Plus ça change

Following on from my post on the latest ABS births data, I just noticed an amusing fact. The reason for the apparent ‘growth’ in fertility isn’t because the 2006 figure is high, it’s because they revised down their 2005 figure.

From the ABS’s October 2006 media release:

Australia’s total fertility rate increased in 2005 to 1.81 babies per woman, up from 1.77 in 2004.

From the ABS’s October 2007 media release:

Australia’s total fertility rate increased to 1.81 babies per woman in 2006, up from 1.79 in 2005, according to figures released by the Australian Bureau of Statistics (ABS) today.

What’s particularly delightful about this is that no-one (myself included) seems to have noticed it before commenting on the data.

We can only imagine the news in October 2008:

Australia’s total fertility rate increased to 1.81 babies per woman in 2007, up from 1.80 in 2006, according to figures released by the Australian Bureau of Statistics (ABS) today.

Advertisements
This entry was posted in Economics of the Family. Bookmark the permalink.

10 Responses to Plus ça change

  1. Damien Eldridge says:

    This is intriguing. I know that some ABS series are often revised, so that one needs to be somewhat careful using recent years data. I believe this happens in the national accounts data in particular. But aren’t births supposed to be registered in Australia? Even allowing for the fact that a small number of births might go unnoticed until the child is older, it is surprising that there are enough of these that they have a sizeable impoact of the overall average. What other factors would cause a revision? As an aside, it would be pretty cool to be the cause of a DOWNWARD revision of the death rate.

  2. Damien Eldridge says:

    On page 90 of the pdf file for “Births Australia 2006” released by the ABS it mentions that not all births are registered in the year in which they occur. Perhaps this is one potential cause of occassional revisions to the numbers? (It is available on the details tab of the ABS webpage that you reach from one of the links in your post).

  3. Sinclair Davidson says:

    I follow the BERD data and also look at the revisions. Some of the changes are massive. It really makes you wonder about making macro-policy on the basis to “new” data.

    Some shameless self-promotion: Brooks, R., G. Berman, and S. Davidson. “The nature and extent of revisions to Australian macroeconomic data”. Applied Economic Letters. 1998 5(3): 169-174.

    Brooks, R., G. Berman, S. Davidson and T. Tan. “Announcements and revisions of Australian macroeconomic data and their news content for Australian financial markets”. Journal of International Financial Markets, Institutions and Money. 1999 9(2): 195-215.

  4. Kymbos says:

    From personal experience, I’ll take incompetence over conspiracy every time when it comes to the ABS.

  5. Though there can be problems because the ABS gets its info from state registrars of births, and parents may fail to do this or do it in on time (the Australian Institute of Health and Welfare data is better, because it comes from hospitals and midwives, who are more organised than some of their patients), it hasn’t changed its number of births for 2005. What it has done is altered its calculation for fertility, which is an estimate of likely babies per woman based on current age-specific fertility rates. Because they now assume that women in the their 30s will have more babies, fertility rates have gone up. Perhaps there was a mathematical error in the 2005 fertility rate.

  6. Andrew Leigh says:

    I expect it’s just a late-lodgment issue. My understanding is that the latest year’s birth statistics are based on births registered up to 31 Dec. As AN points out, their data are compiled from all the state & territory BDM registrars, which makes things tricky.

    My guess is that in Oct 2006, the ABS only had data on 2005 births that were registered by 31/12/05, so they made some assumption about the number of 2005 births that would be registered in 2006, and this turned out to be an overestimate.

  7. Andrew L – The ABS hasn’t revised its 2005 number. I have the 2005 report which I downloaded last year. It says 259,791 births for 2005. That’s the same number as in the ABS spreadsheet I downloaded this week. What’s changed is that they have revised down their fertility estimate from 1.806 to 1.789, suggesting perhaps a mistake in their calculations last year. Remember these can only ever be estimates, and indeed the time series for all recent years is misleading, because they underestimate likely births to women in their 30s and early 40s.

  8. Andrew Leigh says:

    Aha – nice sleuthing!

  9. Budz says:

    Hi Andrews,
    What do you reckon about both parties claim that tax cuts won’t add to inflationary pressure? And that they will only entice more workers into the workforce increasing capacity and then decreasing inflationary pressures. I found Swann’s and Costello’s comments in their debate quite interesting in response to a poll in the Eden-Monaro that showed most people would rather an increase in expenditure on health and education as opposed to tax cuts.

  10. christine says:

    rom my memory (and I really should know this off by heart, and many apologies to David Foot if he ever reads this), Andrew N’s on the mark: the birth rates probably don’t change, because they’re for births in year/# women. Fertility rates are sort of expected lifetime # children / woman, which (I think) always involve a bit of guess work for the current year. Only really know a few years later how it actually turned out.

    It is funny, though, that no matter what last year’s fertility rate was, this year’s is always 1.81.

Comments are closed.