Premium
AN EMPIRICAL TEST OF COMPETING EXPLANATIONS FOR THE MIDTERM GAP IN THE U.S. HOUSE *
Author(s) -
LEVITT STEVEN D.
Publication year - 1994
Publication title -
economics and politics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.822
H-Index - 45
eISSN - 1468-0343
pISSN - 0954-1985
DOI - 10.1111/j.1468-0343.1994.tb00082.x
Subject(s) - presidential system , punishment (psychology) , panel data , economics , test (biology) , empirical research , political science , public economics , positive economics , econometrics , psychology , social psychology , law , politics , mathematics , statistics , paleontology , biology
Previous empirical analyses of the midterm gap have been unreliable for two reasons: the limited degrees of freedom inherent in national‐level time series data, and the failure to adequately control for competing explanations. By using district‐level panel data, this paper is able to circumvent those weaknesses. Both withdrawn coattails and systematic presidential punishment at the midterm play an important role in explaining the observed midterm gap. Economic factors have a somewhat smaller impact. Evidence of systematic punishment of incumbent presidents in on‐year elections is also observed.