Karen Bass leads the LA Mayor’s race, but the more revealing story is what happened after Election Night: Raman overtook Spencer Pratt for second place on the strength of late-returned ballots that were both more progressive in composition and more favorable to her within the middle of the Democratic electorate. EMC Electoral Personas show exactly who those voters were, and why the shift was bigger than demographics alone can explain.
What are EMC Electoral Personas?
EMC California Analytics assigns every registered voter in California to one of nine mutually exclusive Electoral Personas based on shared political attitudes and beliefs, not just demographics. Rather than sorting voters only by party registration or race and ethnicity, Electoral Personas capture how voters think about politics and ideology, grouping together people who approach elections in fundamentally similar ways regardless of their demographic background.
That distinction matters enormously for electoral analysis. Two voters who are both registered Democrats can belong to completely different personas if one is a committed progressive activist and the other is a pragmatic moderate. Electoral Personas make those differences visible and measurable at the precinct level, which is what makes them so powerful for understanding vote count movements like the one we saw in this race.
The nine personas in this analysis, from left to right:
Progressive Left⌄
Reliable Democratic Base⌄
Pragmatic Liberals⌄
Diverse Working-Class Democrats⌄
Anti-Establishment Independents⌄
Disengaged Moderates⌄
Conservative Independents⌄
Everyday Republicans⌄
MAGA Right⌄
A note on the ballot-return windows referenced throughout this analysis: “Before 5/29” refers to ballots returned before the final weekend before the June 3 election; “5/29 to 6/1” covers ballots returned over that final pre-election weekend; “Election Day” covers ballots returned on June 3 itself; and “Post-Election” refers to mail ballots received after Election Day, including those postmarked by June 3 and provisionals processed afterward.
Los Angeles Mayor · Final Count Through June 10, 2026 · 847,083 total votes
On Election Night Raman Was in Third Place
When the first results came in on Election Night, the race looked like a contest between Karen Bass and Spencer Pratt, with Nithya Raman running a distant third. Bass held 34.8%, Pratt was at 30.4%, and Raman trailed at 22.3%, sitting more than 8 points behind Pratt in the race for second place.
That picture shifted dramatically as counting continued into the post-election period, when mail ballots postmarked by Election Day and provisional ballots were added. Raman climbed to 38.4% in those later batches, Bass pulled back to 33.6%, and Pratt fell to 18.6%. By the end of counting, Raman had overtaken Pratt by nearly 20 points, a swing of nearly 28 points from where things stood on Election Night.
Two Forces Drove Raman’s Comeback, Not Just One
The instinctive explanation for Raman’s late surge is that progressive voters tend to return their mail ballots later, so of course a progressive candidate did better as the count continued. That is part of the story. But the personas also show that Raman improved her standing within voter groups that do not belong to her natural base, which raises a question about whether some late returners were voting with the general election in mind.
We can separate the two forces at work. The first is who returned ballots: more Progressive Left voters returning later shifted the electorate in Raman’s favor. The second is how those voters split: Raman appears to have won a larger share within several personas than her earlier numbers would have suggested.
About 57% of Raman’s swing came from the first factor, as the late electorate simply included more voters who favored her. The remaining 43% came from winning a larger share within those groups, which points to something beyond favorable timing. One plausible reading is strategic intent: in LA’s top-two primary, only the two highest vote-getters advance to November. Voters aware of the tight race between Raman and Pratt for second place may have cast late ballots specifically to ensure a Democrat-versus-Democrat general election.
The Late Electorate Shifted, and So Did Support Within It
The Progressive Left had the most dramatic shift of any persona across the count. Before May 29, Progressive Left voters made up just 12.6% of ballots returned. By the post-election window that share had grown to 26.6%, a 14-point jump. Since Progressive Left voters gave Raman roughly 60% of their vote compared to 33% for Bass and effectively zero for Pratt, that shift accounts for a meaningful portion of her surge.
But the late electorate was not exclusively progressive. Reliable Democratic Base and Diverse Working-Class Democrats, both large Bass-leaning groups, together made up roughly 29% of post-election ballots. Across those groups, and across Pragmatic Liberals, Raman picked up modest but consistent support in the later windows even though they leaned heavily toward Bass. Bass still dominated those personas, and that breadth is what carried her to first place. But in a race where Raman needed only to pass Pratt for second, small gains across large groups were enough to matter.
What This Means for Campaigns Using EMC Electoral Personas
The LA Mayor’s race shows why understanding voter attitudes produces more actionable insight than demographics alone. Looking only at the party registration, race, and age of late-returning voters, you would conclude the electorate shifted only modestly toward Raman. But when you account for the actual composition of who returned ballots and when, demographics explain only about 4.5 points of a nearly 28-point swing. The rest only becomes visible when you look at attitudes.
Raman did not win second place by converting another candidate’s base. She won it by improving incrementally across multiple persona groups while benefiting from a late electorate that skewed more progressive. The data also raises the possibility that some late returners were thinking strategically, preferring a Democrat-versus-Democrat general election over a Democrat-versus-Republican one.
A note on method: This analysis uses ecological regression applied to 735 LA City precincts to estimate how each Electoral Persona voted, based on the precinct-level mix of personas and observed vote totals. The estimates are best understood as coalition maps rather than claims about any individual voter’s choice. The model incorporates 25 demographic and voter-file controls to approximate the demographic makeup of returned ballots in each phase of the count. Count data reflects precinct results through June 10, 2026.

