Ahead of elections, some California counties conduct “voter outreach” drives. These campaigns, which typically feature television and social media ads, are supposedly designed to boost voter participation and reach “low propensity” voters. Counties often spend thousands of taxpayer dollars on these initiatives.
How are counties actually spending this money? City Journal looked at a week’s worth of social media advertising by Riverside and San Diego ahead of Tuesday’s midterms. The counties, it turns out, variously targeted their ads at users linked to tags like “African-American history,” “Native American history,” and “Gangsta rap”—in other words, proxies for groups that overwhelmingly support Democrats.
Finally, a reason to check your email.
Sign up for our free newsletter today.
Using Meta’s public database, we reviewed Riverside and San Diego’s social-media ad spending for the week ending November 1. The database provided information on each county’s spending on “detailed targeting” (ads directed at users “based on their interests, behaviors and demographics”) and on ads directed at specific demographic groups, such as young people or Spanish speakers.
In Riverside, where Donald Trump defeated Kamala Harris by just 1.3 percent in 2024, the bias was blatant. The county’s Registrar of Voters targeted more than 10 percent of its Meta advertising dollars at a zip code in downtown Riverside where Harris won an estimated 58 percent of the vote. Remarkably, it directed “detailed targeting” dollars at users linked to tags like “Wiz Khalifa,” “Soul music,” “Rick Ross,” “2 Chainz,” and “Gangsta rap.”
The county’s Registrar of Voters also targeted users who spoke a language other than English. Across the week we reviewed, Riverside directed more than 40 percent of its Meta ad spending at Spanish-speaking audiences. In California, Latino registered voters in Spanish-speaking households favored Harris over Trump by nearly 40 percentage points.
San Diego’s turnout campaign took a similar approach. The county directed nearly 65 percent of its Meta ad spending at viewers who spoke Filipino, Chinese, and other non-English languages. It also funneled taxpayer dollars to ads targeting users that Meta linked to “African-American history,” “African-American studies,” “African culture,” and “Native American studies.”
These content tags are proxies for groups that favor Democrats. Black California voters backed Harris by more than 50 percentage points. Native American voters nationally backed Harris by 18 percentage points. “Asian American and Pacific Islander” voters in California supported the Democrat by a 32-point margin. Neither county, of course, poured money into content tagged “white culture,” or anything of the kind.
Lawsuits may be driving these decisions. In 2022, Inland Empire United, a group represented by the ACLU, sued Riverside County for supposedly failing to give Latinos equal access to “the political process.” The parties settled, and Riverside agreed to engage in outreach efforts in low-turnout areas. The ACLU later boasted that the settlement “opens more doors for our Latine community members in Riverside to participate in the future health and success of their home regions.”
Riverside and San Diego did not respond to our requests for comment.
Supporters may claim these drives are intended to reach citizens who might not otherwise vote. For progressives, however, that is precisely the point. By targeting low-turnout, Democratic-leaning constituencies, counties can use taxpayer dollars for partisan gain.
Whatever the officials’ rationale, these counties were not conducting neutral civic-information outreach. They were, instead, engaged in a microtargeted turnout operation built around cultural markers, language, and neighborhood geography—precisely the tools that political consultants use to locate low-propensity Democratic voters. These choices, preserved in Meta’s own advertising records, seem to reveal a government turnout apparatus that has abandoned even-handed outreach for partisan voter mobilization.