Tuesday, July 14, 2009

Demographic Link Dump

-In a truly fascinating article from last fall, Mara Hvistendahl looks at the causes and consequences of China's gender imbalance. The gender imbalance in Asia in general and China in particular will be one of the most important social phenomena in this century and this article does a great job of exploring it. Highly recommended.

-Ulster's Dooomed is a blog devoted to what, according to the blogger, is the inevitable nationalist and Catholic majority in Northern Ireland. Although he obviously writes from a certain point of view, if you dig in the archives there's plenty of interesting political demography there, including district-by-district demographic analysis.

-For those who don't know about it, the UN's World Population Prospects Database is not a perfect resource for global demographic data, but it's the best we've got. It is more reliable than the CIA World Factbook and even when you have doubts about the accuracy they make it easy to figure out what their source is. With regards to their projections, they do have a track record of overestimating population growth and underestimating fertility decline, so caveat emptor.

-The South Koreans are getting very worried about their demographic prospects. Although the South Korean population is still comparatively young, they're aging quicker than anyone. Key quote: "The percentage of people in South Korea aged 65 and over will reach 11 percent next year, lower than the 15.9 percent average for advanced nations, but wil soar to 38.2 percent by 2050, surpassing the 26.2 percent average for advanced nations by a huge margin" Although they're aware of the problem and done some modest attempts to adress it, their pro-natalist policies have so far been spectacularly unsuccessful.


Anonymous said...

For which side did the Ulster Romanians flee. Was it the protestant side or the catholic side

Aslak said...

From what I gather, it was extreme loyalists -They used slogans like "White Brits only", which obviously the nationalist side wouldn't use. I don't know how representative for the loyalists that really is though.

Anonymous said...

I ask because i wonder which side immigrants choose. This would indicate the nationalistic side is more friendly

Anonymous said...

Maybe the immigrants are smart enough to get involved in the local stupidity. As a descendant of an southern protestant Irish immigrant I can honestly say that any pride I had in that disgusting little island has long ago been bombed and marched out of me. But the real diseased IRA and the prancing Ulster marching/rock-throwing society do serve a purpose; they teach sensible Irish young people of both denominations that they need to escape that venomous land and they give immigrants a completely uncivilized land to colonize.

I think I mentioned this on the Ulster is evil blog but I still remember what was in the news in the days before 9-11. The Protestants were throwing rocks and the odd bomb at little grade one girls who dared to put their Catholic feet on Protestant sidewalks and the IRA has recently been wanting to get back into the bombing business because they are sick of AQ having all the fun.

Down with Ireland. Republic and UK chunk.

Anonymous said...

Thanks for this post. I've been wondering about the UN Pop vs. CIA factbook thing - I follow the updates for both, esp. regarding TFR's. Obviously their numbers differ somewhat, and both sources do shift them around a good bit, so I was wondering, how much margin of error should we assume in TFR's, esp. for developing countries? And if the UN figures are better, why? It seems like the variance between US and UN numbers are pretty random - for some countries one gives higher values, for others the other.

Aslak said...

Most demographic data for developing countries come either from surveys made every 5th year or or censuses which are usually held every tenth year. This Demographers then make projections based on these data. In theory, the data should be pretty reliable and in most countries I think they are, but there's a lot of uncertainty in certain countries. In very diverse countries it's hard to get representative samples for surveys and some countries haven't held a census in decades. Politics often intervene, not to mention wars. When you see data for countries like Somalia or Afghanistan for instance, the data are really little more than educated guesses, but it's the best we've got. In general, data for the least developed countries represent more a general sense of what the situation is than really accurate data.

The UN collects its own data while the CIA gets its data from the US Census Bureau's International Data Base (http://www.census.gov/ipc/www/idb/) Accessing that database directly is by the way better than the factbook, since, among other things, it lets you look at sources.

Now, both are pretty good sources but I prefer the UN database for a couple of reasons: First, it updates all of its database every two years while the census bureau has irregular and partial updates, ie. they update just a few countries at the time, which makes things harder to follow and more difficult to know hen the estimates end and the projections begin. It also means that for certain countries, the census bureau might be more up to date. I get the impression that the Census Bureau is generally better with more "important" countries, than with less significant ones.

Second and more importantly, the UN works with five-year averages while the Census Bureau does not and makes separate estimates for each year and each cohort. In theory, the latter may sound more accurate, but your estimates are only as good as your data. It's a lot easier to get a five-year average right than single years, especially when you're often working with very approximate data.

Aslak said...

Short version: The UN is generally a little bit more updated and less reliant on overly detailed projections.

Anonymous said...

Thanks a lot, Aslak. I also note with interest your saying that the UN numbers tend to underestimate fertility decline - I can see where there might be political and professional reasons to be cautious predicting such declines - after all, for the developing world a prediction of lower fertility is going to count as "good news", and one doesn't want to over-promise or make it look like the battle has already been won. I think the UN sees world TFR hitting 2.1 around 2040 - do you reckon it might go significantly faster? It seems to me that developing-world fertility outside of Sub-Saharan Africa is headed rapidly down, but in places like Nigeria or Congo it's stubbornly high and stable, making global sub-replacement fertility hard to see anytime soon.

I'm very glad to hear your thoughts and appreciate the blog very much

Aslak said...

Well, I want to be cautious. Predicting future fertility is extremely difficult and I think the most honest answer is that noone really knows. If you look at the Middle East, you had, and still have in some countries, fertility rates that were high and stable and then plummeted really fast. I think there is good reasons to expect fertility to decline in Subsaharan Africa, but I wouldn't want to venture a guess about how long it'll take. I think the UN projections are fairly reasonable, but a lot of reasonable projections have completely missed their target in the past.

Also, you should note that the replacement rate in Subsaharan Africa is significantly higher than 2.1 because of mortality rates. The actual replacement rate varies from country to country, but I've seen estimates for African countries (I forget which ones) that were somewhere around 2.7

Anonymous said...

Hi Aslak,

What do you reckon replacement fertility is world-wide? Is 2.1 only really valid for high-income low-mortality countries?

Aslak said...

You're right that 2.1 really is only for the developed world. It actually is usually a little bit lower, in the 2.06-2.08 range depending on the country. As for the global replacement rate, it should be somewhere around 2.3-2.4 not that far off from global TFR at 2.6! Obviously, global TFR might go down if infant mortality decreases. I'll put up a post on this sometime next week.