The 1992, 1994 and 1996 elections: A comment and a forecast

A model of the two-way relationship between elections and the economy, previously estimated on historical data for 1916–1988, is applied to the United States elections of 1992, 1994, and 1996. The 1992 result was a surprise to the model since the economy had performed reasonably well that election year. The midterm elections of 1994 were accurately forecast. The Republicans took control of Congress not because of unusual circumstances but because of a normal midterm cycle. President Clinton's chances in 1996 look dim given the current modest growth rate and an electoral bias favoring Republican presidential candidates. But an alternative model, keyed more to the voters choosing Clinton to balance the Republican Congress, gives him a reasonable chance of reelection.     

Some Remarks on the "Generalized Event Count" Distribution

King (1989) presented the "Generalized Event Count" (GEC) modelas a means of dealing with event count data when the analystis unsure whether the data are "underdispersed" or "overdispersed."Here I establish several useful properties of the GEC modeland make some practical suggestions for estimation.     

Estimating Legislators' Preferred Points

This paper shows that agnostic spatial models that simultaneouslyattempt to estimate legislators' preferred points and ideologicallocations for the proposals on which they vote, such as thewell-known NOMINATE model of Poole and Rosenthal, are not identified.The problem arises because the agnostic estimators inherit thegranularity of the voting data and, so, cannot recapture theunderlying continuous parameter space. I propose an alternativeestimator that achieves identification by modeling the agenda.