Members of Congress may not look like their constituents, but they do look like their parties

By Charlie Hunt

Does the “People’s House” look like the people? This question has been addressed from many different perspectives and the answer is usually the same: not really. On just about every salient demographic, cultural, and economic measure, House members do not “look” much like their constituents. For example, Casey Burgat and I have written  about the stunning wealth gap members of Congress enjoy compared to the American people; a new Roll Call report confirms our findings. And as Figure 1 indicates, groups like women and Americans of color remain sharply underrepresented in the current House, while Christians, married people, and the well-educated are highly overrepresented. Where can those concerned about descriptive representation turn for encouragement?

 Statistics in Figure 1 drawn from Pew Research Center, the U.S. Census, and (for member characteristics) the author’s original research.

Statistics in Figure 1 drawn from Pew Research Center, the U.S. Census, and (for member characteristics) the author’s original research.

A new analysis of original data on personal and occupational characteristics of House members offers one, perhaps surprising answer: our political parties. If the demographic and occupational backgrounds of House members are broken down by party, members start to more closely resemble their own party’s voters -- even if they don’t look like the nation as a whole. While this finding may reflect other problematic causes and effects -- hyperpartisanship, cultural polarization, and policy gridlock, to name a few -- it puts descriptive representation into a more realistic frame for today’s Congress, and more accurately reflects both how voters have come to think about their representatives as party members, and vice versa.

How do the two parties differ on demographic measures?  Figure 1 shows national averages of the public on these measures, but it fails to capture the substantial differences that persist between the parties. Exit polls from the 2016 presidential election, for example, show clearly how split Democratic and Republican voters are on characteristics like race, gender, age, education, and even marital status. Democratic voters are more likely to be female, nonwhite, unmarried, and well-educated; the converses are true for Republicans. This data reflects previously documented shifts : In “Partisan Hearts and Minds,” for example, Donald Green and his co-authors demonstrate that American voters are sharply dividing such that their partisan identities are overlapping more and more with these social groups. So, when the parties became more polarized, so did their representation among these subgroups.

But are these same patterns reflected in Congress? My data on member characteristics from 111th-115th Congresses indicate that, for the most part, they are. Figure 2 summarizes the attributes of each party’s members in the House from 2009-2018. The data show non-random differences between the parties’ members, nearly all of which point in the directions suggested by patterns among the public. For example, there have been about three times as many female Democratic House members as Republicans since 2009, and nearly ten times as many nonwhite Democratic House members. Meanwhile, more Republican members of Congress have been Christian and/or married. Age is the only non-conforming characteristic, with Republicans being younger on average than Democrats; however, this is likely due to the fact that more Republicans are “newer” members during this time period than Democrats.

Hunt - Figure 2.png

Another area where the parties show meaningful differences is in occupational representation, and in ways that are similar to patterns among partisans in the general public. Figure 3 breaks down the numbers: Republican House members are more likely to have backgrounds in business, the military, blue-collar trades and farming, while Democrats are more likely to have backgrounds in law or education. These reflect the trends uncovered in an analysis from Verdant Labs of FEC campaign donation data that show voters with teaching or legal backgrounds are far more likely to donate to Democrats, while business owners, farmers and military service members are just as likely to donate to Republicans.

Hunt - Figure 3.png

What does this all mean? Clearly, regardless of party, members of Congress still vastly over and under represent large swaths of the electorate. It’s a problem many are admirably trying to fix. But this analysis indicates that at the very least, our Representatives do better when it comes to descriptively representing their supportive party constituencies. For all the (well-deserved) doom and gloom surrounding the parties and their destructive tendencies, it appears they help members of Congress do one thing a little bit better: represent the voters who increasingly elect partisans.

Charles Hunt is a doctoral candidate at the University of Maryland, College Park