The FDNY recently released a list of twenty fire companies that will be closed on July 1 as part of Mayor Bloomberg’s budget cuts. The closings have been looming for years. But unlike in the seventies, when officials used complicated and ultimately flawed computer models that resulted in burned-out South Bronx vistas, the department is trying to keep things brutally simple this time around.
“The budget comes from the City Council and the mayor’s office, and our job is to do the best we can with that,” said FDNY Chief of Operations Robert Sweeney. “No one here wants to close fire companies, but if we have to, we think this is the best way to do it.”
The assumption among critics of the closings was that the FDNY would use the complex, multi-million-dollar computer modeling system it purchased from software firm Deccan International to make the cuts. The software works a bit like a high-stakes version of SimCity, crunching years’ worth of fire statistics and response-time data to create models of how the department responds to emergencies.
But the department ultimately decided to ignore the complicated algorithms, feedback loops, and assumption testing of the expensive modeling system in favor of a much simpler arithmetic: It took three years of response-time data, which tracks which fire companies are due first, second, and third on the scene of a given emergency, and simply calculated what would happen if one of those companies was closed.
For example: Whenever it was the first-due company at a fire, Engine Co. 4 in lower Manhattan has averaged three minutes and 56 seconds from dispatch to arrival. The second company to arrive after Engine 4 took an average of 4:39, and the third to arrive took 5:45. So the department calculated that when Engine 4 is closed in July, the arrival time for the first-due company will be 4:39, and the second-due company will arrive in 5:45.
By looking at the companies that went on the fewest runs and would produce the smallest increase in predicted response time, the department came up with a list of potential closings and sent it out for high-level chiefs and borough commanders to critique, based on their own experience and knowledge of more intangible factors. When the data didn’t reflect how important a company truly was (because it was near, say, a large hospital or natural gas tanks, or had good access to a highway like the Cross Bronx) the company was taken off the list and replaced with a less vital unit.
The last time the FDNY had to make such significant cuts was the seventies, when dozens of companies were closed over a five-year period. Response time was the primary factor in the decision-making then, too, with the department relying on computer models supplied by the RAND Corporation. But the models were deeply flawed: One assumed that traffic played no role in how quickly a fire engine could travel through New York City streets; another assumed the nearest company was always available to respond to fires, which was rarely the case in poor neighborhoods where fire companies averaged 20 to 30 calls per day. Most of the companies the models recommended shuttering were among the busiest in the city, and the closings helped kick off a firestorm that created the infamous burned-out Bronx neighborhoods that the likes of Jimmy Carter, Ronald Reagan, and even Mother Theresa came to gawk at.
The current method for closing companies has some definite advantages over the RAND models and how they were used. There are no confusing contingencies or unexpected feedback loops buried in the computer code. And because the current method is so simple, there isn’t the same inaccurate air of omniscience coming from the calculations. By giving the list to commanders on the ground for a “smell test,” the department is avoiding the biggest pitfall of the seventies RAND models, which led to the closing of South Bronx fire companies even as the neighborhoods around them burned to the ground. “We wanted to use as much data as we could,” says Chief Sweeney. “But we believe just as much in our borough commanders and the guys out in the field … They have the experience, and they know the areas the best.”
That said, the new methods aren’t without potential biases. The remaining companies will have to work harder and be out of the firehouse more often. When that happens, the department will be relying on what used to be the third- or fourth-due company to be the first on the scene, losing seconds and even minutes. The predicted response times are also averages of thousands of runs, which don’t say a whole lot about what actually happened during any given fire. Let’s say a company averaged a four-minute response. Does that mean it responded to nearly every call between 3:45 and 4:15, or did it respond to half its calls in 2:30 and the other half in 5:30? The department’s report doesn’t say. And as important as it is for a fire company to pull up outside a burning building, how long it actually takes to start rescuing victims and putting water on the fire is more vital. Because of hiring freezes and budget constraints, the FDNY has had to cut the number of firefighters on many engine companies, which can mean that fighting the same fire might take more engine companies than it used to.
“Focusing simply on response time, although easy to understand, is too simplistic a measure to completely rely upon,” says Glenn Corbett, a professor of fire science at John Jay College. “Providing the proper resources for firefighting and EMS is more than just how quickly you get there.” Corbett says that the bigger-picture question of whether the FDNY should be closing companies at all is still unresolved. In 2003, he was on a blue-ribbon panel that recommended closing five FDNY companies (Corbett himself was against the cuts, but was outvoted by other members).
“The question I was asking eight years ago was, ‘How do we know we don’t need to be opening more companies?’” says Corbett. “Not that we necessarily should have been. But the department desperately needed a long-term, strategic plan looking at how to deploy resources, and the department still needs that today.”