Voting for president: Elections along the U.S.‐Mexican Border

Abstract

Historically, political discussions in the United States have been laced with phrases such as “the solid South,” “the Bible belt,” and “the industrial Northeast.” The ongoing use of these and similar phrases implies a widespread recognition that there are regional patterns of political behavior. To date, studies examining regional effects on voting behavior have been confined to state level data and broad regional definitions such as “the South,” which hide intrastate variation and obscure regional effects that cross state boundaries. Regional differentials in voting behavior are also apparent in other nations, including Mexico. This paper focuses on U.S. presidential voting behavior in the region surrounding the border with Mexico. Its purpose is to determine whether there is aborder effect on voting that goes beyond any economic or social peculiarities of the border region. Three basic questions are asked: (1) Is there a statistically significant difference between border and nonborder voting behavior, after controlling for differences in demographic, social, and economic variables? (2) If there is such a border influence, does it extend beyond the immediate border region? (3) Are differences in border voting behavior consistent across elections? Hypotheses based on these questions are tested using economic, social, and presidential voting data from 1992 and 1996. The units of observation include the 360 counties of the four border states of California, Arizona, New Mexico, and Texas. The paper provides an overview of border voting patterns, which is followed by a review of the literature on statistical and econometric models of presidential voting patterns. This literature review provides the basis for the third section in which a model is specified to allow statistical tests of various hypotheses regarding presidential voting patterns. Finally, the results are presented and conclusions drawn.