Back in 2019, which feels like a decade ago, I wrote on this blog about the ethics and efficacy of the Guardian ad Litem Program. I started that post by dismissing what I viewed as ridiculous arguments that attorneys at the Program were somehow engaged in unethical behavior because of their client’s marketing slogan, “We represent the best interests of the child.” I found those arguments both incorrect and insulting to the attorneys who faithfully represent the Program in its efforts to help kids. (I was one of those attorneys once and it is a hard job to do well. More on that below.) But I also thought the Program’s slogan was intentionally misleading and chosen for the main purpose of crowding out funding for attorneys for kids. The word “represent” is the problem: lawyers and law firms represent people, government programs and non-lawyers do not. Just pick another verb. There’s nothing wrong with “advocating” for children.
The second half of the post was about the GAL Program’s efficacy. Leadership at the Program had ramped up their rhetorical attacks on attorneys for kids, claiming that attorneys were dangerous because they would thoughtlessly advocate for children to return to unsafe homes or would keep insidious secrets from the court because a 4-year-old put her finger over her mouth and went “shush.” Neither of those scenarios are based in reality: attorneys have as much an obligation to counsel their clients as they do to advocate for them, and children aren’t particularly discerning in who they tell things to. Attorneys also have an ethical obligation to the courts to only raise meritorious claims, which includes only advocating for reunification when the attorney believes the facts and law would support it — i.e., when it’s safe. Maybe that includes situations where the GAL and DCF disagree, but also maybe the state agencies err too far in the other direction and keep kids in care who could really go home. It’s ultimately up to a judge to decide based on the evidence, but the judge can’t decide if nobody puts the facts before them. If an attorney felt the child was asking for something the law or facts would not support, the attorney should counsel the child away from the position or withdraw. End of ethical dilemma.
The GAL Program, however, went even further in its rhetoric. Not only were lawyers for kids bad, according to their press releases and public statements, but a system of community volunteers with 40 hours of training, no prior relationship to a child they see once a month, and a staff turnover rate on par with retail clothing stores was categorically and mathematically superior to any imaginable alternative. They had charts and graphs to prove it. This was just too much. In the post, I went through the Program’s performance report numbers and showed that there were no correlations between their main measure — the percentage of kids with a GAL appointed in each circuit — and any of their reported circuit-wide outcomes. I played with controlling for socioeconomics and demographics in the circuits and got some weak correlations here and there, but not in the positive direction that the GAL Program was claiming. The end result was that their numbers didn’t show anything at all. They’re just colorful charts.
And colorful charts are wonderful things, except that those particular charts had captured over $600 million in funding based on a promise of outcomes that weren’t supported by any evidence. It was a blog post. I got it off my chest and moved on.
Then the Florida Legislature asked OPPAGA to do the analysis for real.
The GAL Program’s Structure
OPPAGA is the Office of Program Policy Analysis and Government Accountability. One of its jobs is to assess the functioning of Florida’s government agencies. The report, which came out in December 2020, is long and thorough. The focus is on how the Florida Guardian ad Litem Program is run, its structure and efficiency. Some of the report’s conclusions about the Program were good — or at least not bad. OPPAGA’s staff found the following:
- The GAL Program follows national standards on best interest representation using a unique multidisciplinary team.
- National organizations and the academic literature recommend attorney representation for kids in dependency proceedings. Using lay volunteers is not an evidence-based practice and has shown mixed or no results in studies on things like permanency outcomes and time in care.
- The GAL training is good, but some stakeholders wish it would focus more on educating volunteer GALs about “the realities of foster care and challenges of disadvantaged parents.” Some GAL local programs wish they had more budget for training.
- The Program represents about two-thirds of kids in the system with wide variations among circuits.
- Stakeholder views on the Program’s efficacy split among groups in the system. Judges thought they were efficient and effective due to their use of volunteers and the value of best interest advocacy in general. Attorneys in the system thought the volunteers lacked expertise and too often joined the Department’s positions on issues. Everyone commended the volunteers for stepping up and trying to help.
It doesn’t appear that OPPAGA was tasked with looking into any of the other structural questions that often swirl around the Program, such as whether a GAL should be a separate party or derive party status from the child. The big question — about who the Program’s attorneys represent — was answered clearly. They represent the Program.
Another big structural question not explicitly discussed in the report can be found in OPPAGA’s data. This question involves the feasibility of reaching 100% representation with a volunteer model. The report data suggest that this will never happen. The report finds that state funding for the Program has increased about 20% over the last 5 years, which has increased the number of staff but not the number of volunteers. Despite the rise in funding and staff, the number of children served by the Program has actually gone down. The core performance measure of the Program is the representation rate. Going down is bad.
The report’s data even brings the “number of children served” measure into question because the Program (discussed below) discharges off of so many cases before their conclusion that it’s not a meaningful measure unless you start using fractions. If we instead counted the number of days children spent in care with a GAL, it would be far more accurate and a smaller representation rate than is currently reported. All in all, the report does not seem to support increasing the GAL budget for the goal of increasing volunteer representation. We may have reached the natural limit of what volunteer-based representation can accomplish. And I think that’s good to acknowledge.
The Problem was Outcomes
OPPAGA’s main point of criticism was how the GAL Program measures and reports its outcomes. First, the GAL Program’s performance data includes circuit-wide measures that do not separate outcomes for children with and without a GAL. It’s therefore impossible to say whether any noted difference was an effect of the GAL Program’s efforts or just a circuit-wide variation. In OPPAGA’s language, the performance measures “may not be indicative of the program’s actual performance.”
What we need for analysis is child-specific data so that we can compare outcomes in light of case and child factors. OPPAGA attempted to compare data from the GAL Program, DCF, and the OSCA (the statewide court administration). This was a struggle because, as they wrote:
Even more problematic for analysis, the GAL Program was unable to provide any of the raw data to OPPAGA. I quote: “When OPPAGA requested that program staff export all raw program data for Fiscal Year 2015-16 through Fiscal Year 2019-20, program staff was only able to provide a vendor programmed report, which made it difficult to determine the completeness and accuracy of the data and hindered analysis.”
A quick web search shows that the vendor in this case is Evinto Solutions, LLC., an Arizona company that maintains the Optima database portal. The GAL Program pays Evinto $7,000 a month for access to the website and storage of the case data. I don’t understand how the data would not be available because the contract, which is available online, has the same clause that every state contract must have: the vendor must maintain public records and provide the customer with a copy of any requests “within a reasonable time” at a cost that is set by law. The deidentified records of a public agency that describe how it handles cases is undeniably a public record. If the GAL Program asked for the data and was told no, then Evinto was in breach of their contract. If the GAL Program never asked for the data, then they stonewalled OPPAGA. There is no good scenario in which the Program cannot access (or understand, as OPPAGA found) its own data.
OPPAGA found other problems with the data, too. There were no unique identifiers for each child in the reports and some data entries were clearly wrong (such as having conflicting dates). As such, they had to rely on the DCF and OSCA databases, and were only able to match up about 80% of kids to the GAL data. This means it’s not a representative sample — it’s just the best sample we’ve got. Those GAL Program performance reports that I said in 2019 had no correlation with outcomes also had no verifiable correlation with reality. This is very, very bad.
A few things jumped out at me in the report, based on what OPPAGA could figure out:
- The GAL Program discharged from 20% of cases because the child was in a stable placement, meaning the case hadn’t closed yet. That number has trended slightly up. Most of those cases ended in adoptions but many did not.
- About 5% of cases closed because of insufficient GAL Program resources.
- About 9% of cases were closed because the child ran away, moved to another circuit, or the case was consolidated by the court.
Permanency and time-to-exit are the GAL Program’s main system outcome measures in its reports, and the Program discharged from a third of cases before they even closed. That suggests the representation rate is even lower than reported and the supposed impact is even more tenuous. This is also very bad.
The rest of the report relays stakeholder views. One theme stood out to me, because I remember it well from my time at the Program: scheduling around volunteers’ time is hard and causes delays in staffings, hearings, and outcomes. I promise you that for every week shaved off one case through zealous advocacy, a month was added to another case because a volunteer could only attend court to do MBI testimony on Tuesdays after 2:00. Maybe Zoom has fixed some of that, but there are real drawbacks to having a part-time volunteer workforce.
The rest of the stakeholder views were mixed, with this conclusion statement:
What’s missing: $60 million a year
I wish the report had looked at how the Program spends its funds. You have to go down to the exhibits at the end of the report to find this information: Florida spends over $60 million per year on CASA/GAL services, almost twice what the next-highest state reviewed by OPPAGA (Texas) spends, six times the median total funding amount, and 20 times the median state funding amount. Our state expenditure per child ($1,354) is significantly more than the state with the next highest state funding outlay, North Carolina ($859 per child) and the state with the next highest number of children served, Texas ($447 per child). And again that is likely an underestimate because the Florida GAL does not stay on all cases all the way through.
I don’t know where the money goes, but I do know where it doesn’t: despite years of increased funding, the GAL Program has intentionally underpaid staff since its creation. That’s had serious implications for children. The Program submitted a budget request in 2017-18 asking for more money to pay its employees, writing: “the Guardian ad Litem pay has remained at low levels, due to a focus on expanding staff to serve rising numbers of dependent children (rather than pay raises).” It listed its CAMs as making $31,459 per year and its attorneys as making $41,267. That’s below the DCF payrates and below the HUD cutoff for low-income in most metropolitan areas in Florida. They wrote the following in the budget request:
Twelve percent of staff were on public benefits and the turnover rate was over 50% for anyone not in a senior position. I can’t get over that the highest turnover was for the volunteer recruiters — the engine of a CASA program. According to the Bureau of Labor Statistics, these rates are equivalent to the retail and construction industries, while state government jobs have a typical turnover rate of 19%. The Secretary of DCF testified in a senate committee on January 12 (at timestamp 1:18) that DCF’s turnover rate for child protective investigators (an incredibly stressful job) was 43%. The OPPAGA report doesn’t address turnover directly, but does state that four GAL volunteers they interviewed reported “many of the supervisors [i.e., CAMs] appear to be overworked.” I feel uncomfortable using the budget request in a negative light, since the GAL Program got the extra money for salaries that year and they’ve made addressing staffing instability part of the Program’s long-range plan — but the numbers show the Program has a serious workforce problem and they acknowledge it impacts kids.
I’ve never seen a full expenditure report for the GAL Program and its support organizations. I’ve requested financial records from the Program but am still waiting for the response as of writing this post. Looking around the internet it seems that they advocated for and got over $300,000 per year to reimburse volunteers for mileage and have engaged in lots of extraneous activities (scholarships, internships, lobbying, very large statewide conferences that I’ve definitely been to) that are not part of the core mission of representing kids. That doesn’t include the in-kind donations and events that are meant to reward volunteers for being volunteers, with no direct benefit to the children. Those are wonderful things to do, but not while also complaining that you don’t have the funds to do the actual job or pay your employees enough to stay.
Meanwhile, along with low staff pay comes very hard work. You are told that your purpose is to save lives and protect children, and that weighs on you constantly. The report notes in a footnote that the expected caseload for GAL Program attorneys is 150 children. And, according to the 2017 budget request, there’s no official paralegal position for support (we had wonderful paralegals in Miami). It’s very difficult to do a good job under those circumstances, and when the job is “don’t let kids die,” it feels impossible. It is no surprise that nearly two-thirds of attorneys leave. I hope the pay increase reduced that number significantly.
I’m bringing up the budget because federal law changed in 2018 to permit the federal reimbursement of up to 50% of state expenditures on “independent legal representation by an attorney for a child who is a candidate for title IV-E foster care or in foster care.” You can read more about it here. The funding is triggered by having an attorney represent the child and includes funding for paralegals, investigators, and social worker supports, which would include CAMs. The GAL Program in its current structure has already been deemed ineligible for that funding. The OPPAGA report again makes clear that the GAL attorney represents the Program, not the child. Until we fix that, we are leaving millions of dollars on the table and our best shot at 100% representation.
Let’s talk about alternatives
OPPAGA staff made a few recommendations which, in analyst-speak, they called “options.” The options included choosing concrete performance measures like other states have, using the DCF-assigned ID numbers for each child to aid in cross-reports, and gaining a better understanding of the Program’s own data. These are all fine, but they don’t address the core problem facing the Program: 100% representation of foster kids by unpaid volunteers is not sustainable, and the Program has scraped by for nearly two decades unsuccessfully attempting to reach that goal.
The GAL Program agrees that resolving the problem of 100% representation is critical to its continuing survival. In its current long-range plan, it writes:
Once the high turnover is addressed, the program will then address a key policy decision with the Office of the Governor and the Legislature; that is, 1) whether the program should request additional resources to pursue representation of all children in the dependency system (out of-home and all in their own home), or 2) work with the Legislature to limit its statutory mandate to representing only children in out-of-home care and children 0 to 3 years living in their own homes (as current funding levels permit).Statewide Guardian ad Litem Office Long Range Program Plan for FY 2017-18 to 2021-22.
The Program has also publicly acknowledged that, in its current form, it is not able to fully advocate for the children it represents. In 2019 the Program lobbied against a foster children’s bill of rights that would have clarified the rights children have in care and required the system educate kids on those rights. In its legislative analysis, the Program wrote that if the bill of rights passed into law the Program would need significantly more funding because it “will be advocating for children’s rights in a more expansive way than it is currently doing now.” It warned that the courts and other parties would likely also see a workload increase — and I agree that would probably be true until those parties stopped violating children’s rights.
The Program’s quiet campaign to kill the bill of rights was a huge disappointment to me, even if they saw it as an act of survival. Even more than the public jabs at attorneys for kids, the Program’s bill of rights analysis convinced me that its current structure was at odds with the rights and wellbeing of children. I don’t fault the people at the Program who are doing the work on the ground at all, but I do fault people who have vehemently argued for this status quo for years even when it’s clear it’s bad for children — namely all the children receiving no or weakened representation. The Program was asked to do the impossible, and nearly 20 years of experimentation is enough. Given the changes in federal funding, it is time to address that “key policy decision” of how to move forward to 100% representation of children.
Here are my views.
I believe 100% representation can only come through 100% funding of paid positions. Volunteers have an important role in working with children, but federal law and the needs of children demand 100% representation first and foremost. Federal funding is available to finally make that a reality.
To lock down the federal funding, the attorneys must represent the children and not the Program. This means the attorneys’ primary obligations are to the child and not to a government agency. This is actually a very good thing because the Rules of Professional Responsibility will require the attorneys to hold themselves, the system and even the GALs accountable to the children they serve. (Just look at the nonsense in Pasco for an example of why the Program should not be the client.)
Attorney representation does not mean the end of best interest representation. I support a hybrid model of child representation with GALs directing the attorney’s representation for young children using a best interests model (with a healthy dose of taking the child’s attachments and wishes into account) and kids directing their own representation when they’re old enough to function as a client (with a healthy dose of mentoring and counseling from the attorney and their former GAL). We can fight about where to draw the age line. I’m also open to expanding the universe of people who can serve as a GAL and relying more on the natural advocates in children’s lives like relatives, foster parents, and fictive kin who could fill the role. This could reduce the number of volunteers needed significantly.
Once the Program is not the titular party in every dependency case in Florida (or 65% of cases for 70% of the time at less than 100% advocacy?), we need to re-align the Program’s goals into something human beings can actually do to provide direct and tangible benefits to children. OPPAGA recommends setting real measures, and I think we should go further and re-task the GAL Program back to its roots as a support for GALs and custodians. And we should get GALs out of the opinion business and give them concrete objectives to accomplish to further the best interests of the child. They can testify about their opinion on the same terms as anyone else under the rules of evidence. The quality of their advocacy should be judged by what they do.
There should be process goals for GAL Program, like “ensure children have representation for 98% of days under DCF supervision,” “ensure children are seen in person every 30 days,” “report to the court factual observations in writing,” and “provide child welfare policy guidance to the child, GAL, and custodian regarding the case.” The measures would then be the percent of days in care a child had representation, percent of visits that were timely made, the number of reports or motions filed, and the number of hours spent educating custodians and children about the system.
There should be advocacy goals for the individual GALs in conjunction with the child’s attorney like “ensure school services are in place,” “ensure receipt of all entitled public benefits,” “coordinate therapeutic and medical services to optimize benefit and minimize duplication of services,” and “eliminate barriers to the current permanency goal and advocate to change the goal when appropriate.” You could measure that by logging referrals and linkage to services, counting participation in IEP meetings and due process complaint hearings, summing the actual dollar amounts in public benefits received as a result of the advocacy, and measuring case plan amendments and other permanency advocacy efforts. None of these goals should change once the child is old enough to direct the representation — the work just gets harder because it also involves counseling the child towards good decisions in each of these areas.
Next, maybe outcome measures aren’t the right test. I am more skeptical of outcome measures than most people seem to be. I’ve spent a lot of time working with time-to-exit and permanency rates in the DCF data and there is wide variation in case outcomes by the age of the child at removal, type of maltreatment, race and ethnicity of the child, availability of relative placements, geography, and an unknown number of other salient factors. Very few empirical studies have managed to isolate the impact of representation on these types of outcomes, and that may be because every case is just unique. I prefer measures that are directly related to advocacy goals like described above or the child well-being measures that other states use. If people want to measure outcomes then we will need to start keeping much better data.
I have thoughts about how the GAL functions on individual cases, including how a GAL is selected and discharged, their ethical relationship to the child’s attorney, and other mechanisms that would hold the GAL and attorney accountable to the child and the court at a higher level. I’ll leave those for another day.
Florida has come a long way in its child advocacy, and there is a path to finally reach the goal of 100% representation. I hope we can have a meaningful conversation about how to give every foster child in Florida a legal advocate who can help guide and protect them through the system and make their lives materially better.