Sunday, April 13, 2014

Permits Past

So I know we all can point to examples of new urban infill in Pittsburgh in recent years, some of which is unlike what we have seen in some time.  But this is the time series for new residential building permits within the city of Pittsburgh over the last 18 years.  

For a city of 300K give or take, those numbers are low. Nothing I have added up, but most years are likely not even enough for replacement for the number of demolitions, major structure fires or plain old unplanned house collapses we have every year. Lots of other things going on for sure (new public housing, rehabs), but most housing markets are driven in large part by private sector new construction which will be reflected in these numbers. Just something to keep in mind before over-interpreting any one or more examples in front of us.

Anyway... I'll get into the reason I'm looking at that later in the week.


Blogger Bram Reichbaum said...

If you're launching a Series, you ought to name it.

Monday, April 14, 2014 2:51:00 AM  
Anonymous MH said...

Game of Thrones [in the Basement]?

Monday, April 14, 2014 10:15:00 AM  
Anonymous MH said...

What happened in 2002? Did all the Summerset at Frick permits get submitted at once?

Monday, April 14, 2014 10:39:00 AM  
Anonymous BrianTH said...

Do conversions of buildings to residential count? Because according to the PDP, during 2012 537 new units came to market in Greater Downtown alone, and there were almost 400 more under construction at the end of 2012. But I believe almost all of those were conversions.

Monday, April 14, 2014 11:43:00 AM  
Blogger C. Briem said...

I do wonder about some city statistics since around 2002 or so in this data series. The relevant answer in their methodology:

"A residential building is a building consisting primarily of housing units. In a new building combining residential and nonresidential floor areas, every effort is made to include the residential units in these statistics, even if the primary function of the entire building is for nonresidential purposes."

Since the data you can poke at directly shows virtually no multi-unit construction in years, you have to wonder about some misreporting in one form or another.

Where I am going with this is that this is the primary data to derive city of Pittsburgh population estimates.So if under-reported it could have an impact on city population estimates.

But before everyone gets carried away thinking there is some under-reported city population. As I said, there has been no multi-unit construction showing in the data (does not mean it is mis-categorized) through the 2000s. So if there was under-reporting going on, it would have popped up in 2010, which it did not for the city.

Monday, April 14, 2014 12:01:00 PM  
Blogger C. Briem said...

It would be interesting to see what the city submitted via this form each year:

I think they should be including new residential units, but maybe that is not how they see it for conversions of existing buildings?

Monday, April 14, 2014 12:13:00 PM  
Anonymous BrianTH said...

Certainly by far most of the big multi-unit projects in recent years have been conversions.

But not all of them, and the balance is shifting. Just looking at projects actually under constructions, there are new multi-units going into the South Side Works, Bakery Square, East Liberty, Lawrenceville . . . so even if the problem to date has been the coding of conversions, we should be seeing a spike coming up. If not, there is some deeper issue.

Monday, April 14, 2014 12:28:00 PM  
Blogger JRoth said...

I was just going to say, that I know that one particular conversion project was permitted in 2008 featured 60 or 61 units, or fully 1/3 of the alleged total for that year. Color me extremely skeptical that one building had 60 new units, but the rest of the city combined for only 125 more. I think it's much more likely that this data isn't what it says it is, whether it's because those 60 units counted as 1 permit (which is, indeed, how many permits would have been issued) or because the conversion was somehow not counted correctly (although I don't really see why that would be the case).

There's also the question of how it gets reported when rehab permits are issued; a fair amount of effectively new residential units are ones that were once residential, but have been unoccupied for some length of time (there's been a lot of this on Penn Ave in the Arts District). I doubt there are enough of those to really move the dial, but it's probably another ~10% a year on top of those shown (there are also probably a number of those that don't get permitted at all, but I have no idea whether that would be a significant number).

Monday, April 14, 2014 12:56:00 PM  
Blogger JRoth said...

Actually, I think this settles the issue: the 127-unit Highland Building was permitted in 2012. Anyone out there care to suggest that there were only 10 additional residential units built in Pittsburgh that year?

I don't recall whether the last occupancy was office or senior housing, but either way, those should show up as 127 new units if that data is to be trusted. They don't, so therefore the data is not to be trusted.

Monday, April 14, 2014 1:00:00 PM  
Blogger JRoth said...

To MH's question above, Summerset housing was actually underway earlier than 2002; those numbers probably reflect Southside Works; Wiki says 400 units in the original push.

Monday, April 14, 2014 1:05:00 PM  
Blogger C. Briem said...

It is pretty clear what a unit is, and the data reports buildings as well. You can see that online.

I think this data should reflect literally new units. So unoccupied for a long time, but with a valid occupancy permit may be at issue, but it impacts the other way as well as density falls in some areas.

But again, much of what came online between 2000-2010 I bet is also unreported and yet city population estimates did not come in way too low. I speculate a bit that under-reporting in larger conversions really may just be offsetting declining density across much of city where census methodology I believe has a bias toward assuming stability. Only demolitions may impact their population estimates and I am not sure the actual demolition counts (even if reported accurately) reflect all the changes going on.

Monday, April 14, 2014 1:07:00 PM  
Blogger C. Briem said...

The 2002 data looks like lots of different big multi-units fwiw.

But just so everyone knows. This is data you often see reported in different ways and the source is the city. Census reports if they have to impute the data, which does not appear to be the case for the city. So a lot of reporting out there might need to be discounted. Just saying.

Monday, April 14, 2014 1:12:00 PM  
Blogger JRoth said...

I guess I'm not clear on what you're getting at, Chris. I get that, if the census numbers seem roughly in line with the reported unit numbers, then no harm, no foul in the larger picture, but surely you agree that the Highland Building* shows decisively that the chart you posted is simply not accurate. It says it's new housing units, but it isn't. Why that should be, I don't know, but I don't see any way around it.

*just checked, the senior housing never happened, so there's no way it had a prior C of O as housing. But either way, that's irrelevant; the building sat empty for 20 years and now has 127 residential units in it; I don't see how failing to count those units as new would be informative unless you were looking at very narrow issues around construction (and even then, it was a gut rehab, so there's no meaningful distinction between what was done there and what was done at, say, the Century Building Downtown, which was a gut rehab from office to residential - rehab is rehab)

Monday, April 14, 2014 1:23:00 PM  
Blogger JRoth said...

My last comment was to your 1:07, btw.

Other projects aside, Southside Works included at least 5 distinct residential buildings in 2002. I wonder what else was being built then. IIRC, everything Downtown was still pretty small scale then, and not much was happening in the East End.

Monday, April 14, 2014 1:28:00 PM  
Blogger C. Briem said...

I've confused the thread a bit since I'm really getting to a population issue which I may not have been clear. But I'll come back to that in a future post.

That being said, I push back on every saying any data is "not to be trusted"... all data requires understanding what goes into it and year I agree that some big rehabs are not included in these #'s.. but the city's reporting does seem to be consistent at least over last decade or more.suspect someone misunderstands what ought to be included in this reporting.

Highland permit was 109 units I take. I do think they should be considered new units. As far as I know previous use was all offices from what history I have read.

All that being said... city is a place of ~150K housing units so in the end, even if this excludes all large structures it still is a low number. I will have to compile something on demolitions and major structure fires per year.

Curious thing though is city is always fighting to say population growth is more than it ever appears. You would want to capture all you want to in this data since it is the primary data for inter-censal population estimates (again I am jumping ahead).

Monday, April 14, 2014 1:41:00 PM  
Anonymous BrianTH said...

Even if the data is being reported consistently over time, if it is missing large categories then it may not be all that useful for judging trends in recent years.

In fact, as applied to population, this gets back to an old issue. We know the Census estimates as of 2010 were not too low--just the opposite. What we don't know is whether the Census was getting the trend right from, say, 2006-2010, because there are many plausible paths the City could have followed after 2000 to get to that 2010 number.

So, say, if you want to hypothesize there was an important change in housing trends that only started in the last few years before 2010, it doesn't appear that this data is of much use in assessing that hypothesis.

Monday, April 14, 2014 2:30:00 PM  
Blogger C. Briem said...

We know the Census estimates as of 2010 were not too low--just the opposite.

Which is my point. The estimates trend through late in the 2000s was trending to come in above 310K give or take but came in below 305K. Since it looks like there are no multi-unit structures in the data since.. looks like 2005 forward, and there clearly were some, I'd really be careful to say there is net population being lost because big units are not in recent data.

On that note, I'm kind of curious if the Heinz Lofts were in that 2002 number. Nothing that big after that. Seems early, but might be when permits were approved.. Or Carlyle after on.

Monday, April 14, 2014 3:06:00 PM  
Anonymous BrianTH said...

"The estimates trend through late in the 2000s was trending to come in above 310K give or take but came in below 305K."

But you have to be careful about making assumptions about exactly what the Census estimates got wrong (and when). For example, it is entirely possible they underestimated losses from 2000-2005, then overestimated losses from 2005-2010, with the net effect being an underestimation of losses as of 2010. Indeed, it is even possible that the City actually started gaining population sometime in the late 2000s, but the losses in the early-mid 2000s netted out to result we observed in 2010.

So it isn't necessarily the case that the hard result in 2010 as compared to the estimates showed that the "missing" multi-unit data starting in 2005 (if that is what is going on) is no big deal. We simply cannot tell one way or another on questions like that because we don't have an actual hard count in 2005 from which we can determine a 2005 to 2010 trend.

Monday, April 14, 2014 5:53:00 PM  
Blogger C. Briem said...

So if all the errors work out in Pittsburgh's favor the population in the city is booming. Hard to imagine given what we know of distressed housing across a lot of neighborhoods of the city.. but sure, since we don't know it is at least possible that 2005-6 was the turning point for the city. Though I think a lot of trends (enrollment gains at local universities for one) stretch across the whole decade.

But I'll address some of the errors in the census estimates later in the week that do not work out so well. To give a hint of that, remember the 1,700+ net new residents in the city between 2010-2011?

Monday, April 14, 2014 6:27:00 PM  
Anonymous BrianTH said...

"So if all the errors work out in Pittsburgh's favor the population in the city is booming."

That is quite the straw man you are choosing to fight.

Anyway, the notion that there would be an inflection point of some sort in the late-2000s is not inherently suspect. For example, there were some pretty dramatic economic events going on around that time, and I even seem to recall someone documenting one of the ways in which the area's relative economic positioning changed as a result . . .

Monday, April 14, 2014 11:50:00 PM  

Post a Comment

<< Home