Race, legislative speech, and symbolic representation in Congress

The forthcoming article “Race, legislative speech, and symbolic representation in Congress” by Arjun Vishwanath is summarized by the author below.

Most blacks, Hispanics, and Asians in America are represented by white members of Congress; in light of this, it is important to understand whether these groups are well represented by white MCs. A variety of studies have tested whether white legislators are more likely to vote in line with minority groups’ interests as they become more populous, but these studies do not point to a firm conclusion in any direction. By contrast, little work has examined whether white legislators are more likely to symbolically represent racial minority groups if they have more minority constituents.

This article addresses the question of whether white legislators are symbolically responsive—that is, if a minority group is more prevalent in the district, does its representative speak more about that racial group or evoke symbols particular to that group on the floor of Congress or in her electronic newsletters? I use supervised machine learning algorithms to detect the symbolic content of over thirteen million sentences with respect to blacks, Hispanics, and Asians. The content identified by the model typically takes the form of group mentions (e.g., “Asian”, “Colombian”, “Hmong”, “AAPI”) or symbolic references to terms relevant to the group (e.g., “Rosa Parks”, “Selma”, “Cinco de Mayo”).

I find that white legislators are more likely to engage in symbolic speech of blacks, Hispanics, or Asians if the member hails from a district with more citizens of the respective group. This is true of both Democratic and Republican legislators, although Democratic legislators are more responsive to black group size in their symbolic speech about blacks than Republican legislators are.

While we observe differences across legislators—that is, a white MC from a 25% Hispanic district symbolically invokes Hispanics to a larger degree than another white MC from a 5% Hispanic district (controlling for legislator party and the ideology of the district)—I find that legislators do not actually adapt to constituency changes over time. I test this by looking at sudden changes to the district composition following redistricting. That is, if an MC’s district changed from 10% Hispanic to 15% Hispanic after redistricting, that MC does not increase her symbolic speech of Hispanics.

At first glance, it may seem obvious that white legislators will be symbolically responsive to minority group size. However, I replicate past work showing that legislators are no more likely to vote in line with minorities’ preferences in Congress if they have more minority constituents. Thus, we see white legislators with strongly conservative voting records in places like Texas who nonetheless speak about and engage regularly with their Hispanic constituencies.

This naturally raises the question of whether symbolic representation is meaningful, or whether legislators merely speak hollow words while betraying their constituents’ interests in policymaking. Although symbolic representation should not be taken as a substitute for substantive representation, the article concludes in discussing the ways in which symbolic representation might matter to citizens.

About the Author: Arjun Vishwanath is a Postdoctoral Research Fellow at Vanderbilt University in the Center for Effective Lawmaking and the Department of Political Science. Their research “Race, legislative speech, and symbolic representation in Congress” is now available in Early View and will appear in a forthcoming issue of the American Journal of Political Science.