In the last weeks, there have been unprecedentedly large protests against police violence and systemic racism. The grievances are familiar. As with the acquittal of Trayvon Martin’s killer, which ignited #BlackLivesMatter, or the Los Angeles uprisings after Rodney King was beaten by police, or anti-lynching demonstrations during Jim Crow, black Americans are demanding an end to police brutality.

But this time seems different. The costs of protesting are higher, and grievances are more intense. Partly as a result, policymakers appear to be doing more to address protesters’ concerns.

In my book, tentatively titled “The Advantage of Disadvantage,” I argue that elected officials are more likely to support the interests of groups who overcome great barriers to protest. Costly protest reveals intense concerns that, if ignored, could hurt elected officials’ reelection prospects.

This argument might seem odd. Certainly, some elected officials treat protests as illegitimate. And research continues to find that policymakers tend to favor white people over black people.

But costly protests are different.

What’s a “costly” protest?

Protesting always draws some costs. Taking to the streets takes time and energy away from work, caregiving and other responsibilities. Protesting involves voicing opinions that others disagree with. Protesting can also bring physical or financial costs including arrest, which may require bail or attorney fees.

But protest costs are not distributed evenly.

In my book, I show how protesters’ race correlates with protest costs. For example, in many places, people, predominantly black, who protested George Floyd’s killing were met with militarized police in riot gear, tear gas, arrests, excessive force or death, even when protesting peacefully. Meanwhile, white people dressed in military gear and carrying guns protested against coronavirus restrictions and police didn’t respond — even when protesters stormed and occupied the Michigan capitol building.

What’s more, the pandemic makes protest exceptionally costly. Black people are disproportionately infected with the coronavirus and are dying from covid-19. That’s in part because of the increased likelihood of having “essential worker” jobs as nurses, food service employees and public servants; higher levels of stress, insufficient health care, inadequate sources of healthy food, unclean water and impure air; and living in more crowded conditions as a result of underemployment. Standing in a crowd means further increasing the infection risk — especially when police use tear gas and pepper spray, which could inflame covid-19 infections.

Black people are literally risking their lives to protest because grievances accumulated over a lifetime are boiling over.

Protest costs matter to people in power

My research analyzes protests reported in the New York Times from 1991 through 1995 from the Dynamics of Collective Action data set. With research assistants, I also evaluated an original data set of protests from 2012 reported in 20 newspapers from major U.S. cities to understand how the Internet influences elected officials’ responses to protests.

Using these data sets, I measure the costliness of protest based on the protesters’ race, ethnicity and income levels, and the tactics protesters employed.

Some protesters face greater costs because of how people perceive their actions. For example, President Trump referred to the mostly black demonstrators in Minnesota outraged over Floyd’s death as “thugs” and threatened violence with a phrase used in the 1960s to repress black protests. Conversely, Trump praised gun-toting white protesters who stormed the Michigan Capitol to protest restrictions meant to control the spread of the coronavirus, calling them “very good people.”

I use the race and ethnicity of protesters to estimate the costs of protesting for groups. I code protests by white groups as less costly than protests by black, Latino, and Asian Americans. Protests by the poor and homeless are more costly than protests by the affluent. Moreover, some protest tactics require more time and effort than others, like marching in the street versus signing an online petition.

To understand whether legislators are more likely to support costly protests, research assistants helped me create a data set of protest-related bills before the U.S. House of Representatives in the same congressional session that the protest occurred. They coded whether a legislator voted in support of an issue raised during a protest in their district.

I find that legislators are 44 percent more likely to vote in favor of black protesters’ demands than white protesters’ demands. They are 39 percent more likely to vote in support of Latino protesters’ demands than white protesters’ demands. Legislators are 32 percent more likely to vote in favor of low-income protesters than more affluent protesters. And they are 10 percent more likely to vote in support of in-person protests, like marches or rallies, versus online protests, like online petitions or social media posts.

Legislators’ greater support of costly protest exists even when evaluating protests of equal size, disruptiveness and media attention.

Elected officials are responding to the costly Floyd protests

For instance, the Minnesota attorney general charged all officers involved in George Floyd’s death. With a veto-proof majority, the Minneapolis City Council voted to disband its police force and replace it with a community-led public safety model. A Louisville city council committee voted in favor of a law limiting no-knock warrants after the death of Breonna Taylor. House and Senate Democrats are sponsoring the Justice in Policing Act of 2020 to limit no-knock warrants, prohibit chokeholds and other carotid holds like those leading to the death of Floyd and Eric Garner, along with other reforms. And police departments and schools districts across the U.S. are changing policies in response to the protests.

To be sure, the costliness of the protests is not the only factor in these rapid actions. Black Lives Matter has been shifting public opinion for several years now, and the Floyd protests have dramatically increased that shift, so much so that a majority of Americans now say they support the protesters’ claims. But given the dangers — the costs — of protesting during a pandemic, massive unemployment, widely circulated stories of brutality, and an election year, officials realize that the grievances are intense — and that black people and their allies are likely to punish officials at the polls if they don’t take action now.

LaGina Gause (@LaGina_Gause) is an assistant professor of political science at the University of California San Diego.