July 14, 2005
First, I want to express my gratitude to the Chairman and members of the subcommittee for your efforts to move legislation patterned on the President’s proposal, “Working Toward Independence.” Since I appeared before you in February to testify on welfare reform results and TANF reauthorization you have reported a reauthorization bill, HR. 240, to the full committee. We look forward to the Congress reauthorizing TANF soon to help even more of America’s TANF families realize the dream of economic independence.
I would like to use my time today to briefly reiterate key results from the first phase of welfare reform as a springboard for discussing the information we have gathered most recently on employment of TANF families.
Since enactment, welfare caseloads declined by 55 percent – from 4.4 million families to just under 2 million families, the lowest number since 1970. Not only did single-parent labor-force participation reach historic levels, but the earnings of those households made significant gains as well. The child poverty rate declined from 20.5 percent in 1996 to 17.6 percent in 2003. Today, there are 1.6 million fewer children living in poverty than in 1996.
I believe there are three primary reasons why these results were achieved:
Evidence of More Employment
This undercount in administrative data may occur because some recipients obtain employment, but do not immediately notify the TANF agency, and when it comes time to recertify eligibility, they simply do not keep the appointment. The TANF agency then closes the case, but because it does not know the reason, it simply classifies the exit due to “failure to comply,” “other,” or “voluntary closure.” In some States, adults who work but do not report their employment may appear to be out of compliance with work requirements and are listed as closed due to a “work-related sanction.” Researchers at the Urban Institute have found differences between the administrative data reported by States and research survey data that indicate that work-related reasons are substantially more important to case closure than the administrative data would indicate.
Further, leaver studies from 20 States and the District of Columbia indicate that 48 percent of former recipients, on average, left cash assistance for employment or because of increased earnings, about three times the percentage reflected in TANF administrative data reported by the States.
In addition to these numerous studies supporting higher than reported employment, I would like to share a new source of evidence showing that more recipients (both on and off the TANF caseload) are working than one might infer from TANF administrative data alone. But first I’d like to provide a brief overview of the current data collection mechanisms.
There are several ways States determine whether recipients are working. The first is recipient self-reporting. Recipients report that information to the State and the State in turn, reports it through TANF administrative data to us. We know that some recipients, for whatever reason – they are too busy; they cannot get through to the agency, or they just forget – neglect to inform the State agency that they have started working.
The second source of information is the “Income and Eligibility Verification System” (IEVS) for public assistance programs mandated by Congress in 1984 in recognition of the need for information to supplement recipient reporting. IEVS is a battery of computer matches that includes State Quarterly Wage data, Unemployment Insurance data, Social Security data, and unearned income data from the Internal Revenue Service (IRS). While IEVS is a great tool for correcting overpayments, it does not provide access to a national quarterly wage database or access to timely new hire data.
States do have access to new hire data in their State Directory of New Hires (SDNH) maintained by the child support enforcement agency but as with IEVS, these data are limited to in-State data only. Additionally, employers with workers and businesses in more than one State are given the option to select one State from which to report all new-hire information for its operations throughout the country. So, when Wal-Mart does not report all of its “new hires” to Colorado, neither Maryland’s IEVS nor its SDNH would be able to find out about those “new hires” in a timely manner. Many employers like Wal-Mart, hotel and food service chains, banks, hospitals, and retail outlets report “new hires” through a single State. Currently, over 22,000 multi-State companies and another 46,000 multi-State subsidiaries report to a single State. Whether we are talking about the Bank of America or the Postal Service, many of these companies employ TANF recipients. A TANF recipient could literally be working next door to a TANF agency and not show up in a State “new hires” match.
The National Directory of New Hires (NDNH)
Another advantage of the NDNH is that it provides Federal employment information, which is not available with State-based matching sources. Through the NDNH, we can get information on possible employment, for example, at the Social Security Administration in Baltimore, civilian employment at military facilities, or Yosemite National Park in California. These are all places where the Federal government employs workers, some of who may be TANF recipients.
Of course, not everyone who is employed is in the NDNH. Individuals would not be included if they are self-employed or are working the cash economy. And, of course, the NDNH would not pick up illegal activity. Even when there is a match, it does not necessarily mean the person is actually working. An individual could have been offered a job and filled out a W-4, but later accepted another job and filled out a second W-4. Or, the individual may have failed to show up for work or left a job shortly after starting.
Nevertheless, the NDNH information is available on a nationwide basis and offers more comprehensive data than IEVS or SDNH. With these advantages we sought to find out what would happen if we ran the Social Security numbers of the TANF caseload against the NDNH.
We began our investigation of the potential to match NDNH and TANF data with a pilot test in the District of Columbia. D.C. was an important test case, because many residents of D.C. work outside the District—in Virginia and Maryland—and many work for the Federal government. TANF officials in D.C. believed that matching with the NDNH could yield valuable new information about their caseload that would enable them to improve services and identify work participation they were missing.
For this pilot study, the D.C. Department of Human Services submitted the unduplicated Social Security numbers of 20,096 TANF adults to the Office of Child Support Enforcement in four matches conducted between June 2003 and September 2004. For each match, the District obtained information on earnings and W-4 reports over the preceding 15 months. The total combined period extended from April 2002 to September 2004.
The NDNH matched 6,681 individuals, or 33 percent of the cumulative, non-duplicated Social Security Numbers submitted. Of these, the District independently verified the employment information in one of three ways. First, it checked against its own records. Second, if no record of employment was found, then attempts were made to verify employment with the listed employer. Finally, if unsuccessful at those two attempts, the recipient was contacted. The District verified that 5,410 or 81 percent of the matches were actually employed, of which only 226 had verified wages reflected in the D.C. TANF case record. An additional 578 cases, or nine percent of matches, are still pending verification. And, for 10 percent of the matches, the verified information showed that the client was not employed.
Of the 5,184 clients verified as employed but who had not reported their earnings, 2,436 (47 percent) were closed because their earnings exceeded the income threshold. The cash benefits of the remaining 2,748 cases (53 percent) were reduced. Stated another way, of the 20,096 unduplicated Social Security numbers, 12 percent of these cases were closed for excess earnings and 14 percent had their benefits reduced as a result of the NDNH match. The estimated savings from the case closures and reduced benefits are nearly $10 million, an amount just over 10 percent of the District’s annual TANF block grant.
We have built on these efforts by working with 10 States, representing nearly half of the nation’s caseload. In this expanded effort, we conducted a preliminary match of the TANF caseloads against both the NDNH new hires data and the Quarterly Wage data for a single month — October of last year. We compared what was reflected about employment in the TANF data reported to us, with what is known from the NDNH.
A total of 552,033 valid Social Security numbers for TANF adults in the 10 States for the month of October 2004 were matched against the NDNH. The findings suggest that in every single State more TANF adults were involved in jobs than reflected in the employment rates of the TANF data reporting system. A comparison of employment rates using average monthly TANF data for the first quarter of FY 2005 to the percent of those reporting any quarterly earnings from the NDNH Quarterly Wage data for the same time period shows a 9.3 percentage point differential between the two data sources (25.1 percent vs. 34.4 percent). Across the 10 States, the employment rate differences ranged from 2 to 22 percentage points. Although the two rates are not measured in exactly the same way, there are important differences between employment reported by the TANF agency and the employment information available in the NDNH.
One reason for the large differences in employment between what TANF agencies report and what is found in the NDNH may be explained by the fact that TANF agencies may not have timely access to all data on new hires. Across the 10 States for the month of October 2004, adding the “out-of-State” and “Federal agency” matches increased the total number of matches by nearly 50 percent -- with increases ranging from 36 percent to 114 percent. The large number of W-4 matches from many of these States is an indication of the multi-State employers who do business in one State, but report their new hire data to other States, or of cross-border employment. Even assuming that many of these matches would eventually show up in Quarterly Wage matches, the full complement of new hires matches found only in the NDNH gives TANF agencies the ability to “locate” new employment after about 30 days, rather than finding the recipient six months after they receive their first paycheck.
Implications for Using the NDNH
Second, regular and timely matching of TANF caseloads with the NDNH can help States learn how recipients are faring – whether they need help, and whether they don’t need help. This information can help States better plan agency resources for helping those employed maintain and advance in employment, and overcome obstacles that could otherwise result in a family returning to assistance.
Third, the NDNH can identify individuals who are working and earning too much to qualify for cash assistance. This can help these families preserve their time-limited assistance, and better target TANF dollars to those most in need.
Finally, we also can use this information to help working families obtain other income supports, such as the EITC and child support.
We believe the use of the NDNH is so promising that we want to give every State a chance to determine the benefits. We are sharing this information with States and inviting them to partner with us in a regularly scheduled computer matching process. Further, in order to facilitate State use of the NDNH in the first year so they can test the benefits of these matches, we are exploring whether there are ways to provide some financial support to states where needed.
Implications for Welfare Reform
As President Bush has remarked on the reauthorization of welfare reform, “This compassionate approach will help many more Americans realize a better life of independence, hope and dignity that comes with have a job. It will also promote strong families…while freeing States to seek innovative ways to improve services to those who are transitioning from the welfare rolls to the workforce.”
Mr. Chairman, the information and plans I have shared with you today represents a next step that States and we are taking in welfare reform. But, the fundamental program changes that are embodied in the President’s proposal and H.R. 240 await the action of Congress. The Secretary and I stand ready to work with you on reauthorization to make economic independence within the reach of all America’s neediest families. I would be happy to answer any questions.
Last Revised: July 13, 2005