Bertrall Ross and Su Li have posted this draft on SSRN (forthcoming, California Law Review).  Here is the abstract:

Which classes are considered suspect under equal protection doctrine? The answer determines whether courts will defer to legislatures and other government actors when they single out a group for special burdens, or intervene to protect that group from such treatment. Laws burdening suspect classes receive the strictest scrutiny possible — and under current doctrine, whether a class is suspect turns largely on whether the court views the group as possessing political power.

But how do courts know when a class lacks political power? A liberal plurality of the Supreme Court initially suggested that political power should be measured according to a group’s descriptive representation in politics. Under that measure, the mostly white, male, wealthy, straight makeup of most of the nation’s decision-making councils would indicate that other groups lack political power. But that measure never received majority support from the Court. Instead, the Court consolidated around a different measure of political power, one that focused on democratic actions favorable to a group. If laws have been enacted protecting the group from discrimination or otherwise advancing the group’s interest, the Court assumes that the group can attract lawmakers’ attention and therefore does not need judicial protection.

In the forty years since the Court introduced this standard, no class has been found suspect under it. In fact, it is hard to imagine any class ever meeting the standard of political powerlessness under this measure. Even the most politically marginalized groups (such as the poor, non-citizens, and felons) have benefited from laws favoring their interests. Is favorable democratic action really an accurate measure of political power? Focusing on the poor, we advance the first empirical test of the Supreme Court’s measure of political power. Our findings suggest that legislators’ support for anti-poverty legislation is not motivated by the political power of the poor — implying that favorable democratic action does not always accurately indicate a group’s political power. Given these findings, we argue that the court should rely on a more holistic, and thus more reliable, measure of political power. The measure should include favorable legislative actions, but also indicators of lobbying activity, political responsiveness, voter turnout, and descriptive representation in politics.

This looks to be a very important piece!