistics for the last two decades or so – “fake statistics” are very very rare.
What we have are often:
- no disclosure of the statistics
- partial disclosure of the statistics
- omitting statistics
- changing the definition of the statistics
- changing the time period of the statistics
- not reporting statistics using international norms
- changing the base population of the statistics
- a combination of the above
Let me try to give some examples to illustrate the above.
“no disclosure of the statistics” – the HDB does not breakdown the price of HDB BTO flats into contruction, land and other costs. All that we know is that land is charged at market rates
“partial disclosure of the statistics” – the employment growth statistics are broken down into locals and foreigners, but not Singaporeans
“omitting statistics” – the GIC’s annualised returns are for up to 20 years in US$ – but not the annualised return from its inception in S$. In contrast, Temasek also discloses its annualised return from its inception in S$
“changing the definition of the statistics” – a part-time worker used to be defined as working not more than 30 hours a week – this was changed to 35 hours. So, arguably, by the stroke of a pen – both the categories of full-time and part-time workers’ median wages increased
“changing the time period of the statistics” – “A three-year moving average shows that out of every four extra PMETs employed in the past three years, three were locals. Five years ago, the figure was lower: one local out of every two extra PMETs” (“Parliament: Local PMETs facing brighter job prospects, says Lim Swee Say”, Straits Times, Nov 6) – Why is it that there was no mention of the actual number of PMET jobs created in the last three years?
If we account for the estimated 90,000 new PRs and 60,000 new citizens granted in the last three years – how many of the jobs to locals went to Singaporeans?
This was the first time that a “three-year moving average” was used in labour statistics in this category – such that it is almost impossible to make any meaningful comparisons over various time periods in the past
“not reporting statistics using international norms” – including Transfers to Trust and Endowment Funds in the Budget reporting, which under IMF fiscal reporting guidelines should not be charged as one-time expenditures in the Budget – the Budget surpluses may actually be much higher
For example, if we add back the Trust and Endowment Funds of $3.6 billion – the Budget surplus would rise from $5.2 billion to $8.8 billion in FY2016.
According to the Department of Statistics’ Monthly Digest of Statistics – the total Budget Cash Surplus/Deficit (M130611 – Government Finance, Annual) in 2015 was $11 billion (IMF fiscal reporting guidelines)
“changing the base population of the statistics” – 99 per cent of MediFund applications (not the number of patients) were successful (what about those who were told that they need not apply because they clearly do not meet the criteria)
HDB owners use less than a quarter of their income (CPF almost enough) to pay for the entire monthly mortgage payment – means very affordable – those who can’t afford to buy or couldn’t service their mortgage would not be in the population from which these statistics were derived
“a combination of the above” – by changing the definition to include the employer CPF contribution in gross wages and focusing on this in the narrative, particularly in the news media, which is also not in line with international norms