« Settlement in Hood County, TX | Main | Coming: A "Tilted" List of Recommended New Titles in Law and Religion [WITH UPDATE] »

Tuesday, August 18, 2015

How Should the Supreme Court Determine Whether a Group Has Political Power?

The starting point for this post is my view (not original to me) that the Supreme Court should step in to provide special protection from democratic politics to groups that lack political power.  These politically powerless groups are the discrete and insular minorities that the Court in Carolene Products footnote four expressed concern about being the target of harmful laws animated by antipathy.  These groups are unable to protect themselves from such laws because they cannot vote, they are unable to build coalitions with other groups, or they otherwise lack representation in the democratic process.  The familiar mode of judicial protection for such groups is the application of strict scrutiny to laws that classify on the basis of the group's status and are harmful to the group.  

A key for the Court in determining which groups are eligible for such special judicial protection is whether the group has political power to defend itself in politics.  As I suggested in the last post, one of the Court's primary measures of political power is whether that group has been the beneficiary of favorable democratic actions.  For the Court, such favorable democratic actions suggest the group is able to attract the attention of lawmakers.  If the Court applied this measure to the poor, it might determine that the group has political power because it has been the beneficiary of favorable democratic actions in the past.  But this determination would be in tension with social science findings that legislators and legislatures are not at all responsive to the preferences of the poor; findings that suggest the poor lack political power.

In a forthcoming article, Su Li and I sought to resolve this puzzle by testing the reliability of the Supreme Court's measure of political power as favorable democratic actions and the results were somewhat surprising. Details below the fold ...

We tested the relationship between congressional roll call votes on favorable democratic actions and the demographic makeup of the congressional district.  Our assumption is that for groups that have political power, the greater the proportion of a group in a district the more likely that the representative of that district will support legislation favorable to that group.  The theory behind the assumption is that congressmembers typically do not know the specific preferences of their constituents on most legislative matters.  Legislators therefore rely on heuristics such as the demographic makeup of the district.  Assuming the group has political power, the greater the proportion of the group in a district then the more likely the representative of that district will be concerned that an opponent might be able to mobilize that group to defeat her in the next election.  

We therefore expected that if the group had political power, there would be a positive correlation between the size of the group in the congressional district and support for legislation favorable to the group.  A lack of a statistically significant correlation between the size of the group in a congressional district and favorable roll call votes would suggest legislators did not support the favorable democratic actions in response to the political influence of the group.

We tested this hypothesis on three groups: union members, farmers, and the poor.  In the test we held constant factors that might bias the results such as the party, race, and gender of the representative, the region of the district (South/non-South), and the percent black and percent urban of the congressional districts. 

The results:  For union members and farmers, there was the expected positive correlation.  A ten percent increase in the percentage of union members in a congressional district was associated with a ten percent increase in the likelihood the representative would vote favorably to union members on union-related legislative actions.  Similarly, a ten percent increase of farmers in a congressional district was associated with a 14 percent increase in the likelihood the representative would vote favorably to farmers on farm-related legislative actions.  These findings suggest that past favorable legislative actions were, at least, in part responsive to the political power of these two groups. 

But then a surprising result for the poor.  For the poor there was a negative correlation between the percent poor in the district and the likelihood of a favorable roll call vote.  A ten percent increase in the percentage poor in a congressional district was associated with an 11 percent decrease in the likelihood that the representative would vote favorably to the poor on poverty-related legislative actions.  All of these results (for union, farmers, and the poor) were statistically significant at a .001 level.

What this suggests is that favorable democratic actions might be a reliable indicator of the political power for some groups (union members and farmers), but not for others (the poor).  Instead, it is likely that legislators support bills for reasons other than a group's political power, such as for reasons of ideology or morality.  One question that this raises is whether the source of favorable legislative actions matters.  So long as a group is a beneficiary of some favorable legislative actions, why does it matter that ideology motivated such legislative actions?  One concern is that groups who are beneficiaries of ideologically-based favorable democratic action will often have little control over what protections they will receive and when. 

The winds of ideology can shift in unexpected ways as they have with the poor.   The War on Poverty in the 1960s was followed by what some scholars have described as War on the Poor in the 1980s-1990s.  And this War on the Poor has been followed by a renewed concern about Economic Inequality in which the poor are sometimes mentioned.  The vagaries of the ideological winds might suggest that the Supreme Court should still provide special protection to groups unable to defend themselves in politics even if they have been the beneficiaries of favorable democratic actions. 

If not according to favorable democratic actions, then how should the Court determine whether a group has political power?   And does it even matter for the poor?  In the next post, I address a reason why the Court might not be willing to extend special judicial protection to the poor even if the Court considered this class suspect.         


Posted by Bertrall Ross on August 18, 2015 at 01:46 PM | Permalink


(Oops, sorry -- I just realized that you mentioned this. I should have read more closely.)

Posted by: Orin Kerr | Aug 20, 2015 2:22:05 PM

Thanks, Bertrall, really interesting.

One way to measure the ideological makeup of each district might be to use the percentage of voters in that district who voted for the Democratic or Republican Presidential nominee in the last election (or some set of recent Presidential elections). The higher the percentage of voters in that district that voted D, the more liberal it is relative to other districts; the higher the percentage of voters that voted R, the more conservative it is relative to other districts. That could let you control for the ideology of each district. The data is readily available for recent elections; it shows, for example, that NY's 12th district has voted about 80% D, 20% R in the most recent two elections.

Posted by: Orin Kerr | Aug 20, 2015 2:20:34 PM

Orin, that is a great insight and something we do not entirely account for in this project. What we do account for is the party of the representatives, which we assume is usually correlated with the partisan makeup of the district in the sense that Democratic representatives are usually elected in districts that have more Democrats, etc. We could not find a data source on partisanship of constituents at the district level. There is state-level and county-level data, but no district-level data. The best that we can do is look to presidential vote shares, but we did not consider that a reliable indicator of the partisan makeup of districts because of the many unique qualities of presidential elections. We can plug those numbers in to see if they bias the results. Our next project will attempt to tease out more precise demographic characteristics of districts from census and other data sources so that we can use a matching methodology that will allow us to move closer to making causal inferences about representatives' roll call voting behavior on poverty-based and other types of legislation.

As for your hypothetical about New York's 12th, if that is true about the roll call voting behavior of representatives of those districts then it would seem to support out conclusions that ideology might offer a better explanation for democratic actions favorable to the poor than the political influence of the poor. It seems doubtful that the representative of the 12th district is concerned about losing the vote of the poor if he/she opposed anti-poverty legislation. But he/she might lose the support of other Democrats who have an ideological orientation toward protecting the poor.

Finally, JHW, we do account for racial makeup of the district but do not account for the extent of the class differences in the districts. We hope to get at the specific class composition of the districts in our next project to see whether the degree of economic inequality in the district is associated with roll-call voting behavior on legislation impacting the poor. But Andrew did identify an interesting phenomenon that is related to the Thomas Frank puzzle about why the poor supposedly vote against their interests.

Posted by: Bertrall Ross | Aug 20, 2015 11:14:18 AM

There is also a complicated relationship between the percent poor in the district and the extent to which the voters supporting of the representative includes those poor people. In Andrew Gelman's book Red State, Blue State, Rich State, Poor State, he finds that poorer people vote for Democrats more across the country, but poorer *states* vote for Republicans more on balance because they have larger class differences in voting patterns. (Race is obviously an important part of this story.)

Posted by: JHW | Aug 19, 2015 6:55:29 AM

This isn't my area, so my apologies if I'm missing something. But do you control for the ideological makeup of each district? If not, I wonder if this might explain the results you find. For example, the five richest congressional districts are reliably Democratic. They include New York's 12th, the richest district, which represents the upper east side of Manhattan. The Representatives in those very wealthy districts presumably tend to vote for legislation that support the poor. But they do that because the voters in those districts are liberal, not because of the degree of political power of the poor. In such cases, your model's assumption that voters support legislation only based on self-interest would seem to break down.

Posted by: Orin Kerr | Aug 19, 2015 3:49:41 AM

The comments to this entry are closed.