Are concerns about an arson hotspot in the Mission smoke or fire? After high-profile blazes in the rapidly-changing neighborhood promoted David Campos to hint at the possibility of nefarious shenanigans at play, we took a look at city data to uncover whether the supervisor's worries about an arson cluster are hot air.
"I’m a survivor," says José Montoya, sitting in the basement of the Salvation Army building in the Mission. Cane in hand, he talks about how he and his brother were among the 67 people who lost their homes after a five-alarm fire tore through six buildings on the corner of Mission and 29th Street on June 18, including the Greywood, a single-room occupancy that they lived in. "The vents started smoking up, and I said ‘get out.’"
On a recent Friday night, he’s here to pick up a check for his share of the over $151,000 in donations—roughly $2,400 per person—that have poured in since residents lost their homes, beds, clothes, and possessions.
When the fire started in a building that housed Cole Hardware, Montoya, who served several terms in prison, helped elderly and mentally-disabled residents escape—back into the building to look for those left behind. Now he and his brother are "bouncing around like a beach ball," looking for a permanent place to live.
The fire was a tragedy, especially since its toll disproportionately fell on people like Montoya struggling to hold on in a rapidly-changing Mission. The fire was just one of several recent blazes that have displaced 194 people, according to Mission Local.
But what kind of tragedy was it?
"It doesn’t make sense to me why fires have gone down in every other part of the city but the Mission," said Campos.
Campos has refused to rule out the possibility of arson. "It doesn’t make sense to me why fires have gone down in every other part of the city but the Mission," he said in a speech at a Salvation Army event.
Is Campos right? Is the Mission burning? Is arson to blame? We decided to find out.
San Francisco’s fire and police departments keep records online that date back to 2003. Anyone can access them on the city’s open government platform.
We set out to answer two questions. Do buildings in the Mission burn at a higher rate than those in the rest of San Francisco? And, has the Mission suffered from a rash of building arsons—perhaps set by greedy landlords trying to cash in?
The answer to the first question is yes, while the number of building fires has dropped across San Francisco, given its population, the Mission’s buildings burn at a rate higher than expected. So do buildings in the Financial District, the Tenderloin, South of Market, and the Western Addition. The Mission is part of a geographic cluster—not an outlier. Across time, the number of building fires in the Mission is more or less flat, not rising.
As for arson, the answer is no. Of its yearly average of 37.5 building fires per year, the Mission has 5.5 building arsons per year. Those are each serious crimes, but we compared to the rest of San Francisco, ranks near the middle of neighborhoods. The data does not show that building arson plays a large role in causing building fires in the Mission.
There were 4,322 building fires in San Francisco from 2003 to 2016, according to the Fire Department. Of those, 508 were in the Mission, from the beginning of 2003 to July 15, 2016.
That was second to the Tenderloin, which saw 560 over the same period of time. South of Market had 327, Bayview/Hunter’s Point 225, and Hayes Valley 171. (The lowest was in Seacliff, which had a grand total of one.)
We found the data here, and filtered the cases to show only those categorized with a primary situation of code 111 - building fire. (Sometimes the dataset skipped the dash, writing "111 building fire." We counted those too.) The Fire Department notes which neighborhood each fire occurred in, so it was pretty easy to build a workbook in Google Sheets to run the numbers.
It’s hard to see the trend lines with every neighborhood displayed, so Chart One shows the number of building fires in the five neighborhoods with the highest number of building fires from 2003 to 2016.
The values for 2016 in this chart, and in all other places where noted, are the projected total at year’s end. We calculated it by taking the reported value on July 15 and multiplying it by 12/7.5 to produce a simple linear estimate of the total at year’s end.
The Mission averaged 37.5 building fires per year. That’s second to the Tenderloin, which saw 40.8 building fires per year, but ahead of the South of Market, which averaged 23.7 per year, the Bayview at 16.7 per year and Hayes Valley at 12.3 per year. Since 2003, the Mission and the Tenderloin have generally traded off the number one and two slots for most building fires per year.
The number of building fires in the Mission has generally been increasing since 2012, when it had 23, but the recent trend is still below its maximum, 64 building fires in 2004. The Bayview has also had a similar upward trend since 2012, whereas the Tenderloin has been declining, and those in SoMa and Hayes Valley flat.
How much should we worry about the projected increase in 2016? Well, the standard deviation of building fires in the Mission, a basic measure of how tightly clustered the data is around the mean, was 9.8. That means that 68 percent of all yearly observations lie between 27.7 and 47.3. The most recent increase, in which the Mission is projected to reach 46.4 building fires in 2016, is 0.9 standard deviations above.
Let’s see how the Mission stacks up to San Francisco as a whole.
Chart Two shows the number of building fires in the Mission compared to the rest of San Francisco. (The numbers for 2016 are estimates.)
Building fires as a whole in San Francisco decreased from 492 in 2003 to an estimated 224 in 2016, down 54 percent. By contrast, building fires in the Mission increased 29 percent from 36 in 2003 to a projected 46.4 in 2016.
Qualitatively speaking, we’d say a reasonable gloss on the data is this: As building fires in the rest of San Francisco have dropped, those in the Mission either stayed flat or increased slightly. We don’t think that this chart shows the kind of dramatic increase in building fires in the Mission that would warrant the conclusion that something has drastically changed in the last few years in the neighborhood.
But wait a minute. We haven’t quite been making the right comparison—because different neighborhoods have different populations. The Mission is pretty crowded. If the risk of building fires increases as population does, is it seeing more than we’d expect or less?
Chart Three shows a prediction of how many building fires we’d expect in each neighborhood given its population (the line), and how many actual building fires there were (the dots). The further the dot is to the line, the more anomalous the neighborhood is.
For the Mission, we predict 283 building fires over the time period. In actuality, it had 508. That’s high, but is it a little high or a lot high? And how does it compare to other neighborhoods?
Here’s one simple way to find out. We calculated the standard deviation of the difference between observed and expected values for all of the neighborhoods It was 107. Then, for each neighborhood, we calculated how many standard deviations the difference between the two values was. For some neighborhoods, the model hit dead on like in Chinatown, where the standard deviation of the difference was just 0.03.
The standard deviation of the difference for the Mission was 2.10. That’s higher than all other neighborhoods except for the Tenderloin, which had 3.14. SOMA was 1.52, the Financial District 1.26, and the Western Addition 0.98. All other positive values were around 0.5 or less.
So let’s say any neighborhood with greater than one standard deviation above the prediction line is as a high-building fire neighborhood.
By that definition, the Mission is a high-building fire neighborhood. So is the Tenderloin. SOMA, the Financial District, and the Western Addition. The Tenderloin is worse—much worse—and the others are a little better. In other words, there’s a cluster of neighborhoods with higher than expected building fires, centered around the middle of Market Street.
So what’s to blame? Is it arson? Arsons set by landlords?
That’s a very reasonable question to ask. Those of you who’ve been around for a while remember the 1977 World Series, when the camera panned onto a buildings set ablaze by greedy landlords and announcer Howard Cosell said, "There it is, ladies and gentlemen. The Bronx is burning." It’s a parallel that many have drawn to the building fires in the Mission.
As Campos wrote in the Examiner, "Reasonable people see that fires in low-income buildings almost always result in large profits for building owners and landlords and are disturbed and afraid the uptick in fires is more then [sic] just a coincidence."
It’s a reasonable question to ask, but the data doesn’t bear out the fear. It turns out that arson isn’t why the Mission is burning—and, funny enough, nor why the Bronx was burning.
First of all, it turns out Cosell never said those words. Nobody has ever been able to find it on the tape of the game. And it turns out that arson was not the primary cause behind the fires in the Bronx, although there were plenty of fires there.
Between 1970 and 1980, "seven census tracts in the Bronx lost more than 97 percent of their buildings to fire and abandonment," the popular narrative that attributes the blazes to arsons set by landlords doesn’t square with the data. (For more, see this article in the Post, this book, or this podcast from 538.) Short version: At its peak, arson only accounted for seven percent of building fires, most of them in abandoned units. That was higher than normal, but the much larger culprit was a faulty computer model that led New York to close several fire stations in the South Bronx, increasing emergency response times dramatically.
Judging by data collected by the San Francisco Police Department, the answer is that although the Mission does experience a slightly elevated amount of arson, the total number of arsons is not sufficient to explain the variance in building fires between the Mission and the rest of San Francisco. Like the "Bronx is burning" story, fears of arson are overstated.
That’s not to say there are no arsons in San Francisco. From 2003 to 2016, San Francisco is on track for 697 cases of building arson. (We projected the 2016 numbers the same way we did in the first section.) That’s the sum of what the police code classifies as arson of a commercial building, arson of an inhabited dwelling, arson of a vacant building, and arson of a police building. It excludes those cases classified as arson, arson of a vehicle, arson with great bodily injury, attempted arson, and unlawfully causing fire.
For this section, the city’s geography is analyzed differently than above. The Police Department divides San Francisco into the Bayview, Central, Ingleside, Mission, Northern, Park, Richmond, Southern, Taraval, and Tenderloin districts. You can find the map of the boundaries here. It’s important to note that the SFPD Mission subsection includes the Mission District, as well as Bernal Heights, Noe Valley, the Castro, and others.
So that means, first, we have to be cautious about comparing numbers in this section to the ones in the first section. We can do it, but we should be aware that the data in this section is overcounting building arsons in the Mission District. Second, because the Police Department doesn’t give population figures for its districts, we can’t run the per capita numbers the way we did above
Chart Four displays the information on building arsons.
Of those building arsons, 79 were in the Mission, 11.3 percent of the total. Of the districts for the full data set, the Mission comes in fifth of ten. The Southern District is on track for 98.8 building arsons, Bayview for 91.2, Northern for 82.2, and Ingleside for 79.2.
What’s important in Chart Four is what you can’t see: Where the Mission is. Compared to other police districts in San Francisco, the Mission’s share of building arson does change from 2003 to 2016. In 2003, it had five of San Francisco’s 66 building arsons—7.6 percent. In 2016, it’s on track for 8.4 of 46.4—17.2 percent. But in no year other than 2004 does the Mission have the most number of building arsons of any police district. The most noticeable recent change isn’t in the Mission, either. It’s the jump in building arsons in the Bayview, from two in 2015 to an projected 11.2 in 2016.
Chart Five shows the five police districts with the most building arsons.
Chart Five shows in detail what was implicit in Chart Four—the trend of building arsons in the Mission is not particularly different from the number in other parts of the city. Although it does appear to be rising slightly from 2013 to 2015, it is on track to have fewer building arsons in 2016 than it did in the year before.
How does the Mission compare to the city as a whole?
San Francisco as a whole averages 48.2 building arsons per year. Because we don’t know the population of each police district, we’ll assume an even distribution of expected arsons per district—4.8 per year each.
Chart Six shows this information over the entire data set.
The Mission sees 0.8 more building arsons per year than expected, or 11.2 more building arsons than predicted across the fourteen years in the data set. By contrast, the Bayview had 22.4 more building arsons than predicted from 2003 to 2016 and the Southern District 18.7 more.
So does the Mission have a building arson problem worse than the city as a whole? Yes, but at a statistical rate only slightly elevated, and one less than the over-share in the Southern and Bayview Police Districts.
So are arsons to blame for the rate of building fires in the Mission? They certainly contribute, but not at a very high number Nor do they appear to be dramatically increasing over time. Nor is the Mission share of building arsons worse than several other parts of the city.
Any building fire — especially one that displaces people like the one in the Mission on June 18 did, is a cause for concern. Just ask Montoya, or the other 66 victims of the fire. And any building arson is a crime.
But we fail to find support for the hypothesis that the Mission is experiencing building fires at a rate of crisis.. We do find that it is one of several neighborhoods experiencing building fires at a rate higher than predicted by population, including the Financial District, Tenderloin, the South of Market, and the Western Addition.
We suspect that older buildings, which tend to have fire sprinklers at lower rates, are not configured for heavy electrical use, and may be overcrowded, may play a role in elevating the rate.e. The median year of structure construction in the Mission is 1939, suggesting that age may play a role. But the median year of structure construction in Nob Hill is also 1939, suggesting the story may be more complicated (presumably mediated by socioeconomic status and race).
One thing does jump out. Fire detectors were present in only 49.9 percent of building fires in the Mission, compared to 52.5 in the rest of San Francisco. Automatic extinguishing systems, including sprinklers, were present at the site of 45.9 percent of building fires in the Mission, compared to 50.8 percent in the remaining neighborhoods.
What we do fail to find support for, however, is the allegation that the Mission is in the midst of a cluster of building arsons. The available data suggests that this problem, however reasonable it is to pose as a question, is not driving the elevated rate of fires in the Mission.