The Second Consideration

Judicial reform isn’t just about the courts. It’s also about getting Congress to do its work — and keeping courts out of the work they can’t do well.

Most conversations about reforming the federal courts are conversations about the courts. More justices or fewer, term limits, jurisdiction limits, ethics codes, supermajority requirements: all worth debating, and all questions about who sits on the courts and how much power they have, not about what the courts should be doing in the first place. That gap is where much of the trouble lives. What has gone wrong in the modern relationship between Congress and the courts has less to do with court structure than with the substantive work each institution is doing — and, more often, failing to do.

The framework I’ve been developing in this series is meant to shift the ground of the conversation. It starts from what courts are institutionally built to do and what Congress is institutionally built to do, and asks whether a given reform pushes each branch toward its proper role. Judged that way, a serious reform accomplishes two things: it encourages Congress to legislate, so that courts aren’t left as the default institution for making substantive policy, and it keeps courts out of areas where they lack the competence to act responsibly. Behind those goals sit the two considerations that anchor the framework. Political equality, the argument of the previous post, tells us when courts should push against the political process and when they mostly should not. Institutional competence, the argument of this one, tells us what courts can do well and what they can’t.

Two federal cases make the point. Meta’s antitrust case shows the failure mode; Lilly Ledbetter’s pay-discrimination case shows the balance done right. In structural terms they pose the same question and receive opposite answers. What separates them is the second consideration.

The case Congress left alone

The FTC’s antitrust case against Meta has crossed four presidencies. Obama’s Federal Trade Commission approved Facebook’s acquisition of Instagram in 2012 under prevailing merger doctrine, and approved the WhatsApp acquisition in 2014 under the same doctrine. In December 2020, Trump’s FTC — chaired by Joseph Simons, on a bipartisan vote — sued Meta to unwind both acquisitions on a theory the doctrine had never really entertained: killer acquisitions, in which a dominant firm buys nascent competitors precisely to keep them from growing. The Biden administration inherited the case, and the second Trump administration inherited it back. In November 2025, Judge James Boasberg, an Obama appointee, ruled for Meta on the ground that “personal social networking” is not a coherent monopoly market, given the growth of platforms like YouTube. The FTC appealed in January 2026. The appeal is pending.

Nothing in that thirteen-year history involves an act of Congress.

It couldn’t, because Congress has not passed a substantive antitrust rule in fifty years. The last significant piece of federal antitrust legislation was Hart-Scott-Rodino, in 1976, which created pre-merger notification but said nothing about which mergers should be allowed. The framework the government and the courts have been working with is older still. The Sherman Act, from 1890, runs thirty-four words. The Clayton Act, from 1914, adds provisions on price discrimination, exclusive dealing, and mergers, but leaves the substantive tests general. Everything you associate with modern antitrust (e.g., market definition, monopoly power, per se rules and the rule of reason, tying, refusal to deal, the tests for monopolization) is judge-made doctrine, layered onto those two statutes over more than a century.

The operating rule for the whole apparatus is the consumer-welfare standard, and it is worth being precise about where it came from, because that is where the argument lives. Consumer welfare came from an academic book: Robert Bork’s The Antitrust Paradox, published in 1978. Reagan’s Justice Department adopted the standard in 1981. Every administration since, Republican and Democratic, has run federal antitrust enforcement under it, and federal judges have written it into circuit doctrine. Congress has never voted on it. No bill has ever said consumer welfare is the test; no bill has ever said it isn’t. A standard that governs the substantive shape of the American economy is the product entirely of academic writing, prosecutorial choice, and judicial elaboration, without a single endorsement from the elected legislature.

Under the consumer-welfare standard, harms are measured by their effect on consumer prices, output, and quality. That is the doctrinal reason the FTC has had such difficulty with Meta. Instagram and WhatsApp are free. Whatever else the acquisitions did to the digital economy, they didn’t raise the price of Instagram. A doctrine built to catch price-raising conduct is a poor tool for evaluating platform gatekeeping, and the poor fit is not something a district judge can repair on his own initiative. Judge Boasberg applied the doctrine he had. He was always going to.

What courts are good at (and not good at)

The claim here is not that Judge Boasberg got the case wrong. Reasonable people can disagree about whether “personal social networking” is a coherent market, whether YouTube counts as competition, whether an independent Instagram or WhatsApp would be thriving today. The claim is that a federal court is a strange institution to be answering the question at all.

Courts decide cases one at a time, on the record the parties present. They read the parties’ expert reports and hear the parties’ witnesses. If more evidence would help, they can order more discovery, but typically only from the parties. They cannot commission independent economic studies, convene a panel of experts of their own choosing, or wait five years to see whether TikTok displaces Instagram before deciding whether Instagram is entrenched. And once a case is decided, it’s decided. The judge doesn’t monitor whether the ruling produces sensible market outcomes; if a technological shift changes the picture, no institutional mechanism updates the doctrine. The next case has to arise, be litigated by different parties on different facts, and work its way back through the system.

None of this is a criticism of judges. It’s what judging is. Courts are designed to resolve disputes between specific parties on specific facts, not to make policy for a rapidly evolving economy — and the design constrains what they can produce when they are handed policy work anyway. Judge Boasberg had to decide in November 2025, on the record in front of him, with the tools a judge has. Those are different tools, built for a different job.

A legislature holds the other set. It can commission studies, hold hearings, take testimony from a range of experts and stakeholders, watch markets evolve, and rewrite rules that turn out to be wrong. It can also make the political-economy judgments — how to weigh consumer welfare against innovation incentives, worker outcomes, small-business viability, platform gatekeeping — for which there is no doctrinally correct answer, only choices a legitimate polity gets to make. Set the two toolkits side by side and the reform question rearranges itself. It stops being “what should the courts do?” and becomes “what should Congress have been doing that would have kept the courts from having to?”

The Ledbetter model

The counter-case is Lilly Ledbetter, who worked at Goodyear Tire and Rubber for nineteen years and was paid less than the men doing the same job. She filed a Title VII complaint in 1998, and in 2007 the Supreme Court ruled against her, 5–4. Justice Alito, writing for the majority, held that Title VII’s 180-day filing deadline runs from the original pay-setting decision, not from each subsequent paycheck. Ledbetter hadn’t sued within 180 days of the decision that first set her pay low, so her claim was untimely — never mind the smaller paychecks that had kept arriving for nearly two decades.

Justice Ginsburg read her dissent from the bench, a step a justice reserves for decisions she considers especially wrong. Her closing line: “the ball is in Congress’ court, as I see it, to correct this Court’s parsimonious reading of Title VII.”

Congress took the ball, though not quickly. A response bill was introduced within weeks and passed the House in July 2007 on a mostly party-line vote; the Senate version was filibustered by Republicans in April 2008, and the bill died. Then the 2008 election changed the configuration. Democrats picked up seats, Obama won the presidency, and the bill was reintroduced in the new Congress in January 2009. This time, with a larger Democratic majority and enough Republican senators willing to invoke cloture, it passed. It was the first bill President Obama signed into law, on January 29, 2009. The statute provides that each discriminatory paycheck resets the 180-day clock. The Court’s reading is gone; Title VII pay-discrimination doctrine now runs on the rule Congress wrote.

Two features of that sequence deserve attention. First, the correction took twenty months, an election, a change in Senate composition, and a redoubled political effort. That is what the institutional balance looks like when it works — not fast, not easy. Congress has the tools to override the courts on substantive policy, but they are political tools, and the political system has to be willing to use them. Second, the Ledbetter Fair Pay Act didn’t ask Congress to design a regulatory regime from scratch. It asked Congress to fix a discrete statutory ambiguity, a task well within the modern legislature’s bandwidth even when the politics are hard.

Meta’s case is thirteen years old and has crossed twice as many administrations. Congress has not written a substantive rule. The political system, so far, has not been willing.

The second consideration

Institutional competence supplies the second half of the framework’s test. Judicial substitution for the elected branches is legitimate only when two conditions hold. The first is that political equality is at stake. That is, the political process is failing to deliver the conditions of equal citizenship. That was the argument of the last post. The second is that the Court is institutionally positioned to address the failure. When both hold, judicial intervention is warranted. When neither holds, the courts should stay out. And when only the first is present (political equality is at stake, but the courts cannot address it well) the answer is not to leave the failure unaddressed. It is to legislate, and to write the legislation so that courts aren’t handed work they aren’t built to do.

Ledbetter met both conditions. Pay discrimination is a canonical political-equality concern, and reading a statute is squarely within judicial competence. The Court got the reading wrong, but Congress had the tools to step in. And, after an election, Congress did just that. Meta fails the second condition. Whether killer acquisitions by dominant platforms are anticompetitive is a policy question that demands economic expertise, industry monitoring, and the capacity to update, none of which federal courts possess. Congress possesses all of it, and has done nothing for fifty years, under either party.

The reforms that follow from the framework pull the system toward the Ledbetter mode and away from the Meta mode. Condition judicial deference on the quality of the legislative record, so Congress is rewarded for doing its substantive work. Restrict courts from making empirical determinations they cannot reliably make, so the judiciary isn’t asked to design policy without the tools. Return the responsibility for updating substantive law to the elected branches, so a question can’t sit with the courts for fifty years while everyone looks the other way.

None of these proposals is on the loud edge of the debate, and that is itself diagnostic. The loud edge is occupied by plans to reallocate power between the political parties inside the judiciary. These reforms are about something else: reallocating institutional roles between the courts and the legislature. That is the debate worth having, and it is not, so far, the one we have been having.