David Friedberg: California’s Voting System Looks Fraudulent, But It’s Working Exactly as Designed

## A Statistical Anomaly Hiding in Plain Sight

The 2024 Los Angeles elections produced voting patterns that don't behave like ordinary aggregations of individual choices. David Friedberg lays out the numbers: post-election mail-in ballots for one candidate (Pratt) declined by roughly one-third compared to pre-election mail-ins. For Nithya Raman, post-election mail-ins surged by 80%. For Karen Bass, they fell by 10%.

If millions of independent voters were each casting individual ballots based on personal preference, the ratio of support a candidate receives before election day should look broadly similar to the ratio they receive after. Late-arriving ballots shouldn't dramatically over-favor one candidate unless something is shifting the composition of *who* is sending ballots in late — not just *how many*.

The geographic concentration deepens the puzzle. Raman's incremental votes clustered heavily around the Skid Row area of Los Angeles — a population not typically associated with high-volume, late-cycle mail-in ballot participation. Friedberg's point isn't that any specific person broke a law. It's that the statistical signature of these results is hard to reconcile with the standard model of millions of citizens independently filling out and mailing their own ballots.

## How California's Legal Architecture Enables the Outcome

The provocation in Friedberg's argument is that the system producing these anomalies isn't broken. It's working as designed. Three legislative changes, layered together, create the conditions:

**Unlimited ballot harvesting.** California Assembly Bill 1921 legalized the practice of any individual collecting ballots from any other individuals — regardless of relationship — filling out the envelope paperwork, and submitting them. There is no cap on how many ballots a single person can harvest. One activist can legally walk in with hundreds or thousands of ballots.

**Universal automatic mail ballots.** Roughly 18 months later, California made it permanent that every registered voter receives a mail-in ballot, whether they asked for one or not. Tens of millions of physical ballots get distributed across the state every cycle, sitting in mailboxes, apartment lobbies, and shared addresses.

**Frictionless registration.** A subsequent series of laws removed citizenship verification from the registration process. A gym membership card can serve as identification. There is no ID requirement when a ballot is received, and no verification step confirming that the person filling out the ballot is the person to whom it was issued.

Stack these together and the result is structural: a massive supply of unsolicited ballots circulating in the physical world, no identity check at the point of voting, and a fully legal mechanism for third parties to collect and submit those ballots in unlimited quantities.

## Why "No Fraud" and "Fraudulent-Looking" Can Both Be True

The most counterintuitive part of Friedberg's argument is that nothing illegal is happening. He's emphatic on this point:

> There's nothing illegal or fraudulent going on. In fact, the system is operating exactly as intended.

This is the conceptual move that matters. The American political vocabulary around elections tends to collapse into a binary: either an election was "fraudulent" (laws broken, results illegitimate) or it was "free and fair" (everyone voted, the count was accurate, results legitimate). Friedberg argues there's a third category that California has built — an election where every transaction is legal, but the legal framework itself has been engineered to allow outcomes that look nothing like the aggregation of independent citizen preferences.

In this third category, accusations of "fraud" miss the point and are easily rebutted. No laws were broken. No ballots were forged. But the framework permits operations — mass harvesting from concentrated populations, ballots filled out by people other than the named voter, registration without citizenship verification — that, taken together, can determine outcomes through organized effort rather than individual choice.

The right description, in Friedberg's framing, isn't "fraudulent." It's "appointed, not elected." A candidate can be installed into office through a fully legal process that doesn't depend on persuading a majority of voters to independently fill out their own ballots and submit them.

## The Statistical Test That Matters

The before-and-after election day comparison is a quietly powerful diagnostic. In a normal election, you'd expect the partisan or candidate breakdown of mail-in ballots to be roughly stable through the voting window. There may be small shifts — late deciders, undecideds breaking in patterns — but the *composition* of late ballots shouldn't differ dramatically from early ones.

When one candidate's late ballots collapse by a third while another's nearly doubles, you're not looking at late-breaking sentiment. You're looking at a different *process* producing the late ballots. Whatever mechanism filled and submitted those envelopes operated under different incentives and access than the mechanism producing early ballots.

This is what makes the geographic clustering significant too. If late ballots represented genuine late-deciding voters, you'd expect them dispersed across the city in patterns roughly matching the broader electorate. Concentration in a specific area — particularly an area with a vulnerable, transient population that receives mailed ballots but may not actively manage them — points instead to organized collection activity.

None of this proves wrongdoing by any specific actor. It demonstrates that the system's outputs no longer carry the informational content people assume an election carries — namely, the aggregated independent preferences of citizens.

## What This Means for How to Think About Elections

The deeper lesson Friedberg points toward is about institutional design. Election integrity isn't a binary property that exists or doesn't. It's an emergent property of the rules, frictions, and verifications baked into the voting system. Each individual rule can sound reasonable in isolation:

- Why shouldn't a caregiver be able to drop off a ballot for someone they look after? (Ballot harvesting.) - Why shouldn't every registered voter get a ballot automatically? (Universal mail.) - Why should ID requirements create barriers to voting? (Registration loosening.)

But the *combination* of these rules produces a system in which the link between the registered voter and the ballot cast on their behalf has been severed at every checkpoint. There is no point in the chain where the system verifies that the person who chose the votes on a ballot is the citizen to whom that ballot was legally allocated.

Once that link is severed across the entire chain, the election's output stops being a measurement of citizen preference and starts being a measurement of organizational capacity — which campaigns, NGOs, or political operations can deploy the most labor to register, distribute to, and collect from the most ballots.

Friedberg's framing — that California's system "looks fraudulent but is working exactly as designed" — is meant to refocus attention. The question isn't whether laws were broken in a particular race. The question is whether a state has built a legal framework in which the strongest organized effort, rather than the largest base of independently motivated voters, determines who holds office. If the answer is yes, then suspicious-looking statistical patterns aren't anomalies. They're the predictable output of the machine.