Thursday, March 10, 2011

Desperately Seeking Malthus

For those who are quibbling with my miasmic look at City of Pittsburgh population trends (gee thanks Tim); it may be worse than even I was thinking of when I wrote that.  I apologize if others have noted this, so much Census chatter all around. If you look at the population under age 18 for the city, this is what you get.

18 and over268,055255,905-12,150-4.5%
Under 1866,50849,799-16,709-25.1%

If that does not shock you. Keep looking at it until it does.

Why is this really important?  The overall population loss in the city masks some other things going on. Namely I think all would agree there has been a net gain in college enrollment and student residents in the city.  So the loss in the remainder in the population is higher than the top line number would have you thinking.  The population under age 18 is a decent proxy for what is going on in the household population in the city of Pittsburgh. 


Anonymous The Wiz said...

A couple of guesses as to why the under-18 demo is dropping so much. First off, most people of child bearing age look for great school districts for their kids. Ask any realtor what the main Q is when selling to young couples with kids. Thus they leave the city schools.

They also want a safe neighborhood and decent sized yard for the little angles to play in.

As you become a daddy or mommy you have much less time for enjoying all the amenities the city has to offer.

Many professionals leave the entire area for better climates and better pay after their college years are done. They may stay a couple of years to get experience and then vamoose.

Older "empty nesters" move back into the city after the kids are gone, displacing the younger generation.

How does the county compare as to that demographic? I'm guessing a lot better.

Thursday, March 10, 2011 6:36:00 PM  
Anonymous MH said...

I wonder how much they've spent on Pittsburgh Promise compared to the subsidies on condos? But, the census is kind of confirming my personal experience. People who aren't rich or poor move to Mount Lebo or something when their kids are four. Even the old superintendent felt compelled to get four or five hours away. I'd bet the portion of Catholic kids is increasing.

Thursday, March 10, 2011 7:13:00 PM  
Anonymous BrianTH said...

So have we now just completely abandoned the view that comparing the 2000 and 2010 Censuses doesn't necessarily tell you about recent trends? We now think that everything that happened in average over the last 10 years is also happening right now?

Anyway, the first thing to consider with any Pittsburgh-related age statistic is whether the steel bust might have anything to do with it. A person who was, say, 18 in 1982 would have been in peak child-birthing years from about 1988 to 1998, I believe. By 2000, kids 0-17 still extend somewhat before that range (2000-17=1983). By 2005, you are right in that range. Eventually, you will start coming out the other side for these 18 year-olds--but you still have the kids who left with their young parents during the steel bust to contend with.

Is that the only reason for the baby bust effect in the 2000s in the City? I don't know. But I do know that you would expect the second-generation effects of the steel bust to peak at a considerable distance from the original effects, and this is probably around the right time period.

Thursday, March 10, 2011 7:50:00 PM  
Anonymous n'at said...

I thought we covered all this with the school closings and East Liberty white washing?

I suspect the mass losses were in the first 6-7 millennial years followed by respectable gains in aging hipsters and punk-turned-rockabilly white eagle-types.

though, that's based on the anecdotal uptick in "you cant get there from here" recitations.

Thursday, March 10, 2011 7:59:00 PM  
Blogger C. Briem said...

I thought the aging hipsters were going off to Carrick when they start contemplating procreation. I read that somewhere. I guess that is still the city. They better not move into Brentwood or Baldwin or we will really be in trouble.

Again, city and region are very different stories. There is at least some evidence on what the trends have been intra-decade for the county and the region. There isn't much comparable evidence at all that there has been any change in trend for the city. The only possibly positive sign I am aware of is that the census folks adjusted upward some of the estimates for the city in the latter part of the decade.

Which is a fascinating point. Were they wrong? Seems like they were way off. Yet not exactly. I was told that the reason they adusted upward is because they adjusted upward their group quarters popualtion in the city. I bet that was quite accurate. What they didnt catch was this basic housing abandonment which masked the greater popualtion loss. The scary part is we generally are seeing the scale of abandonment spread of late and into neighborhoods many may presume are stable. Then there is the issue of post NORC abandonment as folks who have been in their homes for decades move on.

If there is some evidence the city's household population trend has reversed, or even inflected, I just don't see it. If nothing else the school enrollment trends through this year confirm it pretty well.

25% loss.. over a decade is huge. Funny arguing the point actually. Had an odd conversation recently arguing against someones presumption that the 3% regional population loss over the decade was not cataclysmic. Not sure what there is to say about minus 25%. At some point decline means decline and bad means bad.

Thursday, March 10, 2011 8:15:00 PM  
Blogger joe said...

Maybe it's the absence of the grandchildren of the 25-30 yr olds who left in 1980?

post NORC abandonment - that's us

Thursday, March 10, 2011 8:27:00 PM  
Blogger C. Briem said...

yes, the grandchildren. I think that was part of BrianTh's point which I agree with in a sense. I believe the first factoid out of my mouth with every journalist this week was precisely about how the peak migration of 30 years ago impacted us becuase those folks took with them their families and their future families. I think that is a regional factor, but hard to extrapolate that into an explanation of the city's trend.

but for the sake of semantics.. it actually has a name. It is called the echo boom when applied in the sense of the baby boom... and per the parsing of dates I also agree that the peak migration of the 80's.. young working age cohorts, were indeed right in the middle of the baby boom in many ways.

Thursday, March 10, 2011 8:33:00 PM  
Anonymous MH said...

I found these numbers on the internet (with a CMU address), so they must be true.

year    enrollmnt % decline.
2001    37,612     NA
2002    35,146     6.56
2003    34,658     1.39
2004    34,131     1.52
2005    32,506     4.76
2006    30,885     4.99
2007    27,680     10.38
2008    26,649     3.72

While it isn't impossible, it is hard to square those numbers if Brian's pattern is that big of a factor.

Thursday, March 10, 2011 9:17:00 PM  
Anonymous MH said...

I tried to find a series for kindergarten enrollment, but all I saw was an article saying it increased by 39 kids last year. That would seem to support Brian's thesis, except that the sentence goes on to say that there was also an increase in kindergarten enrollment four year earlier. Which suggests that kindergarten enrollment numbers might be more variable.

Thursday, March 10, 2011 9:31:00 PM  
Anonymous BrianTH said...

On inflection points:

First, are we just not considering ACS data?

Second, here is City population per Census estimates:

2000 333703
2001 330439
2002 327429
2003 325091
2004 320394
2005 316206
2006 313306
2007 312322
2008 312119
2009 311647

2000-2006, that is an average loss of 3400 people per year. 2006-2009, it is 415 people per year.

But wait--the difference is all college kids! Really? College kid population in the City is growing 3000 per year?

More importantly--that enrollment trend started just in 2007? I seem to remember some article called "Higher Education Enrollment and Graduation Trends in the Pittsburgh Region," where the author showed the trend of increasing enrollment in the area started in 1996, not 2007.

But, but--the estimates were off! Yes, but WHEN? As I discussed in the last thread, odds are most of the error accumulated in 2000-2006, which means the change in loss rate after 2006 may have been even more dramatic.

But seriously, what happened to this guy?

"The Census folks just dumped out the county estimates of the 2010 population. . . . They also provide year over year data that is pretty important to understand the trends you miss if you just look at data points 10 years apart."

That guy was pretty smart. Maybe the guy who wrote the enrollment article can find him for us.

Thursday, March 10, 2011 9:59:00 PM  
Anonymous BrianTH said...


Of course not all kids are enrolled in public schools. Moreover, looking at school-age kids cuts off the youngest kids. So if you are thinking about when the second-generation effects of the steel bust are likely to start fading away, with school-age kids you would have to expect it to happen a bit later than all kids.

Personally, I don't quite understand why we are looking at that data versus ACS data.

Thursday, March 10, 2011 10:07:00 PM  
Blogger C. Briem said...

as I suspect many readers here know.. the municipal level census estimates are not really comparable to the county and regional level estimates. All the city level population estimates are reflecting is building permits. It really isn't telling you all that much and can fail where you have issues with demolitions. I'm pretty sure a lot of the building permits related to those Downtown condos hit the methodology early in the decade and skewed the trend. and Sommerset of course, but I'd have to go remind myself the timing.

but you make the point right. Enrollment is up.. making up ever more of the city population. Thus the population left to stick around, have kids, raise a family and all of that, is that much left. It sure adds up to that 3,400/year loss was consistent through the decade. The non-student popualation must have been declining at a faster rate does it not?

I'll see if I can scare up the pre-revision census estimates for the city.. but I do think the big, or bigger.. they were not that huge, adjustments were 2006 and afterwards. I think that may be skewing the perception of that year over year trend for the city.

on ACS data. The census folks will tell you. It should NOT be used for count data. Use it for characteristics, but it really is not intended to reflect change in total population. In fact, all the ACS data released to date, including the 2005-2009 data is population adjusted to the 2000 census making it less a tool to saying what is going on in terms of total population.

Thursday, March 10, 2011 10:16:00 PM  
Anonymous MH said...

As I discussed in the last thread, odds are most of the error accumulated in 2000-2006, which means the change in loss rate after 2006 may have been even more dramatic.

I wasn't able to follow the argument there. Given that the average loss was 3,000/year, and that is just a bit under what you see in the first part of the decade, why would you assume the error is mostly in the first years.

Thursday, March 10, 2011 10:17:00 PM  
Anonymous MH said...

Or what Chris said. I did the math and got slowed down.

Thursday, March 10, 2011 10:20:00 PM  
Anonymous MH said...

Of course not all kids are enrolled in public schools.

Mine certainly is not.

But, it is pretty obvious that the schools that serve poorer neighborhood are losing kids. Another consolidation was announced this week. As for the private schools that serve the rest of the city, those parents are my peeps. It does seem to be a growing group, but they still tend to move out of the city. The lack of any growth in the east end neighborhoods where students don't often live is a good reminder that my world is a small one.

Thursday, March 10, 2011 11:02:00 PM  
Anonymous BrianTH said...

To sum up Chris's argument at this point:

All the annual data available shows a change in rates mid-decade, but it all has to be thrown out for one reason or another (although I still wonder how an enrollment trend that has been pretty steady since 1997 conveniently only shows up in the data in 2007, just in time to ruin the data). Therefore, we should just assume the trend between the Censuses was consistent each year.

Again, I want back the Chris Briem who agreed "pre-game" that you can't assume a consistent trend between data points ten years apart. I think that Chris Briem would at least admit that if all the annual data available for the City is bad, this becomes an unanswerable question. I don't quite agree with that Chris Briem, but at least he would be making a reasonable argument.

Friday, March 11, 2011 7:13:00 AM  
Blogger C. Briem said...

That's why they play the game.

probably enough here. Note today's PG top editorial. I guess I should say something on that which will restart this thread anyway.

Friday, March 11, 2011 8:02:00 AM  
Anonymous BrianTH said...

"But Mr. Briem believes a look at snapshots from 2000 and 2010 don't give a full picture of what has been happening in the region. The early part of the decade was the tail end of the decline that was precipitated by the loss of the steel industry in the 1980s, but trends for the second half of the decade are different."

Ah, found him again.

Friday, March 11, 2011 8:57:00 AM  
Anonymous Cosmetic surgery said...

I have enjoyed reading. .Thanks for taking the time to discuss this, I feel strongly about it and love learning more on this topic.

Saturday, March 12, 2011 2:00:00 AM  

Post a Comment

<< Home