Unleashing computers to sniff out welfare fraud

Faced with a prospect of losing $1 billion to congressional budget cuts, state and federal welfare officials are developing creative ways to battle wasteful fraud.

The task is not an easy one. Reformers of the current $11.4 billion-a-year system are torn between the desire to weed "ineligibles" from welfare rools and concern that "tightening up" the system inadvertently could squeeze out those who genuinely need help.

Aid to Families with Dependent Children (AFDC) -- the federal program generally synonymous with the term "welfare" -- was designed with broad eligibility requirements to ensure that anyone needing financial help would get it. In that respect, the program has succeeded, most administrators agree. On average, 85 percent of those who are eligible for AFDC participate in the program, according to the US Department of Health and Human Services.

But the agency also estimates that about 8 percent of federal AFDC funds either is paid to ineligible recipients or is overpaid to eligible recipients.

Methods for defrauding the program have become sophisticated:

* Six massachusetts women have been indicted in a phony birth certificate ring uncovered in June. Sixty forged certificates, used to gain $400,000 in fraudulent AFDC payments, have been discovered so far.

* A couple in bridgeport, conn., are due for sentencing July 15 after pleading guilty June 2 to conspiracy. The case was a major fraud discovery, involving more than $50,000 in AFDC payments for nonexistent children.

* In Pasadena, calif., a woman recently was indicted on fraud charges for allegedly using aliases and claiming children she didn't have to obtain $337,000 in AFDC payments.

The relatively infrequent, large-scale fraud by so-called "welfare queens" usually makes the headlines. But many federal and state fraud investigators increasingly are turning their attention to the thousands of "nickel-and-dime" cases -- for example, recipients who claim desertion by a spouse who is actually still in the home, or who fail to report income from investments, property, or a job.

But investigators say the tide finally is turning in their favor -- and none too soon, given the budget crunch that looms over most state welfare departments. Many welfare officials report an increase in tips from angry neighbors of people who collect welfare while owning cars, boats, and property.

The newest tool in the escalated battle against welfare fraud is the computer. Officials in Massachusetts anticipate saving $15 million this year by using a computer system that matches the names of AFDC recipients against quarterly employee rools that employers must submit to the state. Names appearing on both lists are investigated, and many are found to be ineligible for welfare because they earn unreported income. A handful of similar computerized double-checking systems in Califonia netted more than $12 million in 46,000 investigations during 1980.

So-called "liquid asset fraud" is also under attack in California. A bill now in the state Senate would permit AFDC lists to be checked against state income tax records. If the computer finds a match, it would give investigators a lead on a potential violation of the law -- illegally earning more than $600 a year in "concealed assets" (interest of dividends) while on welfare.

California's experimental "Project Sacramento" detects unreported social security or supplemental security income payments through computer matching. Only a few months old, the program has yielded 29 fraud indictments, totaling $ 250,000 in payments which must be repaid to the state. "Project Missing Kids," started last year in Connecticut and Massachusetts, compares school attendance records and town birth records with AFDC rolls. So far, 45 cases of "nonexistent children" have been discovered.

Many state fraud detection rpograms are coordinated under the federal "Project Match," which compares all state welfare lists against federal civil-service rolls. It also crosschecks AFDC lists from every state to clamp down on those who collect duplicate payments in different states.

"Since 1978, we've found over 5,000 people who were getting too much from AFDC, and 3,000 who shouldn't have been receiving it at all," says Bob Wilson, of the US Department of Health and Human Services inspector general's office.

Critics of the current efforts to crack down on fraud say it's not worth the expense.

"The amount of fraud in AFDC, compared with other programs [such as medicaid] is small," says a former welfare official with the Carter administration. He says attempts to tighten eligibility requirements and income reporting procedures will make the application porcedure so complicated that those in need will become frustrated and give up. He likens AFDC to a type of "wage support," and says measures designed to catch phony AFDC applicants only tends to punish truly needy recipients who miss a reporting date with a welfare officer or make an error in reporting their earnings.

Fraud investigators disagree.

"Treditionally, the concern has been to make sure that everyone who needs help gets it -- to see that no one gets left out. And that's still important," says William Ferullo an attorney with the Massachusetts Department of Welfare fraud. But now, he says, recipients who stay on the rolls indefinitely and unnecessarily must be weaned from dependence on welfare.

In terms of dollars, fraud investigation pays off. With computer matching, "We spend one dollar to save two," says Larry Harrison, manager of California's Fraud Prevention Bureau.

When fraud is detected, "you save not only the money which is returned to the state, but the amount of future payments that wouldm have gone to that person indefinitely," explains Mr. Ferullo.

Along with budget cuts to AFDC, some major eligibility changes are included in the umbrella "recondiliation bill" in Congress for the federal budget. Senate and House versions differ in their particulars, but both will serve to bump many current recipients from the rolls.

Many of the AFDC changes concern the so-called "30 and a third" formula, enacted in 1967, which was supposed to encourage recipients to find work.

The premise of the formula is this: If someone could earn $500 a month in welfare payments by doing nothing, or work to earn $500, which will he do? Most likely, he'll apply for AFDC. So the dangling carrot is that if a recipient gets a job, his welfare payments will diminish only slightly -- tied to a formula that allows generous deductions for child care and work expenses.

Supporters of the formula say it makes working more profitable than welfare, and that it's the only way low-income working mothers can survive.

But critics say that by exaggerating their expenses, working recipients are eligible for nearly the same amount of aid as those who don't or can't work.

"One Massachusetts woman with three children is earning $1,264 a month in her job, and is eligible for $394 a month in AFDC," explains Jim Scanlan, an investigator for the Massachusetts Department of Welfare Fraud. "That's only $ 50 less than if she didn't work at all."

He'd like to see AFDC payments halted four month s after a recipient takes a job.