or Connect
Styleforum › Forums › General › Current Events, Power and Money › Daily CE Musings of Piob
New Posts  All Forums:Forum Nav:

Daily CE Musings of Piob - Page 91

post #1351 of 5110
Thread Starter 
Quote:
Originally Posted by Gibonius View Post

That's interesting. A bunch of my progressive friends have been railing about the evil of gentrification, and there's a total lack of discussion about how to solve this. The solution (as always, it seems) is just to shame rich whites for moving into poorer areas and mucking things up. There's really no realistic proposals that solve the issue of high demand, reasonable prices, and maintaining whatever factors make people want to live in these areas in the first place.

A liberal friend of mine I sent this to dismissed the article saying it was "crap" and made no attempt to identify causes. I told him that maybe he should go read the article because it did in fact attempt to isolate some causation.
post #1352 of 5110
Quote:
Originally Posted by Piobaire View Post

http://www.theatlantic.com/business/archive/2014/10/why-are-liberal-cities-so-unaffordable/382045/

An article in The Atlantic makes the case that the more liberal a city the worse the environment is for the poor and middle class to live. Income inequality actually worsens as a metro area become more liberal. The most liberal large US city, San Francisco, is the hardest place to be poor and/or middle class.

His actual study (if you can call what he posts to the trulia blog a study) doesn't really attempt to make that case. He's pointing out interesting correlations, not trying to draw a causation link. For a harvard economist, the whole thing is a little weak (some undergrad level analysis here), but I get that his job is to drive people to Trulia, which posts like this do a very good job at (and posts with regression results do a very bad job at).

There are a few key issues that need further followup before anyone can even start to draw conclusions from this.
  • Need to better account for population density. Right now, there are no controls for population density (and the density difference between the Sarasota MSA and SF are on completely different orders of magnitude). Population density has already been shown to correlate strongly with these things...and his graph basically looks like a population density chart.
  • Control for housing size/land availability. Current results use $/sq.ft. This is a reasonable metric for some analysis, but it is not ideal here. People in sprawling southern metro areas have cheap land and can build out big spaces while you could find a 2BR apartment in SF the size of Piob's master suite. I don't know how much of a factor it is that liberals might think they are "doing good" by living in more compact housing, but I suspect a price per bedroom or some other "housing-unit" metric would be more informative--whether you are living in recycled shipping containers, or a stucco mcmansion, if you've got the money, you are going to want a bedroom for each kid.
  • Rental prices. I get that that's not what Trulia wants to focus on, but you can't talk about affordable housing and then not analyze the type of housing in which most poor people live, especially in areas where the rent vs own equation can be very awkwardly balanced post-crash.
  • This is a minor issue, but how an area voted in the most recent presidential election is not a great liberal vs conservative indicator (unless you are willing to just define liberal as what the democrats want and conservative as what the republicans want). Detroit is solidly blue, but it is nothing like the liberal paradise of SF. Again, this is minor, but it needs to be checked for robustness against other metrics. The Cook PVI is better than results in a single election...and there has been some work out of MIT to build more of an ideological index (with issue-based surveys rather than election result proxies).
  • Finally, you have to look at the time-series. This is not a new issue, so the question becomes...did residents of the high density cities see their poor neighbors suffering and decide to implement a bunch of successful and failed attempts to help them? Were they already liberal and they just made it worse? Is today's situation an anomaly that the market will correct? This sort of housing data is well cataloged over time, so the question is not unanswerable.
post #1353 of 5110
Thread Starter 
Agree with some of your points, agree less with others. For instance I'm not sure size is important but completely agree on the population density. Also I'd toss Detroit as a simple outlier and I think there's plenty of reasons to justify that decision.
post #1354 of 5110
And while that guy's work seems fine (other than those areas for further analysis), what the fuck is this chart that the Atlantic made using his data?

Quote:
The line isn't smooth—and there are exceptions—but the relationship is clear: In general, richer cities have less affordable housing.

You know why the line isn't smooth? Because you are trying to draw a line chart using categorical data. Yes, I know they are rank ordered, but they are not related data points, and there is no reason to interpolate a line between them. And wtf is that axis? As best I can tell, they are using the same axis to both represent a % and a dollar figure which happens to fit it all on the chart...but is awful.

Still though, the metric they are reporting there (from this guy's May 2014 work) seems more appropriate--they are looking at homes for sale and whether or not they are "in reach" of the median income. Not counting bedrooms, but they are allowing home size to vary. If somebody has 300k to buy a home in a dense area, they are simply going to have less square footage if they buy a comparable quality home in Atlanta. If you just want a home in your area, you are probably indifferent to the fact that you could get more space somewhere far away.

For example, a friend of mine just bought a house in Columbus, IN (which is actually a nice little town with a lot of great architecture in a much prettier part of the country) for less money than it would take to own the condo I currently rent. That's great and all...but I don't have work or friends out there, and if I were to find work, it would pay a lot less. So, I am perfectly content paying more for the physically smaller home even though they are probably both in the same "class" (i.e. homes appropriate for a young single person or couple that would still want to move again before having children).
post #1355 of 5110
That chart makes no sense at all.
post #1356 of 5110
The only reason the median income line is smooth is because they rank ordered the cities by median income..."Hey guys, look, I arranged these things in descending order and the values descend!"
post #1357 of 5110
Don't let the man hold you down with axis labels.
post #1358 of 5110
If you look at the graph where he really plots affordability vs political slant


It's basically just six cities that are huge outliers. There would be a much weaker trend without those handful of cities. They're not all labeled, but one of the sources pulled up a chart of the top 10 least affordable cities. Seven of them are in California.

One could likely reduce this thesis to: why are liberal cities in California so expensive?
post #1359 of 5110
Quote:
Originally Posted by otc View Post

And while that guy's work seems fine (other than those areas for further analysis), what the fuck is this chart that the Atlantic made using his data?

You know why the line isn't smooth? Because you are trying to draw a line chart using categorical data. Yes, I know they are rank ordered, but they are not related data points, and there is no reason to interpolate a line between them. And wtf is that axis? As best I can tell, they are using the same axis to both represent a % and a dollar figure which happens to fit it all on the chart...but is awful.

Still though, the metric they are reporting there (from this guy's May 2014 work) seems more appropriate--they are looking at homes for sale and whether or not they are "in reach" of the median income. Not counting bedrooms, but they are allowing home size to vary. If somebody has 300k to buy a home in a dense area, they are simply going to have less square footage if they buy a comparable quality home in Atlanta. If you just want a home in your area, you are probably indifferent to the fact that you could get more space somewhere far away.

For example, a friend of mine just bought a house in Columbus, IN (which is actually a nice little town with a lot of great architecture in a much prettier part of the country) for less money than it would take to own the condo I currently rent. That's great and all...but I don't have work or friends out there, and if I were to find work, it would pay a lot less. So, I am perfectly content paying more for the physically smaller home even though they are probably both in the same "class" (i.e. homes appropriate for a young single person or couple that would still want to move again before having children).

I think they realized that actually plotting income on one axis and %-affordability on the other would have produced something with a very hazy correlation, so they plotted it up in a way that led you to draw in the best-fit line yourself and overemphasize the relevance of the trend.
post #1360 of 5110
Quote:
Originally Posted by otc View Post

And while that guy's work seems fine (other than those areas for further analysis), what the fuck is this chart that the Atlantic made using his data?

You know why the line isn't smooth? Because you are trying to draw a line chart using categorical data. Yes, I know they are rank ordered, but they are not related data points, and there is no reason to interpolate a line between them. And wtf is that axis? As best I can tell, they are using the same axis to both represent a % and a dollar figure which happens to fit it all on the chart...but is awful.

I don't understand your criticism of the chart. It shows that adjusted house prices are generally higher where incomes are higher, but it also shows that there are significant outliers, which is what the article itself is about...
Quote:
Still though, the metric they are reporting there (from this guy's May 2014 work) seems more appropriate--they are looking at homes for sale and whether or not they are "in reach" of the median income. Not counting bedrooms, but they are allowing home size to vary. If somebody has 300k to buy a home in a dense area, they are simply going to have less square footage if they buy a comparable quality home in Atlanta. If you just want a home in your area, you are probably indifferent to the fact that you could get more space somewhere far away.

The idea seems to be that a house is a house.
post #1361 of 5110
Thread Starter 
Quote:
Originally Posted by Ataturk View Post

I don't understand your criticism of the chart. It shows that adjusted house prices are generally higher where incomes are higher, but it also shows that there are significant outliers, which is what the article itself is about...

I think it attempts to show more than that, namely, to show the relationship between how many homes a family of median income can afford relative to other cities. IMO there probably is an impact if 70% of a city's housing it out of reach for median income families.
post #1362 of 5110
Quote:
Originally Posted by Gibonius View Post

If you look at the graph where he really plots affordability vs political slant


It's basically just six cities that are huge outliers. There would be a much weaker trend without those handful of cities. They're not all labeled, but one of the sources pulled up a chart of the top 10 least affordable cities. Seven of them are in California.

One could likely reduce this thesis to: why are liberal cities in California so expensive?

What strikes me about it is the weird graduation on the y-axis. It's not very effective for that reason if nothing else.
post #1363 of 5110
Quote:
Originally Posted by Piobaire View Post

I think it attempts to show more than that, namely, to show the relationship between how many homes a family of median income can afford relative to other cities. IMO there probably is an impact if 70% of a city's housing it out of reach for median income families.

That's what I meant when I said "adjusted" income. Sorry.
post #1364 of 5110
The R^2 for the correlation between income and housing price is only 0.22, if you actually plot that thing out. It's statistically significant, but not very strong.







The other factors that I can't believe they didn't discuss are population(!) and geographic area. The least expensive cities have an average population of 210k and are nearly all in the Midwest/Rust Belt (five of ten in Ohio). The most expensive cities have an average population of 406k and are mostly in California (seven in CA), then NYC and Fairfield CT, and then Honolulu which is much more expensive for entirely external reasons. Same deal with the liberal cities, average population is 670k. Conservative average is 380k.

Hard to figure that you could do an honest analysis without considering some pretty basic factors like that.
Edited by Gibonius - 11/3/14 at 12:54pm
post #1365 of 5110
Thread Starter 
Quote:
Originally Posted by Gibonius View Post

The R^2 for the correlation between income and housing price is only 0.22, if you actually plot that thing out. It's statistically significant, but not very strong.

Is that bottom axis labelled correctly?
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Current Events, Power and Money
Styleforum › Forums › General › Current Events, Power and Money › Daily CE Musings of Piob