The presidential pulse and the 1994 midterm congressional election

The president's party lost more seats in 1994 than in any midterm election since 1946. The revised theory of surge and decline, which had successfully explained midterm losses since 1868, expected Democrats to lose only about half as many seats as they actually did. This research attempts to determine why the theory failed to anticipate the large 1994 losses. The aggregate national-level analysis examines 13 midterm elections from 1946 to 1994. Four hypotheses are examined: (1) that 1992 presidential short-term forces had been mismeasured because of the large Perot vote; (2) that partisan dealignment had weakened surge and decline effects to the point that they no longer accounted for interelection change; (3) that an increased linkage of presidential and congressional voting restored surge and decline effects to their prior levels; and (4) that a Republican realignment had shifted long-term political forces, obscuring the surge and decline in short-term political forces. The analysis supports the Republican realignment hypothesis. Incorporating the realignment into the surge and decline model accounts for 93% of the variance (adjusted R²) in midterm seat losses since 1946 and accounts for the large presidential party losses in 1994. These findings offer further support for the revised theory of surge and decline.