Which is the best model for charging for access to court records: a rest stop, a bus pass, or a bake sale?
What (if anything) should the judiciary charge for public access to records, and how should that decision be made? That question is now squarely facing the federal courts and Congress.
I have blogged periodically about the 2016 class action lawsuit alleging that the federal courts overcharged users for access to its electronic public records system (known as PACER), and used the surplus to fund a variety of internal projects. Last spring, a federal district judge granted partial summary judgment to the defendants as to liability, but concluded that some of the project funding had indeed exceeded Congressional authorization. The decision is now on appeal.
Although no decision will be coming for a while, a number of recent events have returned the case to the public eye. In late January, several prominent, retired federal judges filed an amicus brief arguing that the courts should not charge any fees for public access to court records. That brief led to a story in the New Republic entitled “The Courts Are Making a Killing on Public Records.” All the while, the five-week federal government shutdown forced the courts to use up all of their “rainy day” resources and put them on the verge of operating without funding, illustrating the relative financial fragility of the courts as an organization.
I take as a given that the federal court system, as a whole, is committed to providing public access for all. But it is also a given that on an organizational level, the court system feels an obligation to protect its core activities from environmental disruption, including financial disruption. The current lawsuit provides an excellent illustration of the underlying tension between those values, and also suggests a solution. More below.
The players in the lawsuit
In April 2016, the National Veterans Legal Services Program, National Consumer Law Center, and Alliance for Justice filed a putative class action in the United States District Court for the District of Columbia, alleging that the Administrative Office of the U.S. Courts (AO) was charging external users fees for its Public Access to Court Electronic Records (PACER) service which “exceeded the amount that could lawfully be charged, under the E-Government Act of 2002.” At the time (and still today), the AO required that a user register an account with the PACER system and pay 10 cents per page for access to court documents, with a hard cap of 3 dollars per document. Certain documents (e.g., court orders and opinions) could be accessed for free, and any any individual registrant’s fees would be waived if they were under 15 dollars per quarterly billing period.
While the named plaintiffs were all nonprofit organizations, and the putative class included any member of the public who had paid for PACER access anytime during a six-year stretch from Fiscal Year 2010 to Fiscal Year 2016, the real force behind the lawsuit came from two sources: the news media, and certain individuals whose work in information technology convinced them that the court system’s fee structure was improper. The media had a vested interest in lowering PACER fees: court filings are often excellent resources for news stories, and reviewing many files on a regular basis could be prohibitively expensive–especially with ongoing financial stresses on traditional media posed by digital journalism. Technologically savvy opponents (including legal academics and the researchers who developed RECAP, a system to mine PACER documents and make them publicly available for free) also argued that the AO’s fee structure violated the law and stifled research.
The nominal defendant in the suit was the United States of America, a designation that masked the deep interdependence between the three branches of the federal government. The federal court system–particularly the AO and the Judicial Conference of the United States–were the true defendants. But they were represented by the executive branch (the Department of Justice) against charges of overstepping the bounds of legislative authorization. The question of what Congress actually authorized was left to the courts to decide.
Put differently, the case required the Third Branch of government to determine what the First Branch intended, guided in part by the arguments of the Second Branch. And the decision–whatever it as–would carry major financial consequences for the courts themselves. It was a peculiar setup, one of the most organizationally self-referential decisions facing the federal courts since a lawsuit over federal judicial pay in 2012.
The district court decision
The case was assigned to Senior U.S. District Judge Ellen Segal Huvelle, a longtime veteran of the D.C. District Court. After Judge Huvelle denied the government’s motion to dismiss and certified the class of “all individuals and entities who have paid fees for the use of PACER between April 21, 2010 and April 21, 2016,” the parties filed cross-motions for summary judgment. Both sides agreed on the relevant facts, leaving only a question of statutory interpretation: what restrictions did the E-Government Act of 2002 place on the amount the courts could charge for PACER access?
The historical context of this question matters, because the federal courts began developing PACER — with the full blessing of Congress — more than thirty years ago. In 1988, the Judicial Conference authorized an “experimental program of electronic access for the public to court information,” an action that was supported by various Senate and House committees in 1990.* Throughout the early to mid-1990s, Congressional committees explicitly noted that appropriations bills would authorize the Judicial Conference to “prescribe reasonable fees for public access to case information, to reimburse the costs for automating the collection of the information.” A 1996 House Appropriations Committee report used language that could fairly be characterized as broadly authorizing the collection of “electronic public access fees … to make information and services more accessible to the public through improvements to enhance the availability of electronic information.” Internally, the Judicial Conference recognized the power of the additional fees even more clearly, observing in 1989 that “such fees could provide significant levels of new revenues at a time when the judiciary faced severe funding shortages,” although the Judicial Conference never appears to have framed the issue quite that way to Congress.
As Internet access became more common in homes and businesses, demand for electronic court records rose. By the late 1990s, the court system had developed a web interface to access electronic documents and was charging users 7 cents per page. That fee would later rise to 8 cents a page, then 10 cents a page, with a maximum charge per document of $3.00.
The PACER fees appear to have made up a substantial amount of what the court system calls “other funding sources” — those funds that do not stem directly from Congressional appropriations. According the Judge Huvelle’s findings, the federal court system brought in $146 million in PACER fees in 2016; this amounted to almost 70% of the total fee collections ($213 million; see p. 32) allocated in the federal court system’s budget for the following year. In all, the court system brought in more than $920 million in PACER fees between FY2010 and FY2016 .
The funds derived from PACER fees were put to use on a variety of internal court projects related to public electronic access and information technology. But only some of those projects were directly tied to the implementation and management of PACER itself. Other projects (such as Electronic Bankruptcy Noticing (EBN), web-based juror services, and improving courtroom technology) arguably enhanced public access to the courts, but were not directly related to PACER or its internal documents management counterpart, CM/ECF. Still, the federal judiciary openly reported these projects to Congress (along with the courts’ intent to use PACER fees to fund them), and no Congressional committee raised any objections.
In 2009 and 2010, things changed. In each of those years, Senator Joe Lieberman voiced concerns that the then-current collection and use of PACER fees violated the E-Government Act of 2002. Section 205(e) of that Act amended existing law to provide that “The Judicial Conference may, only to the extent necessary, prescribe reasonable fees … for collection by the courts … for access to information available through automatic data processing equipment.” (Emphasis added.) The collection of hundreds of millions of dollars in PACER fees each year seemed inconsistent with that directive, given that PACER itself only costs a few million dollars a year to operate.
This “only to the extent necessary” language formed the crux of Judge Huvelle’s opinion on summary judgment. Rejecting both the plaintiff’s position that “the statute prohibits the AO from charging more in PACER fees than is necessary to recoup the total marginal cost of operating PACER” and the defendant’s position that PACER funds may be used to fund “any dissemination of information through electronic means,” the judge struck a middle ground. So-called “Program Requirements” (PACER itself, as well as CM/ECF, EBN, Court Allotments and Telecommunications/Communications Infrastructure) were deemed acceptable uses of PACER fees under the Act, while more remote programs like courtroom technology upgrades and web-based juror services were not. One distinction between the permissible and impermissible programs, the judge noted repeatedly, was that the former had a clear “nexus to the public’s ability to access information on the federal court’s CM/ECF docketing system”, while the latter did not.
At the request of the parties, Judge Huvelle’s order was stayed and certified for interlocutory appeal last August. It is now pending before the U.S. Court of Appeals for the Federal Circuit.
* All quotes in this section of this post are taken from the Factual Background of Judge Huvelle’s March 31, 2018 Memorandum Opinion.
“Public access” to court information: narrow and broad views
The core of Judge Huvelle’s inquiry was focused on Section 205(e)’s stated purpose of “requir[ing] federal funds to provide greater access to judicial information over the Internet.” (Op. 36). Through this narrow textual lens, her opinion makes good sense. Nevertheless, the court system’s diversion of surplus funds to projects just outside that ambit could be seen as a good faith effort to increase public access more broadly. First, the projects themselves, while lacking a nexus to PACER or CM/ECF, were designed to improve the court’s electronic communications and presentation of electronic information as part of its larger work. Beyond that, using the surplus money in this fashion freed up other funds for more general use, at least in theory. There is no question that the funds derived from PACER fees — $146 million in FY2016 alone — were substantial, amounting to about 2% of the court system’s most recent annual budget request. Those additional funds might allow the courts to provide better facilities, better staffing, and faster processing of cases — all of which go to a broader form of public access: access to justice.
I am not suggesting that this broad view of access is permitted by any reading of the E-Government Act of 2002. The point is simply that all organizations try to create buffers against budget volatility, and the surplus PACER funds offered the federal court system exactly that sort of buffer. It was money that, at minimum, might be used toward technological advancement that otherwise would have to be sought through the ordinary legislative appropriations process. It would have been human nature for the Judicial Conference and the AO to interpret the E-Government Act as broadly as possible, especially given that the same statute imposed a significant mandate on the courts to make information accessible, without providing any funds to do so.
Which “public” is served by PACER?
The federal court system is a paradigm public organization. It serves essential public functions (resolving disputes and developing the law) in a manner that aims for openness and transparency. And the vast majority of its funding comes directly from the public coffers. Its work is dependent on inputs from private actors (such as case filings and court documents), but the products of the court system are by and large public goods.
The question is, what type of public goods are electronic court records? To say that all services provided by a public organization–even a branch of government–must be “free” (in the sense of covered only by tax revenues) just because they are public goods is too much. Very few of the services that we consider to be “public” are fully available at no additional cost; rather, patrons are expected to pay something for the privilege of using and enjoying them. City buses and subways are not free. Neither are national parks, or (in communities like mine) services like trash collection. For most public entities, the question is not whether it should charge for a service, but rather what financial model it should follow.
There is a model that presumes no additional charge: the public restroom. It’s free, and it will do in a pinch. But that’s about it. Don’t expect frills or much in the way of quality. And if a public facility does impress, expect it to be crowded with free riders.
A second model is the city bus or subway. There, a heavily taxpayer-subsidized transportation system solicits at least a small additional revenue stream from its riders. The user fee is roughly equivalent to the marginal benefit afforded to each rider.
The third model is the school bake sale. Taxes continue to fund the core organizational infrastructure and services, but additional user fees are employed also to raise money above and beyond project costs. Like a parent who pays $3 for a homemade brownie in order to raise funds for a school’s new playground equipment, users of certain public services knowingly and willingly pay extra because they have a particular investment in the health of the organization.
The plaintiffs in the class action litigation essentially argued that Congress intended PACER to run like the city bus, charging only what was necessary to defray the expense of providing the service. In response, the government didn’t go in fully on the bake sale model, but it certainly viewed the use of PACER fees to promote court technology as within the spirit of Congress’s authorization. Judge Huvelle cautiously split the difference, a sensible result in light of the statutory history and text.
It would hardly be outrageous, however, to structure PACER fees according to a bake sale model. When PACER was first in development in the early 1990s, the expected core audience for the service was attorneys and their clients, who had a strong interest in a robust and healthy court system. Most members of the public will never encounter or need access to federal court records, and even if they do, the first $15 in fees each quarter is waived. Placing both the marginal costs of running PACER, and an additional charge to maintain and grow court technology, on the most frequent users (and beneficiaries) of those systems is entirely defensible.
The position that PACER services should be entirely free to all comers, by contrast, is considerably more complicated. Free PACER access would attract high demand from journalists, academics, and “big data” analysts, all groups that were not originally among the anticipated high-volume users of the system, and none of which have the same direct interest in the court system’s overall health as do the system’s everyday users. Moreover, in all public goods, high demand combined with no cost is a dangerous mix. The court system (and Congress) would be forced into one of three choices: (1) let the quality of PACER suffer because taxpayer funding alone was insufficient to meet user demands, (2) maintain PACER quality by diverting additional funding from other court operations, or (3) maintain PACER quality by increasing Congressional outlays to the courts for that express purpose. The first scenario would result in a suboptimal product. The second could compromise the financial health of the court system. The third would saddle taxpayers with the costs of a service enjoyed by a relative few.
Make no mistake: I am among those academics who have used PACER extensively for research, and I freely and frankly acknowledge both the benefits of the system and the frustrations of paying for it (or of seeking fee waivers). I am sympathetic to arguments, like this one advanced by Professor Carissa Byrne Hessick, that using fees to access public records as a means of funding government is generally bad policy. And I would love to see PACER become a cost-free source for academic research. But those funding decisions are best when they come from the legislature, not the courts. We can (and should) criticize the courts if they cynically raise fees simply to fund their general coffers, but when you are dependent upon another entity for 95-98% of your operating budget (as is the case for the federal courts), it is not hard to understand how the Judicial Conference and AO saw PACER fees as a golden goose.
Going forward, the best policy for PACER access — whatever it may be — should be determined by Congress. If this lawsuit is any indication, the court system is in a poor position to make the right determination on its own, as its internal funding needs would conflict with values animating free public access. Congress, by contrast, is institutionally suited to weigh the benefits and burdens of each funding model. If, as advocates suggest, free PACER access is a particularly important common good that promotes democracy, Congress should make that point explicit and reap the political fallout of spending more tax revenue to provide the service free of charge. If Congress determines otherwise, at least the courts and the public would know where they stand.