Kestin first filed another FOIA request with FEMA, this time for material it had been willing to supply earlier—ZIP code data on individual assistance grants. She asked for all the ZIP code information on disaster relief nationwide. While waiting for the second batch of ZIP code data, Maines went back to where he had started the previous September: FEMA’s website. There, he learned that the federal government had declared over 300 disasters since FEMA had begun keeping records of aid distribution in 1999. With Demma, Kestin, and O’Matz, Maines selected 20 of these to investigate, all of them cities in which their sources had told them fraud had occurred. They agreed that, as with the inspectors story, they would try to verify a handful of instances before deciding whether and how to dig deeper.

The FEMA data arrived in March, and Maines set about combing through it for evidence of fraud. As he had done for Miami-Dade County, Maines wanted to compare the aid data to meteorological reports from the National Weather Service. If he found no reports of severe weather or damage in areas that received a great deal of FEMA disaster relief, that could indicate fraud. He constructed maps demonstrating concentrations of aid, which he overlaid with other maps based on NOAA data showing locations of disasters. He then measured distances between where disasters occurred and where FEMA was distributing money. For example, he found that a great deal of aid for the California wildfires of 2003 was concentrated in poor Los Angeles neighborhoods up to 30 miles from the fires.

Another source of information was local newspapers. One disaster which sources had identified as a site of massive fraud was a severe storm in Detroit in 2000, for which FEMA had paid $168.5 million in individual assistance claims. But Maines found scant coverage of the event in articles published at the time, which seemed unusual for any disaster that could cause over a hundred million dollars worth of damage.

But for every breakthrough, there was a dead end. Because of the amount of time it took to gather and analyze weather data, and compare it to FEMA’s grant figures, Maines could research a disaster for days before realizing that fraud was either unlikely or would be difficult to prove. He says:

You look at the storm data and the money data, you call public officials or disaster relief [coordinators] to make sure there’s no damage, but then you also have to look a few days before and after the storm, and sometimes you’ll find that... these poor areas did get hit with [severe weather]... Basically it's probable that there's a lot of fraud going on here but there are blemishes; you can’t [prove fraud] nice and cleanly.

An example was Hurricane Lili, which hit southern Louisiana in October 2002. Like Miami-Dade during Hurricane Frances, New Orleans had dodged the worst of the storm but had generated a suspicious-looking number of disaster relief claims. It seemed more than plausible to Maines that fraud had occurred there. But when he researched declared disasters in New Orleans before and after Lili, he found that Tropical Storm Isidore, which had hit the city a week or so before the hurricane, had caused extensive flooding and damage in the same poor areas he had suspected of fraud. “So we had a reasonable explanation for all the FEMA assistance,” Maines recalls. “FEMA assistance from the two declared disasters was overlapping. Bad bookkeeping, maybe, but not necessarily fraud.” ^[42] Several disasters Maines examined had similar characteristics; in these cases, he had to accept the days he had wasted and move on to the next disaster on his list. Maines contrasts reporting this story to his early days as a police reporter. In the latter case, he says, “It might be a long day sometimes, but the day ends and you’re done. Here the day ends and you’ve just got another day to keep on going.”

Government. The Sun-Sentinel team persevered, however, not least because government investigations were validating the paper’s reports. On May 18, 2005, the Senate Homeland Security and Governmental Affairs Committee held a hearing on FEMA’s performance in Florida, which coincided with the release of the Department of Homeland Security’s audit of the agency, begun in November. The hearing was titled “FEMA’s Response to the 2004 Florida Hurricanes: A Disaster for Taxpayers?”

FEMA Director Brown continued to defend distributing aid in Miami-Dade County and stated that property in the area had been damaged by the storms, and further that mistakes were inevitable in a relief operation on the scale of that required for Florida’s 2004 hurricane season. Committee Chair Senator Susan Collins (R-Maine) told Brown: “I don’t think there is a tradeoff between responsive, swift assistance to those who are truly victims and protecting taxpayers against waste, fraud, and abuse.” ^[43] Kestin and O’Matz wrote of the hearing:

The committee confirmed the [ Sun-Sentinel ’s] findings and uncovered more problems, such as criminals hired as inspectors before completion of background checks and inspectors submitting claims without visiting applicants’ homes. ^[44]

Senate Hearing