Resolve Philly and Spotlight PA partnered to investigate how county foster care agencies across Pennsylvania take Social Security money owed to children in their care.
This story is a follow-up to the 2022 Resolve Philly and Philadelphia Inquirer investigation that focused on Philadelphia Department of Human Services. That investigation revealed the city was taking nearly $1 million every year from children and depositing the money into its general fund.
The story inspired legislation from Philadelphia City Council and within the state House to curtail this practice statewide. Local reporters and legislators across the state asked reporters Steve Volk and Julie Christie: What about the other counties in PA? How much do they take, and how many kids are impacted?
What did we ask for, and how did we get it?
For the statewide investigation, Resolve Philly and Spotlight PA wanted to address these key questions, and asked every Pennsylvania county the following:
How much money did the county take in?
How many kids does the county collect benefits from?
Does the county keep track of how it spends kids’ Social Security benefits?
Does the county do anything to tell the kids or their court advocates that they have this money?
To answer these questions, throughout 2023 and 2024, Resolve Philly sent records requests to all 67 counties in Pennsylvania. Each request sought the same four things from the beginning of 2020 through June 2023:
Records that show the individualized accounting conducted for each child for which the county acts as a representative payee, which is what the Social Security Administration calls the recipient of benefits owed to minors.
Records that reflect the amount of Social Security benefits the county collected as the representative payee for foster children and other children in the county’s custody, and the number of children obtaining these benefits.
Records that reflect other sources of funding used to provide "maintenance care" or "cost of care" for these same children. For example, for youth whose Supplemental Security Income or Old-Age, Survivors, and Disability Insurance benefits were obtained, the newsrooms wanted the other potential programs or funding sources used to pay for their care, such as Title IV-E benefits, Medicaid, or child support collections.
Proof that youth, or their advocates or attorneys, are notified that the county is applying to or has become their representative payee.
Whether counties keep track of individual accounting isn’t just a best practice — it’s required by law. All representative payees for these benefits are supposed to provide annual reports to the Social Security Administration outlining how they spent the money.
Notifying children or their lawyers was a key point in the class action lawsuit in Alaska that resulted in the federal ruling that if foster agencies don’t tell kids about the money, it’s a violation of the constitutional right to due process.
Further, the data collected would help Resolve Philly and Spotlight PA understand just how much the money means to foster care agencies. Child welfare systems are expensive to operate, and Pennsylvania’s Department of Human Services has told the federal government that this revenue source can be vital, especially for smaller counties that don’t receive as much state or federal funding.
For more than a year, counties slowly sent the newsrooms their data. Some were hesitant to hand it over, citing concerns about the privacy of the children in their care. However, the information was easily redacted. Some counties weren’t sure what the requests were asking for, so reporters ended up sending examples to them or negotiating for records in several mediations with the Office of Open Records.
How did reporters analyze the data?
No two counties kept records in the same way, nor did they all produce records that could be easily analyzed. Fifty-four counties sent data in a format that could be opened and analyzed on a computer. Among those, 47 had data that could confidently be processed into a single database.
Some counties provided PDF documents or scans, which allowed reporters to use scraping tools like Tabula to extract data and export them as .csv files. When documents were scraped, spot checks were used to confirm that the data were not copied into the wrong rows or columns, and to manually fix any issues. For some counties, reporters used Google Pinpoint to read through each page and convert the document into a data table. This was done with Philadelphia’s bank statements, helping condense 40 pages of documents.
To process the data, reporters verified with each county what time period was covered, how the county grouped or indicated individual people, and how each column was calculated.
Each county used different acronyms for the same benefits, or added other funding sources to the benefit totals. Some counties combined SSI and OASDI benefits together, and we determined that a combined number showing the total in Social Security benefits was sufficient, rather than trying to split them up. While processing, reporters made a note of each dataset’s specific setup and how that translated into the combined database.
Some counties provided very detailed data: multiple rows would show each expenditure and deposited check for a single child, along with the date of the transaction. This made it easy to get accurate numbers for 2020-2023 and see the totals for each calendar year.
Other counties provided data that aggregated information at different levels: some were monthly totals for each child, some were annual totals for a fiscal year, and others were only the total amounts across all four years requested. Six counties (Adams, Clearfield, Clinton, Dauphin, Northumberland, and Tioga) did not disaggregate other funding sources (Veterans Affairs benefits, State Supplement Programs, stimulus checks, etc.) from some of their Social Security totals.
These different versions of the data provided two clear ways to compile and then compare the information for those 47 counties. For 33 counties, the newsrooms had enough detail to make a database showing how much money was taken and the number of affected kids from each calendar year from 2020 through at least June of 2023. The other 14 counties provided total dollars taken and children affected from across all four years, with no way to accurately drill down further. Those counties were compiled into a separate data table.
For the detailed counties, reporters worked mostly in Excel to clean and prepare the data for analysis. Reporters manually added columns to identify the calendar year of a Social Security payment. They then used a formula to check whether the individual had already received a payment that year. These two additions helped Resolve Philly and Spotlight PA group the payments by year and count the unique individuals who received them. Reporters then used PivotTables to aggregate the total dollars and people for each year and collated the results into a single database.
How did reporters calculate the budget proportions?
Resolve Philly and Spotlight PA sought to understand how much the Social Security benefits money compared to the overall funding for foster care. If the dollars could be life-changing for the children, what did the dollars mean for the counties? Reporters reached out to Pennsylvania’s Department of Human Services and asked for data on how much funding each county received to operate its children and youth services. The state gave the newsrooms total allocations for fiscal years 2020-21 through 2024-25.
The allocation data included the two ways the state funds county child welfare systems: Act 148 and youth development centers, which cover things like foster care, adoption, and youth incarceration. The data also included the amount of federal dollars each county received, which the state decides how to divvy up. For simplicity, Social Security benefits were compared to just the state funding.
The state funding is not a total picture of each county’s budget. Counties also get those federal dollars, set and add to their own child welfare budgets, and make some revenue from actions like receiving Social Security benefits. This means that when comparing state and Social Security dollars, Resolve Philly and Spotlight PA are comparing one slice of the county’s budget to another slice.
For the 33 counties that had data for individual calendar years, the newsrooms compared those tallies with the fiscal year totals from the state. Resolve Philly and Spotlight PA had two full calendar years of data, 2021 and 2022, which were compared to Fiscal years 2021-22 and 2022-23, respectively. While time ranges don’t overlap for the exact same dates, both periods have a full year’s worth of data. Reporters then calculated the percent that Social Security benefits equate to in state funding.
The analysis showed that in 2021, the largest comparison was 2.75%, with a median of 0.22%. In 2022, the largest was 2.09% and the median was 0.28%. This means that the portion would shrink even further if federal and county funding were added, showing that Social Security benefits make up a tiny fraction of the money a county has to spend on kids in their care.
How do counties compare?
Resolve Philly and Spotlight PA created a database of how much money 47 counties received, how that compared to their state funding, and how they account for and notify youth of these benefits.
Reporters reached out to each county before publishing this investigation to make sure the data were correct. Three replied and helped confirm the data or shared corrections.
Will we get to see the data for every county?
Yes, but it will take a while. Resolve Philly and Spotlight PA received data from all 67 counties in the state, but have only fully processed data from 47 of them.
Many of the counties with unprocessed data provided data as PDF files, which contain scanned images of tables. Computers can recognize this text, but aren’t always accurate. That means reporters need to copy over the data by hand, a lengthy process. One county printed more than 100 pages of documents and mailed it, so reporters will have to manually turn those pages into data.
There are also counties that haven’t answered critical questions about the structure of their data that are needed in order to process them correctly.
Resolve Philly and Spotlight PA won’t publish any of the raw data because some counties handed over documents that include personal information like names, Social Security numbers, and birthdays (even though reporters asked them not to.)
