HEARING ON THE USE OF DATA
SUBCOMMITTEE ON HUMAN RESOURCES
COMMITTEE ON WAYS AND MEANS
U.S. HOUSE OF REPRESENTATIVES
ONE HUNDRED TWELFTH CONGRESS
March 11, 2011
Printed for the use of the Committee on Ways and Means
COMMITTEE ON WAYS AND MEANS
ADRIAN SMITH, Nebraska
LLOYD DOGGETT, Texas
JIM MCDERMOTT, Washington
JOHN LEWIS, Georgia
JOSEPH CROWLEY, New York
JON TRAUB, Staff Director
C O N T E N T S
Advisory of March 11, 2011 announcing the hearing
The Honorable Patrick P. O’Carroll, Jr., Inspector General, Social Security Administration
Sundhar Sekhar, Principal, National Health and Human Services Practice Leader, Deloitte Consulting
Joseph Vitale, Director, Information Technology Systems Center (ITSC), National Association of State Workforce Agencies (NASWA)
Elizabeth Lower-Basch, Senior Policy Analyst, Center for Law and Social Policy
Ron Thornburgh, Senior Vice President of Business Development, NIC
INTEGRITY, AND TAXPAYER SAVINGS
Friday, March 11, 2011
U.S. House of Representatives,
Committee on Ways and Means,
The subcommittee met, pursuant to notice, at 10:01 a.m., in Room B‑318, Rayburn House Office Building, Hon. Geoff Davis [chairman of the subcommittee] presiding.
[The advisory of the hearing follows:]
*Chairman Davis. The hearing will now come to order. Before we begin the official proceedings, as many of you may be aware, tragedy has struck the Pacific Rim with a record earthquake and tsunami that has devastated our friends in Japan, and is sweeping across the Pacific as we speak. And I would just like to ask you all to join us here in the dais in a moment of silence for the victims and their families.
[Moment of silence.]
*Chairman Davis. Today’s hearing is about how the government can use data and information technology to better prevent fraud and abuse, increase the efficiency of benefit programs, and produce savings for U.S. taxpayers. That’s an ambitious set of goals.
We are going to start by asking how current efforts to use data and technology are working to improve program administration and benefit accuracy. Then we will expand on that by asking public and private experts how we can use data to provide better services for benefit recipients and at a lower cost to taxpayers.
One key goal would involve preventing improper payments. And, as the chart shows, we’ve got a lot of work to do. In 2010, total improper payments by the Federal Government reached a staggering $125 billion. That reflects payments that went to the wrong recipient, in the wrong amount, or that were used in a fraudulent manner.
It reflects many different streams of thoughts and issues related to payments, not singling out any single cause, but we have disconnected processes, disconnected systems that don’t communicate effectively together, and it’s a disservice both to the taxpayer, to the employees and the agencies who try to manage these difficult programs, and also to the recipients of benefits.
I am alarmed to note that $125 billion in improper payments is an average of over $1,000 per household in the United States. Two of this subcommittee’s programs, unemployment insurance and supplemental security income, accounted for almost one‑fifth of those improper payments, costing taxpayers over $23 billion last year. To address those types of errors and improve administrative efficiency, government needs to work a lot smarter.
So, we have asked lots of smart people here today to help us learn about the current state of data matching and its potential for making major strides in program efficiency and effectiveness in the future. For example, we have seen the private sector find ways to use data to more efficiently detect patters of misuse, such as when credit cards are lost or stolen and streamline backend payment processing. We want to apply those same sorts of lessons, proven private sector concepts, in our programs, as well.
We have seen some of those lessons already applied in states like Utah and Florida. They are using data matches to fill application forms with reliable and verified data, reducing the manual burden on case workers, and increasing payment timeliness and accuracy. This also allows caseworkers more time to spend with their beneficiaries, handling more complex cases, as they should.
On the federal level, a data match success story involves legislation crafted by this subcommittee related to prisoners who should not be collecting disability checks. As a result of that legislation, the Social Security Administration now has a system by which they collect timely prisoner data from state and local jails, rather than relying on the honesty of inmates, literally, to end their own benefits.
From 1997 to 2009, the system helped identify over 720,000 incarcerated individuals who should not have been receiving SSI benefits, contributing to billions of dollars in savings each year. It has been so successful that this data is now shared with the child support enforcement and food stamp programs.
Looking forward, we are interested in promoting the development of a more common set of data elements across all programs in the government. This will improve efficiency and savings in our programs, as well as other costly benefit programs like food stamps and Medicaid that many of our program recipients collect simultaneously.
These issues stretch beyond our subcommittee’s borders to include laws like the Computer Matching and Privacy Protection Act of 1988. That means we will have to work with other committees to achieve real and value‑adding changes, like making updates for current technology, and allowing for computer matching agreements to be completed in a more timely manner.
Ultimately, improving data matching will help us to better measure the effectiveness of multiple programs, and more efficiently target resources to achieve goals like promoting more work and earnings, reducing poverty, and ending dependence on government benefits. These are goals that we should all agree on.
We look forward to all of our witnesses’ testimony. Without objection, each member will have the opportunity to submit a written statement and have it included in the record at this point. And I will now yield to my friend, Mr. Doggett from Texas, if he would like to share an opening statement.
*Mr. Doggett. Thank you, Mr. Chairman. And I believe these are goals that we do all agree on. Use of government programs, whether done by a pharmaceutical manufacturer or a defense contractor — I will try that again.
These are goals that we all agree on. And abuse of government programs, whether a pharmaceutical manufacturer, a defense contractor, or a food stamp recipient, are all unacceptable, especially when there are so many Americans in need of genuine help. Taxpayers have a right to expect that public benefits go only to those to whom they are entitled, and that we seek to eliminate all types of improper payments, misuse of the taxpayers’ monies.
Today we are appropriately exploring the extent to which improved sharing of data can help in achieving that objective. Most public assistance programs already use data from a variety of sources to verify an applicant’s eligibility.
For example, welfare and unemployment agencies routinely check wage data which is collected both by state and national databases in determining initial and continued eligibility. Another example is the Social Security Administration, which cross‑references bank account information for those who are applying for Supplemental Security Income, or SSI.
Such information is obviously sensitive, so we need to ensure that, as we data‑share, we have safeguards to maintain appropriate confidentiality and prevent use for unauthorized purposes.
Additionally, applicants and recipients need to be given an opportunity to correct any incorrect, any false information or out‑of‑date information.
Just as data‑sharing can detect individuals who should not be receiving benefits, I believe they can also be used to improve outreach to Americans who are eligible for assistance, but who are not receiving it. We still have a significant number of poor seniors, for example, who have never accessed the assistance that they need, the extra help that they need, on prescription drugs under Part D of Medicare. I favor using data‑sharing to both reduce fraud, and increase access to those who need help.
One example of where this appears to be working is in the City of Philadelphia, where seniors who may be eligible for but not receiving both food assistance from the SNAP program and help from the Medicare prescription drug coverage, are checked on the basis that they are enrolled in other programs with similar eligibility standards.
A couple years ago, in 2009, the President issued an executive order directing federal agencies to intensify their efforts to reduce improper payments of the type to which the chairman referred. One element of this effort is a new partnership fund to help the states establish pilot programs to identify new and innovative ways to reduce fraud and abuse, and to test better methods of improving program integrity, such as reducing overpayments in the Earned Income Tax Credit and in the TANF program, as well as unemployment insurance.
Unfortunately, the Republican spending plan that is before Congress at present for the remainder of this year would cut funding for this very worthwhile effort to reduce fraud and abuse. This is reminiscent of our first subcommittee hearing on unemployment. Since that time, the same Continuing Resolution that has been proposed by the Republican Leadership would, according to the folks I talked to in Texas, eliminate about two‑thirds of our workforce centers in Texas, and I’m sure have a similar effect in the rest of the country.
I look forward to hearing from each of our witnesses about how to ensure that these public assistance programs assist only those who are intended to benefit from them, and do so in the most effective and efficient way, free of abuse, that we possibly can have.
Thank you very much, Mr. Chairman.
*Chairman Davis. Thank you very much, Mr. Doggett. Before we move on to our testimony, I would like to remind our witnesses that you are limited to five minutes of oral testimony. However, without objection, all of the written testimony will be made part of the permanent record.
On our panel this morning we will be hearing from a distinguished group of people who are living in the real world on this issue from a variety of perspectives in government, the private sector, and bridging both. And we appreciate your valuable ideas and insights.
Our first is The Honorable Patrick O’Carroll, Jr., inspector general of the Social Security Administration; Sundhar Sekhar, Principal and National Health and Human Services Practice Leader at Deloitte Consulting; Joseph Vitale, Director the Information Technology Support Center at the National Association of State Workforce Agencies; Elizabeth Lower‑Basch, senior policy analyst at the Center for Law and Social Policy; and Ron Thornburgh, senior vice president of business development at NIC.
Inspector General, please proceed with your testimony.
STATEMENT OF PATRICK P. O’CARROLL, JR., INSPECTOR GENERAL, SOCIAL SECURITY ADMINISTRATION
*Mr. O’Carroll. Good morning, Chairman Davis, Mr. Doggett, and members of the subcommittee. Thank you for this invitation to testify today.
Data matches have proven to be effective tools for SSA to improve payment accuracy and protect government funds. For many years, my office has recommended that SSA pursue data matches among Federal, State, and local agencies, to make sure that the right person receives the right payment at the right time.
SSA and agencies across the government have renewed their focus on reducing improper payments since President Obama signed the Improper Payments Elimination and Recovery Act of 2010. To comply with the act, my office is working with SSA, OMB, and other inspectors general to identify program vulnerabilities and develop solutions to reduce improper payments.
One of our earliest reports on data matching involved prisoners receiving Social Security benefits. SSA’s data matching with prisons has prevented billions of dollars in overpayments. We determined SSA lacked agreements with thousands of local and county corrections facilities to obtain prisoner information. The absence of these agreements led to significant overpayments to prisoners who were not eligible to receive benefits.
On our recommendation, SSA pursued legislation that eliminated the need to enter into data-matching agreements for prisoner records. Today, SSA receives prisoner information on a monthly basis, and matches it against benefit records. SSA’s most recent estimate puts the savings from this initiative at over $580 million per year for the title II program alone.
SSA’s Access to Financial Institutions project, or AFI, is another data‑matching initiative we recommended years ago that helps the Agency prevent payment errors that had been commonplace. AFI allows SSA to receive financial account information electronically, rather than rely on beneficiaries to report assets that may reduce or eliminate their benefits. Self‑reporting is a leading cause of payment errors. The Agency expects to save $100 million in Fiscal Year 2011 because of the AFI program. The system is present in 25 states, and SSA plans to implement AFI in the remaining states this year.
Those are two success stories, and my office has made other data-matching recommendations to SSA. Those recommendations include: working with State bureaus of vital statistics to obtain death information electronically, as well as information on beneficiaries’ marital status; exploring exchanges with states that maintain automated workers’ compensation databases; and assessing the costs and benefits of obtaining vehicle information from states to verify resources of SSI recipients.
We also have planned reports on potential matches of SSA beneficiary information related to unreported property, pensions, and marital status. We in OIG use data matches in our work, as well, but the Computer Matching and Privacy Protection Act requires formal computer matching agreements that can take years to complete. This prolonged process can delay or derail time‑sensitive audit and investigative projects.
In 2010, the Department of Health and Human Services obtained a legislative exemption for data matches designed to identify fraud, waste, or abuse. We are pursuing a similar exemption, which could serve as a vital tool to our organization as we combat fraud in SSA’s programs and operations.
In conclusion, data matching serves as one piece of a large integrity puzzle for SSA and other agencies. As Chairman Davis has suggested, data matches across the Federal Government could reduce improper payments and improve service to the American public. Just as SSA strives for payment accuracy, so too should all other government agencies.
My office will continue to work with this subcommittee and SSA in an effort to improve customer service, ensure program integrity, and increase taxpayer savings.
Thank you again for this invitation to testify, and I will be happy to answer any questions.
[The statement of The Honorable Patrick P. O’Carroll, Jr., follows:]
*Chairman Davis. Thank you very much, Inspector General.
STATEMENT OF SUNDHAR SEKHAR, PRINCIPAL, NATIONAL HEALTH AND HUMAN SERVICES PRACTICE LEADER, DELOITTE CONSULTING
*Mr. Sekhar. Good morning. Thank you, Chairman Davis, Mr. Doggett, and distinguished members of the subcommittee, for inviting me to testify today. As I explained in detail in my testimony, there are three primary challenges in today’s human service daily exchange environment. And I believe the data exchange concepts and models followed in the private sector could offer opportunities for human service programs to consider. I will go over them briefly now.
Number one. In the administration of human service programs, often caseworkers spend significant portions of their time in collecting and verifying information manually of the client benefit application, reviewing their proof of verifications and validations such as income assets.
In the private sector, institutions such as banks and health care companies rely on advanced data exchange models using consumer‑to‑business and business‑to‑business exchanges that minimize workers’ manual activity in the initial application processing and the verification steps. In a typical bank model, the majority of these verifications and validations are performed in an automated fashion, relying on sophisticated data brokers that are available with information about a client.
This model has really good parallels in the human service environment. By automating data exchanges based on information available from federal and state exchanges, the human service systems can pre‑fill application information already known about a client or a household, and verify some of their proof automatically.
Number two. While every human service programs shown on the chart use some form of data exchanges for verification and validation, there is no single data standard across these programs. In addition, how the data exchange information is defined, processed, and how automation is applied to use these results are not consistent, either.
In the private sector, many of the data exchange transaction formats have been standardized. This allows for them to collaborate across the private sector entities such as employers and banks, and also rely on credit check processes as the basis for verification. Usually their underlying infrastructures are able to handle real‑time exchanges. And each entity determines how to apply the data exchange information that they receive. As a result, they are able to use event‑based processes, and also some predictive techniques that can trigger automatic events instead of worker action.
This also has many parallels in the human service environment. The state and the Federal Government could define standard code sets for commonly‑transacted human service data elements, such as change in income or change in address. By doing so, they bring consistency to data standards, and also common expectations on what needs to be done, based on those changes. And this can be done not just within a state, but also across states at a federal level.
Using that standard as a base, the states could consider moving to a human service collaboration exchange, as shown on the chart, that shares federal, state, and other publicly‑available information exchange for human service programs. The human service programs operating at a state level working with the federal agencies could subscribe to that exchange, and also contribute to that exchange. And their access would be limited, based on what’s allowable for security and privacy controls. Ultimately, this helps the state agencies gain access to a common set of data exchange information that they can use to maintain program integrity.
And, number three, in human service programs, often the service delivery model is still high‑touch, meaning case workers often interact with clients, irrespective of whether they follow a normal business process or they need additional assistance for their benefit processing. This causes a significant workload impact to the case worker.
In the private sector, the prevailing model is most of the common transactions are automated, using data exchanges, and performed without worker intervention. Whether you want to shop online or check your bank accounts or report change in information, the initial interaction is really with that worker intervention. Workers are only assigned to cases that require further review.
Again, this has parallels to the human service environment, as well. They face similar challenges in terms of shortage and case workers, and also increases in workload. As a result, a high‑touch model is expensive and not really practical for all consumers when you’re serving. Automating federal and state data exchanges could drive normal day‑to‑day transactions directly to customers, using a citizen‑to‑government model or using business‑to‑business transactions.
And finally, as you see in private sector, there are additional data mining, data predictive modeling, and other newer concepts that are being explored, which could have parallels to the human service environment. This will ultimately help to proactively manage program integrity, reduce worker time, and improve customer service, ultimately resulting in taxpayer savings.
Thank you. And I will be happy to answer any questions you may have.
[The statement of Sundhar Sekhar follows:]
*Chairman Davis. Thank you very much, Mr. Sekhar.
*Mr. Vitale. Thank you. Good morning, Chairman Davis, Ranking Member Doggett, and members of the subcommittee. NASWA represents the workforce of development agencies of all 50 states, the District of Columbia, and Puerto Rico. Today, states face aging IT systems processing UI claims and collecting wage data. And in the past few years, workloads in the unemployment insurance agencies are at an all‑time high. Consequently, customer service and program integrity have suffered. And the UI overpayment rate has not improved.
The U.S. Department of Labor estimated the overpayment rate at 10.6 percent for fiscal year 2010. As Figure 1 highlights, the major types of overpayments are: lack of timely or accurate information on reasons for separation; claimant failure to timely report a return to work; and unmet work search requirements. These account for almost 70 percent of all overpayments.
To help reduce the first two types of overpayments, U.S. DoL, with NASWA, funded a consortium of six states, multi‑state employers, and employer agents, to create a technology solution: the State Information Data Exchange System. SIDES enables states and employers to securely transmit requests and responses for separation information over the Internet, using a standard data exchange format. Currently, most states request separation information from employers using a manual and paper‑based process through the mail. SIDES automates this process. States receive more timely and accurate, detailed information from employers, resulting in more timely and accurate benefit determinations.
As Figure 2 shows, SIDES is in production in four states: Colorado, Georgia, Ohio, and Utah. Eighteen additional states have received funding from USDOL to integrate SIDES into their UI IT benefit system.
A second SIDES data exchange format, the earnings verification, has the potential to reduce overpayments resulting from a failure of claimants to timely and accurately report their return to work. The SIDES earning verification, format will enable states to augment hire information received from the National Directory of New Hires with information from employers on an individual’s start date and earnings.
SIDES is an example of a data exchange and matching technology that will address several UI areas: administration, customer service, administrative costs, and overpayments. NASWA’s National Labor Exchange Initiative offers the promise to reduce overpayments stemming from a failure to meet the work search requirements. The NLX is a free advanced job search engine used by employers and job seekers nationwide.
The NLX has been adopted by 49 state workforce agencies and the District of Columbia, offered in partnership with Direct Employers Association, composed of 550 Fortune 1,000 employers, the NLX has provided more than 9,000,000 job postings since 2007. NLX helps UI claimants meet their work search criteria, and hopefully return to work more quickly. Further, NLX uses USDOL’s occupational coding system. States coding UI claimants’ most recent work experience are able to generate matches to NLX‑provided jobs.
Both SIDES and NLX offer great potential in reducing UI overpayments and improving customer service. However, many states will be slow to adopt these technologies, because of their aging core UI IT systems. Figure 3 shows that the average state UI benefits and tax system is 23 years old. Many states use outmoded, less flexible 1970s mainframe technologies. Systems over 40 years old are still in operation today.
States urgently need to modernize their core IT systems. However, undertaking this effort as a single state has shown to be challenging, resource‑intensive, and very expensive. Recently, USDOL awarded two groups of four states each funding to explore the feasibility of building a common UI IT system. The pooling of resources through state consortia potentially offers states a more cost‑effective option to upgrade their UI systems, and participate in data exchange initiatives, such as those discussed here.
In closing, I would like to inform the subcommittee of an exciting proposal for an applicant director and exchange system that NASWA recently submitted to the OMB Partnership Fund for Program Integrity Innovation. Based on the SIDES technology and architecture and standard data exchange format, this system would create a potential index of applicants for predefined social programs such as UI, TANF, SNAP, and Medicare, etc.
Operating as a data exchange system and not a data warehouse, it would serve as the single source of customer data for use in determining program eligibility. The goal is not only more accurate benefit eligibility, but also better customer service.
I appreciate your time, and I am happy to respond to your questions.
[The statement of Joseph Vitale follows:]
*Chairman Davis. Thank you, Mr. Vitale.
STATEMENT OF ELIZABETH LOWER‑BASCH, SENIOR POLICY ANALYST, CENTER FOR LAW AND SOCIAL POLICY
*Ms. Lower‑Basch. Thank you. I am honored by the opportunity to testify today. I am at CLASP, a national non‑profit engaged in research and advocacy for policies that improve the lives of low‑income people. We appreciate your holding this hearing. We share your concern with reducing error rates and fraud in order to save taxpayer funds, preserve funding for those who are truly eligible, and protect public support for programs.
Data matching can also reduce administrative costs and improve customer service. All states are already required to participate in certain data exchange systems, including the Income and Eligibility Verification System, and the Public Assistance Reporting Information System, or PARIS, to match against federal and state public assistance records, as well as federal wage and veterans records.
I am going to highlight a few programs that are taking it to the next level, and using data matching proactively to help ensure that eligible people are getting benefits.
Washington State uses the PARIS system to identify Medicaid recipients who are eligible for veterans health insurance and vet coverage and benefits, but aren’t getting it. For example, disabled veterans who are in a nursing home receive a reduced benefit of just $90 a month. Upon discharge from the nursing home, they are supposed to go back to their usual benefit. But that sometimes doesn’t happen. And Washington can look in the PARIS system and identify these cases, and make sure they get their full benefit restored.
Another example is the Benefits Data Trust, which you mentioned before, which cross‑references data from a range of sources to identify senior citizens who appear to be eligible, but are not enrolled in public benefit programs, and then can do targeted outreach and application assistance to just those individuals. And this is one of the most cost‑effective ways to enroll seniors in the low‑income supplement program under Medicare.
I also did want to mention the OMB Partnership Fund for Program Integrity Innovation, which is designed to identify innovative ideas like this, and conduct rigorous demonstrations of their ability to reduce administrative costs and error rates without denying access to qualifying individuals. This fund has spent about a year now soliciting and refining proposals, and they have just started to fund the first projects. And the first one they have selected is that the IRS is going to work with at least one state, maybe more, to look at the public assistance information to validate EITC eligibility, because that has the information about family relationships that Treasury does not always have.
I did want to draw attention to some cautions that need to be kept in mind. Data matching is only as good as the data that goes in. And we all know that people can have similar names. And that’s how late Senator Ted Kennedy got stopped on the no fly list. Social Security numbers are unique, but we all know people make mistakes entering them in, and that can cause errors.
When a matching system flags a discrepancy, this should definitely be a basis for further investigation. But it doesn’t automatically disqualify someone, or mean that they were trying to do fraud. And the CHIPRA match for Social Security records to verify citizenship offers a good model for due process protections. If Social Security doesn’t report a match, clients get 90 days to prove their citizenship through another mechanism before they lose their benefits. And this is important.
Alabama reports that in the first year of doing this, they got over 1,000 applications where SSA did not find a match on the first try. But all but 28 of those did get documented as citizens, they just either needed to fix errors and resubmit or document it in a different way.
It’s also worth noting that income can be highly volatile, particularly for hourly workers. You can earn different amounts each week, depending on how many hours you work. And so, someone might say $280, the data match is going to come back with $292. And that’s not fraud, and it shouldn’t also trigger constant adjustment of benefits, because that’s just an administrative nightmare for both programs and the recipients. It makes sense to ignore variations under a certain amount, and most states use their policy discretion to do so.
So, thank you.
[The statement of Elizabeth Lower-Basch follows:]
*Chairman Davis. Thank you very much.
STATEMENT OF RON THORNBURGH, SENIOR VICE PRESIDENT OF BUSINESS DEVELOPMENT, NIC
*Mr. Thornburgh. Mr. Chairman, Mr. Doggett, members of the subcommittee, thank you for the opportunity to discuss how well‑designed technology systems are helping government agencies match data to improve customer service, uphold the integrity of programs, and save taxpayer dollars.
My name is Ron Thornburgh, I am the senior vice president of business development for NIC. NIC partners with 23 states around the country, providing official government portals, as well as online services. Prior to joining NIC, I served as the Kansas Secretary of State for 16 years, and was very involved in my home state’s drive to enhance states’ digital government services at that time.
I commend the subcommittee for examining how government can use data matching to more efficiently and effectively deliver services to its citizens. It’s important for you to know forward‑thinking leaders are doing this at all levels of government today, as we speak.
The states we serve focus on using cost‑effective means of bringing together key data sets that are managed by different agencies, housed in IT systems that often do not talk to one another effectively and, quite frankly, if at all.
For example, we have helped the State of Montana build an e‑government solution called Montana Connections. This service allows Montana residents in need of public assistance to apply with the single online application for Medicaid, children’s health insurance, temporary assistance for needy families, and supplemental nutritional assistance.
Prior to the use of this new online service, approximately half of all paper applications were rejected due to ineligibility or unanswered questions. Montana Connections ensures that every application is 100 percent complete before it is sent to the appropriate state and county office. These actions alone have dramatically reduced the incomplete and misrouted application submissions that needlessly take up agency caseworker time.
We also built a technically similar system in Arkansas to help the state’s department of higher education more effectively make financial aid available to students. This service aggregates the state’s 21 scholarship, grant, and loan programs, and allows citizens to provide basic screening information to determine eligibility, and submit applications to any of the programs through a single online form.
As a result of this data matching solution, financial aid applications increased 440 percent, and more than $150 million was distributed in the program’s first year. By comparison, the state was unable to match all of the money in the program with the deserving students before the online system was in place.
Now we need to talk about overcoming barriers. These are just two examples of successful data matching programs. Like others, they have proven that the structural, cultural, technical, financial, and design barriers to interagency cooperation can be and have been addressed successfully.
First, structural. Any program involving more than one agency in a single IT system will require collaboration. Agency leaders, while ensuring financial and efficiency benefits to their own agency, must agree to work together to reach a common goal. This is an absolute requirement for any data matching program to succeed.
Next, cultural. Online technology solutions are removing the perceived stigma of applying for social services. People who previously may have been too uncomfortable or unable to go to a government office to apply for support in Montana now do so, thanks to the privacy and security afforded by the online system.
Technical. Shared business rules are an essential component of a successful data matching initiative. In Montana, for example, all the agencies simply work together ‑‑ I say “simply” ‑‑ work together to identify a common language and set of requirements ‑‑ and this is important ‑‑ without sacrificing their own unique agency requirements.
Financial. Paying for a new system is a challenge every government faces. Many of the states we work with have used a self‑funded approach to build systems and services without requiring any appropriation. Modest transaction fees applied to a limited number of commercially‑valuable services, primarily business‑to‑government, are used to fund the development of e‑government systems like the data matching solutions referenced in Montana and Arkansas, without cost to the citizens or the agencies. We have successfully used this model with another departmental level federal data system, and believe the similar funding approach could support the types of data matching solutions the subcommittee is discussing today.
Lastly, design. Data matching systems are only effective when constituents use them, and successful solutions place a high priority on developing straightforward, user‑friendly interfaces on a variety of delivery platforms.
Mr. Chairman and members of the subcommittee, states are using data matching successfully. I believe you can, too. The projects that I have described will continue to provide opportunities to link diverse systems together in ways that provide real‑time eligibility screens and approvals that improve service levels and save money, increase constituent satisfaction, and, very importantly, eliminate fraud, waste, and abuse.
Thank you, Mr. Chairman. I look forward to taking your questions.
[The statement of Ron Thornburgh follows:]
*Chairman Davis. Thank you. Your time has expired. We are going to move on to questions now. And just before we get into that, I want to comment on one perspective.
As often happens in the government, Washington, D.C. is the lagging indicator with legislation versus where technology in the rest of the country is. The Computer Matching and Privacy Protection Act of 1988 went into action at a time that we lived in a different technology world, with different methods of sharing information. The fax at the time was the radical new concept for rapid sharing of information, business‑to‑business, and at a personal level, as well.
And realistically, when we look at this, and trying to tie this information together ‑‑ and I am going to highlight something that Ms. Lower‑Basch had shared ‑‑ that matching done right, in an integrated fashion, will free capacity to manage by exception, instead of having to spend an inordinate amount of time. My own wife, in fact, is on one of those same lists that the late Senator Kennedy was on, after being through numerous security clearances in the military with me.
We have disconnected processes, and that can’t be fixed in the current data environment. And we have many of our citizens, many frustrated agency workers that are trying to be good stewards of the taxpayers’ money that lose this in process.
And I am simply going to throw out, for those who are here and for our fellow Members, there are three basic kinds of activities: those that add value, those that add business value, and those that add no value. Unfortunately, businesses learned this in the competitive transitions of the 1980s and the 1990s, that there are more non‑value‑adding activities than we realize in our day‑to‑day lives. Often, 80 or 90 percent of the things that are performed, often out of necessity, to get the job done don’t really add value to our customer at the end of the day, to our client, or serve the taxpayer necessarily, as well as possible.
Let’s take somebody who is a social worker. I spent many years involved with an organization known as CASA [Court Appointed Special Advocates], working with children, trying to be kept from falling through the cracks as a result of neglect and abuse. A volunteer or a social worker, case worker, is dealing directly with that client. That’s a value‑adding activity, being able to counsel, to directly document clinical information that is necessary to help that young person move forward.
However, we move into business value adding, those are the statutory required measurements that have to be submitted. And, yes, some of those may be questionable, but those are the things that can’t necessarily be changed in the near term.
But what we find with many of our folks in the agency community, as well as those who measure and try to account for this, as well as the clients themselves in many cases, is that they’re chasing data, trying to find that lost information, spending hours and hours and hours of time. And every hour that is spent trying to find a missing piece of information is one hour that is not adding value, or one hour that could be given back to the country, to the taxpayers, or dollars that would not necessarily be wasted.
So, as Mr. Doggett and I have talked, we have common ground on this, we want to work together to find ways to integrate this so that we can have a comprehensive discussion.
This week we learned that government payouts, including Social Security, Medicare, and unemployment insurance, make up more than a third of total wages and salaries of the U.S. population. It’s a record figure that will only increase in the years ahead. I ask unanimous consent to insert an article providing more detail about that in the record.
*Chairman Davis. Without objection, that is so ordered.
[The information The Honorable Geoff Davis follows:]
*Chairman Davis. The committee has jurisdiction over some of the largest of those programs, including Social Security and Medicare. This subcommittee has jurisdiction over somewhat smaller, but no less significant programs like welfare, unemployment, and SSI.
Let me be clear. I am not making an evaluation of the recipients of those benefits, or the benefits that are paid out. That is a separate discussion from what we are talking about today. What we are talking about is a process that largely, across much of our economy, has a significant impact if we have these data problems that can contribute to waste, poor accounting, or improperly matched information.
My question pertains to the idea that programs should use a common set of data, programs in our jurisdiction that use that common data set today, and always verify data provided by applicants to ensure we’re paying the right people for the right benefits. Do you feel that the systems that we have under our jurisdiction are accomplishing that mission?
In addition to that, for example, is the way that we ask for and confirm someone’s identity a best practice in each of our programs? How about their current work and earnings or savings and other resources? Or a place of residence, citizenship, and even continued presence in the U.S.?
In short, I would like the panel to think about what we do today across the range of programs under the Ways and Means Committee’s jurisdiction, and especially this subcommittee, and help us review whether the data that we collect to administer the programs is the right data, whether what we collect can be and is confirmed in a systematic way, and whether those programs share that data to ensure we’re paying the right people the right amount of benefits across programs and states.
Would anyone care to comment? And since this is a big question, I welcome responses for the record describing needed improvements in significant detail. Inspector General?
*Mr. O’Carroll. Mr. Chairman, I will take the first crack at it. There are multiple facets to this issue. Probably the one that you’re talking about is the sharing of information across government agencies. You also mentioned the need for computer matching agreements. I think these issues are parallel.
Each government agency has to apply for the computer matching agreements. And, as a result, each agency, every two‑and‑a‑half years, is renewing individual matching agreements. There is not any coordination among government agencies. And, under the Computer Matching Act, one agency can’t share with another agency without an agreement.
And, as you said, I think it would be better if there was a way that we could allow all federal agencies to share data back and forth, at least if the purpose is for making sure the right person gets the right benefit, and to make sure that there isn’t any duplication across the government. So ‑
*Chairman Davis. Great, thank you. Anybody else? Mr. Sekhar?
*Mr. Sekhar. Mr. Chairman, I have two concepts, based on your questions, that might be relevant here.
One is when you look at the application information that is required for the different human service programs, there is a fair bit of commonality on the kind of questions that is being asked of a client. So, if there is a way to standardize the common elements across TANF ‑‑ child care, child welfare, or even, in some cases, Medicaid ‑‑ so that will reduce some strain of the data capture on the worker side.
And the second piece is, back to the exchange with SSA, I think there is an opportunity for the states to consolidate their request of SSA to exchange, as opposed to each of the programs exchanging independently. So that also brings a level of standardization for what they would do with that information.
*Chairman Davis. Great, thank you. Ms. Lower‑Basch?
*Ms. Lower‑Basch. Yes. I would say there are certainly places and examples where it’s working well. But, by and large, there is a lot of challenges, and people having to bring the same information that they have just told to one case worker to the next worker two weeks later, and no talking. So I would say more gloomy than positive, overall.
*Chairman Davis. In the current, you’re saying.
*Ms. Lower‑Basch. In the current. In the current, yes. I think there is certainly potential, but we’re not there yet.
*Chairman Davis. I think about how we can cross data across organizations when we buy things currently in the retail environment. It’s probably a more ideal model of where we would like to be at the end of the day. Mr. Vitale?
*Mr. Vitale. Well, validation of the data in the unemployment insurance program varies from state to state. And many of the same agencies within the state are validating the same identity of that individual.
For instance, in my home state of New Jersey, we validate the individual by going against the motor vehicle system and the Social Security Administration. And once we have that information validated, that should be available to other agencies within the state. And currently, it is not. And that’s the same with state‑to‑state.
*Chairman Davis. Thank you. Last, but not least, Mr. Thornburgh.
*Mr. Thornburgh. Thank you, Mr. Chairman. If I might, I may go back to my service as secretary of state in the state of Kansas, because I faced many of these very same challenges at that time.
We were in the process of developing a system, just a ‑‑ what I viewed as a simplistic one‑stop business services. The thing that always amazed me is that a business person would want to come and hire people and create jobs and do great things for my home state, and we would make them march from agency to agency to agency. And the really neat thing was that we all asked the same questions: who are you, where do you live, what do you want to do? But we treated it like nuclear secrets, and then we’re unable to share that information across agencies. So we finally got everybody together and we were able to do that.
The second example would be motor‑voter. We matched the state voter registration database with the state motor vehicle driver’s license database so that when an individual applied for a driver’s license, they automatically updated their voter registration status, as well. So when they moved, their voter registration moved with them as well.
It was a vastly more difficult process than I thought it should have been at that time to create the incentives for all the different agencies, because incentive has to be ‑‑ you’ve got to make it better for that agency, as well as for the constituent, in order for them to want to come along and work together with that.
So, there is a lot of work to be done with that. But I go back to my opening statement. It is being done time and time and time again right now. It’s certainly time for us to continue at this level, as well.
*Chairman Davis. Great. Thank you very much. I would like to yield to my good friend from Texas, Mr. Doggett.
*Mr. Doggett. Thank you, Mr. Chairman. And thanks to our witnesses for your helpful responses to the important questions that the chair just raised.
Our role here is, of course, not just to legislate, but to exercise oversight and to try to nudge along some bureaucracies that are sometimes a little lethargic and slow‑moving.
And I gather, Mr. Thornburgh, just to pick up where you left off, that while there are a number of things that can be done, none of them are free. They require allocating resources to accomplish these objectives when people hire the services of your company in Montana and the other states that you mentioned.
*Mr. Thornburgh. Thank you. Thank you, Mr. Doggett. I would be happy to answer that, in that I could take quite a bit of time talking about the self‑funded model that we use at the state level. And I won’t take all of the committee’s time talking about that ‑
*Mr. Doggett. Actually, I want to ask you one specific question about that.
*Mr. Thornburgh. Okay.
*Mr. Doggett. But all I’m asking you now is we would always want there to be a cost benefit ratio that would yield a reduction in cost for the money spent. But to undertake the initiatives that you’re talking about require the expenditure of funds, don’t they?
*Mr. Thornburgh. Well, no, sir.
*Mr. Doggett. They’re free?
*Mr. Thornburgh. No, sir.
*Mr. Doggett. Okay.
*Mr. Thornburgh. The ‑
*Mr. Doggett. In Montana, for example, you mentioned that one way that you financed this was to charge a transaction fee to the businesses involved.
*Mr. Thornburgh. If I could expand on that ‑
*Mr. Doggett. Sure.
*Mr. Thornburgh. ‑‑ just for a moment, because we have to look at the entire statewide enterprise. The Access Montana, which is the state government portal, essentially what happens is we will have a multiple of hundreds of different applications working through a number of different agencies.
Let’s say ‑‑ and I apologize, I don’t know the exact number in Montana, but let’s say there are 400 applications in Montana. Of those 400 applications, probably 20 will be associated with some kind of financial transaction. And then, those 20 different transactions, or those 20 different applications, will provide the funding for the other 380 applications.
So, in an instance like this, with a data‑sharing model, the enterprise would fund the development of that model, so there is no cost to the agency, there is no cost to the citizen using those services. There are commercially viable transactions throughout the enterprise of government in which businesses make a business decision as to whether or not they want to file or retrieve data electronically. When they do so, there is a convenience fee, a small fee, that is attached to that. And then that is what is reinvested to the other applications.
*Mr. Doggett. Increase a fee, then, to the business that access this service to help pay for this?
*Mr. Thornburgh. Not necessarily driven to that ‑‑ for instance, with this data sharing, it may not be a fee directly associated with this particular data set.
*Mr. Doggett. I think I understand. And, Mr. Vitale, you indicated that you have some ideas already underway, and one of them is proposed to this new fund. Right? And I gather from what you’re saying, and as you describe the states, that it’s not so much a matter of our passing new laws here ‑‑ though some may need to be tweaked ‑‑ as it is having adequate resources to do the things that the states would like to do.
*Mr. Vitale. Let me address the two questions.
*Mr. Doggett. Sure.
*Mr. Vitale. First, the application to the fund. Yes, we have a proposal in to the OMB Partnership Fund for Integrity Innovation, and I think some of the questions from the chair could be addressed by that fund ‑‑ by that proposal, as an interim step in getting to this common database or common definitions.
Our proposal calls for going to the agency that first collects that data, and making that the main source of the data, and not bringing it into a common repository, but have a pointer to that as sort of an index file housed centrally, so the next agency that comes in looking for that data knows exactly where to go. They hit that file, and they know that this person ‑
*Mr. Doggett. Do you think the chances of accomplishing that will be improved by slashing the Partnership Fund by a third?
*Mr. Vitale. I ‑
*Mr. Doggett. And I also received a message from the organization that you are here representing, indicating their great concern about the proposal in the same Continuing Resolution to eliminate all funding for the Workforce Investment Act. I know you’re principally in the technology field, but I gather you join your agency ‑
*Mr. Vitale. Sure.
*Mr. Doggett. ‑‑ in opposing that.
*Mr. Vitale. Sure, yes. That would have a dramatic impact on the one‑stop career centers that currently serve the hard‑to‑employ ‑
*Mr. Doggett. Right.
*Mr. Vitale. ‑‑ those with barriers to employment.
*Mr. Doggett. That’s why I’m ‑
*Mr. Vitale. Today our unemployment insurance offices no longer exist in most states. So the one‑stop career centers are the only place people that are not readily job‑ready have to go to.
*Mr. Doggett. Mr. Chairman, since my time is up, I was going to suggest that perhaps the Inspector General might advise the Committee. He said he had a number of recommendations in this data sharing area. If those are being accepted ‑‑ I know you visited with him ‑‑ if those are being accepted, or perhaps ‑‑ some of them are relatively new, and I haven’t had time to review, but I think it would be helpful for us to know whether these various ideas that he wasn’t able to explore in full are getting adopted. And perhaps some of them provide us models for other agencies, too.
*Chairman Davis. I agree. I think there are good benefits ‑
*Mr. Doggett. Okay, just follow up in writing.
*Chairman Davis. If you could get back to us ‑
[The statement of The Honorable Patrick P. O’Carroll, Jr., follows:]
*Mr. Doggett. Thank you.
*Chairman Davis. ‑‑ and potentially sit down with us for a follow‑up meeting that would be quite helpful. I think when we get into this question of cost associated with it, as we fund legacy programs, those ‑‑ and I’m speaking of the information technology disconnects that we have ‑‑ it’s kind of like pumping blood into somebody who has got a bleeding artery. What we want to do is clamp that artery and get it fully integrated.
Mr. Thornburgh’s point, I know professionally I have seen many of these systems, if they’re properly implemented, pay for themselves very quickly. The real issue, though, is process change within government, that will be our problem, from a statutory standpoint. But I appreciate your question.
Now we are going to turn to Ms. Black from Tennessee.
*Ms. Black. Thank you, Mr. Chairman. Mr. O’Carroll, I want to go back to a statement that you made just a few moments ago, and make sure that I heard you right when you talked about there being a sharing ‑‑ that there were some concerns about the privacy issues. Can you talk about that a little bit further?
*Mr. O’Carroll. Yes, Congresswoman. One of the biggest issues that we’re having is that, under the Computer Matching and Privacy Protection Act, agencies have to enact single-purpose agreements to gain access to the data. To give you an example, the Department of Transportation has a significant file on anybody with a commercial driver’s license.
Well, as an example with SSA, we would like to be able to access that commercial driver’s license database, and run it against SSA’s disability and SSI records to see if the people are, in fact, in need of that type of a benefit. And because of the Computer Matching Act, we can’t access that type of data. It takes an application ‑‑ it usually takes several years before it’s approved. And that’s one of the issues with the Matching Act and the privacy concerns that we would like to be able to streamline.
And in the case of HHS, Health and Human Services, their inspector general was able to get a waiver on that type of a matching agreement, so that when the data match was going to be to determine eligibility for a program, or detect fraud, waste or abuse, that HHS can match the data and be able to see if the person was, in fact, entitled to it.
*Ms. Black. And since I’m not familiar with that act, is that act just on a federal level, that this only applies to those issues on a federal level of the data matching? Is that correct?
*Mr. O’Carroll. Yes, Congresswoman. It’s a federal law, and it only affects executive offices in the government. One agency can’t share with another; it’s a federal act.
*Ms. Black. Okay. And I think maybe we need to visit that particular issue as well, as we’re talking about access to information that will help you to do your job.
I want to turn to Mr. Thornburgh then, and ask, as you are dealing with states like Montana ‑‑ and I know you are doing work in the State of Tennessee ‑
*Mr. Thornburgh. Yes, ma’am.
*Ms. Black. Do you have that same barrier there, that there is not an ability to be able to share this information from one department to the other?
*Mr. Thornburgh. The short answer is no. But not quite, in that there are certain restrictions that certainly will apply. But to be quite candid, it seems to be more difficult at the federal level than at the state level to share data effectively. We have a number of cases in which we move data between the states up to the federal level, and the structures and requirements are significant to allow that to happen.
*Ms. Black. I go to Mr. Sekhar. And I am very impressed by your model of being able to share information between all these departments. Have you had any experience in any states where this model has been applied?
*Mr. Sekhar. The model you are looking at is more of a model of each of the human service programs on how they perform data exchanges today.
But I think the challenges we typically face at a state level is raising it one level above, and getting a level of standard. And states have made, for example ‑‑ and I work in the Commonwealth of Pennsylvania ‑‑ they do share information across the programs. But our suggestion is more on having a standard that can be applied across human services.
*Ms. Black. I know that in our state of Tennessee, that there was a significant change when the Department of Labor shared with the Department of Human Services folks who had jobs, and then paying for child support. And it was very effective, and that has been done.
But I know that also in our state I have been very concerned about the amount of money that is spent on IT, and then it goes on for years and years, that it’s not complete yet and we have to put more money into it and, oh, we have to upgrade it and it’s just never quite right. And there is a tremendous amount of money that is spent, I know, at the state level. I don’t know how much is being spent at the federal level with this data mining and sharing information.
Can any of you talk about how the dollars are being spent, and whether you believe that the dollars are being spent in a way that is financially good for our state, and the dollars that are being spent?
*Ms. Black. Maybe Mr. O’Carroll. Do you have that experience with IT and the money that is being spent ‑
*Mr. Vitale. So one suggestion ‑‑ in our presentation we talk about the model of a consortium. Instead of every state trying to build their own unemployment insurance system, and we have to spend somewhere between $30 million and $60 million times 50, if we get the states together and we build it as a group, and then they can share a common code base, and then that code base can be added on to customize for your 20 percent that’s unique to your state, so that would be a good model to implement, to help bring down the cost, and at the same time upgrade the infrastructure of these core UI systems.
*Chairman Davis. Thank you. The gentlewoman’s time has expired. I would like to recognize Mr. Berg from North Dakota.
*Mr. Berg. Thank you, Mr. Chairman, and welcome. I ‑‑ you know, this is a great quest. Obviously, it is a bipartisan quest, it’s a quest to try and become more efficient with our dollars so they are going to, again, the people that are ‑‑ need those, and also to prevent those that don’t deserve them from getting them. I mean it’s pretty simple.
There is two things that I want to talk about from North Dakota. One is there is a ‑‑ I will call it a scam that’s been going on recently where people are filing income tax in multiple states, and they’re filing, like, $25, paying $25 of income tax. The next year they’re applying for a refund of $200, or $500, or $1,000. And a lot of the states are trying to very rapidly get the refunds back out to people. And so, mistakenly, a lot of checks are going out. And again, they are going out with fraudulent ‑‑ I shouldn’t say fraudulent addresses, but addresses that allow these people to collect the money, but then kind of disappear.
So, I mean, I kind of raise that because I think this problem is not only at the real big picture that we’re talking about, but also at the small level. And, you know, it kind of occurred to me we’ve got an issue with the funding that I’m not quite sure ‑‑ you know, years ago that was passed, and we said we want to really link workforce with ‑‑ or, excuse me ‑‑ education with workforce. And some of the feedback I’m getting back from my state are we’re tracking the education part but, because of privacy, we can’t get their social security numbers. And so, we can’t really track whether or not they’re working.
And, you know, I’ve spent a lot of time trying to bring agencies together and, you know, we’ve got all these different silos that are asking business and people for the same information. The next one is asking for the same information. And so, I guess I’m kind of going around about the way, but it really comes down to the crux, in my mind, of this issue is getting this information, whether it’s a social security number or something very basic, you know, across party lines.
And so, two questions, quick questions. One is, do you agree with that as being the core problem here? And if so, how would you propose to fix that?
*Mr. O’Carroll. Since you brought up Social Security, Mr. Berg, I will answer first. I agree that it is a sharing issue. As you brought up, it’s that every agency is in its own silo. We’re not sharing, amongst other things, the wage information, address information, all the other information that is inter‑related.
And I’m thinking that, in many cases, the whole purpose of the Privacy Act was to protect everybody’s privacy to keep your social security number and your personal information out of the public domain. But there are so many other issues to consider. I think, as an example, with any of the benefit programs, you are giving up some of that privacy to receive the benefit.
And maybe with some of these things, at least on the benefit side, there could be a waiver for anybody who completes that type of an application, that you’re giving up some privacy, and that we will be going to other government agencies, asking for your information.
So, from my perspective, we’re looking for more freedom with regard to sharing information when you’re going to be receiving a benefit from the government.
*Mr. Berg. Please.
*Mr. Thornburgh. If I may, Mr. Berg, to simplify the question a little bit, “How do you make this happen,” I think it needs to go back to the agency level.
There has to be an incentive for the agency to make the system better. And that incentive not only has to be financial, they have to be able to show that they’re going to save money and be more effective and more efficient during that time. But at the same time, they also have to make sure that services are delivered in a more timely and effective way, as well. Ultimately, what we all want to do is provide the services to those who are in need of services.
This functionality makes it work for both ends. I guess, in my experience, what I have seen is that the sledge hammer is not very effective in requiring agency heads to ‑‑ “Thou shalt go forth and cooperate” has not been very effective. But when you find the incentive and provide the opportunity for them to be more efficient and save taxpayer dollars, that’s a huge benefit for everyone.
*Mr. Berg. The sledge hammer only works in Kansas, I think.
*Mr. Thornburgh. Yes, sir.
*Mr. Berg. Well, if you were king for the day, what incentive would you create for the Agency?
*Mr. Thornburgh. I think the incentive would have to be financial and beneficial. They have to ‑‑ we all know the giant wrestling match for dollars appropriations. And so there has to be a financial incentive that allows them to save taxpayer dollars, and ultimately they have to have the opportunity to provide benefits more effectively.
*Chairman Davis. Thank you. The gentleman’s time has expired. If you would like to submit some more information in writing specifically outlining this in detail, you are more than welcome to do so.
[The statement of Ron Thornburgh follows:]
The chair would like to recognize the gentleman from Washington, Mr. McDermott, for five minutes.
*Mr. McDermott. Thank you, Mr. Chairman. I commend you on having this hearing, because it’s a real problem. And I am pleased to hear systems being suggested that sound like Denmark and Norway and Sweden, where they have identity and they can collate data, and whatever.
My problem is ‑‑ and I want to ask you if this is the crux of the problem ‑‑ I went into the veterans hospital in Seattle and was talking to some doctors. And you’re sitting in a doctor’s office, and he has two computer screens. One of them is the military, the Defense Department’s health care record. And the other is the Veterans Administration health care record.
The Veterans Administration health care record was designed by and built by the Veterans Administration. Very efficient. Doctors like to use it. The military, the Defense Department one, was done by a private contractor. And there is no way to connect the two. So you have to sit with two computer screens.
I spent more than a year fighting ‑‑ here we’ve got kids coming back from Afghanistan, blown all to pieces. They go to a hospital in Ramstein, Germany. They are taken care of. They are clearly not going back to active duty, so they are transferred over to the Veterans Administration. Their records don’t go with them, except in paper form.
Now, I said, “What in the world is wrong with a country that has all the capacity we do, and we will not take care of our veterans?” And they said, “Well, we have this private contractor who made this Defense Department program, and somehow they can’t figure out how to connect it to the VA.” Are you telling me that this law, this privacy law, is what they’re hiding behind?
I had generals and admirals sitting in front of me, and I couldn’t get any straight answer out of why they couldn’t fix this. And kids were getting poor treatment because when they left Ramstein it wasn’t immediately transferred by wire to Seattle Veterans Hospital. I could not ‑‑ they couldn’t give me a decent explanation. So I want to hear if this is what you think is the reason for that.
*Mr. O’Carroll. Well, first, Mr. McDermott, I applaud your concern for veterans and our armed forces. I do hope that they get the best of treatment.
I’ve got to tell you that you’re hitting it on the head. I can’t so much talk about Defense and veterans, obviously, because that’s not under my purview. But I do know, as an example, SSA’s sharing information with Veterans Affairs is very difficult, because of these matching agreements that I had mentioned before. A person can be on VA benefits, and be qualified for SSA benefits, and not even know it.
So, there are a lot of data exchanges between the two agencies that are not only going to help identify benefits that go to people, we’re also trying to make sure it’s the right person getting the right payment.
*Mr. McDermott. Sounds like what you’re talking about, a matching contract, or whatever that thing is ‑
*Mr. O’Carroll. Matching agreement.
*Mr. McDermott. ‑‑ is really an unmatching, they have an agreement not to match, so that they will never come together. Is that what you’re ‑
*Mr. O’Carroll. I think a few years ago, the thought was, for the sake of privacy, they didn’t want agencies matching data with each other because it could infringe on privacy. But as we’re seeing here in this hearing, it’s not so much a privacy issue you’re eligible for, but in many cases, it’s that you’re not receiving your benefits you’re eligible for. The government is missing information that could help, as well as detect people that are getting benefits that shouldn’t be.
So, I agree. I think the whole Computer Matching and Privacy Protection Act has to be looked at again. We’ve got to be considering the idea that all federal agencies should be able to match with each other.
And then the other issue, which is a much more difficult part ‑‑ and Ms. Black brought it up before ‑‑ is that funding is also a big factor, in that the states all have different systems. The federal agencies have different systems. And trying to merge them all is a major undertaking.
*Mr. McDermott. I was a state ways and means chairman in the state legislature, and I saw us put out millions of dollars for computer systems that never went into effect. And I wondered what was ‑‑ but you’re saying it’s all ‑‑ it’s fundamentally privacy questions that stops the government ‑
*Mr. O’Carroll. From talking to each other.
*Mr. McDermott. ‑‑ from talking to each other.
*Mr. O’Carroll. And then the second step is, once I think agencies started talking to each other, the next step would be talking in the same language, which would be the matching of the systems.
*Mr. McDermott. COBOL probably.
*Mr. O’Carroll. Well, unfortunately, that’s a concern for SSA, is that they’ve been using COBOL for quite a long time, almost too long.
*Mr. McDermott. My brother works for Boeing, and is one of the last living COBOL people.
*Mr. O’Carroll. If he wants to talk to a COBOL programmer, I will give him a number of somebody at SSA.
*Mr. McDermott. Thank you.
*Ms. Lower‑Basch. I do think the technical issues are real at the state level, that it’s not just laws, that we’ve got a lot of legacy systems.
*Chairman Davis. Regarding this issue that Mr. McDermott brought up, the one thing I would say ‑‑ and this is just as an observation ‑‑ systems don’t implement effectively if the processes are not changed to be able to conform to the system. And that’s usually the root of the problem.
And the statutory limitation is one problem that contractor faced ‑‑ having been very involved in that specific issue prior to joining Ways and Means ‑‑ and the other part of the problem is, the requirement that the Agency gives to the contractor is so precise that they are not allowed to deviate outside of that when, in many cases, they recognize this. It led to some of the challenges that we had with the Walter Reed situation a few years ago, in fact.
The chair now recognizes Mr. Boustany from Louisiana.
*Mr. Boustany. Thank you, Mr. Chairman. I appreciate this hearing, and I want to thank our panel for being here today.
I want to focus on the unemployment insurance program for a moment. Earlier this year there was a newspaper article in my home state. It was the Advocate, a Baton Rouge newspaper, and it talked about the Louisiana unemployment insurance fund being highlighted as being one of the best in the nation. And, in fact, the National Association of State Workforce Agencies listed Louisiana as having one of the healthiest funds in the country. That’s the good news.
Now, despite that, I am very concerned about the amount of overpayments. And we have got some additional reports out there ‑‑ there are a series of them ‑‑ that list Louisiana, for instance, as having the ‑‑ as being the worst state in the union with regard to overpayments in the UI program.
So ‑‑ and in fact, I will give you some statistics. 2007, Louisiana’s overpayment rate was 46.5 percent. And I believe, Mr. Vitale, you said overall, nationwide, it’s about 10.6 percent. So this is a significant overage. In 2008 it improved a little bit, it went down to 34.9 percent, then went back up to 41.5 percent. And just to sort of put it in perspective, the 2008 overpayments were estimated to be around $69 million.
*Mr. Vitale. Correct.
*Mr. Boustany. Now, this is really unacceptable. And in effect, it’s penalizing hard‑working businesses in our state and in other states who are seeing these kinds of overpayments.
So, Mr. Vitale, I was listening to your testimony, and you talked about modernization being needed, but being expensive when looking at our IT systems. And in the discussion we’ve had today it’s sort of like we’re always chasing a moving goal, you know. You spend more money on IT, and then you still don’t have what you need, and you go further and you go further, and this continues.
I want to talk a little bit about ‑‑ and I want your perspective on ‑‑ the cost versus the overpayments, and sort of that equation. And give us some perspective on that. I mean, you know, if Louisiana is $69 million, what would be the cost, in your mind, basic general terms, to get to an IT system that the state would need that could interface, you know, with other different programs to prevent these kind of overpayments?
*Mr. Vitale. Sure. So ‑‑ it’s not an exact cost. Louisiana does have one of the old UI IT systems. So they would need to upgrade their entire core system. These technologies that we talked about today are peripheral to the core system.
The core systems reside in the states that pay unemployment insurance benefits and collect UI tax. The technologies that we talked about today to help in the overpayment area need to interface with those core systems. And because of the old technology that is in place in the states, it’s difficult and costly for them to integrate, for instance, an imaging system to old mainframe technology system.
I would estimate that if a state wanted to do it by themselves, it would take somewhere in the neighborhood of $30 million to $50 million to rebuild their entire system and re‑engineer their business processes, etc.
So I hope that answers your question.
*Mr. Boustany. Yes, yes. And what ‑‑ and you mentioned pooling earlier in your testimony. What would be the cost impact if we had some sort of pooling mechanisms?
*Mr. Vitale. Sure, that dramatically reduces the cost. You can pool resources, you can pool funding. If you take four states and each one would take $30 million to $50 million to build it separately, you can build one system that is the Cadillac, probably, for around $50, $60 million ‑‑ I mean, I’m giving ballpark figures here ‑‑ and that would address 80 percent of the functionality in the 4 states. Then each state would have to customize the core system to address their unique needs, about 20 percent of the functionality is unique.
So, you are leveraging the resources, you are leveraging the shortage of business subject matter experts and IT experts in the states by pooling them all together, instead of each state building their own system.
*Mr. Boustany. And how do you stimulate the states to do this?
*Mr. Vitale. Well, USDOL has a ‑‑ recently awarded two grants to four different groups of states: Arizona, Wyoming, North Dakota, and Idaho is one group; and North Carolina, South Carolina, Georgia, and Tennessee is the second group. Those two groups of states got funding to determine the feasibility of building a common system and determining if they work together. And can they develop common requirements for a large part of the system.
They’re at the point now where they’re almost finished that two‑year project, and they have discovered that they can work together, and their differences are not that great, and that they have documented their common requirements.
*Mr. Boustany. It took them two years to get to that point to agree to work together.
*Mr. Vitale. But it’s not that easy. So the next step is they need the funding to go on to actually build the common system, which, at this point in time, is up in the air.
*Mr. Boustany. Thank you. I yield back.
*Chairman Davis. I thank the gentleman. Now Mr. Smith from Nebraska is recognized for five minutes.
*Mr. Smith. Thank you, Mr. Chairman. Mr. Thornburgh, thank you for joining us from America’s Heartland. The ‑‑ I know that you have talked about electronic filing or, you know, using technology, online filing versus paper‑based. Now you generally handle the online filing and you don’t have much say ‑‑ your company doesn’t have much say over the paper‑based. Would that be accurate?
*Mr. Thornburgh. Yes, sir.
*Mr. Smith. Okay. Where do you find, or how often do you find kind of a bias within public policy that taxpayers would absorb the cost of paper‑based filing, but taxpayers would not absorb the cost of electronic filing? Do you see where I’m going with this?
*Mr. Thornburgh. I think so. And so I will take a swing at it. And if I don’t, I am sure you will correct me.
And so you’re right, there seems to be a ‑‑ I won’t even say “institutional” ‑‑ perhaps statutory bias, as the open record statutes and kind of the structure behind that was written, quite frankly, prior to the electronic age, in many cases. And so, while there has been an acceptance of the difficulties of paper filing, to craft a policy that encourages electronic filing ‑‑ I mean again, I’m going to go back to my service as secretary of state.
I can tell you that when someone filed a uniform commercial code document by paper, it cost me approximately $9 to $10 to process that piece of paper. If they threw bits and bytes my way, it cost me about $1.27. So I wanted to create policies to encourage people to file electronically. And, in doing so, we were ultimately able to get to a 90 percent adoption rate for those uniform commercial code filings, simply by a policy change in charging less for electronic filings than we charged for paper filings. If someone wants to throw paper our way, they had to pay full freight for that thing.
So, there are some policy discussions that can certainly craft electronic filing incentives that will encourage agencies to move in that direction.
*Mr. Smith. And then, moving further on ‑‑ in terms of accuracy and errors, how would you be able to point to the difference in the error rate?
*Mr. Thornburgh. Well, I will use two examples for that. One is I have always thought ‑‑ and again, in the case of uniform commercial code, the banks certainly had an incentive to make sure the filing was correct. And they perhaps have a greater incentive than the clerk who was working for me to ensure that that was correct.
And then, the Montana Connections. What we have found is that we can place edits within the software development within the code that will ensure that every line is complete, every line is accurate and consistent, before it’s applied to the system, before the application actually takes place.
What we have found with a paper‑based system, if there was an error, it will be returned two, three, four times. So that same person is going to be handling all of those times. In an electronic system, it gets handled once and it’s correct.
*Mr. Smith. Okay. Thank you, Mr. Chairman. I yield back.
*Chairman Davis. I thank the gentleman. The chair now recognizes the gentleman from New York, Mr. Crowley.
*Mr. Crowley. Thank you very much, Mr. Chairman, for the hearing. I apologize for not being here for your testimony, but we have your written testimony, and we have perused it prior to coming today. And I want to just piggy‑back a little bit on my colleague from Washington State in reference to the VA.
And one of the key areas that can benefit from data matching is veterans care. Our veterans, I believe, and I think everyone on this panel believes, deserve the best of possible health care. And we know that health IT has the potential to greatly increase the quality of the care provided to our nation’s veterans.
Much of the medical information that veterans provide serves dual purposes for both their doctors, as well as for the Department of Veterans Affairs. And that’s why I have supported efforts to encourage electronic medical records to include questions on whether a patient is a veteran.
John Rowan, who happens to be the president of the Vietnam Veterans of America, is not only a constituent of mine, but a long‑term friend. He also happens to be someone who believes very strongly that including veteran information in electronic health records can have a great benefit.
Connecting medical records to veterans status helps doctors to diagnose certain health complications that may only be veteran‑oriented, such as the Gulf War Syndrome. It can also help the VA to match up claims information with beneficiary records, as well as track health trends that may be developing among veterans of a certain conflict. The VA itself is clearly aware of the benefits electronic medical records can provide, as in November 2010 ‑‑ as of 2010, they announced a pilot program to speed the process for veterans to collect their private‑sector medical records. Under this new initiative, a contractor would retrieve the veteran’s records from the health care provider, scan them into a digital format, and send the material to the VA on a secured transmission.
I am interested in hearing from a number of you ‑‑ and I have an additional question, so if you could, be short ‑‑ to hear your thoughts on how you think data matching could be further used to improve the connections between the veterans the VA and, very importantly, the private sector medical care they’re receiving, as well. Does anyone have any comment on that?
*Ms. Lower‑Basch. I will just note that a number of states are copying the Washington State model that I referenced in my testimony of using PARIS to flag people who look like they should be getting veterans coverage and are not.
*Mr. Crowley. Anyone else?
*Mr. Crowley. Ms. Lower‑Basch, since you chose to answer the question, you actually are the focus now of my second question.
You mentioned in your testimony several examples of data matching programs already in widespread use. One promising new initiative is the administration’s Partnership Fund for Program Integrity Innovation, which is designed to help states create pilot projects to reduce improper payments without reducing participation amongst eligible populations. Every project must save at least as much as it costs.
Ironically, the House‑passed CR for the remainder of this fiscal year would cut funding for this fund by nearly one‑third, $10 million rescinded from 37.5 million appropriation. Can you talk about the promise of this new initiative, and the detriment to data matching if these cuts go forward and go into effect?
*Ms. Lower‑Basch. Sure. I think the fund does two things that would probably not happen in the absence of it. One is it does provide some of this little seed money to get things started because, as we have discussed, that even if things wind up saving money down the road, it usually does require some up‑front investment.
It also includes rigorous evaluation, which, while I think highly of a lot of the things that are already happening, they have not been rigorously evaluated. It would be great to actually capture some of the data on what the payoff to the investment is. And that will lead people forward.
I would also say it probably brings people to the table, these sort of interstate things which I think everyone agrees, in theory, makes sense. But getting everyone to do it is sometimes a challenge.
*Mr. Crowley. Thank you. Thank you all for your testimony, and I yield back.
*Chairman Davis. I appreciate the gentleman’s comments on veterans issues, something I have been involved in for many years.
And one thing I would point out. The VA has state‑of‑the‑art data systems in their medical records. One of the challenges is that the VA itself was its own worst enemy, and the very sharing thing that Mr. Crowley and I would like to see happen, when its general counsel issued an opinion on privacy protection. It prevented their doctors from, in fact, collating some related records on some very critical issues related to prescription medication.
And the reason I bring this up, before we go to our last questioner, is as our dialogue continues, I think it’s very important that we come back to the root issues, which are not partisan, they’re not ideological. These are just simply processes, where sometimes the left hand, with very good intentions, puts in place a process that the right hand doesn’t know, and it creates secondary and tertiary effects that create additional costs, and the folks we want to help don’t get helped in that process. So we appreciate your counsel and perspective on that.
For our final question I would like to recognize the gentleman from Minnesota, Mr. Paulsen, and thank him for his Job‑like patience as we have gone through this.
*Mr. Paulsen. Thank you, Mr. Chairman. And, Mr. O’Carroll, I was going to ask you a question, actually. You had, I think, recently ‑‑ I guess your office had recently completed a request by a member of the Ways and Means Committed to review SSA’s online application system, iClaim.
And I want to ‑‑ just might expand on that. I think there was some concern that having an online application, claimants might not be receiving the necessary level of service from SSA to complete their applications. And I think your first review that you went through focused on retirement applications, in particular. And presumably, I mean, that’s, you know, an age group that doesn’t have as much access to the Internet, for instance, or might not have as much exposure to the opportunity for those types of applications.
But you found a pretty healthier 96 percent, I think, return or rate of the online filing experience as being excellent or very good. Can you elaborate on that review? And what are some of the lessons, I guess, learned from implementing a solid online application? How does it complement the existing face‑to‑face or telephone services that the Agency already offers?
*Mr. O’Carroll. Yes, Mr. Paulsen. At a recent hearing with one of our committees here, that issue came up ‑‑ there was some question as to whether or not, by using the online system, potential beneficiaries would be getting the same level of service as if they came into an SSA office. Everyone is so concerned with the backlogs, and the waiting time in offices, that really, the future is going to be through electronic service.
So, we examined the iClaim process. We looked at a sample of people who applied using iClaim, to ask what their experiences were. We found a very, very high ‑‑ in the 95 percentile — rate of satisfaction on it. We asked how easy was it to use, did you find it difficult, did you have any questions on it. Applicants were the most satisfied with the follow‑up that Social Security Administration did.
So, in other words, if applicants had any doubts when they were doing it, if they didn’t have the right type of identification or information or anything else, and there was a question left in the electronic application, SSA contacted them. And they were very happy with those SSA contacts.
One interesting thing we found from talking to them and from talking to SSA employees in a second study that we did, was the telephone numbers that most people gave when they made their initial application weren’t always good. And one of the suggestions from the employees was to have multiple contact numbers so that when they try to reach out and talk to the person during business hours, that they would be able to get a hold of them.
I think that is going to add even more to the success of this program, if SSA can contact claimants easily and quickly, it will help a lot. So I think this is a great success story for SSA, in terms of the service to the public.
*Mr. Paulsen. And from your perspective, can you elaborate if there were any concerns, as a part of that study, at least initially, where you saw that maybe fraud or abuse concerns from online applications were a component? Or, you know, is there worry about that? Or are there advantages or disadvantages from other methods of filing for benefits?
*Mr. O’Carroll. I will tell you on that one, of course we always have a great concern. We work closely with SSA as they are rolling out their programs, to see if they are going to have any vulnerability to fraud.
The retirement side of SSA has probably the lowest level of fraud of the programs because, pretty simply, SSA has all of your earnings information, it’s a relationship that you have had with the retiree for years. There is a lot of trusted information, so you know who the person on the other end of the application is.
So, in SSA’s retirement programs, we don’t have very many concerns in relation to fraud. We are continuing to monitor that. But at the moment, our level of trust is pretty high.
When we start taking a look at disability iClaims, where there are going to be more documents and more information provided, and it is harder to double‑check information, we may have more concerns. I will let you know what we find.
*Mr. Paulsen. Thank you very much. Thank you, Mr. Chairman.
*Chairman Davis. I thank the gentleman. I would like to thank all of you for taking the time, investing the time for preparation, and coming in and patiently walking through the hearing process. Some of these issues can appear to many viewing as awfully esoteric. But as Yogi Berra said, “Baseball is just a simple game of throwing and catching and hitting,” and it’s in those basics that you all have worked in for so many years that, I think, lie the seeds of our solutions.
If Members have any additional questions, I would ask that they submit them to you directly in writing. And we would appreciate your responses to them, so that we can insert them in the official record, as well, for others to read.
I thank you again. I thank my friend from Texas, the ranking member. And with that, the committee stands adjourned.
[Whereupon, at 11:29 a.m., the subcommittee was adjourned.]
MEMBER SUBMISSIONS FOR THE RECORD
The Honorable Geoff Davis