Back to First Principles: Fixing San Francisco’s Public Agencies

A first principle is a basic, foundational truth that does not depend on other assumptions. When you engage in first principles thinking, you go back to fundamental truths and build up to your solution or decision from there. No shortcuts allowed. You’re not following other people’s advice just because they’re “experts” or because “things have always been done that way.”

Elon Musk is famous for first principles thinking. When Musk started Tesla, people told him he’d never be able to make batteries cheaply enough to mass-produce electric cars. To prove them wrong, Musk went back to first principles.

“First principles is kind of a physics way of looking at the world,” he told Kevin Rose in a 2014 interview. “You boil things down to the most fundamental truths and say, ‘What are we sure is true?’ … and then reason up from there.… Somebody could say, ‘Battery packs $600 per kilowatt hour. It’s not going to be much better than that in the future.’

“With first principles, you say, ‘What are the material constituents of the batteries?’ … It’s got cobalt, nickel, aluminum, carbon, some polymers for separation, and a seal can.… If we bought that on the London Metal Exchange, what would each of those things cost? It’s like $80 per kilowatt hour. So clearly you just need to think of clever ways to take those materials and combine them into the shape of a battery cell, and you can have batteries that are much, much cheaper than anyone realizes.”

Now, of course, public policy isn’t battery manufacturing. CEOs can hire and fire at will; public agencies can’t. Private companies are led by CEOs with expansive powers who answer to elected boards of directors; public agencies are led by administrators with circumscribed powers who answer to publicly elected politicians. We wouldn’t want public officials to act entirely like private‑sector CEOs—but we should want more first‑principles thinking in government.

As an experiment, let’s apply first‑principles thinking to two San Francisco agencies. Think of this as a tasting menu—just enough to show what’s possible when we interrogate tough problems at their roots.

The San Francisco Municipal Transportation Agency (SFMTA) manages transit, parking, and street operations. Its Muni division runs buses, trolleybuses, light rail, historic streetcars, and cable cars. Most of SFMTA’s ~$1.5B annual budget supports transit operations.

Muni’s costs have risen for years. Ridership plunged during the pandemic and has only recovered to roughly four‑fifths of 2019 levels (as of Oct 2025). Federal relief temporarily propped up the system, masking underlying structural gaps. The agency now projects a $300M+ shortfall in FY 2026–27 (agency estimate, Oct 2025).

Before floating a new tax, let’s start from first principles:

• Unit cost: What does it cost Muni to move one passenger today? How does that compare—after inflation—to ten years ago and to peer agencies?

• Cost drivers: Which inputs (labor, maintenance, pensions/benefits, procurement, overhead) account for most of the growth? Which can be managed in the near term, and which require structural change?

• Route economics: Which routes are the lowest‑ and highest‑cost per boarding and per passenger‑mile? Are there lower‑cost ways to provide equivalent mobility on the weakest corridors (e.g., frequency adjustments, microtransit, or TNC partnerships for first/last mile)?

• Fare performance: How does fare revenue today compare with a decade ago and with peers? Which policy levers (all‑door boarding enforcement, pass design, means‑tested discounts) move the needle most per dollar of cost?

The answers to these questions would likely leave most San Franciscans skeptical of a new Muni tax in 2026.

Let’s briefly consider one other San Francisco public agency, and how first principles thinking could help it find better solutions to its major problems.

San Francisco public schools educate about 50,000 students. Persistent structural deficits have triggered deep reductions (on the order of $100M+ this year), with more likely ahead—a painful process that leaves many families and educators worried about district decline.

This is a perfect moment for first‑principles analysis. According to the San Francisco Unified School District (SFUSD), its mission is to “provide each and every student the quality instruction and equitable support required to thrive in the 21st century.” Start with questions that clarify value for students and taxpayers:

• Per‑pupil spending: District revenue of roughly $1.3B divided by ~50,000 students implies about $26,000 per student. How does that compare—after inflation—to ten years ago, when enrollment was higher? What explains the change?

• Comparative performance: How do high‑performing charter networks or microschools deliver results at different spending levels? What practices could SFUSD adapt (staffing models, scheduling, curriculum, procurement, technology use)?

• Redesign: How could SFUSD offer better education for less? Where can technology reduce teacher burden without diluting human connection? Which community partners (libraries, nonprofits, City College) could expand learning time or career exposure at modest cost?

This isn’t rocket science. Citizens and leaders make better choices when they return to first principles—stating the mission, measuring what matters, and redesigning around value. The habit of asking fundamental questions doesn’t solve every problem, but it reliably exposes assumptions, surfaces trade‑offs, and points the way to better ways of achieving common goals.