[San Bernardino County] DA: Welfare fraud ‘rampant’

As fewer cases go to prosecutors, questions arise about increase in administrative hearings

Blog note: this article references a 2011-12 grand jury report.

As San Bernardino County leaders have sought to identify the scope of welfare fraud that the county's top prosecutor has described as "rampant," the investigative unit tasked with detecting and deterring such abuse has pushed fewer cases to the District Attorney's office for criminal prosecution.

From 2012 to 2015, the county's Program Integrity Division (PID) Fraud Investigation Unit has increasingly preferred administrative disqualification hearings, calling the hearings a more attractive option due to changes in state law, apparent quicker restitution and relative cost savings.

The DA's office has seen a 70 percent reduction in the number of cases referred to them by PID since 2013, according to Gary Fagan with the DA's Specialized Prosecutions Division, which prosecutes welfare fraud. Last year, less than 1 percent of PID's completed investigations were sent to the DA for prosecution, he said.

An October report by county Human Services Department, which oversees PID, underscores that the unit's 32,312 completed investigations in fiscal year 2015, however, were the most of any county in California and did lead to more prosecutions than any other county in the state except Los Angeles.

But while figures provided by Human Services show closer to a 25 percent decline in referred cases between FY 2013 (227) and FY 2015 (169), the drop, nonetheless, has been acknowledged.

"Everybody knows when you want to commit welfare fraud, you come to San Bernardino County," said one county official, speaking on the condition of anonymity.

In elaborating on this downward trend late last year, PID Chief Virginia Lugo said the pivot toward handling violations administratively was a conscious effort in response to myriad factors, but largely Assembly Bill 109 and Proposition 47. Lugo added that "egregious" cases and some repeat offenders, however, were still prioritized to be pushed to the DA.

Signed into law in 2011, AB 109 eliminated jail time for non-serious, non-sexual and non-violent crimes, the county said, adding that the 2014 voter-approved Prop. 47 reduced the severity of welfare fraud from a felony to petty theft.

County spokesman David Wert said in an email Tuesday that Human Services' thinking is that these state laws have "reduced the ability of prosecutors to prosecute and get meaningful sentences for welfare fraud, making administrative hearings the quickest, least-expensive way to recover ill-gotten gains."

Administrative hearings are presided over by an administrative law judge, requiring "clear and convincing" evidence as the standard for proof needed to obtain a conviction.

Earlier this month, Wert said "the ultimate goal here is to ferret out fraud and get the money back," pointing to the state laws as having "largely removed criminal prosecutions from the counties' toolkit for combating welfare fraud."

But this notion and resulting PID approach also appear to be in disharmony with the philosophy of prosecutors.

"AB 109 and Prop. 47 do not affect our ability or prohibit us from prosecuting welfare fraud," Fagan said. "Both only affect the sentencing options for the court."

The county has recouped about $4.3 million between fiscal years 2013 and 2015 through the combined forces of prosecutions and administrative hearings, yet the latter's frequency has nearly tripled during that time frame, according to a Daily Press analysis of figures provided by the county. Also, prosecutions still account for roughly $2.9 million of that total restitution, about two-thirds, despite there being 78 less criminal than administrative convictions during the period.

It's worth noting that welfare fraud losses don't directly impact the county's budget since it's all federal money, Wert said, but the annual cost of the Welfare Fraud Investigations Unit is approximately $3 million.

According to Wert, this function is almost entirely funded — 96 percent — by state and federal funding through the CalWORKs and CalFresh funding allocations. Only about 4 percent of the cost, or $120,000 of county general fund dollars, is attributable to this function and the county general fund piece is part of a required maintenance of effort the county is mandated to pay for these programs.

The reduced emphasis on seeking criminal penalties, which could have the potential to limit a key welfare fraud disincentive, comes amid a 2011-12 grand jury report that was critical of the county's lax oversight and light penalties for Electronic Benefit Transfer card misuse. It's also occurring at a time when county 1st District Supervisor Robert Lovingood said top officials are about two years into an analysis on pinpointing the actual reach of welfare fraud in the county, which will be followed with an action plan to fight it.

Lovingood said Monday that zeroing in more on administrative hearings has "weakened" the county's efforts. He said the Board of Supervisors was reviewing internal policies with the intention of reverting to more prosecutions.

"I want to see, and I believe District Attorney Mike Ramos wants to see, the increased prosecution of those that are stealing from taxpayers," Lovingood said. "When we see intentional fraud and criminal actions, we want that action prosecuted."

Ramos on Wednesday agreed, saying welfare fraud was "rampant" within the county and vowing that prosecuting violators was "a high priority."

"We stand ready, willing, and able to prosecute every welfare fraud case. We are currently working closely with Supervisor Lovingood and the County Administrative Office to address the lack of referrals," Ramos said. "Our goal is to determine the best course of action to ensure that anyone who commits welfare fraud in our county is prosecuted to the fullest extent of the law."

Compared to FY 2013, the 25 percent decline in welfare fraud cases referred to the DA in FY 2015 resulted in roughly 37 percent less criminal convictions and around 70 percent less restitution, according to a Daily Press analysis of figures provided by the county.

During FY 2013, there were 227 cases referred to the DA, 222 convictions and more than $1.7 million in restitution. Two fiscal years later, only 169 cases were sent to prosecutors with 139 convictions returning less than $520,000, although case referrals and convictions were slightly up from the year prior.

Meanwhile, administrative convictions have skyrocketed from 99 to 268 during that time frame.

The course of prosecutions and hearings in this county also has been synchronous with the advent of a risk model tool the county has employed to weed out non-compliant welfare cases. Lugo has said the predictive model, launched in December 2012, has allowed PID to be more proactive in sniffing out fraud and resulted in a noticeable uptick in the ability to finger case discrepancies.

But it hasn't necessarily been a pathway to prosecutions.

Officially known as the Statewide Automated Welfare System Consortium IV Known CalFresh Non-Compliance Risk Model, but more plainly referred to as Analytics, the risk model tool's case referrals result in 200 to 300 investigations each month.

Regarding those investigations, "between 22.1 percent and 27.2 percent per Analytics reporting period are found to have eligibility and/or benefit discrepancies," the report said. "This is a significant increase from the 6.1 percent identification of discrepancies prior to the use of Analytics."

Before the tool was implemented, cases were referred by eligibility workers, We-Tip or other anonymous sources. The modus operandi now is to combine those past efforts with the power of the predictive model.

Lugo has said the number of potentially fraudulent cases identified by Analytics allows the on-average 22 PID investigators to visit more recipients' homes, ask questions and observe. Essentially, it's helped to put problematic cases on notice.

Wert also reiterated Tuesday that the key goals of the tool were prevention and detection, with Human Services generally rejecting the premise that prosecutions and restitution were the particular measures of its success.

"The only thing the Analytics program can do is detect discrepancies," he said, "and it seems to be doing that quite well."

At the same time, just two cases identified by and attributed to the tool have been sent to the DA's office in the last three years, resulting in less than $13,000 in restitution, according to data supplied by the county.

That underwhelming outcome has raised questions about the quality of the cases being identified by the tool and whether PID may be better off simply reverting to its previous investigative methods.

And the number of prosecutions and restitution overall also have dropped since the tool's implementation, yet the county has explained that the tool was coincidentally introduced at a time when officials were already consciously moving away from prosecutions, signaling it would be unfair to evaluate the tool amid a trend that was underway.

They maintain that the software is effective.

In the 34 months before its introduction, 635 cases were submitted to the DA's office, with the county collecting more than $1.2 million in pending and adjudicated matters, the data showed. During the post-implementation period of January 2013 to October (33 months), just 408 cases were sent to the DA's office although the county collected more (nearly $1.9 million.)

"What we were searching for (with the tool) is what can we do that would give us some indication of someone who may not be telling us everything they ought to be?" Lugo said. "That was the whole idea behind it and how quickly can we get to them? Part of that is not going to show in the (restitution) dollars."

Analytics has been renewed by Supervisors each year since 2012, costing more than $3 million, although all but 15 percent (about $457,000) has been grant funded. The Human Services report said the tool continues to be refined to "improve performance and provide on-going referrals."

"If the current system is not working, it's the board's responsibility to address and make those changes," Lovingood said. "We've had conversations, it's been a point of emphasis and it's ongoing. It's now rising to the top."

The county also has implemented other programs to address welfare-related issues, with one aimed at helping families achieve economic dependence and another putting surveillance on markets that are allowing customers to misuse their Food Stamp benefits, according to Lovingood's policy advisor Don Holland.

The percentage of families in the county likely to become self-sufficient has since increased from 12.8 to 30.3 percent, Holland said, and the county has so far investigated three markets for abetting welfare fraud. Two new investigators were hired in April for the market surveillance pilot program, a cooperative with federal and state agencies, and potentially up to four more staff members could be hired.

In April, a Victorville-based brokerage firm reported the number of people receiving government aid in local cities had grown over the last 15 years, exploding to $1.1 billion in total value of assistance, finding that 50.7 percent of residents in the High Desert and Needles were on some type of aid. But since "aid" had been defined as any combination of CalWORKs, CalFresh and Med-Cal — and amid loosening eligibility standards for MedCal in recent years — Lovingood said he planned to determine what he believed would ultimately be a more realistic figure for the region's welfare population.

The number of county residents receiving MediCal jumped 50 percent between 2014 and 2015, Holland said.

Meanwhile, Wert deferred questions to elected officials about whether substituting prosecutions for hearings had the potential to remove a deterrent.

He added there had been no talks within PID about changing the way things were done, "but clearly questions are being raised that might lead to discussions."

January 16, 2016

Desert Dispatch

By Shea Johnson