June 20, 2013

Crime, Big Data, and real estate investing

Everybody talks about Big Data nowadays, but I never see mentioned one straightforward use: real estate investing.

The police have slowly been ramping up their information technology over the last 25 years to look for "crime hot spots," as recommended by Bill Bratton.

Can you use this data for real estate investing purposes? Scan the data for "crime cold spots" -- low crime blocks in low cost neighborhoods -- because they might be promising for gentrifying. 

Traditionally, cops have been interested in second careers in real estate -- see, for example, Harrison Ford in 2003's Hollywood Homicide -- because they drive around all the time and can notice neighborhood trends first. But this kind of data would make real estate trend-spotting even easier. 

31 comments:

Anonymous said...

http://www.buzzfeed.com/daves4/pictures-that-prove-society-is-doomed

Rob said...

This information has been made available in the UK.

Anonymous said...

I actually used this technique in deciding to move into my current house. Three years in, I think I made a pretty good choice.

FWIW, I would love to see more pieces on how to make money with HBD. (Besides, the more money we make, the more money we can funnel back to Steve's fundraising drives, right?)

bjdubbs said...

By the way, HBDers would have predicted the collapse of the SA Rand. Paul Tudor Jones has recently recommended being permanently short the Rand.

Anonymous said...

I know Trulia offers a crime "heat map" that you can overlay on your search through their app. Maybe orter realtors do too.

Anonymous said...

I'm not sure what to make of this, but the Lavender Mafia has its eye on Sarah Palin's Whitopia.

I wonder whether Sarah could re-invent herself as an uber-SWPL in time for the 2016 election.

LemmusLemmus said...

Wouldn't you want to look for spots that have recently been cooling off, rather than cold spots per se?

Anonymous said...

Traditionally, cops have been interested in second careers in real estate -- see, for example, Harrison Ford in 2003's Hollywood Homicide -- because they drive around all the time and can notice neighborhood trends first. But this kind of data would make real estate trend-spotting even easier.

I don't know, regular beat cops who drive around neighborhoods tend to be young guys and rookies. They don't seem to smart to me, or atleast capable of coming to patterned assertions that well.

The police also have a IQ gradient. Where the beat cops tend to be on the lower end, and then there is a progression to detectives, investigators, forensics, crime lab, etc.

From my conversations with my police friends, most beat cops are more concerned with getting home safely than figuring out patterns where to invest in real estate.

Nonetheless, the rest of the article is quite interesting, in fact I would like to see auto insurance claim data by race, but that will never happen, too racist.

Simon in London said...

The Metropolitan Police map of London shows my immediate neighbourhood as amber high crime in a sea of average yellow, but I know that's because it includes a big hospital, a crime hotspot. Likewise a chunk of leafy Wimbledon shows up as very high crime - because it happens to be in the same ward as the train station & high street. So I'm not sure there's any substitute for local knowledge - eg if you go down my local high street you can see by the new flats and SWPL coffee shops that it's finally gentrifying, but just north of there the giant Deobandi mosque is likely keeping house prices depressed. I doubt that shows up readily on the data.

peterike said...

Trulia already provides crime maps for explicitly real estate purposes.

http://www.trulia.com/crime/#

They are a lot of fun to play with.

Anonymous said...

I bet you could track crime by zipcode using percent of children born to unmarried women.

Dave Pinsen said...

It would also prevent/reduce some of the appraisal fraud that was rampant during the housing bubble.

OT, Felix Salmon makes the case here for a $15 minimum wage. It would be fun to see what would happen if someone like Sen. Jeff Sessions proposed this as an amendment to the Gang of 8 bill. The Chamber of Commerce wouldn't like the taste of that.

Neither would Democrats, for that matter, as this would, presumably, reduce demand for illegals and temporarily make poor workers less dependent on government (at least until inflation drove down the new minimum wage in real terms to its previous level).

MoscowEast said...

I made good use of www.police.uk when I bought my latest house last year. Remarkable how low crime correlates quite well with lack of vibrancy (i.e. youths and our own version of dreamers). The solution is coming to all corners of the UK soon, though. Phew.

Anonymous said...

"this kind of data would make real estate trend-spotting even easier"

Of course... crime and schools is information that every buyer wants. I believe that's why there are laws here in the U.S. that specifically restrict this information. Real estate agents are not allowed to answer questions about them. But even if they were, they often would avoid the subject anyway, as it usually includes a mention of race.

A crime map by itself isn't useful, you have to relate it to the locations of other things, everything from apartment buildings to zoning laws to police and fire stations to bus routes. The data available usually isn't detailed enough. And I doubt details will become available anytime soon, since there are significant disincentives to local real estate professionals if it becomes possible to use Big Data instead of hiring them.

There *are* companies attempting to use data this way, to the extent possible. There are offices in some of the bigger boom-and-bust cities where dozens of people comb through real estate data full time, looking for bargains, and also some who do the same from home. But they still need real live people with intuition to interpret all the soft information like crime rates and local politics, not to mention all the misinformation in real estate listings themselves.

Anonymous said...

http://www.guardian.co.uk/media/2010/jul/08/octavia-nasr-cnn-tweet-fired

McCarthy was never this obnoxious.

Zoink said...

A decline in crime is the last thing that happens in gentrification.

At least at first, gentrification increases crime because it produces a neighborhood that is mostly poor, but has pockets of high value crime targets.

The traditional advice is correct to to invest where the younger and poorer gays and hipsters are moving to buy their first house, they are the first wave of gentrification willing to live in 80%+ NAM neighborhoods.

The people they replace in the first wave of gentrification aren't the ones committing crimes, but rather elderly homeowners, or their children who sell after they pass away.

The big drop in crime at the end of the gentrification process happens when even the smaller rental apartments become too expensive for the old NAM population. This last stage is currently happening in Hollywood, West Hollywood, and the eastern sides of Venice Beach and Santa Monica in LA, SOMA in San Francisco, and the neighborhoods around Williamsburg in Brooklyn. Some of these areas have saw their black population drop by 50% between 2000 and 2010.

Lisa said...

For an investment this is fine but I don't understand why folks want to live in these areas since the areas are still surrounded by crime and dysfunction. I geuss it's mainly made up of homosexuals and childless couples who want to live in the city limits but want neighbors like themselves. But that doesn't keep the riff raff out of their grocery stores.

el supremo said...

There is a ton of data already online but it is scattered and not very illustrative in its pure state. It needs someone with a combination of a discerning eye for how the world works with an inventive approach for finding data sets and mapping them against each other (like Mr. Sailer perhaps?)

For example, many people have noted quality schools as a driving factor. http://www.greatschools.org/ provides racial splits of school populations - they are dated a few years but directionaly accurate. However, its simple demographic categories like "Hispanic" are useless is differentiating between Venezuelan professionals in exile (good) and Nuyorician hooligans (bad). Layering on use of free school lunches would help. Similarly, if everyone with a professional job in the region sends their kids to Catholic school then the racial split of public schools is a poor metric for tracking high end property prices.

You also need a good bit of local knowledge to flag what indicators like business names from permit applications really mean (ie a recently opened church named Tabernacle of Zion Bible Ark is not a good sign for property values).

It would actually be an intellectually interesting exercise . . .

NOTA said...

Anon 9:11:

When we were buying a house in Maryland, our real estate agent told us she was forbidden to discuss racial mix, crime, or school quality. She told us precisely how to get that information ourselves, but wouldn't talk about it or provide it directly. I'm not sure if that was the law or some policy her agency had to deflect lawsuits, but it was interesting to see.

The odd part about this is that I broadly sympathize with the goals of that policy, but think the policies adopted probably do more harm than good--much as with affirmative action, or racially gerrymandered congressional districts.

NOTA said...

If technology makes lunkhead-variety crime harder to get away with over time (when you bust out my windows and steal the gps from my car, I and every other victim has timestamped video footage of it, and even overworked police can find the perp and arrest him), one likely outcome is more willingness to live in racially and class mixed neighborhoods. One result of that: annoying but not seriously criminal prole behavior (catcalling women, littering, loud stereos) will become more of a problem--you move to the prole/mixed-race neighborhood for the low rent, but now have to put up with the annoying neighbors you'd never have dared associate with in a higher-crime world.

This suggests that both stricter HOAs and technology to cover over the annoyingness of your neighbors have a big future. Tall fences with sound-absorbing/sight-obscuring mesh and white noise generators make good neighbors. HOAs can also include litter pickup and lawn maintenance in their services, so that the proleness of the neighbors doesn't matter so much. Yes the neighbors don't care about the lawn, but Diego and Jose come through once a week and pick up the litter/mow the lawn, so it can only get so annoying.

peterike said...

Of course... crime and schools is information that every buyer wants. I believe that's why there are laws here in the U.S. that specifically restrict this information. Real estate agents are not allowed to answer questions about them.

True. Here in the land of the free, as a real estate agent you cannot say ANYTHING about the racial makeup of a neighborhood even if asked directly(at least that is the case in New York -- and they actually send out "inspectors" who try baiting agents into uttering crimethink!). Which is why every real estate agent uses the euphemism "good schools."

"This neighborhood has good schools" means the school and the neighborhood as mostly white.

Whiskey said...

Predicting real estate values is quite hard because you need to accurately predict the future ethnic/racial composition of the area. As more and more immigrants flood into the country, you get more and more ethnic/racial churn.

You can already see this in OC. Massive Mexican influx in traditionally Black areas: Watts, South Central, Compton, Wilmington, parts of Long Beach, have pushed Black residents out and south to Orange County. While OC is still quite low in percentage of Black residents, they stand out in very concentrated areas: Orange, Santa Ana, Tustin, and Irvine. The composition of these residents is different too, back when the El Toro Marine Airbase and the Tustin MCAS helicopter station (the Blimp hangers) were operational, you'd see dependents of Black military families -- pretty well squared away, middle/working class, disciplined and respectful, looking often to join the military which was all in all, not a bad deal then: secure pay, decent benefits, working with monster machines, generally little chance of combat.

The group of Black residents now, concentrated within specific blocks/areas of cities within OC (again Santa Ana, Orange, Tustin, and Irvine) are mostly Section 8 housing and very different from the military kids thirty years ago.

David Davenport said...

... and even overworked police can find the perp and arrest him...

The police will probably also be nonwhite, so why should they arrest the perp instead of arresting you for inciting racial hate?

...one likely outcome is more willingness to live in racially and class mixed neighborhoods....


That is completely wrong.

...

This suggests that both stricter HOAs ...

Courts will rule that cultural diversity requirements trump HOA's.

...and technology to cover over the annoyingness of your neighbors have a big future. Tall fences with sound-absorbing/sight-obscuring mesh and white noise generators make good neighbors.

Very ugly and nightmare-ish.

Are you the nerd who, on an earlier thread, proposed to defeat fifteen muggers with a can of oven cleaner and a folding baton carried, I suspect, in your Batman utility belt?

dsgntd_plyr said...

There was a recent "Lionoftheblogosphere" post/comment thread about making money off HBD. I don't think anyone brought up real estate!

NOTA said...

David:

So your prediction is that police will stop enforcing laws on black criminals when there is a white victim, even when the white victim provides video surveillance footage of the crime? And that in fact those police will arrest crime victims for racial hatred? Do you have some evidence of that happening so far? I mean, we have lots of blacks locked up from majority black cities with heavily black police, prosecutors, and judges, so this looks kinda implausible to me.

I guess you are also predicting that HOAs will be forbidden from hassling residents into keeping up their property on racial grounds. Again, are there any examples of that being a trend now?

Both of those predictions require black elites and white elites to injure their own well-being in the interests of sticking it to whitey and keeping the black underclass out of jail. What do you see happening now that makes this seem plausible to you?

Here's my premise: Lots of relatively well-off people would like to live in houses or condos or apartments that currently have low rents or prices. Perhaps they don't have kids so don't care about the schools. Perhaps they just want to live close to their job and good places to meet single women. One thing keeping them out of those houses and condos is the fear of crime, both serious crime like that bicyclist who got the helll beaten out of him, and minor crime like having stuff stolen out of your car or off your front porch.

If technology changes in such a way that the threat of crime decreases significantly, then more of those people will move into those neighborhoods. If the area is safe but prole/lowbrow, well, that might be worth it for a 30% lower house payment and a 15 minute shorter commute each way. Indeed, whatever is behind the big drop in crime over the last couple decades has already led to some of that--white flight would have been a much smaller phenomenon if not for the very high crime that happened around the same time as desegregation.

As a thought experiment, what do you think would happen wrt gentrification if the threat of muggings and gang beat downs just went away. My prediction is that we would see massively more gentrification, and more poor blacks and hispanics priced out of cities, because there are lots of urban professional couples with no kids who would live in the city if they weren't worried about their wives' and kids' safety.

My guess is futther that modern technology will be a net win for decreasing crime and taking criminals off the street. So barring some kind of collapse of law enforcement that stops arresting people and opens up the prison gates, I expect more gentrification, And that will mean more middle class people rubbing elbows uncomforablly with proles of various shades. And so technology that makes that rubbing elbows easier to bear will be valuable.

BTW, I was on of the guys pointing out that oven cleaner and karate is a poor defense against a beat-down by half a dozen thugs.

sunbeam said...

I have to wonder about these mob incidents. I suppose the beatdown of the bicyclist could be ten or so youths that got a wild hair.

But the big ones, whatever it is called, when a minimart or something is flashmobbed and everyone walks away with something has to be organized.

It'd definitely be non-PC but I'm willing to bet cell phone providers have ethnicity data on their customers. You might have to get data from several providers, but since these cellphones seem to never be cut off for any reason...

It's pretty simple. Just have a database trigger. When subjects meeting the profile are in the same spot at the same time, an alert goes out. Just to let the cops know.

The same thing can be done with cameras, and that is coming. I mean if you can recognize faces, you can recognize a broader class like ethnicity. Any time you get 40 or 50 black kids outside a minimart, an alert goes out to the nearest police station.

I'm also kind of surprised there has been no discussion of the Alex Hernandez case in Boston yet. Whatever the result of that, I have to wonder what that guy made on the Wonderlic, because he sure seems like he is as dumb as a tree stump.

Unknown said...

You also need a good bit of local knowledge to flag what indicators like business names from permit applications really mean


Jim Zaspel

Portrait of a Man said...

Actually, I don't need big crime data for this one. 5 minutes finding out the demographics gives a suprisingly accurate prediction for that.

Anonymous said...

"5 minutes finding out the demographics gives a suprisingly accurate prediction for that."

Much better really as fiddling the stats is skewed towards the worst areas.

All in all i don't see watching this kind of churn being a good money-spinner at a cop level.

If a neighborhood is going down you don't want to buy till rock-bottom and gentrification isn't likely to start straight away after an area hits rock-bottom so if you were a cop the only places would be those which were already really bad and you noticed the first few gentrifiers.

Even then, at least the places i know, cops need to live a significant way away from the people they're arresting so even the above option wouldn't be suitable unless they were single and/or had the cash to have a family home and a spare.

I think to make money from ethnicity based real estate churn you'd need a portfolio of property and treat it like spread betting with lots of hedging like either neighborhood A, B or C will be going down and either neighborhood D, E and F will go up and have property in D, E and F.

sunny singh said...

When we were buying a Condo Pattaya, our real estate agent told us she was forbidden to discuss racial mix, crime, or school quality.

Property Management Atlanta said...

I agree. The Metropolitan Police map of London shows my immediate neighbourhood as amber high crime in a sea of average yellow, but I know that's because it includes a big hospital, a crime hotspot. Likewise a chunk of leafy Wimbledon shows up as very high crime - because it happens to be in the same ward as the train station & high street.